Call me old-fashioned, but I like my seals belonging to biblical characters to mention the biblical character to whom they purportedly belonged. The quartz seal of present interest reads yzbl.1 Marjo Korpel divines the name ‘yzbl (Jezebel), however, which the seal does not in fact read.2 She then confidently takes that unnamed Jezebel to be the biblical Jezebel, whom our modern cultural vernacular has since memorialized as a pejorative for shamelessness. Korpel has effectively taken us back in time 44 years to revisit the seal’s publication by the late Nahman Avigad, who (as the expert nonpareil of West Semitic seals) considered the biblical Jezebel connection and wisely rejected it.3
I will let other biblical scholars dispute whether the preponderance of evidence requires us to restore this unprovenanced seal to the Jezebel of lore. As an archaeologist and palaeographer, I prefer to focus on the material culture’s bearing on the question of date. For me, the threshold of proof is simple (assuming the seal is not a modern forgery, which is always a possibility). If the seal dates as early as the 840s BCE, then there is a chance, however remote, that the queen of Israel owned this seal. If the seal dates any later, then it categorically has no literal connection to the biblical character. A corpse masticated by dogs has little need for a signet.
Read Janet Howe Gaines’s article “How Bad Was Jezebel?” online for free as it appeared in Bible Review.
Since Christopher Rollston has written a thoughtful, authoritative analysis of Korpel’s claims, which he finds unpersuasive, there is no need to duplicate his powerful critique of the purported Jezebel connection.4 Suffice it to say that I strongly concur with Rollston on the following observations. First, he makes a persuasive palaeographic case for an eighth-century date, which I will also address in due course. Second, he notes that the root zbl is common in West Semitic names from the second millennium onward. This weakens the statistical probability that the two names ‘yzbl and yzbl need refer to the same person. I might also note that either name (‘yzbl or yzbl) may be female or male, a fact not without statistical weight against the biblical identification. Third, he notes that Korpel’s reconstruction of the name is just that: a reconstruction of a broken seal. Is it possible that the fractured cavity originally bore additional letters? Yes, of course. Might there instead have appeared different letters or rather additional iconography? Yes, of course. For some, the cavity transforms the reader into a diviner whose vision is limited only by the imagination. For others, the cavity is a warning to proceed with caution and respect for ambiguity. Fourth, Rollston cites the absence of either a patronymic or title of any kind. The seal’s owner claims no royal father or husband. This silence neither proves nor disproves anything, of course, unless one considers that Korpel bases much of her argument for a royal owner on some rather bizarre, artistic eisegesis. “The flower at the bottom of the seal might be a rose or lotus,” Korpel writes, “pointing to a vain lady, which Jezebel was.”5 Five, Rollston recognizes that Korpel’s iconographic dating of the seal depends a priori on her supposition that the seal belonged to the ninth-century queen of Israel. Finally, Rollston, Amihai Mazar and others have accurately noted that there are no stratified Phoenician or Hebrew seals with incised names from any secure ninth-century archaeological context.
In other words, if this seal (with its inscribed personal name) dates to the ninth century BCE, it would be the very first one of its kind! It would represent a statistical outlier, moreover, further compromised by its lack of provenance. All provenanced seals from this period are anepigraphic, i.e., they may feature iconography but no Semitic letters and certainly no names. To bolster the case for the Phoenician princess, Korpel interprets the script as Phoenician, but she does not supply any palaeographic evidence. Making the case for a definitive Phoenician identification of the script becomes even more difficult when one considers that of all the published Phoenician seals with personal names from the Iron Age, no more than three or four are provenanced. Only one unequivocal exemplar from Khorsabad meets any rigorous stratigraphic standards and it dates, expectedly, no earlier than the eighth century. Extravagant ninth-century seals of the so-called Egypto-Phoenician variety more commonly resemble the specimen from Hurbat Rosh Zayit,6 which is not terribly far from Jezreel. Perhaps admirers of Jezebel might have a look around the vicinity for personal effects better representing the material culture of the relevant era.
