1 For a complete, annotated list of al-Zamakhsharī's works, see my A Traditional Muʿtazilite Qur’ān Commentary, Appendix 3, pp. 267 ff.

2 The edition of the Kashshāf used here is that of the Dār Iḥyā’ al-Turāth al-ʿArabī (4 vols, Beirut, 1997).

3 J. Robson, “al-Bayḍāwī,’ EI ² 1: 1129.

4 This scholar is Nāṣir al-Dīn Abū l-ʿAbbās Aḥmad b. Muḥammad Ibn al-Munayyir al-Judhāmī al-Iskandarī (d. 683/1284). His work on the Kashshāf is the Kitāb al-Intiṣāf min al-Kashshāf, described as a counterblast “against the heresies and some opinions on grammar” (GAL ¹, GAL) – GAL ¹ 1: 291, 416, 431; GAL 1: 346; 529–30; GAL.Sp 1: 509, 738.

5 Shams al-Dīn Muḥammad b. Aḥmad al-Dhahabī (d. 748/1348), Mīzān al-iʿtidāl fī naqd al-rijāl, ed. ʿAlī Muḥammad al-Bajāwī (Cairo: Maṭbaʿat ʿĪsā l-Bābī al-Ḥalabī, [1963]), 4: 78 [no. 8367]; Aḥmad b. ʿAlī Ibn Ḥajar al-ʿAsqalānī (d. 852/1449), Lisān al-mīzān, ed. Muḥammad ʿAbd al-Raḥmān al-Marʿashalī (Beirut: Dār Iḥyā’ al-Turāth al-ʿArabī, 1995–96), 6: 651–3 [no. 8313].

6 Nöldeke, Th., Geschichte des Qorâns 1^st ed. (Göttingen: Verlag der Dieterichschen Buchhandlung, 1860), xxviiiGoogle Scholar; Kashshāf (Calcutta: Maṭbaʿat al-Laysī, 1856–59), 1: 7Google Scholar; Goldziher, I., “Aus der Theologie des Fachr al-dīn al-Rāzī”, Der Islam 3, 1912, 220CrossRefGoogle Scholar; C. Brockelmann, “al-Zamakhsharī”, EI ¹ 4: 1205.

7 Smith, J. I., An Historical and Semantic Study of the Term “Islām” as Seen in a Sequence of Qur’ān Commentaries (Harvard Dissertations in Religion. Missoula: Scholars Press, 1975), 92–3Google Scholar; McAuliffe, J. D., Qur’ānic Christians: An Analysis of Classical and Modern Exegesis (New York: Cambridge University Press, 1991), 53CrossRefGoogle Scholar; W. Madelung, “al-Zamakhsharī”, EI ² Suppl., fasc. 11–12: 840–1. Versteegh (EI ² 11: 434) passes over the Kashshāf in silence, referring his readers to the EI ² Supplement for al-Zamakhsharī's contributions in theology, exegesis and adab.

8 Jansen, J. J. G., The Interpretation of the Koran in Modern Egypt (Leiden: E. J. Brill, 1974), 63Google Scholar; D. Gimaret, “Muʿtazila”, EI ² 7: 786; A. Rippin, “Tafsīr”, EI ² 10: 85; Saleh, W., The Formation of the Classical Tafsīr Tradition. The Qur’ān Commentary of al-Thaʿlabī (d. 427/1035) (Texts and Studies on the Qur’ān, 1. Leiden and Boston: Brill, 2004), 23Google Scholar, n. 40. On al-Zamakhsharī's comments on Q93:7, Saleh writes: “The praise from modern scholars showered on al-Zamakhsharī, however, has more to do with their intellectual affinity with his Muʿtazilite theology than with a balanced analysis of his approach to the Quran. Al-Zamakhsharī is as doctrinally motivated as any other medieval scholar” (Formation, 148).

9 J. Robson, “al-Bayḍāwī”, EI ² 1: 1129; GAL ¹ 1: 418; GAL.Sp 1: 741; GAL 1: 532–3. Al-Dā’ūdī was one of al-Suyūṭī's (d. 911/1505) students and wrote, besides the aforementioned Itḥāf on al-Bayḍāwī, Ṭabaqāt al-mufassirīn in 941/1534 (see GAL ¹ 2: 289; GAL 2: 373; GAL.Sp 2: 401). Aḥmad al-Nūbī lived in al-Ṭā’if and composed five works, including one on al-Bayḍāwī, another in praise of his hometown of al-Ṭā’if (1027/1617), and a third against smoking tobacco written in the year of his death (see GAL ¹ 2: 385; GAL 2: 504–5; GAL.Sp 2: 520).

10 For al-Ṭūsī and al-Ṭabarsī (al-Ṭabrisī), see Mohammad Ali Amir-Moezzi, “al-Ṭūsī”, EI ² 10: 744–6 and E. Kohlberg, “al-Ṭabrisī”, EI ² 10: 40–1. Neither author comments on the Muʿtazilite nature of the Kashshāf, although Kohlberg notes that al-Ṭabarsī wrote three commentaries on the Quran, one of which, al-Kāfī l-shāfī min kitāb al-Kashshāf, was a one-volume abridgement of al-Zamakhsharī's.

11 For a more detailed study of al-Zamakhsharī's commentary on Q44 and Q54, see my A Traditional Muʿtazilite Qur’ān Commentary, Chapter 3, especially pp. 118 ff. These two sūras were originally chosen because they were of average length and therefore susceptible to a detailed study within a reasonable time frame.

12 To call a Quran commentary a tafsīr musalsal (“chained commentary”) means that it “begins with the first sūrah of the Quran and comments verse by verse on that sūrah and all subsequent ones. Exegetical chronology has, therefore, its own autonomy, following the sequence of text rather than that of revelation … Within the sūrah each verse is quoted separately and then broken into exegetical units, what medieval Biblical scholars would call lemmata. Each passage, or lemma, is then analyzed separately and relevant comments are made about the verse as a whole, such as its sabab al-nuzūl. What is frequently absent is any extended consideration of the larger context. Occasionally, a connection will be made with the previous verse or, even more rarely, with more distant parts of the sūrah” (McAuliffe, Qur’ānic Christians, 34).

13 A brief perusal of the Kashshāf will show that there are no long excurses on any topic; the most that can be hoped for is a series of explanations, interpretations or variations of the passage under scrutiny.

