I’m still reading Schniedewind’s A Social History of Hebrew
The influence of Egyptian scribal culture would become quite widespread in early Israel. In addition to learning the practices of accounting (that is, using hieratic numerals) and of writing with ink, the early Israelites borrowed several linguistic terms relating to the scribal profession from Egyptian. To begin with, it is worth noting that the Hebrew word for “scribe” (sōᵽēr) derives from the root spr, which originally meant not “to write” but rather “to count,” reflecting the administrative roots of the scribal profession. We have already mentioned the terms for “ink” and “papyrus,” but other Egyptian (Eg.) loanwords include those for “a scribe’s palette” (Eg. gśty; Heb. qeseṯ), “seal” (Eg. ḫtm; Heb. ḥôṯām), “signet ring” (Eg. ḏbʿt; Heb. ṭaḇaʿaṯ), “ephah” (a certain measurement for grain; Eg. ypt; Heb. ʾêᵽâ), “hin” (a certain liquid measure; Eg. hnw; Heb. hîn), and “zeret” (a span of measurement; Eg. ḏrt; Heb. zereṯ). There are very few Egyptian loanwords in Hebrew, but most are related to the scribal technology and profession, most likely reflecting the continuing work of Egyptian natives or Egyptian-trained locals after the retrenchment following the Twentieth Dynasty.
Schniedewind is annoyingly repetitive and occasionally confusing in his attempts to differentiate between spoken and written language (he’s read the linguists but does not seem to have fully assimilated what they say), but I’m learning enough from him that I’m willing to forgive him his deficiencies. (For a less indulgent response, in Russian, see Anatoly’s review.)
Is the Hebrew root for “count/write” spr or sfr ?
The Schniedewind quote says:
In the first counting and telling post you wrote:
Philogos is quoted as saying:
Stu:
Older forms of Hebrew did not have a phonemic distinction between [p] and [f]. Modern Israeli Hebrew does, because it has abandoned gemination, so that √SPR, the root we are talking about, is pronounced [f], whereas √SPPR > sapphire is pronounced [p] rather than [pp]. (The fact that sapphire is /f/ in English has to do with sound-change within Greek, which picked it up from a Semitic language and passed it to Latin which forwarded it to French and English, à la Tinker to Evers to Chance.)
Direct link to Schniedewind’s article on the “Me illiterate, boss? Hell no” soldier stationed at Lachish (from Anatoly’s blog post). There’s a little snootiness there about how the guy doesn’t know scribal conventions for letter-writing and that makes him only “marginally literate”, which is like saying I’m “marginally literate” for not knowing the rules about Yours sincerely vs. Yours faithfully in a business letter.
John: Older forms of Hebrew did not have a phonemic distinction between [p] and [f]. Modern Israeli Hebrew does, because it has abandoned gemination, so that √SPR, the root we are talking about, is pronounced [f], whereas √SPPR > sapphire is pronounced [p] rather than [pp].
Hat complains that Schniedewind is “occasionally confusing in his attempts to differentiate between spoken and written language”. I am for the same reason confused by your “did not have a phonemic distinction between [p] and [f].” My question is about the root as an artefact of writing, not how it was pronounced then or is now.
When you write “√SPR, the root we are talking about”, I assume you are talking about the letters. That would be an answer to my question: “Is the Hebrew root for “count/write” spr or sfr ?”. The answer is spr.
Why then does Philologos write “the same verbal root s-f-r”, and Labuschagne “one and the same verbal root s-f-r” ? Are these typos, or perhaps more evidence of laxness in differentiating between spoken and written language, as Hat charges Schniedewind with ?
Is there a hard and fast convention that √ means “the written form” ? I’ve often seen /[IPA stuff]/ used to mean “the pronounced form”.
The lack of absolutely clear distinctions between letters and pronunciation must make it relatively easy to speculate about what derives from what
Sorry, there was a bit of professional deformation there in what I wrote (and me not even a professional linguist). “√” means neither more nor less than “the root”, so “√SPR” is another way of writing s-p-r or S-P-R.
Pronunciations are written in square brackets using the IPA, so [p] is “the p-sound’, [f] is “the f-sound”, and “[pp]” is “the sound of [p] prolonged”. Slashes are used when we are only notating significant differences; that is, those which can distinguish words. The English word pin is [pʰɪn] and spin is [spɪn], because there is a difference in pronunciation in the two p-sounds, but it is automatic: we use [pʰ] at the beginning of a word, and [p] (which sounds like French p) after [s]. In order to show that we are disregarding this difference, we write /pɪn/ and /spɪn/ with slashes.
