My wife just told me that William Safire has died of pancreatic cancer at the age of 79, and I was shocked and saddened by the news. Long-time readers may be surprised by the elegiac tone of this post, because in the early years of LH I used to take great and unholy glee in ripping apart his language columns in the NY Times; it still annoys me that the Newspaper of Record handed such a potentially powerful educational tool over to someone with no qualifications other than a love of language and writing, who frequently made the kind of obvious errors that set my teeth on edge. But I came to realize that it was not, after all, his fault that he didn’t have the appropriate background; he did, after all, have a strong love of language and writing; and despite it all he did dig up plenty of interesting information. More importantly, from my own selfish perspective, two years ago Oxford University Press had me copyedit the latest edition of his Political Dictionary, and it was one of the best editing experiences I’ve had. He fully appreciated my pickiness about details and had no hesitation making changes I recommended, even occasionally adding chunks of text I provided; what’s more, he credited me by name in those entries and added this heartwarming text to the acknowledgments: “For this fifth edition, Stephen Dodson provided the kind of creative copy-editing and a lust for historical accuracy and semantic precision that a political slanguist expects in dealing with the Oxford University Press, world’s greatest lexicographic organization.” He took to calling me up and we were soon on a “Bill” and “Steve” basis, and the last time we talked he promised to buy me a beer if we were ever both in New York at the same time. I’m sorry we won’t get the chance to have that beer, Bill, and especially that we won’t get to work together on another book.

The NY Times obit is by one of my favorite Times reporters, Robert D. McFadden, who’s been with the paper since 1961 and covers disasters like nobody else; I winced, of course, at the phrase “a talented linguist,” but appreciated writing like this:

He was hardly the image of a buttoned-down Times man: The shoes needed a shine, the gray hair a trim. Back in the days of suits, his jacket was rumpled, the shirt collar open, the tie askew. He was tall but bent — a man walking into the wind. He slouched and banged a keyboard, talked as fast as any newyawka and looked a bit gloomy, like a man with a toothache coming on.

And I was delighted to learn that he was born Safir: “The ‘e’ was added to clarify pronunciation.” Goodbye, and thanks for the fun of both bashing you and editing you.


  1. So it goes.
    Glad to hear that you found such common ground.

  2. I had no idea you knew him so well. No, he wasn’t a linguist, but he was moved by a love of language, and he wrote well, even though he wasn’t able to analyze what made writing good.

  3. That’s great you worked with him! I had no idea!

  4. I always loved his pieces, it was the one thing I always read it the Sunday paper.

  5. Thank you for this post, Languagehat. I’m sorry you were never able to have that beer — I’m sure the conversation would have been lively!

  6. «And I was delighted to learn that he was born Safir: “The ‘e’ was added to clarify pronunciation.”»
    I don’t think I, in this far-off corner of Canada, have ever heard his name pronounced, and have always assumed it was SA-fire (rhyming with satire). Is that correct, or is it SA-fear? Is the stress on the first syllable or the second?

  7. always assumed it was SA-fire (rhyming with satire). Is that correct

  8. Then what was the pronunciation of the original name Safir?

  9. I think without the “e” there’s a temptation towards a “sa-FEAR” pronunciation, which he or his family was trying to point away from.

  10. Then what was the pronunciation of the original name Safir?
    I assume the same, if the “e” was added to clarify pronunciation — though if you’re asking about the name’s history, I have no idea.

  11. “Clarify” may not be the right word here. No literate English speaker would pronounce “Safir” like “sapphire” (which could well be its equivalent), or spell a word rhyming with “fire” without a final e. Perhaps they wanted to avoid “SAY-fur”?

  12. Wikipedia says, more vaguely, that he “later added the “e” for pronunciation reasons”.

  13. Hmm, in Arabic safeer سفير means ambassador. Do these websites ever have accurate information?

  14. Yes, the Dictionary of American Family Names is a very reliable source. (Ashkenazic Jews don’t have Arabic names.)

  15. Ashkenazic Jews don’t have Arabic names.
    But the similarity between Arabic and Hebrew has sometimes startled me (shalom/salam). Not sure how Yiddish fits in, if at all, but it was the similarity of the sound that sent me looking for a Middle Eastern explanation (assuming “Safir” is pronounced like saFEER سفير). It looks like the Hebrew for ambassador is שגריר pronounced something like shagrir, not sure about the accent. Maybe adding an e shifts the accent to the first syllable.

  16. Arabic and Hebrew (and others) are descended from a common ancestor, just like, for instance, French and Spanish and Italian which have Latin as their common ancestor. This is why many words (but not all) are similar in the two languages, although similar words may not have the same meaning. Yiddish is a Germanic language similar to German, but it also contains many Hebrew words.
    The words “safir” and “sagrir” (assuming that the latter is correct) cannot be related to each other.

