Embedded in the literature of Muslims, Christians, and Jews are historicized narratives that purport to rationalize and contextualize the place of minority and sectarian groups in medieval Islamic society. Among these are those that, at first reading, tell the story of an intentional fictionalizing of history on the part of a minority group with the intent to deceive Muslim authorities and thereby gain advantage. A prototype for this narrative strategy is observed in the Book of Joshua, wherein the “pagan” Gibeonites employ a ruse to secure recognition and protection from the conquering monotheistic Israelites, who had been commanded by God to exterminate pagans. Three case studies (on the Sabians of Ḥarrān, Karaite Jews, and Khaybarī Jews) reveal that similar stories in medieval Islam are often the result of co-production, a phenomenon which constitutes a kind of cultural negotiation between the dominant culture and a sub-culture; between rulers and subject peoples, between Muslims and non-Muslims, and even between competing subaltern groups. Reshaped narratives about the caliph al-Ma‘mūn, the Prophet Muḥammad, or other key figures offered narrativized permission for the dominant Muslim religion and culture to tolerate the existence of groups whose theologies or practices challenged Muslim assumptions of collectivity, and correspondingly, might or might not be otherwise deemed unacceptable. These narratives also provided subalterns a kind of myth of origin for their place in Islamic society. What is at stake in these complex interweavings of memory, history, and literary construction are the rights and duties of the subordinate groups.
For both individuals and groups, the past often illuminates the present through narratives that support and make understandable contemporary notions of social order.1 For minority and sectarian groups, narratives of the past can be used to explain the presence of the group within a dominant society in the face of contradictory assumptions of collectivity inherent in the dominant culture. In doing so, a minority or sectarian group may internalize aspects of the dominant society’s sense of the past in order to harmonize the surrounding culture with its identity.2 Narratives of origin, in particular, can be important for the way minority groups express an understanding of their place in society. The result can be a complicated blending of minority self-definition and dominant-society norms in which adopted and adapted notions become narrativized through complex reflexive processes of co-production. As this paper seeks to demonstrate, stories constructed through co-production are layered, interwoven, and re-written by different authors representing a variety of dominant and subordinate perspectives. This reflexive process may be difficult to identify in the narratives, and its setting is further complicated by confrontational relationships among competing subordinate authorial groups within the dominant society.
Several stories of obscure historical origin, two of which are connected to the time of the caliph al-Ma’mūn (d. 833), exemplify the complexity of minority representations of the past in medieval Islamic society. The stories constitute historicized narrative contextualizations of three medieval non-Muslim groups: Sabians (that is, the “pagans” of Ḥarrān3), Karaite Jews, and Khaybarī Jews. In these narratives, the subordinate religious minority is portrayed intentionally misrepresenting itself in order to seek advantage in relationship to the majority and other minorities. In all three cases the misrepresentation involves the manipulation of known characteristics of the dominant religious and political culture of Islam. This type of ruse appears as a literary-historical topos but, as will be shown, can have effects beyond the literary and imaginative realms of historicized narrative to leave evidence in the historical record. In this paper, the functions of such stories and their continued value by both minority and dominant cultures will be analyzed.
An archetypal example of such a narrative is found in the Hebrew Bible. In Joshua 9, the Gibeonites, a Canaanite people, are portrayed employing a ruse to avoid war and destruction at the hands of the conquering Israelites. Knowing that the Israelites had been ordered by divine decree to destroy the Canaanites, the Gibeonites made great efforts to present themselves as foreigners who resided outside of the land of Canaan, saying, “We come from a distant land; we propose that you make a pact with us” (Josh. 9.6). They even possessed particular knowledge of the Israelite God and recent Israelite history. “For we heard report of Him; of all that He did in Egypt, and all that He did to the two Amorite kings on the other side of the Jordan” (Josh. 9.9–10). Joshua and the Israelites, unaware that the Gibeonites were inhabitants of Canaan, were eager to gain allies among other nations as they established themselves in the land. So the men of Israel accepted the Gibeonites’ willingness to become Israelite subjects (Josh. 9.8), and Joshua “established friendship with them” (Josh. 9.15). The key element of the narrative is that the Gibeonite strategy was based on using specific knowledge regarding the soon-to-be dominant Israelite enemy. They knew both that the God of the Israelites had decreed utter destruction (ḥerem) for the “people of the Land” (Deut. 7.1–2), and that oaths were considered sacred for Israelites and were to be scrupulously affirmed.4
Even when the ruse became known, the Gibeonites were saved from massacre on account of the oath of friendship. The Israelite chieftains declare that, “We swore to them by the Lord, the God of Israel; therefore we cannot touch them” (Josh. 9.19). “They shall live!” (Josh. 9.21). In this instance of conflict between two contradictory divine laws, the oath has priority over the ḥerem.5
Given the literary and historical problems associated with the authorship and composition of the Hebrew Bible, one finds no consensus among modern Bible scholars regarding the historicity and import of this tale. Opinions range from complete fabrication to reliable source, and beyond that, conclusions are highly qualified regarding who the specific players are in the narrative and when events occurred.6
From a literary-historical point of view, Joshua 9 is an etiological tale that explains an anomalous situation in the context of biblical assumptions of Israelite collectivity centuries after the time of the described events—that is, the presence of a non-Israelite idol-worshipping people in the midst of the Israelite nation. The tale harmonizes what otherwise would be a contradiction between God’s explicit command in Deuteronomy to destroy the Canaanites and the continued presence of a Canaanite group in the land after the Israelite conquest and settlement. The poles of this disparity represent the difference between historical realities of the Israelite conquest of the late second millennium B.C.E., which permitted the continued presence of “pagan” non-Israelites in the land, and divinely ordained, utopian geo-political objectives as recorded in the Bible perhaps as late as the eighth century B.C.E. The continued existence of Gibeonites in latter days challenged the authors of the Book of Joshua to speak to the inconsistency of memory and reality.7
Furthermore, after the deception was recognized the Gibeonites were reduced to subservient status in Israelite society, relegated to be “hewers of wood and drawers of water” (Josh. 9.21).8 Here the narrative offers a secondary etiology to explain the subordinate status of latter-day Gibeonites. The function of this secondary etiology demonstrates that the Israelites remained honor-bound to their oath, but there were consequences for the ruse. Both the anomaly of continued Gibeonite existence and the subordinate and marginalized place of Gibeonites in society are rationalized.
