The Qeiyafa Ostracon Again: A Sober Assessment in Light of the New Finds

By Christopher A. Rollston
Toyozo Nakarai Professor of Old Testament and Semitic Languages, Emmanuel Christian Seminary

For those working in the field(s) of ancient history, ancient literature, archaeology, or epigraphy there often seems to be a strong desire to associate some new archaeological find, or some recent epigraphic discovery, with some person or event known from literary texts discussing the days of yore.  This basic phenomenon has a long history with regard to literary texts.  For example, within the Hebrew Bible, the book of Lamentations is anonymous, but through the centuries many contended that it was written by the Prophet Jeremiah.  Similarly, the book of Ruth is anonymous, but through the centuries many argued that it was written by Samuel.  Or again, within the Greek New Testament, the book of Hebrews is anonymous, but many attempted to argue that it was written by Paul.  Similarly, the four Canonical Gospels are anonymous, but through the centuries, many have argued that these books were written by known figures of Early Christianity.  Fortunately, critical scholarship has pushed back against such positivistic assumptions and reasserted the obvious: the evidence for these assumptions is not convincing, but specious.

With regard to the field of epigraphy, there are similar attempts.  For example, W. Shea argued that the Izbet Sarteh Ostracon mentions Hophni, the son of Eli the Priest and that this ostracon gives an account of the movements of the Ark of the Covenant from the Temple of Dagon in Ashdod to Kiriath-Jearim (Shea 1990).  Significantly, the readings of F. M.  Cross  (among others) differ markedly with those of Shea, with Cross not even reading the personal name Hophni in this ostracon (Cross 1980).  Therefore, it comes as no surprise that L. Mykytiuk has argued that Shea’s proposal is simply not convincing at all (Mykytiuk 2004).  Along those same lines, M.C.A. Korpel has argued that a seal with only the letters yzbl preserved should be considered that of the 9th century Queen Jezebel of the Israel (Korpel 2006).  But the evidence cannot carry the weight with which Korpel has saddled it, since this root (zbl) is well attested as a Northwest Semitic root, there is no patronymic, the word “queen,” is not present, Korpel’s restoration (’alep) is not at all certain, and there is a dearth of evidence for epigraphic stamp seals in the 9th century BCE (Rollston 2009).

To be sure, sometimes the literary and historical data do converge nicely with the archaeological and epigraphic data.  For example, it is entirely convincing to argue that the “Mesha King of Moab” mentioned in the Hebrew Bible (e.g., 2 Kgs 3:4-5) is the same as the person who commissioned the Moabite Stone (Dearman 1989; Rollston 2006, 126).  Similarly, the “Bar Kokhba” (Dio Cassius; Eusebius) of the Second Jewish Revolt (132-135 CE), the “Bar Koziba” of the Misnha and Talmud (Yadin 1971, 255-259), and the “Simeon Bar Kosiba” (Yadin, et al., 2002) are one and the same (Rollston 2006).  Here is the main point, however: arguments for the association of epigraphic and archaeological data with events and people attested in literary and historical texts must be based on very good evidence.  It is one thing to suggest that there “may be” a connection, but it is much different to suggest that a connection is “probable” or “certain.”  Scholars must be careful to analyze the evidence in a disinterested manner, so that the possible connections are not overstated.  This leads naturally into a discussion of Qeiyafa.

The archaeological site of Qeiyafa is really very impressive, with some significant monumental remains.  Furthermore, the cultic artifacts which have recently been announced are certainly of some consequence as well (cf.  At this juncture (coming in the wake of these recent finds), I wish to make some comments on the ostracon itself, reiterating some critically important data-points (for more details, see Rollston 2011; Rollston 2012).

(1) The script of this ostracon is definitely not Old Hebrew.  Rather, it is a late stage of the Early Linear Alphabetic Script.  The Old Hebrew alphabet is a descendant of the Phoenician script, not the Early Alphabetic Script.  In short, this ostracon is not written in the Old Hebrew script and this is indubitable.

