Monthly Archives: November 2016

The Problem with Reading the Word ‘Hebrew’ in Sinai 115: An Egyptologist’s Response

23 November 2016

Reading the Word ‘Hebrew’ in Sinai 115: An Egyptologist’s Response
A Guest Post by Prof. Dr. Thomas Schneider
University of British Columbia

As part of Douglas Petrovich’s proposal that the Early Alphabetic Inscriptions (e.g., Serabit el-Khadem, Wadi el-Hol) are Hebrew, he has proposed (as putatively supporting evidence) to read a word in one of the Egyptian inscriptions (namely: Sinai 115) from Serabit el-Khadem as the word for “Hebrew.” This, however, is just not correct. Here are some of the reasons:

(1) the correct reading is in two lines: Jpn, son of Jrw. Petrovich combines J and p from line one with r from line 2.

(2) Egyptian /j/ is used to render Semitic aleph, never an ayin (and one would need an ayin here to have the word “Hebrew”).

(3) Similarly, in the 12th dynasty, Egyptian /r/ is never a transcription of Semitic /r/, but of /l/ and /d/.

(4) Also, /p/ is a regular rendering of Semitic /p/ and not /b/, although this is less objectionable.

(5) Ultimately, therefore, Petrovich’s proposed reading of this word as “Hebrew” is not possible.


Dr. Thomas Schneider

The Early History of the Alphabet and the Recent Claim that the Northwest Semitic Inscriptions from Serabit el-Khadem and Wadi el-Hol are Hebrew: Spoiler Alert, They’re Not.

23 November 2016

serabit-el-khadm-rollstonsdrawingThe Early History of the Alphabet and the Recent Claim that the Northwest Semitic Inscriptions from Serabit el-Khadm and Wadi el-Hol are Hebrew: Spoiler Alert, They’re Not.
By Christopher Rollston
Recently, it has been claimed by Douglas Petrovich that the Early Alphabetic Inscriptions (some of which date to as early as the 18th century BCE) from Serabit el-Khadem and Wadi el-Hol are Hebrew. His volume on these inscriptions will be published by Carta (Jerusalem) in the coming months. Petrovich also presented a paper on this subject at the annual meeting of the American Schools of Oriental Research on Thursday, November 17th, 2016. I was present for his presentation. Alas, I am confident that his proposal for these inscriptions will not get traction with scholars who work in the field, that is, scholars who work in the fields known as Palaeography and Comparative Semitics. I will write a much longer discussion of his proposal fairly soon (around the time the book comes out). But I wish to put pen to paper at this moment as well, so as to help bring some important empirical evidence to the fore.

But before discussing some of the serious problems with the Petrovich proposal, I should like to emphasize first that the core inscriptions that form the basis of Petrovich’s claim are not recent discoveries. (1) In fact, in the case of the Early Alphabetic inscriptions from Serabit el-Khadem (in the Sinai), these have been known for more than a century (with publications on them going back, for example, to 1906 and 1916 by Gardiner). (2) And as for the Early Alphabetic Inscriptions from Wadi el-Hol (also in Egypt), these were published more than a decade ago (Darnell, Dobbs-Allsopp, Lundberg, McCarter, Zuckerman, 2005) in a volume published by the American Schools of Oriental Research for which I served as an external reviewer. Now to some of the salient details about Petrovich’s proposal and some of the empirical problems with it (there are a rather large number of problems, but I’ll focus on some of the most salient ones).