Why did scribes not begin to incise personal names in stamp seals prior to the eighth century, especially in Israel? It is a difficult question to answer, but I believe very strongly that we must situate the gradually epigraphic features of the seal medium within the larger changes in the political economy suggested by the archaeological evidence. Of course, scribes did not begin writing in the eighth century; they merely began writing possessive personal names on seals in Semitic scripts. David Ussishkin and John Woodhead even discovered what appears to be a ninth-century ostracon at the site of Jezreel itself, so there is no question about an active scribal culture in ninth-century Israel.7 The important questions concern the media to which scribes applied their technological craft, when, why, and for whom.8 From the standpoint of material culture, the maturation of the petty states in the southern Levant exhibits some radical departures in political centralization, intensification of trade, international relations, urbanism, and social stratification. That is not to say that these socioeconomic elements did not already exist in some form, but the eighth-century patterns of inscribed stamp seals, market weight standards, entombment and other aspects of the archaeological record indicate more predominantly produced and distributed markers of status, wealth, and cross-cultural contact. The introduction of possessive inscriptions on personal seals is somehow entwined with the other changes we witness during the middle of the Iron Age II.9
Let us begin discussing the palaeography of the seal by noting the opinion of Avigad that the incision of the iconography preceded the incision of the Semitic letters. I agree with him on this point, although I am less confident of his view that “our seal was not manufactured with any intention of inserting an inscription.” I remain open-minded on this point. In either case, it is crucial to remember that the spatial relationship between the iconography and the letters had an important effect on the ductus of the letters. Ductus refers to the logistics of inscribing a character: the number of strokes, the order of strokes, stance, and direction. Not only can the compressed space of tight quarters adversely affect ductus, but so can the medium in which the letters are inscribed (wet clay, fired clay, agate, limestone, basalt, etc.). It may be correct to write a letter in a particular fashion in a particular century, but other factors may affect whether the scribe can correctly execute the letter. Anyone who has used different writing implements on different writing surfaces will understand this. The name on this seal begins with the letter yod, which is somewhat quirky palaeographically, but not quirky enough either to cause concern or secure a precise date. It is important to look at the ductus. The scribe incised the upper and middle horizontal strokes first, which we could confirm with a higher-resolution photograph and/or microscopy. While counterintuitive to those who write with a modern Roman alphabet, it is not uncommon for West Semitic scribes to write the horizontal stroke(s) before the vertical stroke(s). One sees this ductus fairly often with Old Hebrew letters like he (which often exhibits three horizontal bars written before the vertical stroke). I want to stress that, since no one has conducted a comprehensive statistical analysis of stroke order on all provenanced inscriptions, we cannot necessarily say that “horizontal stroke first” is the norm per se. It is common enough, however, and most palaeographers who have taken an interest in ductus will have noticed this — at least I hope so, because it is a fairly important detail!10 The minutiae of ductus tell us why a character appears as it does and, more significantly, it can sometimes explain why we occasionally see inconsistencies in an individual scribe’s handwriting (both ancient and modern). In many instances, inscribing or inking the horizontals first will obligate the scribe to orient the stance of the vertical shaft to match them. So while the stance of a character is very important, we also have to look at features or premature “mistakes” that tied the scribe’s hands when it came to the completion of the overall letter. In the case of our yod, the middle horizontal is a little lower than we might expect relative to the lowest (in this case, oblique) horizontal (the yod will often exhibit three horizontal strokes spaced more or less equidistantly). Is it palaeographically significant? Perhaps, but consider that the top two horizontal strokes were apparently executed first (the third is harder to determine without closer inspection; depending on when it was incised, it may have needed to rise so obliquely to the upper right to avoid the cobra’s neck) with the anticipation that the vertical would follow. Notice now how the scribe could not make the vertical spine long enough to space the horizontal strikes equidistantly; the vertical runs almost smack dab into the tail of the cobra (assuming the cobra was incised first, which while very probable, is not absolutely certain; the tail of the right cobra impression is shorter than that of its left partner, which could indicate either [a] the scribe anticipated that he needed to reserve space for the inevitable letters of the name, or [b] the tail is shorter because it avoids running into a squatter yod already there). Some yods of the eighth century feature a downward tick on the right end of the lowest horizontal, which is not the case here. The situation of the cobra’s head may have affected the ductus here, but the eighth century also gives us a mixed bag of “ticked” and “unticked” yods.
