1. Stephen Humphreys, R., Islamic History: A Framework for Inquiry, revised edition (London, 1995), 187ff.Google Scholar

2. al-Qadi, W.: “Biographical Dictionaries: Inner Structure and Cultural Significance,” in Atiyeh, George N., ed., The Book in the Islamic World: The Written Word and Communication in the Middle East. (Albany: State University of Albany Press. 1995), 93–122.Google Scholar For local histories, see more particularly 107-108.

3. This work has a rather complicated (but not untypical) history. See Richard N. Frye's introduction to the translation in The History of Bukhara (Cambridge, Mass., 1954).Google Scholar The most frequently quoted edition is Narshakhi, Muhammad, Tārīkh-i Bukhārā, ed. Ridawi, (Tehran, 1939).Google Scholar The original version was dedicated to Nasr b. Nuh, Samanid amir, in 322/934. This is—according to al-Qadi—the first certain date we can establish for a local biographical dictionary. Narshakhi's work has been studied in detail by Smirnova, O. I. (“Istoriia Bukhary Ṇarshakhi: K istorii slozheniia teksta i o zadachakh ego izdaniia,” Kratkie Soobshcheniia Instituta Narodov Azii 69 (1965): 155–79)Google Scholar who also gives information on the biographical sections not included in the versions that have come down to us.

4. It is not a dynastic history either, even if dynastic elements are evident. It is astounding that this very early text should differ from most of what followed in being so close to what we would term a “local history:” local perspective, local events and developments, traditions, and so forth.

5. Radtke, Bernd, “Theologen und Mystiker in Hurasan und Transoxanien,” Zeitschrift der Deutschen Morgenlandischen Gesellschaft (ZDMG) 136 (1986): 537.Google Scholar

6. Brockelmann, Carl, Geschichte der arabischen Literatur 2nd edition 3 vols. (Leiden, 1943-49)Google Scholar; 3 Supplements (Leiden, 1937-42) (hereafter GAL); Sezgin, Fuat, Geschichte des arabischen Schrifttums vol. 1 (Leiden, 1967)Google Scholar (hereafter GAS). The oldest local history on Marw is mentioned in GAS 1: 351.

7. Abu Muhammad ^ᶜAbdallah b. Muhammad b. Ja^ᶜfar b. Hayyan, known as Abu l-Shaykh: K. ṭabaqāt al-muḥaddithīn bi-Iṣfahān wa^˒l-wāridīn ^ᶜalayhā. Two editions were published almost simultaneously: ^ᶜAbd al-Ghafur ^ᶜAbd al-Haqq al-Husayni al-Balushi, 4 vols. (Beirut, 1407/1987-1412/1992); ^ᶜAbd al-Ghafur Sulayman al-Bundari with Sayyid Kasrawi Hasan, 2 vols (Beirut 1409/1989). The edition in four volumes is preferred because of its better manuscript basis. See also the announcement and discussion by Nasrullah Purjavadi, “Qadīmtarīn tārīkh-i Iṣfahan,” and “Chāp-i dīgar-i kitāb-i Abī ^˒l-Shaykh,” Nashr-i Dānish 10 (1368/1950): 36–39, 48-49.Google Scholar

8. See Rosenthal, F., “Ibn Manda,” EI² 3: 863–64.Google Scholar al-Balushi, in his introduction to the edition of Abu^˒l-Shaykh, seems to take for granted that Ibn Manda wrote a city history of Isfahan. The work is also mentioned in the city history of Qazwin (see below) written in the seventh century; Ibn Manda's book thus seems to have circulated—or at least been known—down to the Mongol conquest. On the histories of Isfahan, see Nurit Tsafrir, “The beginnings of the Ḥanafi school in Iṣfahān,” Islamic Law and Society 5 (1998): 1–22.CrossRefGoogle Scholar I have dealt with this group of sources in a paper presented to the Oxford seminar on “The Shi^ᶜi century in Iranian history” (May 1998).

9. al-Qadi, W., “Biographical Dictionaries,” 107.Google Scholar Apart from that, it is highly debatable whether Bukhara was a “less central Islamic cit[y]” in the fourth/tenth century. Bert Fragner for one thinks that Bukhara perceived itself as an equal of Baghdad and more important than other cities like Damascus or Cairo and that the main axis of the Abbasid caliphate ran from Baghdad to Bukhara during the Samanids’ heyday. See Fragner, B.: Die “Persophonie”. Regionalität, Identität und Sprachkontakt in der Geschichte Asiens (Berlin and Halle, 1999), 45–46.Google Scholar

10. Ed. Sven Dedering (Leyden, 1931), 2 vols. Abu Nu^ᶜaym died in 430/1038, but the last entries in his work are from the 410s.

