1. Sahlins 1981, 1985.

2. White 1981, 14.

3. Ibid., 20.

4. The center of concern in this study is the frontier between British India and Afghanistan (see map 2). This frontier runs south to Baluchistan, but the area that I am primarily interested in is that bordered by Chitral in the north and the tribal territories surrounding the Khyber Pass in the south. The two dominant commercial and political centers in this region are the Afghan city of Jalalabad in the west and the Indian (later Pakistani) city of Peshawar. At the heart of this area, residing in the bare hills and mountains that straddle the border in this area are a number of politically independent Pakhtun tribes, including the Mahmund, Mohmand, Safi, Shinwari, Khogiani, and Afridi. As one moves further from the frontier and the mountains into the broader plains that ring Peshawar in the east and Jalalabad in the west, one encounters more peasant populations that have been traditionally under the control of the state and linked directly to the national economy.

5. British control over Afghanistan's foreign policy remained a sticking point for years and provided one of the explicit rationales for the third and final Anglo-Afghan war in 1919, a war that was provoked by Abdur Rahman's grandson, Amir Amanullah Khan (see Adamec 1967).

6. On the invention and spread of the nation-state, see especially Seton-Watson 1977; Anderson 1983; Gellner 1983; and Hobsbawm 1990.

7. One of the leaders of the abortive Panjshir uprising was Ahmad Shah Massoud, who would later gain fame during the jihad as the commander of the Panjshir Valley resistance. Other equally fruitless attacks occurred that same day in Laghman, Surkh Rud, and Paktia. All were organized by the Muslim Youth Organization, which was severely damaged by the failure of these attacks and the capture of most of its leaders.

8. The best example of this sort of analysis applied to Afghanistan is the introduction written by Richard Tapper to his edited volume on tribe-state relations in Iran and Afghanistan (1983).

9. I encountered one notable exception to this rule while conducting research in an Afghan refugee camp in 1984. That exception was the Ahmadzai tribe, most branches of which come from the region south and east of Kabul. The Ahmadzais are a numerous and successful branch of the Ghilzai confederacy, which essentially ceased to exist as anything more than a name following Amir Abdur Rahman's suppression of a Ghilzai tribal revolt in 1888. Due in part to their proximity to Kabul and an effective leadership, the Ahmadzais continued to wield considerable influence in national politics at a time when other tribes were fragmenting and losing ground. Following the Marxist coup, the tribe was split as some members (the most prominent of whom was Dr. Najibullah, the leader of the Parcham party, who succeeded Babrak Karmal as president of the Democratic Republic of Afghanistan) sided with the government and others took up arms against it. Again, because of their proximity to Kabul, many of those who resisted were forced to flee the country early in the war, and those who did were scattered in refugee camps throughout the North-West Frontier Province of Pakistan.

Despite these ruptures, the Ahmadzais have continued to maintain a degree of coordination that few other tribes evince. Within the camp I studied, the Ahmadzai groups were unique in having maintained their coherence as segmentary tribal sections—that is, all of the male heads of house in a given residential group were related by descent from common (paternal) ancestors. Likewise, the Ahmadzai khans in the camp where I worked met periodically in assembly with the khans of other tribal sections in other camps, and the tribe kept an office in Peshawar, as they once had done in Kabul. These and other acts of group solidarity indicated that the tribal idea was alive and well, at least for this one group. (For background on the economic and social situation of the Ahmadzais and other Ghilzai tribes prior to the war, see Jon Anderson's excellent ethnographic articles on the Ghilzai of Logar and Ghazni, especially “Tribe and Community among Ghilzai Pashtun” (1975) and “Khan and Khel: Dialectics of Pakhtun Tribalism” (1983).

10. See Azmat Hayat Khan, “Afghan Resistance and National Leadership” (1981).

11. In noting the greater attention paid to organizational than to symbolic dimensions of political relationships, Steven Caton has argued that more research is needed that would “focus on the significant individual and not just segmentary groups—using culturally laden signs in concrete acts of communication” (Caton 1990b, 99). In Caton's view, concern for the nature of political groupings has obscured the role of individuals and clouded as well the nature of Ibn Khaldun's own model, which “hinges crucially on a Weberian notion of charismatic personality, in the guise of either the desert chieftain or the prophet” (Caton 1990b, 89–90).