Most of our stratified eighth-century seals and seal impressions come from 701 BCE destruction layers, meaning they represent the very end of the century and allow for (as yet) unknown antecedent typological permutations in the preceding decades, any specimen (discovered or undiscovered) from which could exhibit such a yod (so too the zayin, as I’ll discuss in a moment); the ninth century is hardly the only era in which we might situate this particular yod or zayin. The Samaria ostraca are no exception (some graphemes with ticks, some without); nor can we exclude the possibility that this cache includes mixed epigraphs from the early to the mid-eighth century bound together forever beneath the Assyrian destruction material from 720 BCE. Much of what we know about eighth-century material culture in general is informed by destruction layers in Israel and Judah from the final third of the century, the consequence of which is a murky knowledge of late ninth- and early eighth-century typologies of ceramics as well as palaeography. Our knowledge of palaeographic typology, in other words, must be contextualized within our broader knowledge of material culture, which is often less clear than we should prefer. We simply do not have enough stratified examples of epigraphs from this relatively long era to determine what a ninth-century seal script might look like (if any existed).
One might choose to date the zayin earlier than the eighth century on the basis of the absence of a tick on the right end of the lower horizontal stroke. This tick is not an uncommon palaeographic characteristic of mid- to late eighth-century Hebrew zayins. This absence is not compelling evidence to place this zayin in the ninth century, however. In addition to the comments above, consider a few variables, each of which seriously compromises the certainty with which Korpel assigns this seal an early date. First, there are several late eighth-century zayins without the tick. The palaeographic exemplars are mixed. Second, the iconography of this seal is executed in such a way that it privileges the head of the bird at the very place where the scribe would have incised a tick were a tick even intended (which we cannot and must not assume). There is arguably an execution under duress of space relative to the iconography that compressed the available room in which the scribe could squeeze the zayin between the bird and the solar disc. The effect on the ductus is clear. The center vertical shaft (again incised last) transects the horizontals through the luxury of an available space denied the right end of the lower horizontal stroke vis-à-vis the avian icon below it. Third, seals tend to exhibit a more conservative palaeographic ductus than ink-brushed ostraca. Careful palaeographers distinguish between lapidary scripts (found on monumental inscriptions and most seals) and cursive scripts (found on ostraca, papyri and leather), which tend to be more progressive, precisely because the latter succumb to the rapid ad hoc changes of efficient handwriting or shorthand;11 lapidary texts tend to be inscribed with more archaic features. This commonly produces a result in which two texts (one lapidary and one cursive), which date to the same time period, may exhibit different scripts: one conservative and one more progressive. Stamp seals tend to preserve older palaeographic features that appear earlier than the cursive trends of their era. Thus it is possible for a seventh-century seal to bear graphic characteristics of lapidary script attested in the eighth century.12 Consider the relevance that this phenomenon might have on a seal script that bears affinities with ninth-century characters. It most certainly could be an eighth-century seal with conservative graphemes. This underscores the importance of using only provenanced, stratified inscriptions to inform typological dates. The unprovenanced seal, which Korpel would date to the ninth century (without any palaeographic analysis, by the way), hardly fits the bill.