The socio-historical function of this biblical narrative and its particular features constitute a literary topos that is found elsewhere in traditional literature of the ancient and medieval Near East. The important features are: 1) that the Gibeonites, a subordinate group at the time of the writing of this story, had possessed and exploited particular knowledge of the dominant Israelite culture; and 2) that they used this knowledge in a gambit to trick the representatives of dominant Israelite power into adopting a more favorable policy toward their subordinate group than might otherwise have been the case.
In medieval Islamic history, three examples of similar literary-historical topoi open up differences between Islamic assumptions of collectivity and contradictory historical realities.
With the “Gibeonite gambit” narrative in mind, a similar ruse is uniquely documented in the tenth-century Fihrist of Ibn al‑Nadīm, whereby the non-monotheistic inhabitants of Ḥarrān in upper Mesopotamia were able to justify their existence in the face of monotheistic Muslim assumptions of collectivity.9 It is reported that on his way to campaign against the Byzantines, the caliph al‑Ma’mūn (813–33 C.E.) encountered the non-Muslim Ḥarrānians. He asked them, “Which of the subject peoples (dhimmīs) are you?” When they were unable to provide a satisfactory answer, the caliph declared that when he returned from campaigning they would be required to convert either to Islam or one of the recognized religions or be slaughtered. Some promptly converted to Islam or Christianity. Others who did not benefited shortly thereafter, when a Muslim shaykh provided the Ḥarrānians a piece of particular knowledge with which to represent themselves to the caliph and thereby trick the Muslims. After receiving a large sum of money, the shaykh told the Ḥarrānians that they must represent themselves as “Sabians” (al‑Ṣābi’ūn), a pre-Islamic “people of the book” mentioned in the Qur’ān.10 Although al‑Ma’mūn died on campaign and would never return from this journey, the Ḥarrānian community, now identified as Sabian, continued to exist and even flourished for some time afterwards.11
A comparison of this narrative with the Gibeonite story is instructive in several ways. First, as in the Book of Joshua, it offers an etiology for the challenge to Muslim assumptions of collectivity represented by the continued anomalous existence of “pagans” in the heart of dār al‑Islām, the “abode of Islam.”12 Among the assumptions of collectivity of mature medieval Islam was the premise that the presence of polytheism had not and would not be tolerated in society. Muslim collectivity sanctioned the presence of Jews and Christians, who were members of an ahl al-kitāb, or “people of the Book,” and thereby were granted dhimma, or “protection.” Dhimmī communities were permitted to practice their religions and enjoy communal autonomy in exchange for certain limitations and the payment of taxes applicable only to dhimmīs.13 In contrast, idolaters and “pagans” were not to be tolerated. This prohibition was, in part, premised on an image in Muslim tradition of the seventh-century Muslim conquests, when idolaters would either have been exterminated or saved themselves by choosing to convert to Islam. Nonetheless, the historical unfolding of the conquests at a time when Islamic assumptions of collectivity were not fully developed meant that non-monotheistic groups such as the “pagans” of Ḥarrān or Zoroastrians were offered dhimmī status.14 In fact, it is recorded by the thirteenth-century ‘Izz al‑Dīn ibn Shaddād that when ‘Iyād ibn Ghanm conquered Ḥarrān in 639, he took the Ḥarrānian temple and made it into a mosque, but then granted the Ḥarrānians another site at which to build a new temple.15 As in the case of the ancient Israelites, the history of conquest and later assumptions of collectivity presented contradictory notions.16
It must be observed that both stories hinge on a seemingly preposterous element that would lead us to believe that in both dominant cultures, biblical Israelite and early medieval Muslim, the ruling elites were completely unaware of the presence and character of minorities in their midst. It is difficult to believe that after conquering all of Transjordan and then establishing a foothold in Canaan with victories at Jericho, Bethel, and Ai, the Israelites would not have known of the Gibeonites.17 More preposterous is the assumption that al‑Ma’mūn would not have known of the identity of the Ḥarrānians. In fact, after Ḥarrān surrendered to Muslim forces in 639 or 640, it was a site of some importance in early Islamic history.18 Under the Umayyads it was a center for the Banū Qays tribal confederation.19 The caliph ‘Umar ibn ‘Abd al‑Azīz (717–720) is credited with relocation of the Alexandria school of medicine to Ḥarrān and Antioch.20 In 744 it became the caliphal capital for six years under Marwān ibn Muḥammad ibn Marwān.21 Al‑Mansūr spared its walls in the reduction of Mesopotamian fortresses after the ‘Abbāsid revolution,22 and under Harūn al‑Rashīd (786–809) a canal was constructed for the city’s water supply.23 Significantly, according to Bar Hebraeus, Ibrāhīm, the uncle of al‑Ma’mūn, “permitted the pagans of Ḥarrān to perform their mysteries openly.”24 For a century Ḥarrānian religion was tolerated by the highest Muslim authorities.
Even if the Gibeonite and Ḥarrānian narratives are based on kernels of historical truth, the credulity attributed to the dominant cultures in the face of historical evidence to the contrary suggests a high level of literary reshaping. The confused historicity of the Israelite tale and the lack of corroboration for al‑Ma’mūn’s encounter with the Ḥarrānians begs that the stories be analyzed using literary-historical approaches. The credulity of the dominant culture in these narratives, in fact, places blame squarely on the dominant culture itself for the anomaly in its midst. In the case of Ḥarrān, this critique of the dominant culture’s easily-manipulated rulers can be read either as minority accusation against Islam or majority self-criticism, depending on the context. However, this observation by itself is insufficient for understanding the text’s origins or its multiple functions.
