1. For instance, the scribe who copied the London manuscript of 841/1438, of which the sigla in Khaleghi-Motlagh's edition is L3, tends to maintain archaic readings of his exemplar because he is neither very literate nor particularly meddlesome. He therefore, copies what he sees in his exemplar even if he does not understand it. This is helpful to the collator because this scribe's corruptions, even when mutilated in terms of sense, are orthographically close to the correct reading and are helpful in restoring the text. By contrast, the copyist of the Paris codex of 844/1441, i. e., the codex P, is quite learned, and his changes are often done so well that they are difficult to detect without help from other manuscripts.

2. Although a number of medieval Persian and Arabic texts have more than one authorial revision (e.g. Mirṣād al-^ᶜibād, Ṭabaqāt al-aṭibbā^˒, and Āthār al-bilād), the thorny problems of deciding between various authorial revisions is not nearly as difficult in medieval texts as it is in editing of modern authors. See Thomas Tanselle, G. “Editorial Problem of Final Authorial Intention,” in idem, Textual Criticism and Scholarly Editing (Charlottesville and London, 1993), 27–72Google Scholar; “Classical, Biblical, and Medieval Textual Criticism and Modern Editing,” idem, Textual Criticism and Scholarly Editing, 274-325; Haravi, Mayil-i, Naqd wa taṣḥīḥ-i mutūn (Mashhad, 1990), 288ff.Google Scholar The matter of Firdawsi's revision of the Shāhnāmah, in my opinion, has been highly exaggerated for reasons that have nothing to do with scholarship. There are no textual reasons to believe that he revised the Shāhnāmah to any great extent beyond including his praise of Mahmud.

3. For a theoretical discussion of this notion see Greetham, David, “A Suspicion of Texts,” in idem, Textual Transgressions: Essays Toward the Construction of a Bibliography (NY and London, 1998), 198–219.Google Scholar

4. Stemmatics or the Lachmannian method, also called the “genealogical method,” is most closely associated with the name of Karl Lachmann (1793-1851). This school of editing attempts to reconstruct the archetype of all existing witnesses by a three-step process that involves recensio, namely the construction of a stemma that shows the genealogical relationship of all extant witnesses; examinatio, which determines the authority of the variant readings, and divinatio, which is the process of removing errors by emendation. Although hailed as the scientific method in the nineteenth and twentieth centuries, the Lachmannian method came under attack from the followers of the “best text” school, championed by the French scholar, Joseph Bédier (1864-1938) who favored following the text of one “best manuscript” except in the case of its glaring errors. See Bédier, J., “La Tradition manuscrite du Lai du l'Ombre: réflexions sur l'art d'éditer les anciens textes,” Romania 54 (1928): 161–96, 321-56.CrossRefGoogle Scholar This piece was later reprinted in pamphlet form in 1970. Italian scholars, who questioned both Lachmann's method and the appropriateness of attributing it to him, launched the most devastating attacks on the genealogical school. See Pasquali, Giorgio. Storia della tradizione e critica del testo (Florence, 1934)Google Scholar and Timpanaro, Sebastiano. La genesi del metodo del Lachmann. (Florence, 1963).Google Scholar Timpanaro's seminal work was translated into German by Irmer, Dieter as Die Entstehung der Lachmannschen Methode. (Hamburg, 1971).Google Scholar

5. Housman, A. F., Selected Prose, ed. Carter, John (Cambridge, 1961), 36.Google Scholar I cannot help but suspect that old manuscripts endear themselves to most editors partly because of their status as “relics.”