11. None of the works written for Isfahan are included in GAS, although both extant works fall into the period covered; the lost works are not discussed either as they are in other cases, such as Marw. For a typology of local histories such as biographical dictionaries and their introduction, see also the first part of Parvaneh Pourshariati's contribution to this volume.

12. For a discussion of Samarqandi histories, see Weinberger, J., “The authorship of two twelfth century Transoxanian biographical dictionaries,” Arabica 33 (1986): 369–82.CrossRefGoogle Scholar I overlooked this article when preparing my own “Histories of Samarqand,” Studia Iranica 22 (1993): 69–92.CrossRefGoogle Scholar Out of a larger body of texts, two fragments survive which belong to two different versions of a work which itself has a complex history; Idrisi's contribution is the oldest layer of the extant fragments.

13. This work is the “Persian Qandīya”, written most probably some time in the later sixteenth century. See my “Histories of Samarqand” cited above. The edition prepared by I. Afshar represents only one of the two versions known (Tehran, 1344/1955).

14. See the articles on Samarqand quoted previously, and Mustaghfiri's entry in GAS.

15. Published as an appendix to al-Sahmi, Hamzah b. Yusuf, Ta^˒rīkh Jurjan, ed. al-Din, Nizam (Haydarabad, 1967), 466–500Google Scholar; new edition ed. Muhammad ^ᶜAbd al-Mu^ᶜid Khan (Beirut, 1407/1987), 510-548: “I saw what Abu Sa^ᶜd ^ᶜAbd al-Rahman b. Muhammad al-Idrisi al-Astarabadi had composed as a history [biographical dictionaiy) especially on the people [ulama] of Astarabad to the exclusion of all others,” 510. But Idrisi does not seem to have written a “History of Jurjan” (as one would expect from Hajji Khalifa and GAS). The works by Sahmi and Idrisi are frequently confused: see the discussion of this confusion in the introduction to Mu^ᶜid Khan's edition of Sahmi, 19-20.

16. He is mentioned in the Ta^˒rīkh Jurjān just quoted (old edition, 219; new edition, 260) and in the Ta^˒rīkh Baghdad (vol. 10, 302f.). His fame may be gathered also from the fact that his death in 405 is mentioned by Ibn al-Athir together with the Ta^˒rīkh Samarqand. (al-Athir, Ibn, al-Kāmil fi'l-ta^˒rīkh (Beirut, 1965) 9: 252Google Scholar (year 405).

17. Ta^˒rīkh Jurjān. This work has remained practically untapped down to the present day in spite of its relatively early date of publication. It is not mentioned in GAS, and in GAL, it is included under the rubric “local history” whereas other works of the same type are described under “hadith.” Idrisi was an important source for Sahmi: see the index of the Mu^ᶜid Khan edition s.v. ^ᶜAbd ar-Rahman b. Muhammad al-Idrisi.

18. Bulliet, R., “A Quantitative Approach to Medieval Muslim Biographical Dictionaries,” Journal of the Social and Economic History of the Orient (JESHO) 13 (1970): 196.Google Scholar The work itself was edited in facsimile together with its continuation by R. Frye (London, 1965); the introduction also includes remarks on the history of the text itself and its transmission (10—16). al-Hakim's work is not mentioned in GAS, either.

19. Published in the volume edited by Frye. The author, ^ᶜAbd al-Ghafir al-Farisi, died in 529/1134-35. There is an abridgement of this latter work, al-Muntakhab min al-siyāq li-ta^˒rīkh Naysabūr, by Ibrahim b. Muhammad b. al-Azhar al-Sarifini, edited by Muhammad Ahmad ^ᶜAbd al-^ᶜAziz (Beirut, 1407/1989).Google Scholar

20. Edited by ^ᶜAzizallah al-^ᶜUtaridi in 4 vols. (Haydarabad, 1984-45). R. Mottahedeh worked from the Istanbul manuscript, Ko^ᶜuşlar 1007. See his “Administration in Buyid Qazwin,” in Richards, D. S., ed., Islamic Civilization 950-1150 (Oxford, 1973), 33–45.Google Scholar

21. See the list in GAS. Ghunjar, the author of a city dictionary on Bukhara (frequently quoted by Sam^ᶜani) likewise died in the early fifth/eleventh century (412/1021). Smirnova raises the question whether Ghunjar's work was instrumental in the deletion of biographical material from the extant versions of Narshakhi. (See O. I. Smirnova, “Istoriia.” The histories of Marw above all have been lost, but could probably be in part reconstructed on the basis of Abu Sa^ᶜd ^ᶜAbd al-Karim al-Sam^ᶜani Kitāb al-Ansāb, ed. Yamani, , 13 vols. (Haydarabad, 1962-82).Google Scholar See also the later edition of ^ᶜAbdallah ^ᶜUmar al-Barudi in 5 vols (Beirut, 1408/1988). Sam^ᶜani's work is not a book on genealogy, but a real biographical dictionary according to al-Qadi's criteria (“Biographical dictionaries,” 96). al-Qadi only discusses dictionaries arranged by first names (asmā^˒), not by ansāb and kunan.