A difference in these two stories will help illuminate features of the medieval Islamic social structure. Whereas the Bible does not tell how the Gibeonites gained knowledge of the Israelite dominant culture, the Ibn al‑Nadīm account is explicit in identifying a renegade, corrupt Muslim shaykh as the source of the key information gained by the Ḥarrānians. This narrative element of an informer functions differently when considered from the points of view of different religious communities.
Some scholars have cited Christian hostility to both Ḥarrānian “paganism” and Islam as an explanatory key for this story. A “Christian reading” of the corrupt shaykh rests on Ibn al‑Nadīm’s cited source for the narrative, a Christian named Abū Yūsuf Īsha’ al‑Qatī‘ī.25 It would advance the interests of Christian opposition to Ḥarrānian “paganism” to lay blame for the city’s “pagan” religion and regional power on Muslim credulity. Such a religio-historical polemic is not merely a one-dimensional argument directed against an opponent, but in this case, the narrative’s social and historical context suggests a maneuvering for position that often occurs as different subordinate groups compete for a place within a dominant society (in this case, Christians and Ḥarrānians). By portraying one’s co-subordinate opponents negatively, one’s own group can be made to appear more favorably. The polemic of al‑Qatī‘ī is representative of an ancient enmity and competition in Syria and Mesopotamia between Christianity and various Hellenistic religious traditions, often expressed through the competition and hostility of the Christian city of Edessa with “pagan” Ḥarrān.26 Because some Christians in the ninth-century caliphate could be perceived as a kind of fifth column with pro-Byzantine leanings, it was inevitably in their interest to try to position themselves favorably in relation to the dominant culture and in opposition to other subordinate groups. In fact, the record of this Christian-Hellenistic contest helps to explain why the caliph Mu‘āwiya (661–80) and other Umayyads selected Ḥarrān as a place of military and administrative importance to counter the perceived power of Christianity, both local and foreign. The city’s religio-strategic position was confirmed in the revolt of the Jazīra during the ‘Abbāsid revolution. While Edessa and other cities supported the Umayyads, Ḥarrān was an ‘Abbāsid stronghold.27
In contrast, the “Islamic reading” of the story provides a secondary etiology to explain how the Ḥarrānians were able to acquire specific knowledge of the Qur’ān and its contents, something that should not occur given strictures against the teaching of scripture to non-Muslims (another assumption of collectivity).28 The informer in the narrative takes care of this contradiction. The Islamic reading is further supported by the report that after the death of al‑Ma’mūn and the subsequent success of the new Sabian identity, many newly-baptized Christians returned to their former religion. Without this last element, the narrative retains a Christian triumphalist conclusion. Interestingly, this element also suggests that the report in Ibn al‑Nadīm comes from early in Islamic history, before the prohibition against conversion to a religion other than Islam became a general (though not unanimous) assumption of collectivity.29
One must note that although this story contains elements that can be read as critical of Islam, it is located in the literature of the dominant Islamic culture. A Christian reading, as an anti-Ḥarrānian story, places value on the unworthiness of Islam exemplified by Muslim credulity and despotic power, but a Muslim reading of the same story functions as the Gibeonite story does for the Bible—providing a necessary etiology for the existence of an anomalous group within dār al‑Islām. Although it appears in this Muslim narrative that Muslims are telling stories about dhimmīs, in fact Muslims are adopting a mimetic strategy to describe their own society.
The double reading suggests that these stories are needed whether or not they are true. One cannot tell whether the story is historically reliable, or what elements of the story may be Christian or Muslim. Many important Ḥarrānians are identified with the nisba “al‑Ṣāb’ī” (“the Sabian”) beginning in the middle of the ninth century. Thus, the Ḥarrānians adopted and internalized the story to create an originary myth for their presence and continued existence in Islamic society. And, in later Muslim historical and other writings, the appellation al‑Ṣāb’ī designates “pagan” Ḥarrānians.30 Whatever the historicity of the narrative, it marks a key transformative historical moment in Ḥarrānian-Muslim relations.31 The tenth-century al‑Khwārizmī manifests Muslim internalization of the story and states in regard to the Ḥarrānians that “they were called Sabians from the time of al‑Ma’mūn.”32 Although Muslim authorities debated the status of Sabians,33 the new usage served both minority and dominant perspectives, because it opened up an historically imaginative space in which the myth of origin could exist.
The composition of this narrative must be understood as a joint production of Christians, Muslims, and Ḥarrānians. Narratives such as this one, which describe Muslim relations with non-Muslims, can be the result of a complex dialogical process between the dominant and the dominated. Chase Robinson has demonstrated that in regard to conquest narratives, this mediated process emerged from the interaction between minority communities seeking to defend or expand their customary privileges and efforts by the ‘Abbāsid administration to establish (or impose) organized and coherent regulation of non-Muslims in the late eighth and early-ninth centuries.34 Even later, the establishment of Muslim legal norms would further impel non-Muslims to produce historical evidence as precedent in support of their communal rights and privileges. In these historical examples, the produced and co-produced evidence took the form of an aesthetic object, specifically a historical narrative or myth. As such, one can imagine the encounter of al‑Ma’mūn with the Ḥarrānians to be a co-production whose contributing authorship includes Christian, Muslim, and Ḥarrānian voices.
The co-production of minority tales in medieval Islam speaks to the contractual character of the social order for minority communities (and thus to dhimma) and to the importance of historical precedent in the formation of social contracts. These tales are expressions of powerlessness and power in the relations between subordinate and dominant communities.