6. Several scholars beside Timpanaro (see note 4) have argued against the genealogical method. See for instance, Colwell, Ernest C., “Genealogical Method: Its Achievements and Limitations,” Journal of Biblical Literature 66 (1947): 109–33.CrossRefGoogle Scholar E. T. Donaldson argues in favor of abandoning the method all together. See “The Psychology of Editors of Middle English Texts,” in his Speaking of Chaucer (NY, 1970), 102–18.Google Scholar

7. Cited in Kenney, E. J., The Classical Text: Aspects of Editing in the Age of the Printed Book (Berkeley, Los Angeles, London, 1974), 130.Google Scholar Housman advised substantially the same thing when he wrote: “Knowledge is good, method is good, but one thing beyond all others is necessary; and that is to have a head, not a pumpkin, on your shoulders, and brains, not pudding, in your head.” See “Application of Thought to Textual Criticism,” Proceedings of the Classical Association, 18 (1921): 84.Google Scholar

8. Throughout this paper, Kh refers to the Khaleghi-Motlagh critical edition of the Shāhnāmah, even if the text is from volumes edited by myself or Dr. Khatibi in Iran. The two Shāhnāmah editions that lack apparatus critici, namely those of Jules Mohl, and the recent Tehran text (1374/1994), are called Mohl and Tehran respectively. The Moscow critical edition is indicated as Moscow. Manuscript sigla are the same as in Kh, but they are indicated by Latin letters. Thus, the primary MSS are: London 675/1276 = L; Istanbul 731/1330 = S; Cairo 741/1341 = Q; Karachi 751/1351 = K; London 891/1486 = L2; Istanbul 903/1498 = S2. The secondary codices are: Leningrad 733/1333 = LN; Cairo 796/1394 = Q2; Leiden 840/1437 = LY; London 841/1438 = L3; Paris 844/1414 = P; Vatican 848/1444 = V; Leningrad 849/1445 = LN2; Oxford 852/1448 = O; Berlin 894/1489 = B. Al-Bundari refers to the Arabic translation of the Shāhnāmah by Fath ibn ^ᶜAli al-Bundari, which was effected in 620-21/1223-24.

9. This is how George Kane defined the term. See his “Conjectural Emendation,” Medieval Literature and Civilization: Studies in Memory of G. N. Garmonsway, ed. Pearsall, D. A. and Waldron, R. A. (London, 1969), 155–70.Google Scholar

10. Older or more authoritative codices tend to preserve older readings in a less mutilated state more often than do the less authoritative manuscripts. They also have a better order of verses and hemistiches and fewer interpolations and lacunae. Even interpolations are almost always found as discrete pieces, inserted at specific points in the narrative of codices that generally belong to the same stemma. As such, one can reasonably suppose that these interpolations found their way into the text from some marginal note on the pages of some lost archetype.

11. All but the first of these examples are taken from the story of Alexander in volume six of this edition, which is still being prepared for publication by myself. The first of my examples is taken from volume five, which was edited and published by Professor Khaleghi-Motlagh.

12. This reading is metrically deficient in accordance with the inadequacies of the scribe of this codex. That is, more often than not, when this copyist attempts to restore a reading that makes no sense to him, he does so incompetently and makes a metrical or orthographical blunder. These errors are usually very good clues that something might be wrong with the text at that spot.

13. Readings of secondary MSS are put in parentheses in order to distinguish them from the readings of the primary codices.

14. Followers of Harvard's tribal religion of “oral formulaic theory,” tend to ascribe such textual variety to an illusory “poetic oral tradition” of the Shāhnāmah. Their position however, is patently absurd, and is not taken seriously by a single student of the text of the poem. Those who subscribe to this bizarre view tend to be entirely innocent of knowledge of the epic's manuscript tradition. However, since I have already discussed the matter in some detail elsewhere, I shall not belabor the point here. See Omidsalar, M., “Yes Virginia, Manuscripts Matter: Orality, Mouvance, and Editorial Theory in Shāhnāma Studies.” Jerusalem Studies in Arabic and Islam 26 (2002)Google Scholar: (in press).