22. GAL 1:333f.

23. Hasan b. Muhammad b. Hasan Qummi, Kitāb ta^˒rīkh Qumm (original version); Hasan b. ^ᶜAli b. Hasan b. ^ᶜAbd al-Malik Qummi (translation), ed. Sayyid Jalal ad-Din Tihrani. On the work, see Lambton, A.K.S., An account of the Tärīkhi Qumm,” Bulletin of the School of Oriental and African Studies (BSOAS) 12 (1948): 586–96.CrossRefGoogle Scholar Her more recent “Qum: The evolution of a medieval city,” Journal of the Royal Asiatic Society (JRAS) (1990): 322-39, does not address the text itself, but uses it largely for the earlier periods. See the very detailed analysis of this work, its contents and history, in Drechsler, Andreas, Die Geschichte der Stadt Qum im Mittelalter (650-1350) (Berlin, 1999).Google Scholar

24. Radtke compares Faḍā^˒il-i Balkh to the great dictionaries concerning Damascus and Baghdad in that after an introduction on topography, in the main part, biographies follow (“nach einer topografischen Einleitung folgen im Hauptteil Biografien,” (“Theologen,” 536). This is generally correct. However, Radtke fails to distinguish between biography and biographical dictionary.

25. Mottahedeh notes that in the dictionary on Qazwin, there are approximately three thousand entries - “a surprisingly large number of entries for a relatively small town” (“Administration,” 33). Balkh of course was much larger than Qazwin.

26. Abu Bakr b. ^ᶜAbdallah… Wa^ᶜiz-i Balkhi (original author); ^ᶜAbdallah Muhammad b. Muhammad… Husayni Balkhi (translator), Faḍa^˒il-i Balkh. For the history of the text and a summary of the biographies, see Radtke, “Theologen.” See also Abu Talib Mir ^ᶜAbidin, Balkh dar tārīkh wa adab-i fārsī and U. Berndt's master's thesis, Stadtgeschichten von Balkh. Die Entwicklung der persischen Lokalge schichtsschreibung dargestellt am Beispiel der Faḍā^˒il-i Balḫ und Tariḫ-i šahr-i Balḫ (Halle, 2000)

27. There is a vast literature on Timurid Herat. On geography, see T. Allen. On social history and “history of notables” as well as religious history, see Maria Subtelny's numerous works. For the end of Timurid dominion over Herat and the beginning of Safavid rule, see the works of Maria Szuppe. For even later periods, see Tumanovich, N.N., Gerat v XVI-XVIII vekakh (Moscow, 1989).Google Scholar In contrast, however, for the pre-Mongol history of Herat there is only Gaube, Heinz, “Innenstadt und Vorstadt. Kontinuität und Wandel im Stadtbild von Herat zwischen dem 10. und dem 15. Jahrhundert,” in Schweizer, Günter, ed., Beiträge zur Geographie orientalischer Städte und Märkte (Wiesbaden: TAVO, 1977), 213–41Google Scholar, and his focus on the evolution of the city's architectural shape. It thus offers little information on political and social affairs. Gaube states that Herat did not play an outstanding political role, but was very prosperous, according to the geographical sources of the fourth/tenth century. Gaube used no biographical dictionaries for his study.

28. They are more or less systematically treated in Masson Smith, J., The History of the Sarbadar Dynasty 1336-1381 (The Hague, 1970).Google Scholar Some specific aspects (notably relations with Central Asia) are analysed in Aubin, Jean, “Le Khanat de Čagatai,” Turcica 8 (1976): 16–90.Google Scholar

29. Zakeri, Mohsen, “The ^ᶜAyyārān of Khurasan and the Mongol Invasion,” in Melville, Charles, ed., Proceedings of the Third European Conference of Iranian Studies (Wiesbaden, 1999), 269–76.Google Scholar See also Paul, Jürgen, “L'invasion mongole comme “révélateur” de la société iranienne,” in Aigle, Denise, ed., L'Iran face à la domination mongole (Tehran, 1996), 39–53.Google Scholar

30. Katib Chelebi (Hajji Khalifa), Kashf al-ẓunūn. 1: column 309. The quotation at the end could mean that he was the first to arrange his material alphabetically (not according to generations or classes as had been common practice in the first centuries).