It’s precisely because the question of whether to say [p] or [f] depends on the environment of the sound, and there are no (older) Hebrew words distinguished by the difference, that it makes no difference whether the root SPR is written SPR or SFR. In older Hebrew, indeed, there is only one letter for both of them; modern Hebrew uses a diacritical mark to distinguish them when necessary.
I am no Marie-Lucie, and I should have gone to bed hours ago, but I hope that helps.
But not to demonstrate what derives from what, of course. What I mean is that the influence of writing and pronunciation on each other over time must give precious hints about developmental changes. However, people without at least some awareness that things have changed over time often use “surface similarities” of writing and pronunciation to make unsound derivational claims.
Stu: A single Hebrew character is used to designate /p/ and /f/. It is פ and when it appears at the end of a word the character appears as ף. (My understanding is that the word-final form [and that of a few other letters as well] is the original form; the form of the letter at the beginning or within a word is due to the tail [descender] being extended to the left by the scribe moving the quill to begin the next letter.) That’s why when transcribing to the Latin alphabet some authors will use F and others will use P — and occasionally PH, as in Phineas.
When diacritic marks were developed (their form was finalized about 800 AD in Tiberias), a decision was made to insert a dot into the middle of the P-F character when the sound it represented was P. I don’t know how to create it directly in the comments window, so I’m pasting from Word and perhaps it will show correctly: פּ
For technical reasons the /p/ sound historically cannot ‘close’ a word, though in today’s practice it occasionally does, as when transcribing words from other languages. When that happens, the form without the descender is used.
Wiki entry here.
The word-final form of the letter is a near-perfect mirror-image of Latin P. That is not a coincidence.
John: my last written before your last, but posted after, without my having seen yours. Your explanation has cleared things up for me, thanks.
Does the word “cypher” have any etymological relationship to the Hebrew S-F-R discussed supra? The wiki entry takes this word back to an Arabic word meaning zero, but only a thousand or so years ago. Obviously, zero has something to do with counting; so too, sending code by cypher might have reflected scribal activity similar to telling a story or sending a message.
Or is this one of those false etymologies that Mr. Languagehat abhors?
Thanks for the further explanation, Paul.
The word-final form of the letter is a near-perfect mirror-image of Latin P. That is not a coincidence.
Heh. Do you mean because they wrote back’ards ?
Does the word “cypher” have any etymological relationship to the Hebrew S-F-R discussed supra?
Nope, that root has a different s. AHD has a good etymology:
[Middle English cifre, from Old French, from Medieval Latin cifra, from Arabic ṣifr, from ṣafira, to be empty (translation of Sanskrit śūnyam, cipher, dot); see ṣpr1 in Semitic roots.]
Do you mean because they wrote back’ards ?
Kinda.
I’m not sure about early Semitic writing, but when the Greeks initially adapted the Semitic alphabet for their own use, the direction of the script was still fluid and people often wrote as the ox plows.
The phenomenon seen with פ P can also be seen with ר r and with ק q.
Sort of like a which-way-is-it-going Studebaker, for those of us who remember such things.
The notion of borrowing scribal terminology seems eminently reasonable, but I wonder about this one: Egyptian (Eg.) loanwords include . . . “signet ring” (Eg. ḏbʿt; Heb. ṭaḇaʿaṯ)
The root טבע teva appears not only in Hebrew, but also in Assyrian, Aramaic, Arabic and Ethiopic. We need a cuneiform scholar to tell us more.
In I Samuel 17, a verbal form of that root is used to describe the stone sinking into Goliath’s head . . .
“signet ring” (Eg. ḏbʿt; Heb. ṭaḇaʿaṯ) … The root טבע teva appears not only in Hebrew … In I Samuel 17, a verbal form of that root is used to describe the stone sinking into Goliath’s head
Thus the German word for brass knucks: Schlagring.
I consider it condescending to append an 🙂 to jocular remarks addressed to people who have not fallen on their heads (Ger.loc.).
On further consideration, David’s deed seems less remarkable if in fact his little pouch contained not just five smooth stones, but also a signet ring – what he really slung at Goliath. Today, boxer’s gloves are inspected to make sure they don’t conceal rocks and nuts.