14 I use Arberry's translation of the Quran (The Koran Interpreted. Oxford: Oxford University Press, 1985Google Scholar, 1964¹); any changes are indicated by square brackets. Since Arberry follows Flügel's enumeration of the verses, both his and that of Cairo will be given where necessary, first Cairo and then Flügel's.

15 Kashshāf 4: 436.

16 The Muʿtazilites had five principles (al-uṣūl al-khamsa) which Gimaret summarizes thus: 1. The uniqueness of God (al-tawḥīd); 2. The justice of God (al-ʿadl); 3. “The promise and the threat” (al-waʿd wa-l-waʿīd); 4. The theory of “the intermediate state” (al-manzila bayna l-manzilatayn) for the sinful Muslim here on earth (neither “believing” (mu'min) nor “disbelieving” (kāfir), they are a “malefactor” (fāsiq)); 5. The obligation laid upon every believer to “command the good and forbid the evil” (al-amr bi-l-maʿrūf wa-l-nahy ʿan al-munkar), i.e. to intervene in public affairs to uphold the Law (EI ² 7: 786–7). The last three principles were derivative of the first two, divine unity and justice, and the Muʿtazilites were known as the People of (Divine) Unity and Justice (ahl al-tawḥīd wa-l-ʿadl). The principle known as “the promise and the threat” (al-waʿd wa-l-waʿīd) said that God must necessarily reward the good and punish the wicked in the next life. However, the accent was on the second part, the eternal damnation of the sinner who did not repent. That is why al-Zamakhsharī speaks only about turning [people] away from the promise and the threat, for he is talking about eternal punishment. Van Ess gives the “eternal punishment of the ‘transgressor’” as the third principle of the founding father Abū l-Hudhayl (“Muʿtazilah”, The Encyclopedia of Religion (New York: Macmillan and London: Collier-Macmillan, 1986) 7: 225); and Gimaret writes that by al-waʿd wa-l-waʿīd “is understood that on account of the ‘threat’ uttered against him in the Ḳur’ān, every Muslim guilty of a serious offence, who dies without repentance, will suffer for eternity the torments of Hell” (“Muʿtazila”, EI ² 7: 786).

17 This line is by al-Aʿraj al-Khārijī, according to the editors’ notes from both the 1997 Dār Iḥyā’ al-Turāth al-ʿArabī (Beirut) and 1998 Maktabat al-ʿUbaykān (Riyadh) editions of the Kashshāf (4: 436, n. 1; 5: 658–9, n. 2). He could be, then, Abū Mālik al-Naḍr ibn Abī l-Naḍr, called al-Aʿraj, a poet at the court of Hārūn al-Rashīd (r. 170/786–195/809).

18 Lane, E. W., An Arabic–English Lexicon (repr.) (Beirut: Librairie du Liban, 1980), 8: 2976Google Scholar, at the root y-s-r, which defines yassara l-faras as: “he prepared the horse for riding, by saddling and bridling”.

19 Abū Isḥāq Ibrāhīm ibn Sahl ibn al-Sārī al-Zajjāj (d. c. 311/923) was a lexicographer, grammarian, and an intimate student of al-Mubarrad Abū l-ʿAbbās Muḥammad ibn Jazīd al-Azdī (d. 285/898 or 286/900), the most important representative of the so-called Baṣran school of grammar of his time. After his studies, al-Zajjāj was tutor in the household of ʿUbayd Allāh Ibn Sulaymān, vizier of the ʿAbbāsid caliph al-Muʿtaḍiḍ (r. 892–902 CE). Later he was employed in the service of Ibn Sulaymān's son, al-Qāsim, when he became vizier; he remained al-Qāsim's secretary until his death in Baghdad at the age of over eighty. For al-Zajjāj, see GAL ¹ 1: 110; GAL 1: 111–2; GAL.Sp 1: 170; C. Versteegh, “al-Zadjdjādj”, EI ² 11: 377–8; Jalāl al-Dīn ʿAbd al-Raḥmān ibn Abī Bakr al-Suyūṭī (d. 911/1505), Bughyat al-wuʿāt fī ṭabaqāt al-lughawiyyīn wa-l-nuḥāt, ed. Muḥammad Abū l-Faḍl Ibrāhīm ([Cairo]: Maṭbaʿat ʿĪsā l-Bābī al-Ḥalabī, 1964–65), 1: 411–3 [no. 825]; and Yāqūt ibn ʿAbd Allāh al-Ḥamawī al-Rūmī (d. 626/1229), Irshād al-arīb ilā maʿrifat al-adīb [Muʿjam al-udabā’ wa-ṭabaqāt al-urabā’], ed. Margoliouth, D. S. (London: Luzac & Co. and Cairo: Maṭbaʿa Hindiyya, 1925), 1: 51–63Google Scholar [no. 9]. This section of the article was part of a larger one in the thesis, in which al-Zamakhsharī's use of al-Zajjāj's Maʿānī l-Qur’ā n was examined. It was a methodological study, an effort to see how accurately what was read in the Kashshāf reflected the earlier sources that were referred to; the results were presented at the 2002 annual meeting of the American Academy of Religion in Toronto. In the Kashshāf, al-Zamakhsharī refers specifically to al-Zajjāj by name on thirty-two occasions at least. In his references, al-Zamakhsharī gives only al-Zajjāj's name; there are no references to particular works. He indicates that he is borrowing from al-Zajjāj usually by the expression qāla l-Zajjāj (“al-Zajjāj said”) (75 per cent of the time). Of the thirty-two references to al-Zajjāj in the Kashshāf under study, all but one could be traced to a corresponding text in the Maʿānī l-Qur’ān. The other was not traced; it must have been a reference to a work of al-Zajjāj that is not yet available or was lost, or else it is an error. Q35:8/9 is the only occasion on which we see some Muʿtazilite tampering with al-Zajjāj's original text.

20 Some translations, however, supply an answer. Pickthall's translation, for example, reads: “Is he, the evil of whose deeds is made fairseeming unto him so that he deemeth it good, (other than Satan's dupe)?” (M. M. Pickthall, The Meaning of the Glorious Qur’ān (Cairo: Dār al-Kitāb al-Miṣrī and Beirut: Dār al-Kitāb al-Lubnānī, n.d.)). The hamza at the beginning of the verse can also be an interjection; Blachère translates it as “Eh quoi (What)!” (Le Coran, Paris: G.-P. Maisonneuve & Larose, 1980). Arberry stays closer to the Arabic.