The bet is consistent with exemplars we see in the eighth century. Chris Rollston places special emphasis on what we call the recumbent stance, i.e., it leans backward to the right. This is especially common in Hebrew epigraphs from the late eighth century. Even a casual observer can see that he identfies the stance correctly. I am less concerned with the stance as a diagnostic marker of date than Rollston is, however, on two grounds. First, the bet is clearly constructed as an integral component of the inner oval of the seal. In fact, the spine of the bet is literally also a segment of the oval, which affords two coterminus lines in one. This may be the result of spatial duress or it may be a stylistic choice, but the fact remains that the stance of the bet is clearly related to its position vis-à-vis the inner oval’s curvature. Second, the combination of the lower spine and tail creates a sharper angle than appears on eighth-century Hebrew bets from Kuntillet Ajrud and some of the Samaria ostraca. The Hebrew bets from that era tend to have curvy lower spines gently producing a rounded tail, but the curt ductus in this instance (note that the tail was incised independently of the spine, which is further evidence in my mind of the spine’s prioritized relationship with the inner oval) could have constrained the scribe’s ability to execute the curving tail. So let us give the scribe the benefit of the doubt on this one. I sympathize with Rollston’s interpretation of the recumbent stance (as well as his observation that the recumbency is not normative for Phoenician scripts), but I am a little less bothered by it palaeographically as a marker of date. So the bet is not necessarily a smoking gun for me personally, but there are indeed powder burns on other features of the seal, which do complicate the possibility of a ninth-century date.
The lamed is fairly angular, which would be more diagnostically meaningful in leatherhard ceramic or a medium affording ink strokes. In this particular instance, however, there is little about the appearance of the character for which the constraints of the quartz medium could not account. There is consequently a fairly long timeframe during which we might find this variety of lamed.
I have saved the alleged alep for last, since it is the most problematic ingredient of the Jezebel reconstruction. Here in this picture of the seal impression is the alep that Korpel needs to reconstruct the name Jezebel (‘yzbl) as it appears spelled in the Hebrew Bible. Of course, the problem lies principally in the fact that the alep is not there. The break in the seal makes it impossible to reconstruct any missing component, which need not include a letter, by the way, except by conjecture. That permits us also to imagine all manner of missing incisions (if any) in the proverbial dark matter; lacunae are very democratizing. The maxim “absence of evidence is not evidence of absence” was archaeological jargon long before the former Secretary of Defense quoted it to explain the elusiveness of certain items of interest in post-war Iraq. As an archaeologist keenly aware that what we unearth only scratches the surface of a broader material culture, which has not survived unscathed, I am hardly averse to this maxim (at least for elusive artifacts). But this maxim was intended to prescribe and honor the exercise of caution, not leaps of faith and cavalier speculation at odds with the empirical patterns that distinguish what is plausible from what is wishful thinking.
The name ‘yzbl means “Where is the Prince?” It is a name type common to the Semitic languages, which indicates concern for a deity’s absence.13 Like their West Semitic neighbors, and unlike medieval theologians, the Israelites did not have a word for omnipresence. When a deity like Yahweh, Kemosh, El, or Baal threatened to leave, abandoning his adepts to whatever misfortunes awaited them, he was evidently taken at his word. When those who swore fealty to a state god displeased him in some way, the deity might simply revoke any promises of patrimonial custodianship. There are many examples of this ancient Israelite belief in the Hebrew Bible, but one stands out as particularly noteworthy for the commemorative name in its epilogue. In 1 Samuel 4, Yahweh abandons the Israelites to the Philistine army ostensibly as punishment for the sins of Eli’s sons, Phinehas and Hophni. He specifically sacrifices the ark, his war palladium or “chariot,” which the Philistines capture as booty. Phinehas and Hophni perish. Teeth gnash. Since the ark seats the visible presence of the deity (kabod, which most English Bibles translate as “glory”), the Israelites attribute the loss of the ark to Yahweh’s departure from the battlefield. When Phinehas’ pregnant wife receives the news of the ark’s capture and her husband’s death, the shock sends her into labor. Exasperated over the deity’s withdrawal, she names her infant son Ichabod (‘ykbwd), that is, “Where is the Glory?” Some later Judeo-Christian theological traditions differ with the Israelites on whether the deity literally leaves, but we must remember that the ancient West Semitic standards of proof were relatively empirical on this point. When the kabod departed, so too went its proprietor. Where is the Glory? Where is the Prince? What we suspect to be absent reveals quite a bit about what we presume to have been present. In the case of one gray quartz seal, fractured at the apex, what do we suspect and what do we presume? And with which empirical evidence do we train a light on the darkness of that cavity in the seal? Indeed, where is the Queen?