15. The type of contradictions and irrational statements typical of the Homeric poems and oral poetry are non-existent in the Shāhnāmah. The Shāhnāmah's “contradictions,” if such can be shown, are contradictions that occasionally mar literary works. For instance, F. Scott Fitzgerald is notorious for making conflicting statements in his novels. I know of no Fitzgerald critic who has interpreted these contradictions to be the proof of the oral character of his prose. For examples of these see Bruccoli, Mathew J., “The Text of The Great Gatsby,” in Scott Fitzgerald, F., The Great Gatsby, preface and notes by Bruccoli, Matthew J., (NY and London, 1995), 191–95.Google Scholar

16. Ishaq ibn Ibrahim al-Nishaburi (5th/11th century), Qiṣaṣ al-anbiyā^˒:Dāstān-hā-yi payghāmbarān, ed. Yaghma'i, H. (Tehran, 1980), 326–27Google Scholar; ^ᶜAjāyib al-dunyā, 430.

17. E.g., Muhammad b. Mahmud-i Hamadani, ^ᶜAjāyibnāmah, ed. Mudarris-i Sadiqi, J. (Tehran, 1375/1996), 3.Google Scholar

18. ^ᶜAbd al-Malik b. Muhammad Tha^ᶜalibi, Ghurar akhbār mulūk al-Furs wa siyarihim, ed. Zotenberg, H. (Paris, 1900), 427–28.Google Scholar

19. The exception is al-Bundari's Arabic translation of the Shāhnāmah, in which the word has been translated as jām. One does not know of course, if this is the actual reading of the MS or the editor's interpretation of the MS's “un-dotted” text. It may be that the Shāhnāmah MS from which al-Bundari worked also recorded jām here, which makes al-Bundari's text an “old error,” but an error nonetheless.

20. During the present author's childhood, it was common in the provinces of Fars and Isfahan to keep yogurt, cheese, syrup, and cured meat in skin-bags. These bags were generically known by the name khām and were usually identified according to their contents, e.g., khām-i māst, “bag of yogurt,” khām-i panīr, “bag of cheese,” etc.

21. Mohl does not have this verse.

22. For a discussion of the varieties of dogs known among Iranians, see Omidsalar, Mahmoud and Omidsalar, Teresa P., “Dog, I: In Literature and Folklore,” EIr 7: 461.Google Scholar

23. It may be that the MS reading is actually which after normalization is spelled as Nasafī, Bāznāmah-i Nasafi, ed. ‘Ali Gharavi (Tehran, 1354), 174.Google Scholar

24. Although we will not discuss the dynamics of the formation of variants here, they fall into several general groupings: visual (including all spelling errors), aural (as distinct from oral), memory interference, and simple trivializations.

25. Since neither the Mohl, nor the Tehran editions have a critical apparatus, one has no way of knowing what resources their editors were able to draw on.

26. These two examples which have been taken from Vis u Ramin, were kindly sent to me by Professor Khaleghi-Motlagh.

27. See Vāzhahnāmah. (Tehran, 1369/1990), 294-95.

28. Ed. R. ^ᶜAfifi (Tehran, 1370/1991).

29. This should by no means be understood as Firdawsi's use of “oral formulae.” It is in the nature of Persian prosody that certain words are exclusively used in rhyming pairs with certain other words. For instance, the words tazarv and sarv, marv, and gharv are always paired together in the Shāhnāmah and elsewhere in classical Persian. See Matini, J., “sarv u taẕarv,” Iran Nameh 3 (1985): 475–502.Google Scholar That is because in Persian and Arabic poetry we do have abundant “literary formulae” that poets are either forced to use because of the stringent requirements of the prosody, or which they use out of literary habit. There is no “oral formulaic usage” behind this tendency. For a discussion of the literary formulae in classical Persian see, Omidsalar, M., “Bayān-i adabī va bayān-i ^ᶜāmiyānah dar ḥamāsa-hā-yi fārsi [Literary and Folk Styles in Persian Epics],” Gulistān 2 (1998): 85–113.Google Scholar