21 This statement may look very un- or even anti-Muʿtazilite, for it seems to be stating that God causes those whom he leads astray to be blinded to the evil of their deeds. A closer reading of the passage indicates that the tazyīn al-ʿamal = al-iḍlāl equation is the outcome of a process where individual liberty comes first. After stating that the two expressions have the same meaning, al-Zamkhsharī continues: “That is, the one who disobeys is one for whom benefits are of no avail, so that he merits thereby to be forsaken by God Most High, to be abandoned – he and his lot etc. (wa-huwa an yakūna l-ʿāṣī ʿalā ṣifa lā tajdī ʿalayhi l-maṣāliḥ ḥattā yastawjiba bi-dhālik khidhlān Allāh taʿālā wa-takhliyatahu wa-sha'nihi ilkh)”. Later al-Zamakhsharī writes that the Prophet is not to regret what happens when God forsakes “those who are determined to be unbelievers (al-muṣammūna ʿalā l-kufr)”. It is clear, then, that God leads astray those who are determined to be so treated, who, by their freely chosen disobedience, merit what they get (Kashshāf 3: 609).

22 Kashshāf 3: 609.

23 Ibāhīm ibn al-Sarī ibn Sahl al-Zajjāj, (d. 311/923), Kitāb Maʿānī l-Qur’ān (wa-iʿrābuhu), ed. ʿAbd al-Jalīl ʿAbduh Shalabī, (Cairo: al-Hay'a al-ʿĀmma li-Shu’ūn al-Maṭābiʿ al-Amīriyya and Beirut: Maktabat al-ʿAṣriyya, 1974), 4: 264Google Scholar.

24 Madelung, W., “The theology of al-Zamakhsharī”, in Actas del XII Congreso de la U.E.A.I. (Malaga, 1984) (Madrid: Union Européenne d'Arabisants et d'Islamisants, 1986), 488Google Scholar, where the manuscripts are described as consisting “of about six tightly written folios”; Schmidtke, S. (ed. and trans.), A Muʿtazilite Creed of az-Zamaḫšarî (d. 538/1144) [al-Minhâj fī uṣûl al-dîn] (Abhandlungen für die Kunde des Morgenlandes, 51/4, Stuttgart: Deutsche Morgenländische Gesellschaft – F. Steiner, 1997), 9Google Scholar.

25 Schmidtke translates the titles of the chapters (sing. bāb) as follows: 1. Chapter on the Proof for the Temporality of the Bodies and that they have a Producer (bāb al-dalīl ʿalā ḥudūth al-ajsām wa-ʿalā anna lahā muḥaddith); 2. Chapter on the Knowledge of God and His Attributes (bāb maʿrifat al-qadīm wa-ṣifātihi); 3. Chapter on the Imposition of Moral Obligations (bāb al-taklīf); 4. Chapter on the Facilitating Favours (bāb al-alṭāf); 5. Chapter on Pains (bāb al-ālām); 6. Chapter on Sustenance, Prices and Terms of Death (bāb al-arzāq wa-l-asʿār wa-ājāl); 7. Chapter on the Promise and the Threat (bāb al-waʿd wa-l-waʿīd); 8. Chapter on the Command of what is Proper and the Interdiction of what is Reprehensible (bāb al-amr bi-l-maʿrūf wa-l-nahy ʿan al-munkar); 9. Chapter on Prophecy (bāb al-nubuwwāt).

26 Madelung, “Theology”, 489 ff. passim, 495. Madelung's views are echoed by Schmidtke (Minhāj, 9). Madelung illustrates his analysis of al-Zamakhsharī's position, as it appears in the Minhāj, with references to the first two chapters of the treatise, which deal with the proof of the existence of God and with the divine attributes.

27 Madelung, “Theology”, 492–3.

28 Schmidtke, Minhāj, 43 (E = English), 80 (A = Arabic). Arberry's translation is followed for Quranic passages. My translation of the Minhāj does not always follow Schmidtke's.

29 Q2: 24/22.

30 Q54: 45.

31 Q9: 33.

32 Q5: 67/71.

33 One is reminded of al-Zamakhsharī's presentation of the muḥkamāt and mutashābihāt verses in his commentary on Q3:7. He gives two pairs of verses: in each case one verse, the mutashābih, is referred to the other, the muḥkam, in order to be understood. However, he does not say why these verses are so considered. See Kashshāf 2: 365–6.

34 See L Gardet, “ʿIlm al-Kalām”, EI ² 3: 1141–50, especially section III: Method and problematic (pp. 1146–7).

35 See S. van den Bergh, “Dalīl”, EI ² 2: 101–2; L. Gardet, “Burhān”, EI ² 1: 1326–7.

36 Schmidtke, Minhāj, 38–9 (E, which has “fall into the hell-fire for eternity”), 75–6 (A). I have not followed Schmidtke's English translation exactly for either tradition; in the latter case, the Arabic has: shafāʿatuhu li-ahl al-kabā’ir min ummatī. In these few pages from the Minhāj we see a number of points that reflect Gardet's comments on traditional arguments in Kalām, i.e. arguments from authority: quotations from the Quran, traditions, and positions that were agreed upon.

37 Schmidtke, Minhāj, 16–7 (E), 54–5 (A).

38 Al-Zamakhsharī includes in his explanation a list of impediments to seeing, one of which is the inability to face or be in front of an object (khilāf al-muqābala). Although the word has the meaning being “in front of” or “facing”, there should not be, in theory, any “front” and the only “face” would be that of the one seeing. The main idea is “to take a position with respect to something”, or for something “to have a position” (fī jiha). This leads to the next point.

39 The passage from the Minhāj reads as follows [I have used square brackets to indicate changes to Schmitke's translation]: “Yet it is impossible to face God because He is neither a body nor subsisting in a body. God said “The eyes attain Him not” [(Q6:103)] and He said “Thou shalt not see Me” [(Q7:143/138)] as He said “(those) … shall never create a fly” [(Q22:73/72)] because [seeing Him is as impossible as their creating a body]; and God said “for they said, ‘Show us God openly’. And the thunderbolt took them for their evil doing” [(Q4:153/152)]. If they had asked for something that [was] possible (jā’iz) they would not have done evil and would not have been taken by the thunderbolt … [Mose's statement] – peace be upon him – ‘Show me, that I may behold Thee!’ [(Q7:143/138)] is not binding because he said this [only] in order to silence those who demanded to see (God) by proving its impossibility and in order to stop their arguing when they heard His saying “Thou shalt not see Me” [(Q7:143/138)]” (Schmidtke, Minhāj, 17 (E), 55 (A)).