JC: me not even a professional linguist
I know some of them whom you would put to shame!
I am no Marie-Lucie : and I am no John Cowan!
I was going to say the same thing as Paul Ogden about ṭaḇaʿaṯ “signet ring”. This looks like a clear derivative of the native root ṭbʕ, rather than a loanword.
This is a bit imprecise (uncharacteristically for JC):
√SPR, the root we are talking about, is pronounced [f], whereas √SPPR > sapphire is pronounced [p] rather than [pp]
In Modern Hebrew, roots like spr give forms with both stop and fricative, depending on reasons of historical phonology, some of which are no longer synchronically clear: e.g. siper “(he) told” but safar “he counted”. And there was never a root sppr; there are no Semitic roots of the form XYYZ (and arguably none that contain the same consonant twice in sequence at all), and in any case sap(p)ir “sapphire” is a loanword in Hebrew (according to Wiki, ‘from Old Iranian sani-prijam, from Sanskrit, Shanipriya (शनिप्रिय), from “shani” (शनि) meaning “Saturn” and “priya” (प्रिय), dear, i.e. literally “dear to Saturn”’).
And from ṭbʕ, we get the Arabic word for “to print”.
The case of ḏrt > zereṯ, right next to ḏbʿt > Heb. ṭaḇaʿaṯ, arouses my suspicions though. Why should Egyptian ḏ, whatever its pronunciation was, be reflected so differently in two different words?
An older but more technical reference for Egyptian > Hebrew, Lambdin 1953.
The word ṭaḇaʿaṯ puzzles me. From what s it an Egyptian loan, or does it derive from the Hebrew root ṭb‘ ‘to sink’, of proto-Semitic origin (as Paul O. has noted)? Compare with behemoth, בהמות. Is it the morphological plural of בהמה behemah ‘beast’, i.e. ‘big beast’? Or is it a borrowing of an putative Egyptian form p-eḥe-mau ‘water ox’, i.e. hippopotamus?
This sort of thing is what makes the craft of etymology so difficult. You have a number of competing hypothese, all slightly imperfect but plausible, and it takes a long search for a sliver of evidence that might eliminate one possibility and favor the other.
To add more to this etymological stew, Gesenius relates ספיר ‘sapphire’ to Arabic šfr~ṣfr ‘to scratch’, and hence ‘to polish’. He then connects ‘to scratch’ with sfr ‘to write’, etc. The Aramaic forms (search for ‘sapphire’ <a href="http://cal1.cn.huc.edu"here) are quite varied, and include forms like sanprinon. Is the middle n old? Is it epenthetic?
I don’t like the semantic path ‘to write’>’to tell’, but that’s just me. count>tell (‘recount’) and count>write make more sense, as does count>tell>write.
Lameen, are there any Berber roots which might be cognate to this whole sfr complex?
The word ṭaḇaʿaṯ puzzles me. [I]s it an Egyptian loan, or does it derive from the Hebrew root ṭb‘ ‘to sink’, of proto-Semitic origin (as Paul O. has noted)?
Well, it could, in a sense, be both: i.e. a loanword eggcornized into a native-looking form.
…And sapir occurs in Exodus, much earlier than when most Persian loans entered Hebrew.
Gesenius relates ספיר ‘sapphire’ to Arabic šfr~ṣfr ‘to scratch’, and hence ‘to polish’. He then connects ‘to scratch’ with sfr ‘to write’, etc.
But again we have the wrong sibilant in Hebrew, right? That Arabic root could be cognate with Hebrew ṣippōren “fingernail”, but Hebrew /s/ should equal Arabic /s/. (There’s also the semantically close Hebrew ṭofer “claw, talon”, but I don’t think that can be related either given the usual correspondence sets.)
TR, you’re right, I was not convinced either. Incidentally, I found a paper by Blau, which discusses in detail supposed irregularities in Semitic sibilant correpondences. He argues against most of them, and in particular he thinks Early BH /s/ almost never corresponds to Arabic /š/.
And from ṭbʕ, we get the Arabic word for “to print”.
The printing press came late to the Arab world — using the Arab alphabet, not until Napoleon set one up in Egypt. I wonder when the Arabic word was coined. The Hebrew word for printing is דפוס dfus, borrowed from the Greek tupos, impression. I don’t have my Klein handy, but at a guess it was borrowed in Talmudic times.