I. Douglas Petrovich argued (in his presentation at ASOR) that root-words such as rb (“great,” “big”), ‘l (“God”), yyn (“wine”) are present in the Early Alphabetic inscriptions and that these are Hebrew words. These words are attested in Hebrew, but that is only part of the story. Namely, these words are actually not just Hebrew but rather they appear in many Semitic languages. For example, this root for “big,” “great,” occurs not just in Hebrew, but also in Phoenician, Ugaritic, Aramaic, Akkadian, Old South Arabic (among others). Similarly, the word ‘El for God is also Common Semitic and it occurs in Ugaritic, Phoenician, Akkadian, Aramaic, Old South Arabic (along with some others). Or again, the word for “wine” (sometimes spelled with two yods and sometimes with one) occurs not just in the Semitic languages (such as Ugaritic, Phoenician, Arabic), but even in Indo-European languages such as Hittite! (and it comes into Greek and Latin, and ultimately even into the Romance Languages and even English). In short, there is no word that can be considered distinctively Hebrew in the Early Alphabetic Inscriptions from Serabit el-Khadem, Wadi el-Hol, or any other Early Alphabetic Inscription. Rather, for more than a century, the best that we can say is that these Early Alphabetic inscriptions are written in Canaanite, that is, the language of the Ancient Canaanites (who lived in ancient Canaan and often traveled down into Egypt, a tradition also reflected in Genesis). Thus, to say that these inscriptions from Serabit el-Khadem and Wadi el-Hol are somehow distinctively Hebrew is just not going to work. After all, if a word occurs in lots of different languages, that word cannot be considered diagnostic for one language. For that to be the case, that word would have to be present ONLY in one language. And that’s just not what we have with these inscriptions.

II. We have hundreds of Hebrew inscriptions and the earliest of these dates to ca. 900 BCE (see my article in Biblical Archaeology Review on “The Oldest Hebrew Inscription” for some of the most relevant inscriptions) and the alphabet of these Hebrew inscriptions has twenty-two consonantal letters. The Early Alphabetic Inscriptions (e.g., from Serabit el-Khadem and Wadi el-Hol), however, have twenty-seven consonantal letters. Thus, to claim that these inscriptions from Serabit el-Khadem and Wadi el-Hol are Hebrew is an argument that is strained well beyond the breaking point.

III. The Early Alphabetic Inscriptions are written from left-to-right (dextrograde), right-to-left (sinistrograde), and columnar. The Old Hebrew inscriptions are uniformly written from right-to-left (that is, sinistrograde). Thus, even with regard to the direction of writing, the attested inscriptions that are definitively Old Hebrew do not correspond with the diversity of directions that is present in the Early Alphabetic inscriptions. Again, the problem with the claim that these from Serabit and el-Hol are Hebrew is that the evidence doesn’t line up very well. And only the most strained of arguments can make it “line up.”

IV. In one of the Egyptian inscriptions from Serabit el-Khadem (i.e., many inscriptions are present at this site, most of them are written in Egyptian, not Northwest Semitic…and as for one of them written in Egyptian…), Petrovich attempts to read the word ‘br, a word that is the root for the word “to cross over,” and when used as an ethnicon is the basis for the word “Hebrew.” The first thing that I would say is that this is a very, very strained reading of the Egyptian inscription. I am confident that few trained Egyptologists will embrace Petrovich’s reading of this text as having the root ‘br. Furthermore, even if this root were to be present, the fact of the matter is that even that root word (ayin, bet, resh) is attested in lots of different Semitic languages, including Phoenician, Aramaic, Arabic, Old South Arabic, Akkadian, and Ugaritic. And in these cases, the word can and normally does mean “to cross over,” “to overstep.” In short, one must be careful to understand and remember that if a word occurs in multiple Semitic languages, it cannot be considered diagnostic or reflective of just one of those languages…and that holds true for all words, including the root ‘br.

V. In short, the things that Douglas Petrovich considers to be markers of Hebrew are, in fact, just markers of the Semitic languages in general. We even have a term of these sorts of words that occur in multiple Semitic languages. We call them “Common Semitic,” because they are attested in so many languages. In short, the only thing that can reasonably be said about the Early Alphabetic inscriptions from Serabit el-Khadem and Wadi el-Hol (etc.) is that they are written in a Northwest Semitic language and script. There is nothing in these inscriptions that is diagnostic for Hebrew. It would be interesting if there were features that could be considered distinctively Hebrew, but there are not. So, as has been the case for a very long time, we refer to these inscriptions as Canaanite or Early Alphabetic. And from this script called Canaanite or Early Alphabetic, the Phoenician script will later develop during the final decades of the 2nd millennium (and that’s an interesting story in and of itself). And from the Phoenician script, the Hebrew and Aramaic scripts will come in the early First Millennium (and that’s another interesting story). But that’s a long ways down the line in terms of time…long after the scribes of Serabit el-Khadem and Wadi el-Hol penned their inscriptions. Thus, the Early Alphabetic inscriptions from Serabit el-Khadem and Wadi el-Hol are definitely not Hebrew.