40 L Gardet, “ʿIlm al-Kalām” EI ² 3: 1146. Gardet's reference to “manuals” would indicate that he is speaking about a later period.

41 Schmidtke, Minhāj, 38 (E), 75 (A).

42 In the Minhāj, the text runs: “inna lladhīna yarmūna l-muḥṣanāti” ilā qawlihi “wa-lahum ʿadhābun ʿaẓīmun”.

43 The verse is addressed to Muḥammad.

44 Kashshāf 4: 163. Al-Zamakhsharī supports his interpretation in the Kashshāf with references to Q4:174 and a tradition reporting that al-Ḥasan al-Baṣrī said, “By God! They will have absolutely no intercessor (wa-llāhi mā yakūnu lahum shafīʿ al-batta)”.

45 Kashshāf 1: 165. Al-Zamakhsharī says: “If you were to say, ‘Is there any proof therein that intercession for the disobedient is not accepted?’ then I would say, ‘Yes, because He did not allow one soul to pay another's due, by which it would be free from [what was its due for] either an action or an omission; nor did He allow that intercession be accepted from it. So it is known that the intercession of an advocate is not accepted for those who disobey’”.

46 Kashshāf 4: 123.

47 Kashshāf 2:51: al-baṣr huwa l-jawhar al-laṭīf alladhī rakkabahu llāh fī ḥāssat al-naẓr, bihi tudraku l-mubṣarāt – the last part could read: bihi tudriku l-mubṣarāt (“by which you perceive what can be seen”).

48 Kashshāf 2: 52. For the technical terms used here, see R. Arnaldez, “Hay'a”, EI ² 3: 301–2; S. van den Bergh, “Djawhar”, EI ² 2: 493–4; Tj. de Boer, “Djism”, EI ² 2: 553–5.

49 Kashshāf 2: 52. His commentary can be summarized as follows:

50 The core dialogue here runs:

51 The question is: “If you were to say, ‘How could Moses – Peace be upon him – have asked that, he being one of the most knowledgeable of people as to God and what is and is not permitted concerning him; and as to his being above being seen, which is to be attained by one of the senses? This is true only for something that is in a position, but it is impossible for what is neither a body nor an accident to be in a position. The Mujbirites’ refusing to say it is rationally absurd is not binding, since it is not the first time they act arrogantly or commit a sin. How could he (i.e. Moses) have demanded it, when he had said – when ‘the earthquake seized’ those who had said openly, ‘Show us God,’ – ‘Wilt Thou destroy us for what the foolish ones of us have done? [It is only Thy trial, whereby] Thou leadest astray whom Thou wilt’ (Q7: 155/154). So he washed his hands of their action and called them foolish and lost”. The answer is: “I would say, ‘He asked to see only in order to censure those whom he called foolish and lost, and whose actions he washed his hands of; and in order to silence them. That is because, when they asked to see, he denied it to them and let them know about sin and told them about the truth; but they were obstinate and persisted in their obstinacy, and said, ‘No escaping it! We will not believe you (lan nu'mina laka) until we see God openly.’ So he wanted them to hear the clear revelation (naṣṣ) from God about the impossibility of [what they were asking]. That is his word, ‘lan tarānī (Thou shalt not see me),’ so that they would be sure and the doubt that had entered them would depart from them. It was for this that he said, ‘rabbi arinī anẓur ilayka (Oh my Lord, show me, that I may behold Thee!)” (Kashshāf 2: 144–5). By Mujbira, al-Zamakhsharī means his orthodox opponents in general; see W. M. Watt, “Djabariyya”, EI ² 2: 365.

52 Schmidtke, Minhāj, 17 (E), 55 (A) (“[Mose's statement] – peace be upon him–”show me, that I may behold Thee!’ is not binding because he said this [only] in order to silence those who had demanded to see (God) by proving its impossibility and in order to stop their arguing when they heard His saying ‘Thou shalt not see Me.’”).

53 Al-Zamakhsharī writes: “And his saying ‘that I may behold Thee!’ with the meaning that it has of ‘to be in front of, to face’, which is pure anthropomorphism and crude materialism, is a proof that it is an interpretation of their demands and an account of what they said. The one riding the camel (ṣāḥib al-jamal) was too great to make God something that could be looked towards or that could be in front of the sense of sight, for how [could that be] with one who is more deeply rooted in the knowledge of God Most High than Wāṣil ibn ʿAṭā’, ʿAmr ibn ʿUbayd, al-Naẓẓām, Abū l-Hudhayl, the two shaykhs and all of the Kalām theologians?” (Kashshāf 2:145). The two shaykhs referred to here are Abū ʿAlī al-Jubbā’ī (d. 303/915) and his son Abū Hāshim al-Jubbā’ī (321/933). From what I have been able to discover, ṣāḥib al-jamal is a title given to Muḥammad. Here, however, al-Zamakhsharī seems to be using it with reference to Moses. Otherwise, he either suddenly started to speak about Muḥammad (not to attribute the Quranic passage to him) or else his text was corrupted in transmission and an intermediary passge is missing.

54 Kashshāf 3: 172–3.

55 Schmidtke, Minhāj, 17 (E), 55 (A).

56 Kashshāf 1:618. The Mushabbiha would be those guilty of tashbīh, that is, the anthropomorphists.

57 Schmidtke, Minhāj, 8. Rukn al-Dīn's exact name is uncertain. Following McDermott and Madelung I will refer to him consistently as Ibn al-Malāḥimī.

58 Aḥmad b. Muṣṭafā b. Khalīl Ṭāshkubrī Zāda (d. 968/1560–61), Miftāḥ al-saʿāda wa-miṣbāḥ al-siyāda fī mawḍūʿāt al-ʿulūm, ed. Kāmil Kāmil Bakri and ʿAbd al-Wahhāb Abū l-Nūr (Cairo: Dār al-Kutub al-Ḥadītha, [1968]), 2, 100.