31. Sam^ᶜani, Ansāb, 10: 143. Sam^ᶜani does not mention that Fami is the author of a work on the history of Herat. Since the two were contemporaries and since Sam^ᶜani does not give a date for Fami's death (he begins by “He was born in” but omits the date), this could be explained by the simple fact that Fami's work had not yet been published. This conclusion is corroborated by the fact that Sam^ᶜani nowhere seems to quote Fami.

32. Subki, Ṭabaqāt al-shāfi^ᶜīya (Cairo, 1386/1967) 7: 105f.Google Scholar Subki is quoting his master Dhahabi as having said that Fami's work was incomplete, wa-laysa ta^˒rīkhahu bi-mustaw^ᶜab.

33. Isfizari, Muhammad, Rawżat al-jannāt fl awṣāf madīnat Harāt (Tehran, 1338/1959) 1: 42.Google Scholar

34. GAS 1: 351, no. 3.

35. See Ahmad Dhahabi, Muhammad b., Tadhkirat al-ḥuffāẓ (Haydarabad, 1375/1955) 3: 93.Google Scholar The same author also gives the earlier date, however, see his Mīzān ali^ᶜtidāl, 1: 70 (no. 565).Google Scholar

36. Safadi, Khalil b. Aybak, al-Wāfī bi'l-Wafayāt, vol. 1, ed. Ritter, H. (Istanbul, 1930), 48.Google Scholar Fami has an entry in Safadi vol. 18, ed. Fu^˒ad Sayyid, 155), with references to other dictionaries. Safadi's main source in this case seems however to have been Subki. See also vol. 7, ed. Ihsan Abbas (Wiesbaden, 1969) and vol. 8, ed. Mohammed Yousef Najm (Wiesbaden, 1971), containing the men named Ahmad. Since the authors in question are not found under their given names, it would be pure chance if one came across a quotation from their works in Safadi's gigantic compilation.

37. Dhahabi, Muhammad b. Ahmad, Mīzān al-i^ᶜtidāl, ed. A. N. Jami^ᶜi (Cairo, 1325/1907), 1: 65Google Scholar, no. 531.

38. Subki, Ṭabaqāt, 7: 105–6.Google Scholar

39. A possible explanation for this would be that Katib Chelebi knew Fami's book in a riwāya going back to Abu Rawh. Even a copy made by Abu Rawh could be quoted as a separate work if the copyist had taken such liberties with the text as were common even in the world of hadith.

40. He is included neither under Bazzāz nor under Ḥaddād nor under Harawī. For the Heratis, Sam^ᶜani seems to rely above all on his personal acquaintances and, for earlier generations, on the work of ^ᶜAbdallah al-Hakim on Nishapur. Generally speaking, it is hard to resist the impression that ulema from Transoxiana are better served by Sam^ᶜani since he has made extensive use of works on Samarqand, Bukhara, and Nasaf (by Idrisi, Ghunjar, and Mustaghfiri in that order, see the entries on the corresponding authors in GAS). It is therefore tempting to conclude that in Sam^ᶜani's day, none of the works on Herat had yet gained wide circulation. Another explanation—in the case of Abu Ishaq—may be that this author apparently did not enjoy a very good reputation, see Dhahabi, Mīzān, 1: 70Google Scholar and Tadhkira, 3: 93. Dhahabi's ultimate source in this evidently was Idrisi: āla'l-Idrīsī kāna yaḥfuẓu sami^ᶜtu ahla baladihi yaṭ^ᶜanūna fīhi lā yarḍawnahu, “Idrisi said: He knew the Qur^˒an by heart. I heard his fellow townsmen denigrate him. They did not accept him [as a transmitter of hadith].”

41. GAS, 1: 351.

42. Harawi, Sayfī, Tārīkh-nāma-yi Harāt ed. Siddiqi, M. Z. (Calcutta 1944).Google Scholar For a first approach to this very valuable, but nearly unstudied work, see Potter, Lawrence, The Kart Dynasty of Herat: Religion and Politics in Medieval Iran (PhD diss., Columbia University, 1992)Google Scholar and Mohsen Zakeri, “The ^ᶜAyyārān of Khurasan.”