Respectfully submitted,

Dr. Christopher Rollston (Ph.D. Johns Hopkins University)
George Washington University
Department of Classical and Near Eastern Languages and Civilizations
Washington, DC

The Jerusalem Papyrus: Complementary Notations

1 November 2016

The Jerusalem Papyrus: Complementary Notations

by Dr. Christopher Rollston, George Washington University

Here are the readings of the editio princeps: [‘]mt. hmlk. mn‘rth. nblym. yyn. Yršlmh.
Translation of the readings of the edition princeps: [Maidse]rvant of the King, from N‘rth, (two) jars of wine to Jerusalem.

As the papyrus continues to garner attention, it is perhaps useful to convey certain additional reflections, especially with the release of the editio princeps (i.e., I have now received a copy of it and so can augment some of my reflections). I have also had some time to spend with a very fine high resolution image of the inscription.

Moreover, I should like to emphasize that I would be pleased for this papyrus inscription to be ancient, but the fact remains that forgeries have been produced in steady numbers during the past century (and much earlier), in a number of languages, including Greek, Latin, Hebrew, Aramaic, and Phoenician (for a synopsis of much of this data, see my “Forging History: From Antiquity to the Modern Period,” in _Archaeologies of Text_, eds. Morag Kersel and Matthew Rutz [2014], available on Therefore, vigilance should be the modus operandi of the entire field, as the great Joseph Naveh of Hebrew University often reminded us (e.g., his BASOR 247 [1982] article “Some Recently Forged Inscriptions,” his JNES 27 [1968] article “Aramaic Dubiosa,” and his IEJ 48 [1998] article with Israel Eph’al on the Moussaieff Ostraca) and as the great Frank Moore Cross reminded us (with his Orientalia 37 [1968] article on the forged Phoenician Inscription from Brazil and his IEJ 53 [2003] article on the Jehoash Inscription). Of course, the recent Gospel of Jesus Wife Papyrus forgery should also be instructive for all in the field.

On a different note, it should also be emphasized that we have many Old Hebrew inscriptions from First Temple Period Jerusalem, some of them particularly famous and especially important, such as the Siloam Tunnel Inscription, the Royal Steward Inscription (with N. Avigad’s masterful decipherment), the Ophel Inscription (see the list and full texts of inscriptions from Jerusalem in the volume entitled _Hebrew Inscriptions_ edited by F.W. Dobbs-Allsopp, J.J.M. Roberts, C.L. Seow, and R.E. Whitaker [Yale, 2005]. Furthermore, we have Mesopotamian inscriptions, such as the famed Sennacherib Prism, with its reference to Jerusalem and its King Hezekiah (r. ca. 715-687) during the 701 BCE siege of Judah. In addition, the city of Jerusalem is mentioned in a nice Old Hebrew inscription in stone from Khirbit Beit Lei (as Adam Bean and Nathaniel Greene have emphasized) that dates to the late First Temple Period. In other words, from inscriptions we know a lot about and from Jerusalem.


(1) Carbon dates normally, of course, have a plus and minus range. In the media discussions of the Jerusalem Papyrus, there have been many references to an assumed secure 7th century carbon date for the papyrus….but this is too narrow…as it is without a stated plus and minus range. Important, therefore, is the fact that in the editio princeps of this inscription, a larger range is given (footnote 3, page 248), with a calibrated 2-sigma date being (756- 679 BC [34%], and 671-430 BC [61.4%]). Of course, this is indeed fairly important because it shows that the C 14 date for the papyrus is certainly not necessarily a 7th century date. That doesn’t necessarily change things all that much, but it is important, for a number of (fairly obvious, I think) reasons. Furthermore, Israel Finkelstein, Eric Cline, Benjamin Porter, and Felix Höflmayer have all drawn my attention to the fact that the “Hallstatt Plateau” is arguably to be factored in here…namely, according to the “Hallstatt Plateau,” radiocarbon dates of ca. 2450 BP (Before Present = 1950) consistently calibrate to ca. 800-400 BC, regardless of the measurement precision. The editio princeps of the Jerusalem Papyrus gives 2460 BP +/- 71. In short, the plateau is operative. Thus, it is quite problematic to suggest that the carbon date for the Jerusalem Papyrus is, and can only be, 7th century BCE. The plus or minus must be admitted to be considerably wider for the carbon date, arguably a couple hundred years.