59 Ed. McDermott and Madelung (London: al-Hoda, 1991). According to Madelung, Ibn al-Malāḥimī states in the introduction to the Muʿtamad that his intention is “to condense, complete and update” ʿAbū l-Ḥusayn al-Baṣrī's largest Kalām work, the Kitāb Taṣaffuḥ al-adilla, in which he critically scrutinized (taṣaffaḥa) the arguments of the Muʿtazilī scholars as much as those of their opponents”. However, he adds that Ibn al-Malāḥimī does not go much beyond the end of this book, his main but not only source (Muʿtamad, xi).The present edition of the Muʿtamad is over 600 pages long and includes dozens of references to the Quran, although there is no index at the end of their edition. The edition is divided into two parts (sing. juz’), the first and the third, following the divisions of the manuscripts that were used. According to the introduction, this edition contains most of the first two parts and the complete third part of the Muʿtamad, even if only the first and third parts are indicated in the table of contents. This is because the section heading for Part 2 in the manuscript was omitted, and not because the second part was missing. The fourth part, however, is missing, although the editors have incorporated one chapter (bāb) that was found in a manuscript of another of Ibn al-Malāḥimī's works, the Kitāb al-Fā’iq, described as “a greatly abridged version of the Muʿtamad” (Muʿtamad, xiv). Each part is divided into three large sections of unequal length; each of these has a number of chapters (sing. bāb), some of which are divided into smaller sections (sing. faṣl); there are a few independent sections called faṣl also.

60 There is no need to limit Muʿtazilite sources to this one author. Al-Zamakhsharī had come into contact with the teachings and Quran commentary of the Muʿtazilite Zaydī imām al-Ḥākim al-Jushamī (d. 494/1101) through the latter's student Aḥmad ibn Muḥammad ibn Isḥāq al-Khwārazmī, although he did not draw directly on this commentary when composing the Kashshāf. See Madelung, “Theology”, 487; “al-Zamakhsharī”, EI ² Suppl., fasc. 11–12: 341.

61 The original study was carried out for the author's PhD thesis and involved a detailed comparison of what al-Zamakhsharī and Ibn al-Malāḥimī had to say on approximately twenty verses in their respective works.

62 Ibn al-Malāḥimī, Muʿtamad, 502–3. Ibn al-Malāḥimī compares the first use of the word to its application to the atom (jawhar); see S. van den Bergh, “Djawhar”, EI ² 2: 493–4.

63 Al-Kashshāf 1: 236. The passage here seems a little like a flow of consciousness. Al-Zamakhsharī ends his comments on this verse with a sabab al-nuzūl: “It was said: The polytheists had three hundred and sixty idols around the Kaʿba, and when they heard of this verse, they were amazed and said, ‘If you speak the truth, then bring a sign so that by it we may know your truthfulness’; and the verse was revealed (fa-nazalat)”. As it stands now, this occasion for revelation implies that the verse had been heard of before it was revealed.

64 These passages read: Q2: 210/206: “What do they look for, but that God shall come to them (ya'tiyahum Allāh) in the cloud-shadows?” and Q6:158/159: “What, do they look for the angels to come to them (ta'tiyahum al-malā’ika), nothing less, or that thy Lord should come (ya'tiya rabbukum)?”

65 Ibn al-Malāḥimī, Muʿtamad, 327–8. This is a very Muʿtazilite understanding of the verse that has been placed on the lips of Ibn ʿAbbās, since we see in it references to two of the five Muʿtazilite principles.

66 Kashshāf 4: 754–5: “If you were to say, ‘What is the meaning of predicating [the verb] “to come” of God, since movement and change of place are permitted only concerning those who can be in a position?’ I would say, ‘It is a comparison to the manifestation of the signs of his might and the demonstration of the effects of his power and authority. In this, his state is comparable to the state of a king who, if he is present himself, there appear with his presence effects of awe and command that are not evident with the presence of all of his armies and of all his ministers and leading personalities together without exception.’”

67 The results of this research were presented at the 2005 annual meeting of the American Academy of Religion in Philadelphia.

68 The following is based primarily on Madelung, Wilfred, “The origins of the controversy concerning the creation of the Koran”, in Morewedge, P. (ed.), Islamic Philsophical Theology (Albany: State University of New York Press, 1979), 504–25Google Scholar.

69 See M. Hinds, “Miḥna”, EI ² 7: 2–6 and Nawas, John A., “A reexamination of three current explanations for al-Ma'mun's introduction of the Miḥna”, International Journal of Middle East Studies 26, 1994, 615–29CrossRefGoogle Scholar.

70 Names associated with the created Quran at this time are Jaʿd ibn Dirham (d. 125/743) and Jahm ibn Ṣafwān (d. 128/745), both executed towards the end of the Umayyad period. See G. Vajda, “Ibn Dirham”, EI ² 3: 747–8 and W. M. Watt, “Djahm b. Ṣafwān”, EI ² 2: 388.

71 M. Hinds, “Miḥna”, EI ² 7: 5–6 This view contrasts with that held by Patton, who wrote: “When al-Rashîd died, the matter remained in the same position during the time of his son al-Amîn; but when al-Ma'mūn succeeded some of the Muʿtazilites led him astray and made the doctrine of the creation of the Ḳorân to appear plausible to him”. Patton's source here is the Egyptian historian al-Maqrīzī (d. 1442); the “matter” he refers to is that the Quran was the uncreated Word of God (Patton, Walter M., Aḥmed ibn Ḥanbal and the Miḥna (Leiden: E. J. Brill, 1897), 48–9Google Scholar).

72 D. Gimaret, “Muʿtazila”, EI ² 7:785.

73 Aḥmad ibn Muḥammad Ibn Khallikān (d. 681/1282), Wafayāt al-aʿyān wa-anbā’ abnā’ al-zamān, ed. Iḥsān ʿAbbās (Beirut: Dār al-Thaqāfa, [1968]), 5: 170 [no. 711]. See also Taqī al-Dīn Muḥammad ibn Aḥmad al-Fāsī, (d. 832/1429), al-ʿIqd al-thamīn fī tārīkh al-balad al-amīn, ed. Sayyid, Fu’ād (Cairo, 1959–69), 7: 141Google Scholar; Shams al-Dīn Muḥammad b. Aḥmad al-Dhahabī, (d. 748/1348), Ta'rīkh al-Islām wa-wafayāt al-mashāhīr wa-l-aʿlām, ed. ʿUmar ʿAbd al-Salām Tadmūrī, (Beirut: Dār al-Kitāb al-ʿArabī, 1991–2000)Google Scholar, vol. 521–530H/531–540H, 489 [no. 398]; ʿAbd al-Ḥayy ibn Aḥmad Ibn al-ʿImād, (d. 1089/1679), Shadharāt al-dhahab fī akhbār man dhahaba (Maktabat al-Qudsī, 1931–33) 4: 120Google Scholar. After noting Ibn Khallikān's comments in the preface to his edition of the Kashshāf, Nassau-Lees writes: “It is not at all improbable that al-Zamakhsharí originally did write Khalaqa – first because, instead of being ashamed of his doctrines, he was proud them; and secondly, because it would appear that the word Anzala is not his, for a different form of the same verb occurs in the next sentence, which, though not a breach of the rules of good composition, would here be an inelegancy. As to his having substituted the word J'ala (sic), the idea is absurd. To introduce the identical word in the same line twice in the same sense, would be a gross breach of the rules of good composition, and a blunder, that no man of sound sense would be guilty of accusing the most profound philoger and elegant scholar of his age (sic). The word J'ala (sic) is understood before the words قرانـا عربـيـا which occur lower down, and will there, no doubt, be taken in the sense of Khalaqa. I am satisfied that Ibn Khallikán's statement, which I am surprised to see that al-Soyootí has endorsed, is without foundation” (Calcutta: Maṭbaʿat al-Laysī, 1856–59, 1: 7–8, n.7). This does not mean that Nassau-Lees necessarily accepted that al-Zamakhsharī wrote khalaqa l-Quran but rather that the latter would never have changed it to jaʿala l-Qur’ān, as Ibn Khallikān maintains.