(2) The roots attested in this ostracon (i.e., those that can be read) cannot be considered distinctively Old Hebrew.   For example, the root mlk (king) arguably occurs in line 4. This is certainly a Hebrew root, but this root also occurs in numerous Semitic languages, including Ugaritic, Amorite, Phoenician, Punic, Moabite, Aramaic, Ammonite, Edomite, and even Palmyrene. Because this root occurs in so many ancient Semitic languages, it should be classified as Common Semitic. It definitely cannot be considered as diagnostic for Hebrew.   Similarly, as for the root `bd (serve, servant) in this ostracon (e.g., line 1), it cannot be considered distinctively Old Hebrew either.  After all, while it certainly occurs in Old Hebrew, it also occurs in numerous Semitic languages, including Ugaritic, Phoenician, Aramaic, Nabatean, Palmyrene, Ethiopic, and Classical and even Modern Arabic.  The same thing can be said about the root špţ (judge), and nqm (avenge).  Indeed, the same can be said for every root in this ostracon.  To be sure, a critically important root for the original editors of the text (Misgav, Garfinkel, and Ganor 2009) is `sh (“to do”) in line 1. This root is their primary basis for the contention that the Qeiyafa Ostracon is Hebrew.   Nevertheless, the fact of the matter is that this root is attested in the Moabite language (e.g., in the famous Mesha Inscription, on lines 3, 9, 23, 24, 26; Dearman 1989) and the great Semitist Z. Harris believed that the root ‘åh (śh) might very well be present in Phoenician as well (Harris 1936, 136).  Obviously, therefore, this root is not distinctively Old Hebrew.  In short, the language of this inscription is not something that can be determined with certitude.  The question must be left open.

(3) There does seem to be judicial terminology in the Qeiyafa Ostracon, but judicial terminology is attested throughout much of the ancient Near East (Weinfeld 2000).

(4) It should also be emphasized that within this ostracon there is no reference to a particular city or nation state and this absence must be factored into any putative suggestion about the significance and Sitm im Leben of this ostracon.

(5)  I am very disinclined to accept Puech’s problematic speculations that the Qeiyafa Ostracon should be associated with the coronation of a particular king, be it Saul or David.  Obviously, for this to be considered a cogent understanding of this ostracon there would need to be some reference in the ostracon to Saul or David! (Puech 2010)

(6) Finally, I should also like to emphasize that the decisive manner in which the site of Qeiyafa has been associated with a particular king or a particular “kingdom” (e.g., David) is pressing the data much harder than I would.  Or, to put it another way, even if we could contend that this site was Judean or Israelite, could we definitively state that it is to be associated with a particular king of one of these states?  I would suggest that without decisive epigraphic evidence, the answer must be no.  Rather, we must be content to refer to some possibilities, and to leave it at that.  Archaeologists will continue to debate and discuss this issue (e.g., Finkelstein and Fantalkin 2012 and the bibliography there), but it is caution regarding all such conclusions that I would urge.

In the final analysis, it must be said that there is a sincere human desire on the part of scholars to fill the gaps in our data, to fill in the lacunae.  That is honest and it is sincere.  Nevertheless, it is also imperative that we attempt to be sober, disinterested scholars, restricting our conclusions to the data at hand.  Thus, as for the site of Qeiyafa and its ostracon, I would suggest that both are important and discussions will certainly continue.  This is good, but caution about conclusions must be our modus operandi.  That is, it is imperative that a concerted effort be made to avoid going further than the data would allow.

Select Bibliography

Cross, F. M.
1980       Newly Found Inscriptions in Old Canaanite and Early Phoenician Scripts.  BASOR 238: 1-20.

Dearman, A. (ed.)
1989       Studies in the Mesha Inscription and Moab.  Atlanta: Scholars Press.

Finkelstein, I., and Fantalkin, A.
2012       Khirbet Qeiyafa: An Unsensational Archaeological and Historical Interpretation.  Tel Aviv 39: 38-63.

Harris, Z. S.
1936       A Grammar of the Phoenician Language.  AOS 8.  New Haven: American Oriental Society.

Misgav, H.; Garfinkel, Y.; and Ganor, S.
2009.  The Ostracon. Pp 243-257 in Khirbet Qeiyafa Vol. 1. Excavation Report 2007-2008, ed. Y. Garfinkel and S. Ganor. Jerusalem: Israel Exploration Society.

Korpel, M.C.A.
2006       Seals of Jezebel and Other Women in Authority.  Journal for Semitics 15: 349-371.

Mykytiuk, L. J.
2004       Identifying Biblical Persons in Northwest Semitic Inscriptions of 1200-539 B.C.E. SBL Academia Biblica 12.  Atlanta: Society of Biblical Literature.

Puech, É.
2010.  L’Ostracon de Khirbet Qeyafa et les Débuts de la Royauté en Israël. RB 117.2: 162-184.