Hebrew for coin is מטבע matbe’a, from that same ṭbʕ root.
I don’t like the semantic path ‘to write’>’to tell’, but that’s just me. count>tell (‘recount’) and count>write make more sense, as does count>tell>write.
Count > recount > write makes a lot of sense. I recall that one of the first uses of “writing” in commerce was to indicate the contents of containers, which by nature would have included quantities, aka numbers.
@Paul Ogden, re: “at a guess [דפוס] was borrowed in Talmudic times”: Indeed. Even-Shoshan marks it with his symbol for words first attested in that general time-period, and gives two quotations from the Talmud (one from Menachot 94, one from Bava Batra 16).
@Paul Ogden, re: “Count > recount > write makes a lot of sense. I recall that one of the first uses of ‘writing’ in commerce was to indicate the contents of containers, which by nature would have included quantities, aka numbers”: I guess I see that, but in that case it seems strange that it would apply to the words for “book” and “scribe”, but not to the word for “write” itself (which is from the root K-T-B).
Or the root רשם rasham to record.
@Lameen: I know that Arabic also uses the root כתב katav. Does it use rasham as well?
רשם is an Aramaic word, borrowed into both Hebrew and Arabic.
Although the Arabic root has /s/: rasama, presumably because at the time of the borrowing Arabic did not have a /š/ sound (maybe it was still /ɬ/ as in Proto-Semitic, or maybe it had become /ç/ or the like).
Müteferrika’s 1727 petition to Sultan Ahmet III to allow him to set up a printing press already used the word ṭibāʕah for “printing”, so the extension of the term to this usage is presumably older. I assume dfus meant something other than “printing” in the Talmud?
rasmī is “official”, but the more usual connotation of rasm in modern Arabic is “drawing”.
Signet rings were worn on the finger, so Egyptian ḏbʿt is presumably related to ḏbʿ “finger”, itself related to widespread Semitic ‘iṣbaʕ.
In Arabic, the vocalisation ā-a in ṭābaʕ and xātam “seal”, where one would expect ā-i, is something of a giveaway that both were borrowed from Aramaic. But, judging from Kitab al-Ayn, the root ṭbʕ has a number of other more obscure senses in Arabic that look more likely to be original, such as “to fill completely” and “(of a sword) to get dirty”. The sense of “nature” (ṭabīʕah) is probably derived from “stamp, seal”.
In Algerian Arabic, ṭəbbəʕ means “push” – not sure how that happened…
I assume dfus meant something other than “printing” in the Talmud
In modern Hebrew, dfus is also used along with טיפוס tipus to mean character, as in “He’s a strange character.” This parallels English usage (character=type). That may have been the original Greek meaning too, and hence Talmudic, as back then there was no printing and no printing type.
rasmī is “official”, but the more usual connotation of rasm in modern Arabic is “drawing”.
In modern Hebrew, Arabic rasmi is also used colloquially to mean official; I presume that’s a holdover from Ottoman times. Standard Hebrew uses רשמי rishmi. רישום rishum means drawing.
The root רשם r-sh-m is used to indicate writing/recording in special contexts, like making a grocery list, taking minutes at a meeting and the like. Certain maintainers of official lists, like the government registrar of companies, use the term too: רשם החברות rasham hakhaverot. רשומות reshumot is the title of the government gazette.
Even-Shoshan . . .gives two quotations from the Talmud (one from Menachot 94, one from Bava Batra 16).
From what I can make of the Aramaic, the usage in Menachot 94, where the discussion is about baking bread, indicates that דפוס means something like ‘form’, aka baking pan.
The discussion in Bava Batra 16 is too obscure for me to understand.
דפוס in Bava Batra 16 seems to mean a path that lightning takes. It all has to do with Job, puns, and the resurrection of the dead. Why? I have no idea.
Or molds for shaping raindrops individually (!). That’s what Jastrow says about דפוס at BB16a, at any rate.
The OED s.v. character says that Greek χαρακτήρ meant ‘instrument for marking or graving, impress, stamp, distinctive mark, distinctive nature’, < χαράττ-ειν 'to make sharp, cut furrows in, engrave'.
…which may in turn be a Semitic loan: cf. Aramaic and Hebrew ḥaraṭ “engrave”.
If it’s really the case, it’s jolly awesome that a borrowed -tt- gets interpreted as *-kʲ-, making the verb a velar-stem verb with a verbal noun in -ktēr.