(2a). The Restoration. A striking thing is that the authors of the editio princeps restore an ‘alep to get the word ‘mt (construct form of ‘mh, that is, “maidservant”). The assumption is that it was an ‘alep that was at the end of the preceding line (a few traces of that line are visible, so there was a preceding line, but *no* traces of an ‘alep are present at the end of that preceding line). The authors of the editio princeps do consider other options (noted on p. 241 of the editio princeps), but they settle on [‘]mt, that is, (the construct form of) “maidservant.” However, I would wish to emphasize that restorations are notoriously difficult things. Thus, without an initial ‘alep actually present, the restoration cannot be considered at all certain. Thus, and this is very important, this inscription might not be about “the maidservant of the King” at all, as the word maidservant is based on a restoration.

(2b) The Restoration Continued. In addition to positing a different letter or two that would result in a different noun (i.e., not “maidservant,” but some other nominal), it would also be possible to posit a verbal form (including even the root mwt, “die,” “dead”….which could make this inscription really interesting… although various other possibilities could also be posited). Again, though, that’s the problem with restorations….they are very uncertain, unless we’re dealing with highly formulaic language (e.g., a legal contract). And even then it’s not always easy. In this case, we do not have a highly formulaic text, and we have very little context, so I consider any particular restoration (whether it be ‘mh or something else) to be only a possibility.

(3a) Orthography and Lexicography. There are a couple of potential problems with the orthography in the Jerusalem Papyrus, the first one particularly concerning. Namely, regarding the words of the Jerusalem Papyrus nblym (with yod as an internal mater lections) and yyn, that is, “jars of wine” or “two jars of wine” (if the term for jar is understood as a dual, which is the way that the authors of the editio princeps wish to understand it), it must be conceded that the orthography of the inscription is problematic. After all, within Semitic languages, this construction (this juxtaposition of two nouns or nominals) is called a “construct chain” (i.e., x of y). And in Semitic construct chains (including Hebrew, Phoenician, Aramaic, Akkadian, Ugaritic, Ethiopic, etc.), the noun in construct takes the “construct form.” Sometimes there is no discernible difference (at least in the writing system), but when a masculine plural or masculine dual noun is in the construct form, the plural ending îm and the dual ending ayim (see Jouon paragraph 92 g on this) are replaced by the vowel ê (and, most importantly for our purposes) without the mem. That is, the mem is eliminated. Gone. Note, along these lines, that the very same expression is used in the Hebrew Bible, that is, “vessels of wine.” BUT in Standard Biblical Hebrew it is written: nbly.yyn (1 Samuel 25:18; for additional examples of the plural construct form of nbl “jar” in the Bible, see Job 38:37; Lam 4:2). That is, as is standard in Ancient Hebrew, no mem is present in the Biblical form (the yod in the biblical form is a mater lectionis). This is the way it worked in Ancient Hebrew. In essence, therefore, there is a serious orthographic problem (i.e., mistake) in the Jerusalem Papyrus inscription, and a rather large one at that: the word nblym is in construct and so there should be no mem (note that Ezekiel 47:4 and Numbers 9:20 are arguably corrupt; I am grateful to Jason Bembry for mentioning these texts to me) and arguably are reflective of a deep structure that is different from the context in the Jerusalem Papyrus. Of course, someone might posit that in the inscription we have a “frozen form” of some sort, but this would be a very strained argument to make, since we have the very same expression in Standard Biblical Hebrew and the mem is lost…just as it is supposed to be (across the Semitic languages). Someone might propose that although this syntactically like it should be a construct chain, it is not. But that is a really hard argument to make as this construction is basically a textbook case for a construct chain (and that’s the way Biblical Hebrew has it as well). Also, I suppose that someone might attempt to propose some sort of strained argument to account for the mistake in the Jerusalem Papyrus’s orthography, but I’m disinclined to embrace special pleading all that readily, especially in an inscription on papyrus (and thus ostensibly a first-class scribe). We do have some mistakes in inscriptions at times, but this would be a big one for an ancient scribe to make, and one that would be hard to explain for a *native speaker of Ancient Hebrew and a first-tier scribe (for as Professor Joseph Naveh has noted, papyrus was the medium that the most highly educated and gifted of official scribes would use). Conversely, this is the sort of mistake that someone might make if they didn’t know Ancient Hebrew all that well (note that Modern Hebrew often avoids construct forms…with the use of “shel” being a common substition in Modern Hebrew for something that would have been a construct form in Ancient Hebrew). In short, I consider this to be a howler of an error.