Rollston, C.A.
2006       Inscribed Ossuaries: Personal Names, Statistics, and Laboratory Tests.  NEA 69:125-129.

2009       Prosopography and the Yzbl Seal.  IEJ 59: 86-91.

2011       The Khirbet Qeiyafa Ostracon: Methodological Musings and Caveats.  Tel Aviv 38: 67-82.

2012       What’s the Oldest Hebrew Inscription?  BARev 38: 32-40, 66-68.

Shea, W. H.
1990       The ‘Izbet Sartah Ostracon.  AUSS 28: 59-86.

Weinfeld, M.
2000  Social Justice in Ancient Israel and in the Ancient Near East. 2nd edition. Jerusalem: Hebrew University Magnes Press.

Yadin, Y.
1971       Bar Kokhba: the Rediscovery of the Legendary Her of the Second Jewish Revolt against Rome.  New York: Random House.

Yadin, Y., et al.
2002       The Documents from the Bar Kokhba Period in the Cave of Letters (Hebrew, Aramaic, and Nabatean Documents).  Jerusalem: Israel Exploration Society.


All content provided on this blog is for informational purposes only. The American Schools of Oriental Research (ASOR) makes no representations as to the accuracy or completeness of any information on this blog or found by following any link on this blog. ASOR will not be liable for any errors or omissions in this information. ASOR will not be liable for any losses, injuries, or damages from the display or use of this information. The opinions expressed by Bloggers and those providing comments are theirs alone, and do not reflect the opinions of ASOR or any employee thereof.

13 thoughts on “The Qeiyafa Ostracon Again: A Sober Assessment in Light of the New Finds

  1. Thank you for the article, Dr. Rollston. I am anxious to be with you once again this fall.

    I cannot speak as a scholar in any related field, but I can wonder aloud nonetheless. I wonder, then, if part of the desire in these fields [each of which I have an affinity for] to relate a discovery with a personal name is the need for validation. I.e., of what significance is ‘my’ discovery? Is it significant if not tied to an identifiable (tangible) person(s) or people? I can only imagine the personal sense of fulfillment in being able to draw a legitimate connection.

    But how much of these name associations, would you suggest, are made based on that need for validation-or perhaps recognition is the better word.

  2. Just a couple points of contention: A. In your pt. 2 you say: ‘For example, the root mlk (king) arguably occurs in line 4. This is certainly a Hebrew root, but this root also occurs in numerous Semitic languages, including Ugaritic, Amorite, Phoenician, Punic, Moabite, Aramaic, Ammonite, Edomite, and even Palmyrene. Because this root occurs in so many ancient Semitic languages, it should be classified as Common Semitic. It definitely cannot be considered as diagnostic for Hebrew.’

    Then, when does a Hebrew root mean a Hebrew root? That is like saying the word ‘ballet’ is not French because it appears in English.

    B. You say: ‘Obviously, therefore, this root is not distinctively Old Hebrew. In short, the language of this inscription is not something that can be determined with certitude. The question must be left open.’

    Wouldn’t the context and location of the find help determine its language? Since the Elah fortress is far from Phoenicia and possibly Moab, isn’t it a big stretch to say that root is not Hebrew but Moabite/Phoenician considering the fact that nothing Moabite or Phoenician have been found at the site? (I have looked at their website )

  3. Pingback: A word of caution about Khirbet Qeiyafa « With Meagre Powers

  4. Dr. Tee,

    Respectfully, your example (A) more properly describes a “loan word”—of which English has many. Dr. Rollston’s point, if I unsderstand him properly, is that while mlk *is* in fact a Hebrew root (i.e. it occurs in Hebrew), it does not necessarily make it diagnostic.

    It would be more akin to finding the Latin root pns* (to think) and arguing that it must be French, though the root also occurs in Spanish et al.

  5. Yet that doesn’t answer my question as I knew that already. I want to know when a Hebrew root actually means a Hebrew root and not on of the other semitic languages.

    if we toss every word out because other languages have the same type of root then where do we get the evidence for Hebrew words? My B question sought to limit the multiple sources and zero in on what can be used to determine when we have ancient Hebrew writing.

    I would hate to think that bias against 10th century literacy is playing a role in deciding when a Hebrew root is actually a Hebrew root.