1. The Jezebel seal is not provenanced. Let me state at the outset that BAR has every right to publish provocative stories about controversial items. Transparency of scholarship permits the public to glimpse how we go about our business, even if it means our missteps receive the kind of scrutiny many would prefer to avoid. I also agree with Christopher Rollston, Frank Cross, and other eminent epigraphers that professional journals may publish unprovenanced materials insofar as they identify them as unprovenanced and saliently factor the probability of authenticity into the interpretation. There is a significant difference between treating an unprovenanced inscription on its own palaeographic merits (which BAR has asked me to do here) and using an unprovenanced inscription as a benchmark to date provenanced inscriptions palaeographically (which I am not doing). I have little doubt that the Jezebel seal, while unprovenanced, is authentic, but that is simply my tentative opinion. I reserve the right to recant. A few of the seal’s curiosities (irrelevant to this article) might under greater scrutiny suggest suspect components.
2. M. C. A. Korpel, “Seals of Jezebel and Other Women in Authority,” Journal for Semitics 15 (2006) 349-71; idem, “Fit for a Queen: Jezebel’s Royal Seal,” BAR 34/2 (2008).
3. N. Avigad, “The Seal of Jezebel,” Israel Exploration Journal 14 (1964) 274-76. Avigad considered the spelling yzbl to be a biform of ’zbl, but scoffed at “any basis for identifying the owner of our seal with this famous lady.”
4. C. A. Rollston, “Problems with Proposing That the Seal of Yzbl Was Queen Jezebel’s,” forthcoming.
5. Korpel, “Seals of Jezebel,” p. 360.
6. Z. Gal, “A Phoenician Bronze Seal from Hurbat Rosh Zayit,” Journal of Near Eastern Studies 53 (1994) 27-31.
7. D. Ussishkin and J. Woodhead, “Excavations at Tel Jezreel 1994-1996: Third Preliminary Report,” Tel Aviv 24 (1997) 63-64. I might acknowledge also that the script of the Jezreel ostracon (while discovered in fill material) appears more archaic than that of the Jezebel seal.
8. See an expanded discussion of these questions in R. Byrne, “The Refuge of Scribalism in Iron I Palestine,” Bulletin of the American Schools of Oriental Research 345 (2007) 1-31.
9. I explore the political dimensions of the Iron II archaeological continuities and discontinuities in my forthcoming book Statecraft in Early Israel, Volume 1: An Archaeology of the Political Sciences (Eisenbrauns).
10. One may consult a most fastidious discussion of Semitic ductus in Ada Yardeni’s magnificent Textbook of Aramaic, Hebrew and Nabataean Documentary Texts from the Judaean Desert and Related Material (Jerusalem: Dinur Center, 2000).
11. Cf. the evolution of Hieratic and Demotic from Egyptian hieroglyphs or the gradually simplified cuneiform wedges evolving from pictographs and ideographs.
12. It is easy to overlook this important distinction; cf. A. G. Vaughn, “Palaeographic Dating of Judaean Seals and Its Significance for Biblical Research,” Bulletin of the American Schools of Oriental Research 313 (1999) 62, n. 11.
13. See the compelling study by J. S. Burnett, “The Question of Divine Absence in Israelite and West Semitic Religion,” Catholic Biblical Quarterly 67 (2005) 215-35.
Ryan Byrne is a Semitist, epigrapher, and co-director of the archaeological excavations at Tel Dan. He teaches at Rhodes College in Memphis.