43. Isfizari, Rawḍat al-jannāt.

44. See the brief remarks in Lambton, A.K.S., “Persian local histories,” in Amoretti, B. S. and Rostagno, Lucia, Yādnāma in memoria di Alessandro Bausani (Rome, 1991), 230.Google Scholar

45. The collection was compiled and edited by Fikri-yi Saljuqi (Kabul, n.d.) The first text is called Maqṣad al-iqbāl-i sulṭānīya and the author calls himself Amir Sayyid ^ᶜAbdallah al-Husayni ma^ᶜruf bi-Asil al-Din Wa^ᶜiz-i Harawi. The other two texts refer to much later periods.

46. In the case of Samarqand, for instance, local “historical” writing seems to have evolved at a somewhat slower pace. The “Persian Qandiyya” which dates to the second half of the tenth/sixteenth century is not fully a guide for pilgrims; that stage is achieved only with the “Samariyya” written in the nineteenth century. See my Histories of Samarqand” for a discussion of this. For Balkh, see Ulrike Berndt, Stadtgeschichten von Balḫ.

47. Sayfi, 25; Isfizari, 1: 142. Fami is further mentioned in Sayfi, 142.

48. Isfizari, 1: 87, 94.

49. The list occurs at the beginning of the part devoted to “history” in a narrower sense, 1: 378-87. At the end of this list, the author remarks that up to this point, Fami has been transmitting a text by Abu ^ᶜUbaid-i Mu^˒addib. He adds that there is another version (riwāya) in the book of Abu Ishaq al-Haddad; however, he does not give this version. If he did know it, it may have come down to him through Fami.

Abu ^ᶜUbaid Ahmad b. Muhammad b. Abi ^ᶜUbaid al-^ᶜAbdi al-Mu^˒addib al-Harawi is known to Khallikan, Ibn, Wafayāt al-a^ᶜyān, 1: 79 (no. 35)Google Scholar who mentions him as having authored a Kitāb al-gharībayn. No work connected with local history is attributed to this author. Lists of governors are not unique to this author; there is another list, for instance, in the Tārīkh Jurjān.

50. Isfizari, 2: 49-55. It thus seems that both books, Fami's and before him Abu Ishaq's, consisted not only of biographical notices, but also had an introduction as was current in this literary genre; faḍā^˒il were common stuff in these introductions. Short notices on events are not so frequently encountered, there is, however, a section on ḥawādi. in Ibn Funduq, Tārīkh-i Bayhaq.

51. ^ᶜAbdallah Ansari, Ṭabaqāt al-ṣūfīya. This work is sometimes called Ṭabaqāt-i mashāyikh-i Harāt in our source.

52. This is of course an allusion to Max Weber's “gesatzte Regeln.” Without taking issue here with the way Weber viewed Islam and Muslim history, we should keep in mind that he had set out to explain why the scientific, industrial, and political evolution characterizing European modernity took place in the West (and nowhere else). This made him take the Western evolution as a model and developments elsewhere as aberrations (to grossly oversimplify). In order to achieve this, he proceeded—perhaps a little too quickly—to compare the West and the East on a phenomenological level. But it is important to compare the functioning of societies after having studied them on their own terms. The much debated question of “urban autonomy” is not a fit subject of comparison as long as we have no real understanding of how a town interrelated with the central government, the surrounding countryside and so on, and above all, how it functioned intra muros, how social action and activity was organized. I do not hesitate to say that, in spite of real progress in the last decades, we are still far from such an understanding. For a debate of Max Weber's view of Islam and the “Islamic city” see my “Max Weber und die ịslamische Stadt” (forthcoming).

53. I prefer not to use “patricians” because this evokes “plebeians” on the other side, and thus tends to produce a binary opposition along what could be interpreted as a class divide. Also, because “patricians” is too closely associated with ancient Roman history, I cannot imagine a way of dissociating the term from that context. Bulliet himself introduced it only tentatively. The term “notables” to denote a given section of urban society was introduced into scholarship on Muslim societies by Hourani, A. in “Ottoman Reform and the Politics of Notables” first published in Polk, W. R. and Chambers, R. L., Beginnings of Modernization in the Middle East (Chicago, 1968)Google Scholar and reprinted in 1981 in A. Hourani, The Emergence of the Modern Middle East: Collected Papers. Besides the discussion in Bulliet, Richard, The Patricians of Nishapur (Cambridge, Mass., 1972)Google Scholar, and his “Local Politics in Eastern Iran under the Ghaznavids and Seljuqs,” Iranian Studies 11 (1978): 35–56CrossRefGoogle Scholar, mention must be made of Lapidus, Ira M., A History of Islamic Societies (Cambridge, 1988)Google Scholar with further references to his earlier work, and of Shoshan, Boaz, “‘The Politics of Notables’ in Medieval Islam,” in Asian and Arican Studies 20 (1986): 179–215.Google Scholar I have myself discussed the “politics of notables” extensively in my Herrscher, Gemeinwesen, Vermittler: Ostiran und Transoxanien in vormongolischer Zeit (Beirut/Stuttgart, 1986).Google Scholar