  6. David,

    It would be helpful for you to look at some of the standard works on Comparative Semitic, as these sorts of subjects and issues have been dealt with for a long time, in one form or another. In terms of my recommendations to help you along, I would suggest that you could begin with volumes by Bergstrasser, Bauer and Leander, Moscati, Kaltner and McKenzie (_Beyond Babel_), and even something such as Saenz-Badillos (a nice volume on the history of Hebrew, but with a useful introductory section on Comparative Semitic). Garr’s monograph would be very good for you to read as basic background as well. Of course, Huehnergard has a number of articles on subjects relating to Comparative Semitic, classification of the Semitic languages, etc., which should help you as well. In any case, the basic factor that is in play is the fact that Hebrew is a Semitic language, part of the Northwest Semitic family in particular. For this reason, lexemes are rarely good anchors for making precise determinations about language or dialect…because the same roots are often part of the shared ancestry of language groups. I would encourage you to peruse a good lexicon such as Koehler-Baumgartner and see all of the cognate data for the lexical entries. This sort of thing is often quite revelatory for those beginning to reflect on issues related to Comparative Semitics, linguistic methodologies, etc. As for your query about provenance, that is an important issue, but you need to realize that there are multiple issues in play here as well. For example, with regard to this site, archaologists are still attempting to determine the presumed national or ethnic identity with which this site is to be associated (in addition to the publications by Y. Garfinkel and his teaam, see the recent publications by N. Na’aman, I. Finkelstein and A. Fantalkin, etc.). That is, you assume that it is Judean, but even this is not absolutely certain (note that it is very close to the presumed border with Philistia. Moreover, and along those lines, it is important to remember that sometimes scribes of one region or nation state will write in the prestige language of the region, rather than in their own native dialect. I cite various examples of this in my Tel Aviv article, and Naveh provides examples of this sort of thing as well. Furthermore, inscriptions travel and so do scribes. Also, how much do we truly know about the nation states of this region during the 11th century BCE? How much do we know about their distinctiveness? (some things, I suppose, but there is much that we don’t know). Also, how much do we truly know about the NWS Levantine languages at *this* period? Can we even speak of Hebrew at this time and if so, what is our evidence? Can we really rely totally on the archaic of the Hebrew Bible for these conclusions (e.g., Exod 15, Judg 5)?The answer is basically that we do not know a great deal. And ultimately, I prefer not to speculate. There are too many factors that reside in the gray area. As you may know, N. Na’aman prefers just to refer to the language of this inscription as Canaanite. That’s a rather safe position and I’m quite comfortable with that as the safest of all positions….in light of the fact that we do not, at this time, have very much data. In short, in light of the fact that we do not have diagnostic linguistic features that are distinctive Hebrew elements, then it is necessary to leave the matter of the language open. As for levels of literacy, I draw your attention to my BASOR article in 2006, or to a couple chapters in my recent monograph. Also, I would recommend the articles of Ian Young in VT, although written some time ago, they remain current, authoritative.


    Christopher Rollston

  7. Patrick and Jack.

    Yes, Patrick, I think that you are correct…that sort of thing is definitely part of the equation at times. Jack, thanks for the note and for raising that important issue. When I think of loan words in NWS, I think of the fine methodologies outlined and practiced in Mankowski’s volume on Akkadian loanwords in Biblical Hebrew. In any case, thanks for that note, Jack…very useful.

    All best wishes,

    Christopher Rollston

  8. Dear Dr Rollston,

    You conclude your intereting essay with the following words: “Thus, as for the site of Qeiyafa and its ostracon,… I would suggest that .. discussions will certainly continue… it is imperative that a concerted effort be made to avoid going further than the data would allow.”

    It is impossible to quarrel with this thoughtful suggestion.

    Here are some reflections about the site and then the ostracon.

    The fortifications around this community clearly point to a strong authority with resources to construct them, since no small local community could this on its own.

    Philistine presence is ruled out by absence of any trace there. The Moabites are too far away west to have played any role, as are are the Phoenicians in the north. Further, the absence of pig bones which characterizes a number of Judahite sites like Beth Shemesh would be a strong indicator of early food taboos of early Israelites. However, much more can and hopefully will be dug and brought to light there to further clarify the ancient situation.

    As for the ostracon - the varied readings of epigraphers make it clear that its meaning cannot be certain at this stage of scholarship.

    Let me repeat in slightly different words what was stated above: The argument that many words are attested in other cognate dialects or languages raises the question when does a language or dialect attain a uniqueness of its own? Can one not apply this approach to the entire lexical corpus of the Hebrew Bible and reach similar conclusion: this is nothing but Canaanite, or a NW semitic dialect?