Another example of the “Gibeonite gambit” in medieval Islamic history comes from a story found in the literary tradition of the Karaites, a sect of the Jews that rejected the authority of the rabbis and of rabbinic texts such as the Mishna and Talmud. Like the Gibeonite story, the historicity of narrative elements is difficult to ascertain. And like narratives associated with the rise of Islam in the Middle East, the story of the beginnings of Karaism is obscured by a complex layering of narrative resulting from co-production. Before the early tenth century, Karaite origins are obscure, but by that time a scripturally-oriented, anti-rabbinic, Palestino-centric, and strongly millenarian Karaite movement was active in Palestine.35 Over the following two centuries a flourishing alternative to rabbinic Judaism was centered at Jerusalem.36 In this period, Karaite scripturalism and messianic millenarianism placed the focus of attention on the divinely revealed text of the Hebrew Bible. Self-definition found expression in legal and pietistic activity buttressed by textual exegesis and philosophy, and as a consequence, notions of the Karaite past and history were devalued as a means for understanding the community.37 The immediacy of sanctification and messianic expectation made irrelevant a need for thinking about the interval between events in the Bible and the present.
By the end of the eleventh century, the Karaite center in Jerusalem was destroyed in the aftermath of Seljuq inroads into Palestine and the Crusader conquest. The center of Karaism shifted to Constantinople, where new perspectives on the past emerged as part of subsequent developments in Byzantine Karaite thought and literature in the twelfth and thirteenth centuries.38 This change in Karaite attitudes towards the past is woven into the story of the eighth-century heterodox figure, Anan ben David, who is framed as the founder of Karaism in the twelfth-century Ḥilluk ha-Kara’im veha-Rabbanim (“Differences between the Rabbanites and the Karaites”) by Elijah ben Abraham.39
The Ḥilluk seeks to explain the division of Israel (that is, of Jews) between Karaites and Rabbanites (rabbinic Jews), in part by laying out a historical scenario that begins in biblical times and culminates with the story of Anan. Appended to the narrative is a list of post-Anan Karaite scholars that brings the timeline up to the twelfth-century period of the author. The Anan story demonstrates how Karaites and their interlocutors utilized narrative strategies similar to the Ḥarrānians in order to establish a useful post-biblical past and a foundational myth for medieval Karaism.
To summarize the story, Anan, of the ancient Davidic princely family in Iraq, was passed over in the succession to the princely office of the exilarchate on account of his heterodox ideas. His brother was elected as exilarch (resh galuta, “Head of the Exile”), so Anan gathered both his own followers and the surviving remnants of Second Temple sects, such as Sadducees and Boethusians, and declared himself exilarch.40 As a result, he incurred the wrath and opposition of the rabbis, who persuaded the Muslim authorities to intervene. Anan was thrown into prison and sentenced to death for challenging caliphal authority, since succession to the exilarchate was confirmed by way of caliphal appointment and any tampering with the process would be considered an act of disobedience. While in prison, he met a Muslim scholar who advised him how to approach the Muslim authorities. Following the advice, Anan bribed the vizier to gain an audience with the caliph and presented himself as leader of a religious community different from that of the rabbis, and not merely an unauthorized sectarian group. He also said to the caliph, “The religion of my brother employs a calendar based upon calculation of the time of the new moon and intercalation of leap years by cycles, whereas mine depends upon actual observation of the new moon and intercalation regulated by the ripening of new grain.”41 Since this Jewish law, or halakha, was in accord with Muslim sharī‘a, the caliph looked favorably on his case and Anan was able to secure governmental protection for himself and his followers. Thus, according to this story, Anan’s hoax elicited the caliph’s identification and Karaism was founded.42
Parallels with the Ḥarrānian story are noteworthy. Just as the al‑Ma’mūn story is credited to a Christian anti-Ḥarrānian source, the Anan story is explicitly identified as a Rabbanite polemical tale aimed at demonstrating that Karaism is a false creed whose very existence is the result of caliphal intervention into internal Jewish affairs. It has been attributed to the polymath Saadia Gaon (892–942), head of the rabbinic academy of Sura in Baghdad and active opponent of Karaism.43 Like the Christian reading of the Ḥarrānian story, the “Rabbanite reading” of the Anan story places blame on Islam for the existence of a competing errant religious tradition. Part of the Ḥilluk reproduces the anti-Karaite narrative, but then states that “the account proves the opposite of what the Rabbanites had intended.”44 Elijah then takes up the anti-Karaite accusations one at a time. To give a single example, he argues that Anan could not have successfully bribed the vizier since the Rabbanites, as leaders of the Iraqi Jewish community, could easily have outbid him. In fact, such a claim better reflects Elijah ben Abraham’s twelfth-century Byzantine circumstances rather than that of previous centuries when Karaite wealth and power could be significant.45 More to the point, the wealth and power of rabbinic elites in seventh century Iraq are unknowable.
Both stories, Ḥarrānian and Karaite, engage the plot device of a Muslim informer, a renegade shaykh, who advises the protagonists on how to proceed. This renegade possesses the inside information necessary for the plot to unfold and for the gambit to be risked. In a document from the Cairo Geniza, Anan’s cellmate is identified as Abū Ḥanīfa al‑Nu‘mān ibn Thābit, the imam of one of the four orthodox schools of Sunnī Muslim law.46 In fact, Anan probably lived in the middle of the eighth century, and the dates of Abū Ḥanīfa in Muslim tradition are 700–767. Indeed, Abū Ḥanīfa was imprisoned late in his life, probably for supporting a Zaydī Shī‘ite rebellion.47 Notwithstanding possible historical contemporaneity, Abū Ḥanīfa’s appearance as a trope in a Geniza text that cannot be earlier than the thirteenth century strongly suggests literary reshaping and functions as a plot device that adds an additional dimension to the interpretation. As founder of one of the orthodox schools of Muslim law, Abū Ḥanīfa’s function not only makes the Muslim host culture culpable for the birth of Karaism, but also reframes the entire story. It suggests that the caliph was able to offer recognition to the nascent Karaite community because the problem of Anan and the rabbis corresponded to Muslim jurisprudential notions, whereby a madhhab, or “law school,” can co-exist with other equally orthodox schools of legal interpretation.48 With the cellmate identified as Abū Ḥanīfa, Anan becomes the founder of a jurisprudential school that can be understood to be cognate to the way schools of law are understood in Islam and is worthy of recognition in the eyes of the Muslims. Interpreted this way, the narrative suggests that the caliph could easily understand Anan’s legally defined position. Identification of Abū Ḥanīfa provides for the “Karaite reading.”