  17. I appreciate your comment that the New York Times missed the opportunity with Safire to turn that real estate to someone more erudite. But Safire was fun to read and turned on many of us who wanted to be entertaining and also accurate in what we wrote about language. I did not know Safire as you did, but my tribute is posted here: sinandsyntaxsalon.com.

  18. There is no doubt that Safire’s name is pronounced “sapphire”, and it is clearly a Hebrew name meaning, in fact, sapphire (or at any rate some blue precious stone, perhaps lapis lazuli). Hebrew itself may not be the source of the Greek form σαπφειρος, but some Semitic language was. From Greek it spread into Latin and so the Western languages, and also into Persian and (probably) Arabic. The /pp ~ ppʰ/ > /f/ transition happened independently in Greek, the Romance languages, and Hebrew.
    How the word got into Semitic is more conjectural: the OED’s and Online Etymology’s favored theory is that it is ultimately < Skt sanipriya ‘precious to Saturn’, with the /n/ assimilated to /m/ (as in Aramaic) and then /p/ before /p/.

  19. marie-lucie says

    So Aramaic is the source of the last name Samphire, which I had been wondering about.
    What is the “Saturn” element in the Sanskrit word? it would have to be a functionally equivalent god in Hindu mythology, which I don’t know that well.

  20. Sani is the ‘Saturn’ morpheme, but I don’t know if the planet Saturn or a counterpart to the god Saturn is meant. The Greco-Roman-Celtic gods and the Hindu gods do have, very remotely, a common IE origin, though the theme of the “kingship in heaven” (with the ruling god consuming his children, who then turn the tables on him) may be of Semitic origin, since it shows up in Babylonia but as far as I know not in India.
    I should mention that while sapphire is first recorded in English in the 13th century, it takes French forms: safir (OF), saphir (MF, ModF). Not until the 18th century is the Graicizing spelling sapphire , with “pph” from πφ, found.

  21. From Greek it spread into Latin[.]
    John C., is it possible that some form of the word “sapphire” was borrowed by early Latin speakers at the same time as, or later but independently of, their Greek fellow-Mediterraneans? I wonder if the migration of, for example, Semitic words into Latin need to have been a specifically Greek introduction, and not a (later) Hellenicizing cementing.
    (I love studying ancient Greek and the cultures of the Greek speakers, but, you know, they weren’t the only sea-farers, travelers, merchants, and whiz kids!)

  22. marie-lucie says

    is it possible that some form of the word “sapphire” was borrowed by early Latin speakers at the same time as, or later but independently of, their Greek fellow-Mediterraneans?
    From the Online Etymological dictionary:
    sapphire: from O.Fr. saphir (12c.), from L. sapphirus (cf. Sp. zafir, It. zaffiro), from Gk. sappheiros “blue stone” …… from a Semitic source (cf. Heb. sappir “sapphire”).
    The “pph” in the middle of Latin sapphirus shows that it is borrowed from Greek, ending up with a [pf] sound which became [ff] or [f] in later forms of the language such as Italian, Spanish and French. If it had been borrowed directly into Latin from Hebrew or another Semitic language such as Phoenician, it would have had [pp] or [p] or perhaps [f] (written f, not ph). If [pp], Italian would still have it, and Spanish and French would have [p]. If [p], Italian would still have it, Spanish would have [b], and French would have [v]. If single [f], Italian and Spanish would still have [f], and French would probably have [v].
    Such are the attested correspondences (found consistently in many words) between the sounds of various languages, which allow historical linguists to trace the origins as well as the language-to-language transmission of words through borrowing. A rare and expensive luxury item such as sapphire is the kind of thing for which the word follows the item in its travels, in this case through the ancient trade routes: probably Phoenician > Greek > Latin. The other likely transmission route could be Phoenician > Etruscan > Latin, but the Latin word, if borrowed from Etruscan, would not be sapphirus (I am not sure what it could be, but I am pretty sure that it could not be that).

  23. deadgod, m-l: It is quite possible that Latin acquired safirus or the like directly from Semitic-speakers, and then respelled the word in Greek fashion later. We know this is possible because of what happened in English, where the original French spelling was replaced by a Greek one.

  24. Thanks, marie-lucie, for explaining clearly the (it looks to me) strong argument for the borrowing by Latin speakers of a Semitic word from its Greek borrower-users, rather than directly from Semites, an explanation emphasizing consistency of orthographic pattern over a relevant sample and concomitant cultural-anthropological context.
    And thanks, John C., for explaining a similarly strong counter-argument (to marie-lucie’s) in about a thousand words fewer than I would have.