74 Ibn Khallikān says that it was the modification of the people (iṣlāḥ al-nās) (Wafayāt 5: 170 [no. 711]), Abū l-Fidā’ that it was the work of al-Zamakhsharī's companions (aṣḥāb) (Abū l-Fidā’ Ismāʿīl ibn ʿAlī (d. 732/1331), Kitāb al-Mukhtaṣar fī akhbār al-bashar (Beirut: Dār al-Fikr, 1956–61) 2: 25 [juz’ 5]), and Ibn al-Wardī merely states that the text was changed after his death (baʿdahu) (Zayn al-Dīn ʿUmar ibn al-Muẓaffar Ibn al-Wardī, (d. 749/1349), Ta'rīkh Ibn al-Wardī (al-Najaf: al-Maṭbaʿa al-Haydariyya, 1969) 2: 63Google Scholar). For the meanings of khalaqa, see R. Arnaldez, “Khalḳ”, EI ² 4: 980–8.

75 Wolfson, H. A., The Philosophy of the Kalam (Cambridge and London: Harvard University Press, 1976), 297Google Scholar; Patton, Aḥmed ibn Ḥanbal, 58, 73–4, 90–1. For the traditionist and eponym of one of the four Sunnī madhāhib, see H. Laoust, “Aḥmad ibn Ḥanbal”, EI ² 1: 272–7.

76 The results in this section are based on a study of approximately 250 manuscripts of the Kashshāf, most of which were studied during a nine-week trip to Istanbul in the spring and summer of 2004.

77 The umm al-Kashshāf is the name al-Zamakhsharī gives to the final draft he finished in Mecca in 528/1134; he refers to it in the postscript he added to this copy. My study indicates that it was not the only copy used by medieval copyists. The other was stored at the Abū Ḥanīfa tomb-shrine (mashhad) in Baghdad, and is referred to in the colophon of other manuscripts. The Baghdad copy could well have been the rough draft (sawād) to which al-Zamakhsharī refers in the postscript to the umm al-Kashshāf and from which he copied the umm al-Kashshāf in Mecca. See my A Traditional Muʿtazilite Qur’ān Commentary, especially pp. 70–75.

78 MS. Fatih 344 is a complete copy of the Kashshāf finished on 20 Shawwāl 798H. MS. Fatih 340 is a copy of the commentary on Q1–Q33; it was finished in Dhū l-Qaʿda 732H. Although the glossator on MS. Fatih 340 writes “I have heard some people say” and “I have studied the Kashshāf”, it is highly unlikely that he did. This marginal gloss repeats almost verbatim what al-Yamanī (d. 750/1348–9) says on the matter of khalaqa l-Qur’ān in his sharḥ on the Kashshāf, completed in 733/1332. This gloss and al-Yamanī's comment are confirmed by a more recent source; Nassau-Lees writes: “Firawzabádí (apud the Kashf al-Zonoon) on the other hand says, that he asked his master regarding it, and that he told him he had seen at Baghdad, in the Imám Aboo Hanífah's tomb, the Author's autograph, and that it bore no signs of erasure or emendation (Kashshāf (ed. Calcutta, 1856–9) 1: 7, n.7). Firawzabádí is undoubtedly Fīrūzābādī (d. 817/1415); the “master” may be Taqī al-Dīn al-Subkī (d. 756/1355) with whom he studied in Damascus in 750/1349 and then accompanied to Jerusalem the same year. See H. Fleisch, “Fīrūzābādī”, EI ² 2: 926–7; J. Schacht and C. E. Bosworth, “al-Subkī”, EI ² 9: 743–5. The copy of the Kashshāf referred to here, though, is that of Abū Ḥanīfa's mashhad, and most likely not the umm al-Kashshāf.

79 MSS. Beyezit 575–6 are a complete copy of the Kashshāf (Q1–Q18 and Q19–Q114); they are undated. MS. Köprülü 129 is a first volume (Q1–Q18) from the eighth century H.

80 Qiwām al-Dīn Luṭf Allāh Amīr Kātib b. (ʿAmīd) Amīr ʿUmar b. Amīr Ghāzī Abū Ḥanīfa al-Itqānī was born in Itqān, Turkmenistan (or Farāb according to GAL.Sp) in 685/1286. In 730/1330 (or 720/1320, GAL.Sp) he travelled to Damascus and Cairo, and then was professor and qāḍī in Baghdad. In Damascus again, he was al-Dhahabī's successor in the Dār al-Ḥadīth al-Ẓāhiriyya, but following a quarrel with Taqī al-Dīn al-Subkī (d. 756/1355) and others, he resigned and went to Egypt where he was professor in a new madrasa built in 757/1356 by Ṣarighitmish. He died in Egypt in 758/1357. In GAL.Sp the itinerary differs slightly in that al-Itqānī had only one stay in Damascus, the quarrel having taken place before he left for Baghdad, from where he then proceeded to Egypt. Brockelmann lists seven works attributed to al-Itqānī, several of which are commentaries on other works. These include The Crusher of the Muʿtazilites (Shaddākhat al-Muʿtazila), a risāla “against al-Zamakhsharī's interpretation (gegen Zamaḫšarīs Auslegung)”, that is, his interpretation of the Quran (GAL ¹ 2: 79; GAL 2: 95; GAL.Sp 2: 87; for al-Subkī, see J. Schacht and C. E. Bosworth, “al-Subkī”, EI ² 9: 743–5).