12. Two of the first articles that I wrote on my research dealt with the maintenance of cultural forms by refugees (Edwards 1986d, 1990). While I still hold to the conclusions I presented there, I also recognize now that the situation was more complicated than I realized and that the urge for order in my own mind led me to focus on small triumphs of cultural survival that individual families and groups were achieving while underrepresenting the larger picture of chaos and dissolution.

13. For examples of Rafiq Jan's poetry and an analysis of its significance, see Edwards 1993a. My enthusiastic response to Rafiq Jan's poems was undoubtedly conditioned by my earlier reading of Michael Meeker's book, Literature and Violence in North Arabia. Meeker's poetic analysis offered an answer to a problem faced by Middle Eastern specialists interested in symbolic analysis, and that problem had to do with the relative paucity and stereotypical nature of ritual expression in most Middle Eastern societies. In the late 1970s and early 1980s, many of the influential symbolic studies that were then shaping the discipline centered around ritual, but in the Middle East, ritual did not seem to be at the center of cultural identity and practice, at least not in the distinctive way that it was in other cultural areas. Meeker's focus on poetry and narrative circumvented this difficulty by offering an alternative arena within which to think through symbolic issues. Meeker, it should be admitted, was not the first or only Middle Eastern scholar to call attention to poetry. Both Alois Musil (1928) and Ignaz Goldziher (1967; originally 1889–1890) had written on Arabic tribal discourse generations earlier, while a number of other scholars published significant works having to do with poetry and poetic production shortly after Meeker's book came out. These scholars include Samatar (1982), Abu-Lughod (1986), and Caton (1990a).

14. Such a notion was probably never realistic given that Massoud is an ethnic Tajik (and therefore never one who would have seriously contended for power in the past), while most of those with whom I interacted (and the vast majority of the refugees in general) were Pakhtuns. Pakhtuns constitute a slim majority of the Afghan population, always controlled things in the past, and were unlikely to cede control now to a non-Pakhtun. But the very fact that Pakhtuns of my acquaintance were willing to discuss the possibility of having a Tajik in a position of political power illustrates how little faith they had in their own immediate Pakhtun political leaders.

15. In the mid-seventies, there were not that many people interested in the oral history of early uprisings and fewer still who were focusing on Islam as a dynamic social force in Afghan society. With his idiosyncratic and omnivorous appetite for all things Afghan, Louis Dupree had dealt with both oral history and Islam in a number of articles, and so too had a few other scholars (e.g., Poullada 1973; and Canfield 1973). But the majority of scholars (and I probably would have been among them if I had been conducting research then) seemed to view Islam as part of a quickly receding and largely irrelevant past that had little to offer outside antiquarian interest. Afghanistan scholars were not alone in holding to this view, of course. The same sense pervaded scholarship of the sixties and early seventies throughout the Middle East, and a classic expression of this tendency can be seen in the introduction of Richard Mitchell's classic work on the Muslim Brothers, in which he obliquely apologizes for choosing a subject that “has had its moment in history, and that for very few of [whose] leaders will historians reserve a place larger than a footnote” (Mitchell 1969, xv).

16. There were several reasons why rural elites chose to settle in Peshawar. First, Peshawar was centrally located and contained most of the relevant offices (e.g., the offices of the Commissioner of Afghan Refugees and of the resistance parties) that they would need if they were to secure assistance and employment. Secondly, Peshawar has long enjoyed the status of being the hub of Pakhtun culture. This status goes back centuries, and while it was partially diminished by the imposition of a national border at the end of the last century, that border has always been porous, and Pakhtuns, whether nominally of Afghan or Pakistani citizenship, have never relinquished their affection for the place. A third factor is that whereas the rural elites with whom I dealt were not wealthy on any objective scale, they had greater resources to draw on than most refugees and could consequently afford Peshawar rents. Thus, in addition to whatever resources elites received from either the political parties or the Pakistan government, many also continued to receive income from land that tenants farmed throughout the war. The owner's share was far smaller than it had been before the war, but the continued payment of the share provided some income while also preserving the owner's claim to the land into the future.

17. While I have taken the tack of focusing on just a few individuals in this book, I have availed myself of some of these interviews in other published articles (1986b, 1986c, 1987, 1989, 1993b, 1993c, and 1995).