    Finally, I believe the original publishers of the Ostracon found a few instances of the negative imperative such as “Al Taas”, which they considered a typical Hebrew form. I’m not sure if this is a position they still hold? Also, I do’t remember if you have read it the same way.

    Thanks again for your civilized and elightening contribution.

  9. Thank you Dr. Rollston for your reply but I was just looking for an answer to a question not sent off to the library. I happen to live in a country where those type of books are very difficult to find and it would take along time to get them shipped to me.

    Again, I already know most of the things you told me which weren’t part of the question. I do know that it was possible for the scribe an dinscription to travel but you are speculating if you assume that took place because you have no evidence for that route.

    Mr. Hurwitz has pointed out some of the facts we already know which provide a strong indication that the site was a Hebrew possession. With allthis in mind, it seems to me that those who disagree with the idea that it is a Hebrew inscription is because they don’t want it to be a 10th century Hebrew inscription.

    I come to this conclusion not because I favor it to be a Hebrew inscription but because the effort to discredit it and say otherwise is over the top. Sometimes I think we over-think these matters and judging by the evidence on hand the inscription should be accepted as Hebrew until such a time when contrary evidence is discovered.

    To say it isn’t Hebrew, given the current evidence, is speculation as there is no proof that the inscription was written by a member of those civilizations you mentioned which shared the root.

  10. Dear Uri,

    Thanks for the note and the good queries. Basically, with regard to the provenance of the inscription…someone might wish to say: “Although there are no diagnostic linguistic elements in the Qeiyafa ostracon itself, based on provenance, I believe it must be x or y.” As long as people are willing to come clean on that issue (and concede that there are no definitive linguistic markers), I’m fine with it. However, for methodological reasons, my position has been, and remains, that from the linguistic data in the ostracon itself one cannot make a case for a distinct dialect or language, per se. Thus, for me, the most cogent position methodologically is to say that we cannot know. I’m not ruling out Hebrew, or Phoenician, Canaanite, etc., I’m just saying that from the inscription itself, we don’t get the data we would need to make a clear determination.

    As you probably know, similar issues are in play for the Gezer Calendar…I detail some of this in my recent BARev article, including reference to Dennis Pardee’s forthcoming article (reference in his review of the Zayit volume, in the current issue of JNES) in which he argues that Gezer is indeed Phoenician in language (I have argued, as has Naveh before me, that the script of the Gezer Calendar is Phoenician, that is, it does not reflective diagnostic features for the Old Hebrew national script).

    As for /’l ta’as/, that is something I deal with in my recent _Tel Aviv_ article in more detail (and also in the BARev article as well), but just with a short-hand reference in this blog post…but the difficulty there is that /’sh/ is attested not just in Hebrew, but in other languages, most notably in Moabite, arguably in Phoenician, and probably in OSA. In short, the philological evidence reveals that this root is not a distinctive root in Hebrew, but is attested more widely in the Semitic language family (and thus a shared retention). And, of course, Puech does not even read the root ‘sh here…

    As for whether or not, we can ever make determinations about whether something is Hebrew, or Phoenician, Moabite, Ammonite, etc., yes, there are linguistic features that allow such determinations. I think the best discussion of the diagnostic differences between the Iron Age languages of the Levant is Randall Garr’s _Dialect Geography_. In short, yes, there are dianogstic features that can help us make such determinations….but, alas, none of these diagnostic features is present in this ostracon. And it is also the case that the difficulties of reading this ostracon have compounded the problem (e.g., note the different readings proposed by the authors of the editio princeps, Yardeni, Puech, etc., etc.). Naturally, once the Old Hebrew national script rises (during the 9th century, I believe, following Naveh), this certainly helps matters….but, in the case of this ostracon, we do not have the Old Hebrew script (because this inscription antedates the development of the Old Hebrew script, of course), but rather we have the Early Linear Alphabetic script. Thus, the script cannot help us with regard to making a determination about the language of the Qeiyafa Ostracon.

    With all best wishes,

    Christopher Rollston

  11. Is there a large enough corpus of Phoenician that statistical methods can be applied to estimate the likelihood that a text is in that language? For all I know, being merely a layman, there could be software in use already by those in your field.

    Having no particular axe to grind, I simply appreciate your well-reasoned argument and clarity of the information presented therein.

  12. Pingback: When Does A Hebrew Root Mean A Hebrew Word? | theologyarchaeology

  13. Pingback: A Calm, Cool, and Collected take on Qeiyafa « Thoughts in the Dark

Leave a Reply

Your email address will not be published. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>