81 MS. Shehid ʿAlī Pasha 153 was finished in Muḥarram 704H.

82 Part of al-Itqānī's copy of the Kashshāf exists. It is MS. Lâleli 216, copied in 718H in Baghdad (the copyist's name is given as Ibn ʿAmīd Kātib al-Itqānī); at the top of folio 2A a gloss says: al-rābiʿ min al-Kashshāf li-Jār Allāh al-ʿallāma bi-khaṭṭ Qawām al-Dīn al-Itqānī. At the end, al-Itqānī writes: “[It] was copied from the autograph copy written in [the author's] handwriting, bequeathed to the tomb-shrine (mashhad) of the great Imām Abū Ḥanīfa (May God be pleased with him and his followers)”. MS. Lâleli 216 contains only the commentary on Q15–Q23 and does not have the Shaddākhat al-Muʿtazila in the margins.

83 In the Itqānī manuscript (MS. Lâleli 216) and its two copies (MSS. Ali Emiri Efendi 80–81 and Nurosmaniye 290/399), it is clearly stated that the copy of the Kashshāf used was the author's, which had been bequeathed (mawqūfa) to the Abū Ḥanīfa tomb-shrine (mashhad) in Baghdad. MSS. Lâleli 216 and Nurosmaniye 290/399 both add that this copy was written in his own handwriting.

84 Completed on 09 Muḥarram 1050H by ʿAlā l-Dīn al-Shuhūdī. It was later compared with al-Itqānī's copy and collated with it and another copy, completed on 27 Ramaḍān 1051H. The short muqaddima on folio 1A was probably added at this time; it concludes: wa-nuqilat min khaṭṭ al-shaykh Qidām al-Dīn al-Itqānī raḥimahu llāh taʿālā.

85 For more on al-Zamakhsharī's life and work, see my A Traditional Muʿtazilite Qur’ān Commentary, Chapter 1 and Appendix 3.

86 Abū ʿAlī (or Muḥammad) al-Ḥusayn b. Muḥammad b. ʿAbd Allāh al-Ṭībī's (d.743/1342) commentary on the Kashshāf, entitled Futūḥ al-ghayb (fī l-kashf ʿan qināʿ al-rayb), is one of several works he composed. Al-Suyūṭī describes him as an eminent scholar (ʿallāma) in the sciences, Arabic, and rhetoric; and quotes Ibn Ḥajar's exceedingly positive opinion of him. Al-Suyūṭī says that, in his commentary on the Kashshāf, al-Ṭībī mentions that shortly before beginning to write it he saw the Prophet in his sleep, who offered him a bowl of milk from which he drank (al-Suyūṭī, Bughya 1: 522–3 [no. 1080]; GAL ¹ 1: 290; 2: 64; GAL 1: 345; 2: 76; GAL.Sp 1: 508; 2: 76; al-Fihris al-shāmil li-l-turāth al-ʿarabī al-islāmī al-makhṭūṭ (Amman: al-Majmaʿ al-Malakī li-Buḥūth al-Ḥaḍāra al-Islāmī, Mu'assasat Āl al-Bayt, 1987) 2: 515). ʿImād (ʿIzz) al-Dīn Yaḥyā b. al-Qāsim al-ʿAlawī al-Fāḍil al-Yamanī (d. 750/1348–9), a Shāfiʿī adīb, grammarian and commentator, travelled to Baghdad, Damascus and Khurasān. Al-Suyūṭī says he was familiar with the Kashshāf and wrote a gloss (taʿlīqa) on it. Al-Ziriklī gives the title of his gloss, finished in 733/1332, as Tuḥfat al-ashrāf fī kashf ghawāmiḍ al-Kashshāf, that listed in GAL.Sp. However, it is listed as Durar al-aṣdāf fī ḥall ʿuqad al-Kashshāf in GAL and the Fihris al-shāmil. Interestingly enough, in GAL.Sp Durar al-aṣdāf is the title of al-Yamanī's gloss on al-Ṭībī's Futūḥ al-ghayb. Kaḥḥāla lists both without further comment (al-Suyūṭī, Bughya 2: 339 [no. 2130]; ʿUmar Riḍā Kaḥḥāla, , Muʿjam al-mu'allifīn. Tarājim muṣannifī l-kutub al-ʿarabiyya (Beirut: Mu'assasat al-Risāla, 1993) 4: 110Google Scholar [no. 18100]; Khayr al-Dīn al-Ziriklī, , al-Aʿlām. Qāmūs tarājim li-ashhar al-rijāl wa-l-nisā’ min al-ʿarab wa-l-mustaʿribīn wa-l-mustashriqīn (Beirut: Dār al-ʿIlm li-l-Malāyīn, 1979) 8: 163Google Scholar; GAL ¹ 1: 290; GAL 1: 345–6; GAL.Sp 1: 508; al-Fihris al-shāmil 2: 515) ʿAlī b. Muḥammad b. ʿAlī al-Jurjānī al-Ḥusaynī al-Ḥanafī al-Sayyid al-Sharīf (d. 816/1413) travelled to Harāt to study with Quṭb al-Dīn al-Rāzī al-Taḥtānī (d. 766/1364) who, because of his advanced age, turned him down and sent him to one of his students in Egypt. Tritton writes that he “belonged to an age which wrote commentaries on earlier works” and that in his Taʿrīfāt, “he was not afraid to be simple” (pp. 602–3). He wrote on many subjects: grammar, logic, law and language; and as a theologian devoted considerable space to philosophy. Kaḥḥāla says he wrote over 50 works, and almost all of those that he lists are ḥāshiyas, sharḥs, or ḥāshiyas on other sharḥs or mukhtaṣars (al-Suyūṭī, Bughya 2: 196–7 [no. 1777]; A. S. Tritton, “Al-Djurdjānī, ʿAlī b. Muḥammad”, EI ² 2: 602–3; Kaḥḥāla, Muʿjam al-mu'allifīn 2: 515 [no. 10037]; GAL ¹ 2: 216–7; GAL 1: 346; 2: 280–1; GAL.Sp 2: 305–6 GAL.Sp 1: 508). Al-Jurjānī never finished his gloss on the Kashshāf and this is evident in editions of the Kashshāf that contain it, for it ends in the first volume.