TR, you’re right, I was waylaid by Gesenius again, and by my ignorance of Arabic. Arabic has ršm ‘make the sign of the cross’—that has to be an Aramaic loan. The Hebrew word occurs once in the Bible, in Daniel, so I think it’s likely an Aramaic loan into Hebrew. Proto-Semitic rśm seems a reasonable source for Arabic rsm and Aramaic ršm, if I’m not getting the sibilants mixed up again.
Well, there would have been lots of synchronic models, e.g. prāttō : prāk-sis. (Although it’s not clear that it really is a loanword, apparently; Chantraine thinks the loan hypothesis is “aberrante”, though he χαρακτηρistically fails to say why.)
JC: The OED s.v. character says that Greek χαρακτήρ meant ‘instrument for marking or graving, impress, stamp, distinctive mark, distinctive nature’, < χαράττ-ειν 'to make sharp, cut furrows in, engrave'.
Nice connection!
TR: …which may in turn be a Semitic loan: cf. Aramaic and Hebrew ḥaraṭ “engrave”.
Pokorny finds cognates of Greek χαρακτήρ in what looks like three unnamed Slavic languages but nowhere else.
BDB has an entry for חרת ḥaraṭ but notes that it appears only in the qal construction and most likely not in any other Semitic language.
Turkish has matbaa for printing press. Rotogravure is tifdruk, a brazen theft from German Tiefdruck, “deep-printing.”
Btw, it’s tempting to connect the Aramaic root ḥrṭ “engrave” and its Hebrew cognate ḥrṣ, which gives meanings like “cut, sharpen; ditch, trench”, with ḥrš “plough, till”, which came up a while ago in a thread I can no longer find (in which I tried to relate it to Hittite har(a)š- meaning the same thing, though no one was convinced). A lot of Semitic triliteral roots seem to be built on earlier biliterals, and in this case there seems to be some evidence for such a biliteral in the noun ḥor “hole”.
Pokorny finds cognates of Greek χαρακτήρ in what looks like three unnamed Slavic languages but nowhere else.
No Slavic languages, just Lithuanian žeriù, žer̃ti `scrape’ and its frequentative žarstýti.
Paul, I was referring to חרט ḥaraṭ, not חרת ḥarat; how the latter root fits into the story, if at all, isn’t clear to me.
I was referring to חרט ḥaraṭ, not חרת ḥarat; how the latter root fits into the story, if at all, isn’t clear to me.
BDB indeed has a larger entry for חרט ḥaraṭ and derivatives. With this spelling (I’m not sure we’re really dealing with separate words), there are a number of cognates in Arabic and Aramaic. But the sense of cutting, engraving and so forth is the same.
A lot of Semitic triliteral roots seem to be built on earlier biliterals, and in this case there seems to be some evidence for such a biliteral in the noun ḥor “hole”.
Yup and yup. The lovingly-written but badly-designed How the Hebrew Language Grew explores this notion at some length.
I’ve recently been reading up on controversies in biblical interpretation, one point of which is the question as to whether Biblical Hebrew can be properly divided up into “Standard” and “Late”. Avi Hurvitz wrote a couple of works defending the division by bringing up several examples of vocabulary pairs for which there is good reason to think that each belongs to “Standard” and “Late” Hebrew, respectively. One of the example pairs was the “Standard” seper/sefer (ספר) and “Late” ʾiggeret (אגרת) as terms for “letter; writing; missive”.
Hurvitz notes that sefer is used in the Lachish ostraca, and apparently has some cognate in Ugarit, whereas ʾiggeret is even used in the Targumim as a translation for sefer.
The OP reminded me of this, and makes me wonder what the etymology for ʾiggeret is.
(And what other Hebraists/Semiticists might think of the SBH/LBH issue — is there really any controversy?).
References/Links:
The historical quest for “ancient Israel” and the linguistic evidence of the Hebrew Bible. Vetus Testamentum, 1997
Can biblical texts be dated linguistically? Chronological perspectives in the historical study of biblical Hebrew. Vetus Testamentum Supplements, 1998, Volume 80
ʾiggeret looks like it comes from the root ʔgr “collect, gather” (i.e. words/information collected together?), but apparently things aren’t so clear; Hebrew Wiktionary thinks it’s a borrowing from Aramaic but ultimately has a Persian source.