In terms of leveraging the dominant culture by following the Muslim scholar’s advice, Anan’s (and late Karaite) calendrical determination of the new moon is a legal problem firmly rooted in Scripture, and as such, responds to rabbinic halakha. It does not derive from Muslim counsel, as the story claims, but is a halakhic position that is both characteristic of several Jewish sects in history and is also debated in rabbinic literature.49 The Rabbanite anti-Karaite intentionality of the “original” narrative notes the similarity of Karaite and Muslim calendation in order to claim that the Karaite halakhic position is derivative of Islam, thereby furthering Muslim responsibility for the birth of Karaism and framing Karaite belief and practice to be both false and essentially not Jewish.
In the end, Anan was able to persuade the caliph that he was the head of a community separate from that of his brother, the exilarch. This element of the narrative makes historical sense for circumstances many decades after the tale’s eighth-century setting, when, under the religious policies of al‑Ma’mūn, the caliph might recognize a group with as few as ten adherents as a separate religion. In fact, a decree of al‑Ma’mūn in 825 is recorded in the Syriac Chronicle of Zuqnin and repeated by Bar Hebraeus in regard to a Jewish exilarchic contest of that century, when the grandson of Anan, Daniel ben Saul, was removed as exilarch.50 His followers continued to maintain allegiance to Daniel’s family and his son, also named Anan. Moshe Gil has argued persuasively that the Anan “Karaite gambit” narrative conflates Anan ben Daniel with his great-grandfather Anan ben David as a polemic directed against an entire branch of the exilarchic family.51
Whereas Elijah ben Abraham sought to refute the anti-Karaite story’s particulars in the twelfth century, the very fact that he used this narrative points to the ahistorical perspective hitherto embraced by Karaites. The challenge in twelfth-century Byzantium was for Karaites to present themselves as a legitimate religious entity to two dominant powers: Rabbanite Jews and Byzantine imperial authorities. Byzantine Karaites first came to the empire by way of Greek conquests of Muslim lands in tenth-century Syria and later through immigration from the Islamic world.52 As a consequence, these non-rabbinic Jews—whose wider cultural background was Islamic—suffered in the empire as foreigners with a double disadvantage, being both non-Christian and culturally non-Greek at the same time.
At the time of Elijah ben Abraham’s writing of the Ḥilluk, the descendants of those formerly Islamicate Karaites were engaged in a massive transformation of the linguistic and cultural heritage of Near Eastern Karaism—with its Judeo-Arabic written tradition—into something new in the Greek Orthodox Byzantine environment. By translating old texts and writing new ones in Hebrew (such as the Ḥilluk), a written tradition was inaugurated that shared the language of law and religious discourse with the rabbis. Karaite writing began to display more focused concern for issues of self-definition, partly as an effort to deflect rabbinic attacks. The shift in environment coupled with the trauma of the late eleventh-century destruction of the Jerusalem center forced them to face the fact that they indeed had a past. The ahistorical immediacy of a scripturalist and messianic-millenarian Palestino-centrism would not suffice in the new doubly-articulated, diasporic environment.
A third order of disadvantage resulted from rabbinic opposition to the presence of Karaites in the Empire. In response, the Karaites became acutely aware of their status as a minority within a minority and responded in part by trying to position themselves in a favorable light in relation to other Jewish sects, such as the Mishawites.53 Karaites sought to describe a continuum of Jewish beliefs and practices, in which they themselves represented a lesser aberration than others from assumptions of collectivity held by rabbinically self-defined and Christian-recognized dominant Judaism. Thus, the connection of Anan with the ancient Sadducees and Boethusians, the prototypical heretics of the rabbis, is retained in the Karaite narrative.54 This kind of social maneuvering among subordinate Jewish groups mirrors the competition between Edessa and Ḥarrān in the ninth century and is evident in other Byzantine Karaite works.55 In sum, the Karaites sought to locate themselves positively in relation both to the Byzantine Greeks and Rabbanite Greek Jews by using ideas and images that would be recognizable to the recipients of their message. By representing themselves as a Jewish sect with historical roots sanctioned by the foreign and Islamic past, they describe themselves using the language of heresy (well-known in Byzantine Greek culture). And, to Rabbanites, they portrayed a wide range of Jewish difference, in which Karaism was portrayed as the least offensive to the rabbis. Thus, Karaites began to describe themselves self-consciously by using some of the Jewish terminology of sectarianism, thereby insinuating Rabbanite dominance.
The construction of the Anan narrative demonstrates more transparently than the Ḥarrānian story the stages and internal logic of co-production. Like the Ḥarrānians, the Karaites explicitly adopted and were able to capitalize on what was originally an antagonistic narrative by substituting new interpretations and establishing new meaning for the story. Just as Ḥarrānian-Christian antagonism was contextualized against Muslim authority, this story displays the over-arching Muslim and Byzantine contexts for Jewish-Jewish (that is, Rabbanite-Karaite) co-production. Whereas the Ḥarrānian-Sabian narrative is obscure in regard to internal Ḥarrānian developments and thereby frames the “new” ninth and tenth-century identity of Sabian to have been created from impetuses external to the Ḥarrānian community, the Karaite narrative lays bare the historical fact that Karaite identity was constructed simultaneously both from within and without.
Originally, the Rabbanite reading offered etiological explanation for the existence of an anomalous non-rabbinic Judaism that challenged Rabbanite assumptions of collectivity. By employing the “Gibeonite gambit” as a topos, the dominant, non-Jewish Muslim culture was portrayed as responsible for the “illegitimate” birth of Karaism. Later (perhaps two centuries) and in a different civilizational orbit, the Karaite reading offered a narrative of origins in which the authority of dominant Islam was no longer an interfering evil force that forced a false sect on the Jewish community, but was a wise arbiter of Jewish difference providing acknowledgment and sanction for an ancient and separate Judaism inscribed in the image of Christian heresy. The story of Anan became a permanent feature of Karaite memory.