87 Here I am following Topkapi MS. 1780 A230 (copied about the 11^th/17^th century). The opening words of the Kashshāf are: al-ḥamdu li-llāh alladhī anzala l-Qur’ān kalāman mu'allafan munaẓẓaman wa-nazzalahu bi-ḥasab al-maṣāliḥ munajjaman wa-jaʿalahu bi-l-taḥmīd muftataḥan wa-bi-l-istiʿādha mukhtataman (Kashshāf 1: 41).

88 The Arabic text reads: hādhā huwa l-murād lā mā qīla innahu qāla awwalan khalaqa l-Qur’ān thumma ghayyarahu taqiyyatan li-annahu ṣaraḥa bi-dhālika bi-qawlihi ‘wa-mā hiya illā ṣifāt mubtada’’ ilā ākhirihi. The words wa-mā hiya illā ṣifāt mubtada’ (mubtadaʿ) ilā ākhirihi to which al-Ṭībī refers are the continuation of the passage in the Kashshāf. In it al-Zamakhsharī describes the Quran and refers to it in terms of a created being (e.g. mubtada’ mubtadaʿ, munsha’ mukhtaraʿ).

89 Tajnīs al-ishtiqāqī could be translated as same-root paronomasia. Heinrichs writes: “A general definition of tadjnīs would be: a pair of utterances (mostly, but not necessarily single words), within a line or colon, which are semantically different but phonetically, either completely or partially, identical. The alternative “completeness or lack of such” is the basis for distinguishing the various subtypes that the rhetoricians have discovered. Since words that are only partially identical are very likely to be semantically different anyway, it becomes clear that two notions have merged in the tadjnīs concept: a narrow one which covers only the case of complete phonetic identity (this is the tadjnīs tāmm, which some say, or imply, is the original and “correct” meaning of the term), and a broader one in which the two terms of the tadjnīs show any kind of lesser degrees of assonance, down to root-repetition (ishtiḳāḳ, figura etymologica). Some authors deny that ishtiḳāḳ is a subtype of tadjnīs” (W. P. Heinrichs, “Tadjnīs”, EI ² 10: 67).

90 Here I follow MS. Nurosmaniye 417/563, completed on 03 Ṣafar 997 in Yemen by ʿAbd al-Hādī b. Saʿīd al-Yamanī. The card catalogue gives no title but lists it as a ḥashiya. It is interesting to note that al-Yamanī refers to the “original copy” of the Kashshāf, that people said had khalaqa l-Qur’ān, as nuskhat al-umm. This gives the impression that he is referring to the umm al-Kashshāf that al-Zamakhsharī finished in Mecca in 528/1134, and not the copy at the Abū Ḥanīfa mashhad in Baghdad.

91 Unlike the two previous commentaries, al-Jurjānī's has been printed in a number of editions; here the Muṣṭafā l-Bābī al-Ḥalabī (Cairo: 1385–87/1966–68) edition has been used.

92 Kashshāf (ed. Cairo, 1966–68) 1: 3. The Arabic text reads: yurwā annahu waqaʿa fī umm al-nusakh khalaqa makāna anzala thumma ghayyarahu l-muṣannif fa-in ṣaḥḥa dhālika fa-l-taghyīr li-fawā’id.

93 Nassau-Lees refers to similar results in the first printed edition of the Kashshāf: “I have myself seven or eight good copies, and some of them … are very old; yet in none of them occur (sic) either of the words, khalaqa or J'ala (sic)” (Kashshāf (ed. Calcutta, 1856–59) 1: 7, n. 7).

94 “al-Zamakhsharī”, EI ¹ 4: 1205; Jansen, Interpretation, 62.

95 See for example: Ibn Khallikān, Wafayāt 5: 170 [no. 711], reproduced in al-Fāsī, al-ʿIqd al-thamīn 7: 141 and Ibn al-ʿImād, Shadharāt al-dhahab 4: 120.

96 Makdisi, G., “The Sunnī Revival” in Islamic Civilization 950–1150 (Papers on Islamic Civilization, 3, Oxford: Bruno Cassier, 1973), 157Google Scholar; and Madelung, W., “Abū Yūsuf Qazvīnī”, Encyclopaedia Iranica, ed. Yarshater, Ehsan (London, Boston and Henley: Routledge and Kegan Paul and New York: Bibliotheca Persica Press, 1983–2001), 1: 398–9Google Scholar.

97 For ʿAbd Allāh b. Saʿd (Saʿīd?) Ibn Abī Jamra al-Andalusī (d. 699/1300?), see GAL ¹ 1: 372; GAL 1: 458–9; GAL.Sp 1: 635. His comments on the Kashshāf can be found in his Bahjat al-nufūs wa-taḥallīhā bi-maʿrafat mā lahā wa-mā ʿalayhā (Beirut: Dār al-Jīl, 1972) 1: 46 [juz’ 1]. The Bahjat al-nufūs is a commentary on one of Ibn Abī Jamra's earlier works, the Kitāb Jamʿ al-nihāya fī bad’ al-khayr wa-l-ghāya, a summary of al-Bukhārī's Ṣaḥīḥ.

98 Such examples should be kept in mind when reconstructing earlier texts from references to them in later works.

99 The biographical sources consulted for the original study numbered nearly three dozen and spanned a period of seven centuries, from Ibn al-Anbārī's (d. 577/1181) Nuzhat al-alibbā’ fī ṭabaqāt al-udabā’ to al-Khuwānasārī's (d. 1313/1895) Rawḍāt al-jannāt fī aḥwāl al-ʿulamā’ wa-l-sādāt.

100 Even at this level the selection of Quranic passages (verses or parts thereof) that could have been used at various places in the Minhāj and the Muʿtamad was probably wider than what was used; Ibn al-Malāḥimī says as much himself (see §2.2.1 above). Why certain passages were chosen for specific theological arguments may be hard to explain, beyond the fact that they supported the latter in a general way without getting into the details of the Kalām arguments. If this is the case, the connection was probably too weak for such passages to stand out as the Quranic pillars on which specific theological arguments stood, and this would be reflected in the Kashshāf. In any case, they remained connected to only a small part of the theological treatises, and represented a tiny percentage of the Quran. By al-Zamakhsharī's time, it is possible that the passages he refers to (and maybe others he did not mention) were solidly set in the theological tradition as part of the arguments; something similar could probably be said for the muḥkamāt and mutashābihāt verses of the Quran (n. 33). Nevertheless, it might be possible to trace back the theological tradition to the point where certain passages first appeared.