18. Geertz was not the first or only person to recognize the value of life histories for the study of cultures (Paul Radin and Edward Sapir both come to mind as early pioneers in this field), but he was the one with whom I was most familiar and who has probably had the greatest influence in developing this line of research among anthropologists of the Middle East. Whether the debt is explicitly acknowledged or not by individual authors, Islam Observed clearly anticipates a number of influential works that center on Middle Eastern lives, including Crapanzano 1980; Dwyer 1982; Munson 1984; Eickelman 1985a, 1985b, 1991; Mottahedeh 1985; Loeffler 1988; Friedl 1989; Lavie 1990; Beck 1991; Abu-Lughod 1993; and Burke 1993. For a general review of life histories in anthropological research, see Peacock and Holland 1993; and see Schwartz 1987, for an excellent example of a “Durkheimian” life history focusing on the social uses to which great lives are put.

19. See especially Nagy 1979.

20. While the ethnographic literature on Islam in Afghanistan was relatively limited prior to the war, relevant ethnographic work had been done in the neighboring North-West Frontier Province of Pakistan by Fredrik Barth (1959a). Barth's initial work has also been supplemented by Ahmed (esp. 1976, 1980, 1983), Lindholm (1982, 1986, 1992), Grima (1992), and Jahanzeb (1985), all of whom provide interesting insights into Islam's role in Pakhtun tribal society. Recent work pertaining to Islam in Afghanistan includes Shahrani and Canfield 1984; J. Anderson 1984; Ashraf Ghani 1978, 1983, 1987; M. Mills 1991; Roy 1986; and Edwards 1986b, 1986c, 1986d, 1993b, 1993c, and 1995.

21. The cultural nature of perception and of narrativity is a principle deeply rooted in anthropology, going back most significantly to the work of Edward Sapir and Benjamin Whorf. Others who have foregrounded texts and textual questions in ways that I have found fruitful include Basso (1986), Rosaldo (1986), and Bowen (1993); and—among scholars working in the Middle East—Goldziher (1967), Mottahedeh (1980), Abu-Lughod (1986), Dresch (1989), Caton (1990a), and Messick (1993).

22. As in the case of my decision to focus on “axial figures,” I cannot claim originality for my view of the importance of particular kinds of historical texts for the understanding of other, historically situated minds since I had been aware of Le Roy Ladurie's Montaillou (1978) before embarking on my fieldwork and read works by Darnton (1984) and Ginzburg (1985), among others, on my return.

23. See B. Anderson 1983 for a perceptive analysis of the symbolic and social implications of this division. Also, see Lindholm 1992 for a similarly insightful comparison between Afghan and Moroccan conceptions of the division between tribe and state.

24. It can be argued, of course, that documents like these are incommensurable with oral histories such as the ones I use to discuss honor and Islam. Histories of any sort reflect the times in which they are recorded. Abdur Rahman's autobiography and proclamation are solely of the past and represent its concerns, whereas the oral histories I collected represent the present of the teller along with the past of the told. This objection would have merit if I were concerned either with the performative context in which the oral histories were told or with the way in which they were being strategically deployed in the present. But what I am interested in here are the traditional moral orders that held sway during a significant moment in Afghanistan's recent past, and the oral sources on which I rely are not only the best sources available for considering these moral orders, but they also give every appearance of being relatively fixed.

I will discuss the fixity of these stories at appropriate points in the relevant chapters, but it is worth noting here the general grounds on which I base this judgment. As indicated previously, almost all of the stories told about the Mulla of Hadda are miracle stories, and they are virtually identical to those associated with other Muslim saints from India and Afghanistan. Having read several hundred of these stories from different times and places, I have become convinced not only that the stories are relatively stable, but that stability is one of their key features. With regard to the story of Sultan Muhammad Khan that is the centerpiece of chapter 2, I have less proof that it has not been transformed over time by different tellers in different contexts, but it is clear from certain aspects of the story itself that the story is not Safi's own, that it has been told to him by others in a relatively set format, and that its style and structure offer limited scope for revision. Safi does make occasional remarks in the course of telling the story that are intended to clarify points that might be obscure to me, but the rhetorical bracketing of these remarks makes it clear that they are meant to stand outside the narrative itself.

25. James 1898, 5.