Although the examples of the “Gibeonite gambit” discussed above are rooted in literary traditions as much as history, the third example of this type of narrative moves the story-as-strategy further out of the realm of literature and into historical documentation. As such, it concretely illustrates the interaction of narrative with political power.
The fourteenth-century Sunnī Muslim traditionist and historian Muḥammad ibn Aḥmad al‑Dhahabī (d. 1374) reports in his massive biographical dictionary, Siyar a‘lām al‑nubalā’ (“Biographies of the Most Learned Nobles”), on an incident involving al‑Khaṭīb al‑Baghdādī, one of the great Sunnī traditionists and historians of the eleventh century (d. 1071).56 Al‑Baghdādī’s patron was the vizier, Abū al‑Qāsim ibn al‑Muslim, who decreed that no one in Baghdad was permitted to narrate a ḥadīth (tradition about the Prophet Muḥammad) without first authenticating it with al‑Khaṭīb al‑Baghdādī. It is reported that on one occasion some Jews presented the vizier a document which they claimed to be a letter of the Prophet Muḥammad that exempted the Jews of Khaybar in northwest Arabia from paying the poll tax or jizya, a stipulation of dhimma that was otherwise incumbent upon all Jews and Christians. This document cited as witnesses some Companions of the Prophet who were purportedly present in the year 7 A.H. (629) after the defeat of the Jews of Khaybar by the Muslim army. The Jews claimed the letter had been dictated by Muḥammad and written by ‘Alī ibn Abī Ṭālib, the Prophet’s cousin and close associate. The vizier forwarded the document to al‑Khaṭīb al‑Baghdādī, who identified it as a forgery through disqualification of the signed witnesses, Mu‘āwiya ibn Abī Sufyān and Sa‘d ibn Mu‘ādh. He noted that Mu‘āwiya had not become a Muslim until two years after events at Khaybar (Mu‘āwiya would go on to become the first Umayyad caliph many years later), and that Sa‘d died two years previous to the battle.57 The Jewish claim was dismissed.58
From a literary point of view, this effort at deploying the “Gibeonite gambit” works differently from the others. It is not characterized by credulity on the part of representatives of dominant culture. In fact, although Jews do have particular knowledge of the Muslim past and its prosopography, the gambit fails because they do not know it well enough. They fail in arranging the details. The purpose of such a tale in this biographical setting is to demonstrate the efficacy of Muslim legal methodology and the deep and particular knowledge of tradition that could be brought to bear by someone like al‑Khaṭīb al‑Baghdādī. Jews function merely as a narrative device in this minor representation of dhimma, which teaches that the dominant authorities need be wary of claims from the subordinate dhimmīs. Naturally, like the other “Gibeonite gambit” narratives, the contractual nature of dhimma is central to its cultural assumptions. Most interestingly, the narrative does not offer etiological explanation for the presence of an anomalous group in society—as the Gibeonite, Ḥarrānian, and Karaite narratives—but is witness to a group seeking anomalousness.
The “Khaybarī gambit” is not merely the product of Muslim literary imagination, but also has documentary attestation. A Judeo-Arabic Geniza document that was first published by Hartwig Hirschfeld in 1903 is a letter of protection from the Prophet Muḥammad written to the people of Khaybar and Maqnā.59 In Muslim tradition, after the battle of Khaybar in 7 A.H. the Jews were allowed to retain and work their lands in exchange for half of their annual harvest.60 Maqnā enters the historical narrative two years later during Muḥammad’s Tabūk campaign, when it is recorded that letters of protection were issued to Ayla (modern Eilat/Aqaba) and to three communities in northwest Arabia whose inhabitants are specified or understood to be Jews: Adhruḥ, Maqnā, and Jarba. The terms of each of these letters display a differing variety of payments owed as tribute.61
The reliability of these letters of protection as historical evidence for the seventh-century is debated by scholars.62 The absence of a systematic and well-understood method for exacting tribute in these and others letters suggests that none existed at the time of the Prophet. On the other hand, the presence of anachronistic features more characteristic of later ‘Abbāsid-era administration are present.63 Together, these contradictory features indicate literary shaping over time, which establishes many aspects of the letters as backward projections of Muslim notions of sovereignty and dhimma from a later era. By the time such documents appear in Muslim literature—in the sīra (biography of Muḥammad) and in Arabic historical writing in the ninth century and beyond—they had been reshaped and rewritten in order to support agendas contemporaneous with their time of writing. In the case of the letters to the three communities, they were framed to provide historical precedents for military engagements authorized by Muḥammad outside of the Arabian peninsula and to fully articulate imperial and constitutional concepts of dhimma including the duties and taxes owed by these (and later) non-Muslim communities.64
Furthermore, the Geniza document is addressed to “the people of Khaybar and Maqnā and their progeny as long as the heavens are above the earth,” although in Muslim conquest narratives a letter of protection is witnessed for Maqnā only.65 Neither does the Maqnā letter of Muslim tradition imply that descendants of the Jews of Maqnā could bear the terms of the settlement with them in perpetuity. The Geniza document combines elements of the unique terms granted to the Jews of Khaybar in the sīra and maghāzī literature (some of which were generous in comparison with terms later set for dhimmīs) with elements and the form of the Maqnā letter as it is known from Muslim tradition. This awkward editing together of features from reports of two moments in early Muslim history is a Jewish forgery, whose re-writing of the Muslim past has practical implications for its bearer. The content of this letter includes obvious interpolations and extravagant claims written in the Geniza period of the tenth through twelfth centuries. As Salo Baron noted, the letter is “almost exclusively concerned with the safeguarding of the latter’s [Jews’] rights, rather than the imposition of duties.”66 Its bearer seeks privileges and tax exemption. Contrary to the Maqnā letter of Muslim tradition, found in al‑Wāqidī, al‑Balādhurī, and Ibn Sa‘d, and contrary to classical conceptions of dhimma, the Geniza letter states, “You shall not have the annoyance of land-tax,” and “. . . you shall not be as other poll-tax payers.” Although scholars today might doubt the authenticity, wholly or in part, of Muḥammad’s letters, this document illustrates that in Muslim society of the Geniza era, Muḥammad’s letters of protection were understood to be authentic, and as such, could have legal standing.