54. As a recent example, see N. Tsafrir, “The Beginnings of the Hanafi school.”

55. See Chabbi, Jacqueline, “Remarques sur le développement historique des mouvements ascétiques et mystiques au Khurasan,” Studia Islamica 46 (1977): 5–77.CrossRefGoogle Scholar

56. For a discussion of these attempts, see Humphreys, Islamic History (the passage quoted in note 1 above).

57. See Bayhaqi, Abu'l-Fadl, Tārīkh-i Bayhaqī (Tehran 1324/1955), 588Google Scholar, but see the following note for a general appreciation of this source's biases.

58. This had already been the case before the final disaster. Events at Nishapur some years before Dandanaqan are particularly well documented in a passage in Bayhaqi, translated by Bosworth, C. E. in The Ghaznavids (Edinburgh, 1963)Google Scholar, 252ff. However, there might be reason to read this passage with more caution for it should not be taken at face value. It is presented as the reproduction of a letter written by a faithful servant of Mas^ᶜud's court to his sultan. In his letter, he gives an account of the city's surrender to the Seljuqs. The notables had held a meeting where they had finally decided not to resist the Turkish onslaught. They knew that they could not expect any help from the Ghaznavid governor who had already left together with the garrison. The octagenarian qadi Sa^ᶜid then said that on principle, the city should not resist: Subjects are not expected to intervene in military affairs, the city's master is Mas^ᶜud, and it is his task to defend it. If he is unable to do so, the citizens should accept the new rule. This report must be read in the context of the whole book; as is generally known, Bayhaqi wanted to show that Mas^ᶜud was responsible for the loss of Khurasan.

59. Ibn al-Athir, 9: 483. Also Isfizari, 1: 387ff. The Ghaznavid garrison was practically non-existent. This conclusion was reached already by de Beaurecueil, Laugier, Khwādje ^ᶜAbdullāh Anṣāri (396-481 H/1006-1089), mystique hanbalite (Beirut, 1965), 85.Google Scholar His conjecture is supported by the simple statement in Isfizari that Herat was defenseless after Dandanaqan, 1:388 (quoting Fami).

60. Ibn al-Athir, 9: 488. This cannot be taken to mean that this was a pro-Ghaznavid movement; this would not fit with the Heratis’ previous behavior.

61. Isfizari, 1:, 388 (dated 432/1040-1). […] bar khāst wa bī manshūr wa mithāl āghāz-i taghallub nihād wa bar shahr wa wilāyat mustawlī shud. The Usmi family is represented by Abu ^ᶜAbd Allah b. Abi Dhul al-Usmi in Sam^ᶜani. This is Muhammad b. al-^ᶜAbbas b. Ahmad […] al-Dabbi al-Usmi, a noted hadith transmitter and prominent scholar. He is given the title ra^˒īs. He died in Safar 378 (late-May 988) Kitāb al-Ansāb 4: 204-6. This al-Usmi is probably a forebear of the Abu Muhammad mentioned by Fami. It is evident from Sam^ᶜani that the al-Usmi family had landholdings. This representative of the family allegedly spent all the tithe (^ᶜushr) due from these holdings on charity so that most of the city's paupers were being fed by him. This Bu ^ᶜAbdallah-i Bu Dhul is also mentioned in Ansari, Ṭabaqat al-ṣūfīya, 597 as ra^˒īs-i Harāt, as walī and as “father of the Heratis” (pidar-i Harawigān). The title ra^˒īs does not necessarily denote an official position in my view. See my Herrscher, Gemeinwesen, Vermittler.

62. Ibn al-Athir, 9: 506.

63. Ibn al-Athir, 10: 34. See also Isfizari, 1: 388-89. No date is given for Bayghu's occupation of Herat.

64. Ahmad b. Sahl came from an old family of Iranian noblemen from the Marw oasis. Soon after his appointment to Herat, he rebelled against his Samanid overlords who “had made promises which they did not keep afterwards.” The history of his revolt is retraced in Ibn al-Athir, 8: 117-118. See also Bosworth, C. E., “Aḥmad b. Sahl b. Hāšem,” in EIr 1: 643–44Google Scholar and my study The State and the Military: The Samanid Case (Bloomington, IN: Research Institute for Inner Asian Studies, 1994).Google Scholar

65. Isfizari, 1: 384.

66. Ibid. Both quotations are taken from the list of governors Isfizari transmits from Fami who in turn had it from Abu ^ᶜUbaid-i Mu^˒addib. I have not been able to trace the elected persons, but the nisba “Dabbi” is also given to one of the leading families in Herat, the Usmis (see note 61). The genealogy given for the representative of the family in Sam^ᶜani does not help in identifying these two amirs.