(3b) Orthography and Lexicography. It could be contended that the proposed reading (of the editio princeps of the Jerusalem Papyrus) of the place-name N‘rth is problematic. Note that the form in the inscription (i.e., the spelling in the inscription) is the same as that in Joshua 16:7. At first blush, this might seem good. But this may not be the case, once we look more closely at the context of Joshua 16:7. In the Joshua reference, this place-name has a directive heh on it (note the verb yrd “to go down” is present in the Joshua text, so a directive heh, with its semantic range of “direction toward” makes perfect sense). Note also that the standard dictionary of Biblical Hebrew (HALOT), as well as BDB, presuppose that the lexical form of the word in Joshua is n‘rh (so at the very least, the authors of the editio princeps differ with HALOT and BDB [p. 655] on the lexical form of the place name). Thus, the form in Joshua is a result of the presence of the directive heh (thus, again, BDB page 655). Significantly, the form in the Jerusalem Papyrus is preceded by the preposition min (with assimilation of the nun) and so the directive heh would not normally be used in Ancient Hebrew (as the directive heh signifies “direction toward,” not “direction from” and this is something the Rabbis noted long ago, when they reasoned that the directive heh stood in place of a prefixed lamed “to”…so Waltke and O’Connor, paragraph 2.1b, et passim). In short, in Ancient Hebrew, this place-name is N‘rh, and when the directive heh is attached, the form we have (as demonstrated by Joshua 16:7) is N‘rth. Thus, the form that should arguably have been in the Jerusalem Papyrus Inscription is N‘rh, not N‘rth (the “double feminization” line of thought is pretty strained). However, there are exceptions, with a small number of cases with both the preposition min and the directive heh (e.g., Jer 27:16 and Josh 15:10).

(3c) Orthography and Lexicography. In this connection, I should like to mention that, as both Jason Bembry and Joel Baden have emphasized to me (personal communication) that it is certainly possible to understand n’rth as meaning “his young (female) attendant” (i.e., n’rh [nfs meaning “female attendant,” “female maidservant,” etc.] with the 3ms ms possessive suffix….of course, it could also be a 3fs possessive suffix, but in this context, 3ms is more likely). In any case, if one accepts this lexcial understanding (i.e., a common noun, rather than a place-name), then the meaning of the text is certainly different from the way that the authors of the editio princeps understand it, and this understanding (i.e., the common noun) has the elegance of being a fairly common word and with orthography that is standard Ancient Hebrew.

(4) There are also some problems with the script…perhaps in part related to the morphology of some letters, but especially in the realm of the ductus of some letters (the term ductus refers to the number of strokes that form a letter, the direction of those strokes and the order of those strokes). I’ve spent a fair amount of my life looking through a microscope or jeweler’s loop at Old Hebrew ink-inscriptions from excavations. And I have some ductus concerns (forgers can often get the morphology of a letter pretty accurately, but they struggle more with ductus….because palaeographers don’t discuss it very much. As for ductus of some forgeries, I discussed this sort of thing in detail in Maarav back in 2003, with regard to the Moussaieff Ostraca and Jehoash Inscription, etc.). I will be discussing this in the journal article that is currently in preparation and will be in print during the spring of 2017.

Updated 2:30 eastern (Nov 2).


Christopher Rollston