More to the point, the document demonstrates much more impressive familiarity with Muslim historical narrative on the part of Jews than does the report in al‑Dhahabī, and it manipulates that familiarity in order to bolster its claims of authenticity. For example, the letter stipulates to the Jews of Khaybar and Maqnā, “You shall be held in honor on account of your own high station and the station of Ṣafiyya, the daughter of your uncle.” The Muslim past is used by referring to Ṣafiyya bint Huyayy of Khaybar, whom Muḥammad took as a wife after the campaign there.67 Safiyya is mobilized as a mediating signifier that links the honor of Muḥammad to Jews through marriage. And, like al‑Khaṭīb al‑Baghdādī’s document, Hirschfeld’s Geniza letter is noted to have been written by Muḥammad’s cousin and close associate, ‘Alī ibn Abū Ṭālib (after being dictated by the Prophet). However, it has a different list of witnesses than those disqualified by al‑Khaṭīb al‑Baghdādī, two of whom in Muslim literature are actors in the march to Tabūk. All three witnesses, Ammār ibn Yāsir,68 Salmān al‑Fārisī,69 and Abū Dharr al‑Ghifārī70 are recorded to have died well after the battles of Khaybar and the later Tabūk expedition.71
Most interestingly, the letter uses Qur’ānic terminology and also quotes the Qur’ān itself,72 although canonization would not be complete for at least some decades after the letters to Khaybar and Maqnā (and after the death of Muḥammad).73 However, in Muslim literary tradition quotation of Qur’anic verses by Muḥammad, a Companion, or other speakers (for example, the akhbār of the Jews in Medina) before the redaction is a natural Sitz im Leben. This feature of the text is not itself anachronistic or problematic from a traditional Muslim perspective and would support the letter’s authenticity. Nonetheless, when such a quotation is analyzed from a literary-historical point of view, it stands as testimony to Jewish familiarity with the Qur’ān later in the Geniza period.74 Since in sharī‘a, non-Muslims are prohibited from touching a Qur’ān, much less reading or studying it, the document represents another kind of anomaly set against Muslim assumptions of collectivity.75
Attestation to the historicity and effective use of Khaybarī identity is found in several eleventh-century contexts. An Arabic petition from the Geniza, dated c. 1030, is directed to the Fāṭimid authorities regarding the Jewish taxes (jizya) of Tiberias, in which it is noted that some Jews pay 5 dirhams “at a fixed rate” and others pay 10, while “those who claim they are Khaybarīs” pay none.76 Since Jewish communities were collectively responsible for taxation to the Fāṭimid government, the Khaybarī claim in this document not only supports an exemption vis-à-vis Muslim power, but also is used to obtain release from Jewish communal responsibility. In a Geniza letter a poet named Yākhīn seeks relief from paying the jāliya (as the poll tax was called in Egypt) based on his Khaybarī descent.77
A number of other Geniza documents are pertinent. An eleventh-century petition is concerned with the financial difficulties of Ṣāliḥ al-Khaybarī, “a Khaybarī Jew from Baghdad.”78 A list of contributors dated to around 1035 includes one Ibn al-Khaybarī.79 Another Ibn al-Khaybarī is mentioned to have sent a letter from Egypt to the West.80 A court record indicates that a certain Rayyisa bint Joseph Bīmī, wife of Yeshū‘ā ben Nissīm, brought suit against Saadya ben Benjamin Khaybarī for four dinars.81 In a letter dated 1090, Abraham ben Ḥalfōn, chief judge of Ashkelon, mentions to the addressee in Fusṭāṭ that a letter from the Kayābira (plural of Khaybarī) is on the way.82 Finally, a petition dated Shawwāl 654 A.H. (October–November, 1256) certifies that one Ibrāhīm ibn Ismā‘īl is a “son of one of the ‘pardoned’ Khaybarī Jews,”83 which demonstrates the persistence of Khaybarī identity beyond the eleventh century.
In fact, the preponderance of attestation to the “Khaybarī gambit” comes from the eleventh century, including reports found in the Ta’rīkh of the Melkite Christian historian Yaḥyā ibn Sa‘īd (d. c. 1066), who mentions that on two occasions the Fāṭimid Caliph al-Ḥākim (996–1021) exempted Khaybarī Jews from sumptuary regulations and other restrictions imposed on Jews and Christians.84 This exemption must be understood as a consequence of the difference between Sunnī and Shī‘ite versions of early Islamic history. Sunnī Muslims were forced to rewrite the history of Khaybar in order to harmonize Muḥammad’s assurances to its Jews with the later expulsion of Jews and Christians from the Hijāz that is ascribed to the caliph ‘Umar (634–44). For Sunnīs, the contradiction between authoritative precedent set forth by the Prophet and later caliphal policy led to the generation of ahādīth that provide prophetic foreknowledge of this later development.85 In addition, Muḥammad’s settlement imposed upon the Jews of Khaybar was revised in order to demonstrate that the Prophet was explicit in indicating that its terms could be revoked.86 For the Shī‘ite Ismā‘īlī Fāṭimids, the first three caliphs were illegitimate usurpers who supplanted the rightful caliph, ‘Alī ibn Abū Ṭālib. The Sunnī historian al-Maqrīzī (1364–1442), a major source on the Fāṭimids, characterizes the Fāṭimids, and al-Ḥākim in particular, as “excessive in terms of rafḍ,” that is, refusal to accept the authority of ‘Umar and other Companions.87 Hence, al-Ḥākim’s exemption of Khaybarī Jews constitutes an anti-Sunnī gesture that upholds Muḥammad’s prophetic authority against its later violation by ‘Umar.