67. Hawqal, Ibn, l-Qasim, Abu, Kitāb al-masālik wa'l-mamālik, (Leiden, 1892), 2: 437–38.Google Scholar Isfizari dates this event 341/952-3 (1: 385-86. Ibn Hawqal visited Herat in 357/968 (Togan, Z.V., “Herat,” Islam Ansiklopedisi, 5: 430–31Google Scholar) and found the walls destroyed by order of the Samanid ruler.

68. Strained relations between the central government and its appointees on the one hand and the notables of Herat on the other are alluded to in a brief reference in ^ᶜUtbi/Jurfadhaqani: Abu ^ᶜAli Simjur and Fa^˒iq, two military commanders with all the features characterizing warlords, went to Herat “in order to save this city from its stubborn resistance and to take the regional army with them,” see Tarjuma-yi tārīkh-i Yamīnī, (Tehran, 1334/1955), 105.Google Scholar

69. It is perhaps significant that the Heratis, in their foundation legends, did not ascribe the building of their formidable walls to some king, not even a mythical one such as one of the well-known empire builders of Iranian mythical history. Instead they ascribed them to the designs and strategems of a “notable” woman, thus to a local initiative. From the very beginning, the city walls are thus seen to serve as a defense against “unjust” rulers, and more particularly against the (always or typically) greedy taxation policy of the central government. External enemies are not necessarily the main concern. It should be noted that rulers not residing in a city are more prone to tear down its walls than to repair them.

70. The early Islamic history of this group is now presented in Zakeri, Mohsen, Sasanid Soldiers in Early Muslim Society. The Origins of ^ᶜAyyārān and Futuwwa (Wiesbaden, 1995).Google Scholar For the decisive role ^ᶜayyar groups had in Herat in the period immediately following the Mongol conquest, see Zakeri, “The ^ᶜAyyārān of Khurasan and the Mongol Invasion.”

71. Isfizari, 1:385. The Risāla-yi mazārāt-i Harāt would lead one to believe that there were ^ᶜayyār groups always active.. In the time of Abu Muslim, a woman lived in Harat called Bibi Sitti. “At the beginning of her [mystical] life, she had a husband called Abu Nasr, a man who roamed around at night (shab-raw). Both of them were engaged in ^ᶜayyārīi” (11). In this context, the term probably is to be understood as robbery. It is perhaps worth noting that it was considered possible for women to be ^ᶜayyārān.

72. Their history is told in the Tārīkh-i Sīstān where their activities take up a good portion of the text. See the article by C.E. Bosworth in this issue.

73. See for instance Bal^ᶜami, Tārīkh-i Bal^ᶜamī, 952, 1203.Google Scholar They became a plague after the Seljuq victory at Dandanaqan. (See Ibn al-Athir 9: 483.) Sometimes, they may have been integrated into the “official” army, or at least such an integration may have been attempted, as shown in Bukhara in al-Qadi b. Zubair, Kitāb al-dhakhā^˒ir wa l-tuḥaf (Kuwait, 1959), 145.Google Scholar See also Bosworth's translation of this text in An alleged embassy from the Emperor of China to the Amīr Naṣr b. Aḥmad: A contribution to Samanid history,” in Minovi, M. and Afshar, I., eds., Yādnāma-yi Irānī-yi Mīnūrskī (Tehran, 1969), 17–29.Google Scholar They participated in the defense of Balkh against the Seljuqs, this time at the command of the Ghaznavid governor (Bayhaqi, Tārīkh-i Bayhaqī, 643Google Scholar). They fought at the side of the “last Samanid” against the Qarakhanids at Samarqand, see Gardizi, Zayn al-akhbār (Tehran, 1347/1969), 176.Google Scholar These are just a few examples. See also my Herrscher, Gemeinwesen, Vermittler for more.

74. Gardizi, 165. On Abu ^ᶜAli, see note 68.

75. See above, note 68

76. The question of military slavery cannot be addressed in this context. I am referring here to a remark by Michael Cook that “[I]t is remarkably hard to find in Islamic history instances of what might be called citizen armies—armies locally recruited, by a state identified with the area in question, from a settled population that was not tribal” (Cook, Michael, “Islam: A Comment” in Baechler, Jean, et. al., eds., Europe and the Rise of Capitalism [Oxford, 1988], 133).Google Scholar It seems to me that we are so used to placing all military activities within the realm of the state that such forces as those mentioned for Herat are likely to escape our attention. The Herati “army” certainly was not recruited by a “state,” however identified, since it was at least occasionally actively engaged in fighting against men serving this very state (in the person of the Samanid ruler).