  25. marie-lucie says

    JC: It is quite possible that Latin acquired safirus or the like directly from Semitic-speakers, and then respelled the word in Greek fashion later. We know this is possible because of what happened in English, where the original French spelling was replaced by a Greek one.
    Good point, but there is the evidence of the Latin descendants, especially Italian.
    Another, strong argument is that the sound [f] is very rare in Latin between vowels, although frequent at the beginning of words. One exception is rufus ‘red’, which was considered “rusticus” as opposed to the standard, urban ruber ‘red’. A luxury item like sapphire would not have been within the means of “rustic” speakers, and if the original word had had an [f] it would most likely have been reformed in early Latin: on the model of the [f] of rustic rufus corresponding to the [b] of urban ruber, the hypothetical word safirus (or more likely safirum, as this is an inanimate substance) would have been reshaped as sabirum. In that case, the descendants would have evolved quite differently: Spanish would still have [b], but Italian and French would have [v].
    A person knowledgeable in the history of the Semitic languages would probably be able to say whether there was a form in [f] likely to have been imported in early Latin or not, but both linguistic and cultural features are against the hypothesis of an early borrowing, while a borrowing from Greek (itself a borrowing from a Semitic language) for this luxury item at a later date is the most likely both linguistically and culturally.
    From a more general point of view, “respelling” and “orthographic pattern” refer to the written word, not the spoken one. Until recent times (the Renaissance and later) the spelling of words followed the pronunciation quite closely, and sound change was unaffected by “spelling” (it was the other way around). That’s why I put [p], [f], etc between brackets: following standard practice in linguistics, these symbols refer to sounds rather than letters, as with Latin pph (letters) representing [pf] (sounds). Note that the respelling of Eng saphire (from Middle French) as sapphire on the Greek model did not result in a change to a spelling pronunciation “sapfire”.

  26. That the English saphire was “respelled” sapphire “on the Greek model” is a case of what John C. was talking about: a respelling that did not reflect any difference in pronunciation, but rather the influence on the spelling of an admixed word by a third cultural presence. That is, ‘the Latin re-spelling of a Semitic word in accordance with a Greek model’ might be like ‘the English re-spelling of a French word in accordance with a Greek model’.

  27. marie-lucie says

    deadgod, you and JC are reasoning in the abstract, but there are concrete facts too. The idea of a respelling of an existing word under the influence of a more prestigious language might be true in this case if circumstances were the same in the Latin period as they later were in English or French, but they are not.
    In Latin, a different spelling reflects a different pronunciation (and there are older texts and comments which reflect the changes over the centuries). In English or French, that was true until the Renaissance, when people discovered the classics and respelled a number of words which they saw (or imagined) must have come from Latin or Greek, and added extra, “silent” letters to the existing words: commonly cited examples are the b in debt because of Latin debita, or for French the g in doigt ‘finger’ because of Latin digitus. The point was not to change the pronunciation, but to show off one’s knowledge of classical languages. There was no such reason for a respelling of Latin words of whatever origin.
    Note that these English and French examples are of everyday existing words, not of words fished out directly out of Latin or Greek texts and minimally adapted to the language. In this respect those words are also unlike Latin sapphirus which is an obvious copy of Greek sapphiros, not an everyday word but one referring to a luxury item for which there is no earlier Latin word attested (and I suggested what the word would have been according to Latin rules, if it had been borrowed from a Semitic language at a much earlier date). If there had been such a word, there would have been no reason to borrow a Greek word for the same thing: Greek borrowings in Latin are not replacements for meanings for which there were already Latin words. This follows the general pattern according to which words are borrowed from other languages, either because they denote new or foreign things or concepts, or because the foreign word refers to a different instance of the same thing or concept. For instance, if you look up “cake” in an English-French dictionary, you will find gâteau, but the English word “cake” is also used in France, not as a synonym for just any kind of cake, but only for fruitcake, at first an English specialty.
    For Latin there are many writings on the language by Latin authors, which cite or discuss old or otherwise different forms. For Late Latin there are some lists (called in French “gloses”) of the type “Say X, don’t say Y” through which one can learn some of the differences between Classical Latin (eg Cicero) and a later spoken version, in terms of pronunciation and vocabulary. Of course, there may have been many more such texts which have not survived, but If there had been an older term than sapphirus (or a similar example), that is where it would be found. And don’t forget that the Romance descendants of Latin are unanimous in deriving their word for sapphire from the Greek borrowing: one could have thought that in Spain, where the Romans were hundreds of years before they conquered Gaul, and where there had been Phoenician colonies for much longer, there could have been a more direct borrowing from Semitic into Latin which might have remained in the local language and passed into the ancestor of Spanish, but there is apparently no trace of any such word.

Speak Your Mind