In the context of Fāṭimid Shī‘ite anti-Sunnī ideology, claims to Khaybarī Jewish identity make sense. For Fāṭimid Shī‘ites, acknowledgement of the existence of Khaybarī Jews created an anomaly that challenged Sunnī assumptions of collectivity. Had the Khaybarī Jews not been expelled by ‘Umar, their particular tax regime would in the eleventh century have been local (at Khaybar) and knowable. Neither would there have been a heretical deviation from prophetic command. Mirroring ninth-century Ḥarrānians-cum-Sabians, evidence of the nisba al-Khaybarī, or Ibn al-Khaybarī in the eleventh-century suggests that deployment of the “Khaybarī gambit” likely emerged in the aftermath of al-Ḥākim’s persecution of Christians and Jews on account of its Khaybarī Jewish exemption. Conditions favored those who would claim Khaybarī Jewish descent, providing new opportunity to engage in the Gibeonite strategy by disseminating their version of the terms of an ancient treaty of the Prophet and claiming its tax regime for themselves “as long as the heavens are above the earth.”88
Perhaps as important as documented use of the “Khaybarī gambit” is the retelling of the story in Islamic literature, because the process of co-production is revealed. In the first instance, the “Khaybarī gambit” topos was the work of “Muslimly-informed” Jews who sought tax exemption, if not prestige. Subsequently, Jews sought to establish Khaybarī identity in Egypt, while Sunnī authorities quashed it in Baghdad, whose story was subsequently incorporated into Sunnī Muslim tradition. That is, the complex interplay of Sunnī and Shī‘ite historiosophical polemics included—and contested—events at Khaybar during the life of the Prophet. The contest created an opening for Jewish manipulation of the past which resulted in multi-layered co-production that is evident in Geniza documents and enshrined in the fourteenth-century Sunnī biographical dictionary of al-Dhahabī. With this in mind, al‑Khaṭīb al‑Baghdādī’s repudiation of the “Khaybarī gambit” was also an anti-Shī‘ite move.
These narratives are not unique examples in medieval Islam of the reformulation of memory on the part of minority groups whose purpose was to gain recognition and advantage. The concern for defending against dhimmī trickery that underlies the narrative of al‑Khatīb al‑Baghdādī reflects the contractual character of dhimma, which consequentially opens the door to other non-Muslim minorities with similar claims. The so-called Edict of the Prophet to the Christians is described by Claude Cahen as a “pious fraud of Nestorian monks in the 3rd/9th century.”89 This document, as described in the tenth-century Chronicle of Seert and elsewhere, purports to be a letter of security given to the Christians of Najrān after the city’s conquest in A.H. 10. It bears several features of the “Gibeonite gambit:” topos, including the use of Muslim sources, the maneuvering of minorities (denunciation of idolators and Jews), a meeting with Muslim authorities (the delegation of Catholicos Ishoyabh II to Muḥammad), and claims to tax exemption (for monks).90 Another example is the Bahīrā story, which functions in similar ways in Muslim literature. As a Muslim narrative, it portrays the monk Bahīrā acknowledging prophetically the prophethood of Muḥammad early in his career. Embellished as a Christian story, it depicts Muḥammad granting perpetual recognition and protection to monks and Christians in exchange for a divine promise of Muslim invincibility.91 The extensive use of these memory projects, including those of the Ḥarrānians, Karaites, and Khaybarī Jews, are fantasies of minority relations to power in Islam.
The “Gibeonite gambit” narratives represent “dreams” of subordinate groups in Muslim society that through the motif of manipulation express such themes as anomalousness, powerlessness, subordination, and integration into the larger society. These themes are descriptive of some general conditions of minority life under Islam and consequently are plausible and thereby integral to the self-perception of minority groups. Identifying a fundamental circumstance for minority life in medieval Islam, the narratives are all dependent upon the contractual nature of the relationship between non-Muslim groups and power in the dominant Muslim society and culture. In the historical record, the contractual relationship of dhimma is played out between elements of the minority groups on one hand and tax officials, the caliphal and princely chanceries, the courts of rulers, and religious authorities on the other.92 The enactment of these relationships constitutes rituals of subordination, or what James C. Scott refers to as “public transcripts.”93 The “Gibeonite gambit” stories represent the narrativization of the origins of such relationships in early Islam as imagined through the imbrication of subordinate and dominant memory in the process of literary co-production. They stand between what Scott refers to as “hidden transcripts” and public transcripts, in that they contain elements of inner-minority discourse about dominant power, while at the same time are presented as open and unrestricted demonstrations of dhimmī cooperation with Muslim authorities. Although the fundamental manipulation is literary and intertextual at its base, its impact in terms of official recognition, taxes, and day-to-day life could be concrete.
As hidden transcripts, these narratives are also what Marc Michael Epstein calls “dreams of subversion.”94 As such, the subversive element expressed in the motif of manipulation indicates ironically the depth of integration into the dominant society that these minority groups arrogate to themselves. In terms of literary integration, these acts of manipulation require deep knowledge of Islam, while at the same time they reify a highly developed Islamic literary sub-genre in which the motif of trickery is given distinction. Robert Irwin notes, “The celebration of artfulness or tricksiness, whether in the commission of crimes, in arguing points of law or in inserting puns in poems, is one of the most striking features of medieval Arabic literature.”95 In these stories, characterized by a doubly-articulated strategy, the minority imaginary is intimately connected to that of dominant Muslim society. Thus, identity formation among these minority groups cannot be purely an inner-group development, but must take Islam into account. That is, identity formation is characterized by the dhimmī contract and the group’s ability to negotiate its terms.
Negotiation of dhimma can be a moment of danger when potential loss or gain is risked. As such, the risk of the “Gibeonite gambit” offers an additional and imagined element to the originary moment of negotiation as it is inscribed in Muslim memory. Put simply, the narratives allow the minority group to take responsibility for the risk of dhimma rather than assign its dangers solely and wholly to dominant Islam. As such (and again, ironically), these narratives are dreams of integration.