77. Isfizari, 2: 54. This was perhaps not as exceptional as we might think at first. Lambton tells a very similar story in Landlord and Peasant (Oxford, 1969), 307–308.Google Scholar

78. This term is usually used to denote the most respected person in a given place. No official appointment is necessary. The title shaykh al-islām, used for the father of the hero of the following story, ^ᶜAbdallah Ansari, may be seen as the religious counterpart of mihtar.

79. Aside from this identification, we have little information on ^ᶜAbd al-Hadi Ansari. He is not mentioned in the biography devoted to his father in Jami's Nafaḥāt al-uns. He is mentioned, though, in the Risāla-yi mazārāt-i Harāt, where he is given his own entry that confirms the year of his death as 493/1099-1100. This short biography says, “When the Batinis ruled over Herat, they caused him much grief, and finally killed him in the citadel. They buried him in the same spot but some days later, these disorders came to an end, and his body was exhumed from that burial ground and transferred to Gazurgah” (34). It is difficult to tell whether the author is relying here on material transmitted by Fami; another possible source could be a family and/or a shrine tradition.

80. It is well known that in practically all cities of Khurasan (and in many other places throughout Iran) during the pre-Mongol period, confrontations were continuous, often serious enough to cause bloodshed, between two, rarely three, parties in the city. These parties were sometimes defined by religious markers such as a law school. The fights themselves are called ta^ᶜaṣṣub, ^ᶜaṣabīya being the term used for what united the fighting parties. The most prominent example is the city of Nishapur, which met an untimely end in struggles of this kind. See Bulliet, The Patricians of Nishapur.

81. The Risāla-yi mazārāt-i Harāt states explicitly that ^ᶜAbd al-Hadi is the son of the celebrated shaikh.

82. See S. Laugier de Beaurecueil, Khwadje ^ᶜAbdullah Ansari.

83. This probably is the background to the information that the Heratis stayed loyal to Sultan Mawdud b. Mas^ᶜud. See above, notes 60 and 62

84. For roughly a century preceding the Mongol invasion, the Persian sources provide little information about Herat. This is not surprising: Fami wrote his book about 1130 CE (if we presume that Sam^ᶜani did not know it because it had not yet been published), and other books, if they were ever written, did not survive the destruction of the city at the hands of the Mongols. The principal source for the following section is Ibn al-Athir who does not reveal his sources. Florian Schwarz has shown that Ibn al-Athir must have used some “History of Nishapur” (a regional dynastic history) for information about Ay Aba (a warlord of Nishapur and one of the contenders for power in the second half of the sixth/twelfth century) and his successors. See Florian Schwarz, Der Sultan von Ḫurāsān (masters thesis, Tübingen, 1993). (I am grateful to Florian Schwarz for providing me with a copy of this work.) I consider 536/1141 to mark the beginning of the end for the Seljuqs because in that year, Sanjar was utterly defeated by the Qarakhitay (near Samarqand), and this loss seems to have done irreparable damage to his hayba (authority of rulership). Various regional dynasties—at least in theory subservient to Seljuq overlordship—began their bid for independence at that moment, among them, of course, the Khwarazmshah.

85. Ibn al-Athir, 11: 151. The Ghurids were a regional dynasty based in what is now central Afghanistan and one of the major players in Khurasan in this period. Their history is best followed in Muhammad Abdulghafur, The Ghorids. History, Culture and Administation (Doctoral dissertation, Hamburg, 1960).

86. It is well known that some Ghuzz took Sultan Sanjar prisoner in 1153 CE; he spent three years in captivity. The questions surrounding these tumultuous years are best described in Schwarz. The center of their activities and in the Badghis, too close to Herat to be ignored.

87. Ibn al-Athir, 11: 227 (year: 552).

88. Ibid., 294.

89. Ibid., 311 (year: 559).

90. Such consequences are graphically described by Funduq, Ibn, Tārīkh-i Bayhaq, (Tehran, 1317/1939).Google Scholar See also Parvaneh Pourshariati's contribution to this volume and my Herrscher, Gemeinwesen, Vermittler 118-121.

91. For the whole story, see Ibn al-Athir 12: 225-7, 260-5; see also ^ᶜAta Malik Juvayni, Tārīkh-i Jahān-gushā (London and Leiden, 1912-1937), 2: 62–65.Google Scholar

92. Ibn al-Athir, 12: 207.