¹ Al-Ghazāli, , Al-Iqtisād fi al-Itiqād, edited by Çubuçu, I A. and Atay, H. (Ankara, 1962), pp. 80–99.Google Scholar This work will be abbreviated Iqtisad. This is the earliest of two complementary works Ghazali devoted to an exposition and defense of Ash'arite dogma, the second being Qawā'id al-'Aqā'id (The Principles of Belief) which is included as one of the books of his Ihyā' ‘Ulūm al-Dīn (The Revivification of the Sciences of Religion). In the Tahāfut al-Falāsifa (The Incoherence of the Philosophers) Ghazali states that he is writing the Tahāfut to refute the philosophers, not to affirm “true doctrine,” and that he will write another work for affirming true doctrine, giving Qawā'id al-'Aqā'id as its title. Thus at the conclusion of the debate over the first proof of the philosophers’ theory of the world's pre-eternity, he writes: “As regards the true doctrine, we will write a book concerning it after completing this one - if success, God willing, comes to our aid - and will name it The Principles of Belief. We will engage in it in affirmation just as we have devoted ourselves in this book to destruction.” Al-Ghazali, , Tahāfut al-Falāsifa, ed. Bouyges, M. (Beirut, 1928), pp. 77–8.Google Scholar [This work will be abbreviated Tahāfut]. However, it is the Iqtisād, written shortly after the Tahāfut that best fulfils Ghazali's purpose and is the sequel to the Tahāfut. See Hourani, G.F., “A Revised chronology of al-Ghazali's writings,” Journal of the American Oriental Society, 104, 2 (1984): 239–94CrossRefGoogle Scholar; also the author's “Al-Ghazālī on bodily resurrection and causality in the Tahāfut and the Iqtisād,” Aligarh Journal of Islamic Thought, 2 (1989): 46–75, see pp. 49–51.Google Scholar

² Tahāfut, p. 377.

³ Iqtisād, p. 98.

⁴ Ibid., p. 99.

⁵ Iqtisād, pp. 80–3.

⁶ Ibid., p. 80.

⁷ Ibid., p. 81.

⁸ See note 49 below.

⁹ Iqtisād, pp. 83–6.

¹⁰ It should be reiterated that Ghazali does not deny that every temporal event must have a cause. What he denies is (a) that a causal act proceeds from the essence or nature of the agent, (b) that such an agent can be an inanimate existent and (c) that there is any agent other than God.

¹¹ Iqtisād, p. 86.

¹² Ibid., pp. 86–95.

¹³ The source of this argument goes back to al-Ash'ari himself. See al-Ash'ari Kitāb al-Luma’, ed. McCarthy, Richard J. (Beirut, 1953), pp. 39 ff. (pp. 59 ff. in English translation).Google Scholar

¹⁴ Ibid., p. 87.

¹⁵ Ibid.

¹⁶ Ibid., p. 92.

¹⁷ Ibid., p. 93.

¹⁸ Ibid., pp. 95–9.

¹⁹ Ibid., p. 95.

²⁰ Motion for the Ash'arite Ghazali is an accident.

²¹ Iqtisād, p. 96.

²² Tahāfut, pp. 278–9.

²³ Iqtisād, p. 98.

²⁴ See note 81 below.

²⁵ Iqtisād, p. 99.

²⁶ Averroes, , Kitāb al-Kashf'an Manāhij al-Adilla, ed. Müller, M.J., in Philosophie und Theologie von Averroes (Munich, 1859), p. 105Google Scholar; also Averroes, ' criticism of the Ash'arite definition of act in his Tahāfut al-Tahāfut, ed. Bouyges, M. (Beirut, 1930), p. 158. This work will be abbreviated Tahāfut al-Tahāfut.Google Scholar

²⁷ Ibid., p. 92.

²⁸ See Tahāfut, p. 15: “Let us concede all this to [the philosophers], either dialectically or out of conviction.”

²⁹ For a comprehensive discussion of the Ash'arite doctrine, see Gimaret, D., La doctrine d'al-Ash'ari (Paris, 1990), pp. 131 ff.Google ScholarIqtisād (p. 181, 1. 11) affirms this doctrine.

³⁰ Ghazali affirms the priority of divine power to the world's creation in Qawā‘id al-‘Aqā’id of his Ihyā’ ‘Ulūm al-Dīn, in al-Rukn al-Thālith, al-Asl al-Thānī. Al-Ghazālī, , Ihyā’ ‘Ulūm al-Dīn, 4 vols. (Cairo, no date), vol. I, pp. 193–4.Google Scholar

³¹ Tahāfut, pp. 52 ff. This doctrine is reaffirmed in al-Maqsad al-Asnā. Al-Ghazālī, al-Maqsad al-Asnā, ed. Shehadi, F. (Beirut, 1971), pp. 159–60.Google Scholar See also the author's “The Logical role of the argument from time in the Tahāfut's second proof for the world's pre-eternity,” The Muslim World, XLIX, 4 (1959): 306–14.

³² Tahāfut, pp. 277–8.

³³ Ibid., pp. 278–9; Iqtisād, p. 97.

³⁴ “To deny efficient causes that are observed in sensible things, is sophistical talk,” writes Averroes in his criticism of Ghazali. Tahāfut al-Tahāfut, p. 519. But the point at issue is whether we observe necessary connection between the efficient cause and its effect. Averroes here does not answer this point, although he goes on, as we see it, to give more effective criticisms.

³⁵ Sīnā, Ibn (Avicenna), al-Shifā' (Healing): al-Ilāhiyyāt (Metaphysics), ed. Anawati, G.C., Dunya, S., Musa, M.Y. and Zayid, S., 2 vols. (Cairo, 1960), p. 8, lines 8–9Google Scholar; al-Burhān (Demonstration), ed. Affifi, A.E. (Cairo, 1956), pp. 96, 223, 249–50.Google Scholar For Avicenna the observation of regularities is not sufficient for establishing necessary causal connection. In addition there is the “hidden syllogism” to the effect that if these regularities were coincidental or accidental they would not have happened always or for the most part. From this he concludes that the regularities derive from the inherent nature of things that connect them causally and necessarily. Avicenna's “hidden syllogism,” however, is open to the criticism that it is circular since the accidental and the coincidental are by definition those things that do not happen always or for the most part.

³⁶ Ghazali concludes that the regularity derives from God's decree, not from any causal properties in created things. Al-Ghazālī, Miʽyar al-ʽIlm, ed. S. Dunya (Cairo, 1960), pp. 189-91. Nature's uniformity is decreed by t he divine will, but is not necessary in itself. Hence its disruption when God creates a miracle does not involve contradiction. The events that are normally regarded as causes and effects but are not necessarily connected can best be referred to as “habitual” (to follow Ghazali's language) in two senses of the term, (a) They follow the habitual course of nature ordained by God – the decreed uniformity that is only disrupted when God creates the miracle, (b) They are “habitual” in the sense of our habitually believing them to be causes and effects. These habitual causes and effects, are not real causes and effects. But they do follow an order which includes the relation of the priority of the cause to the effect, a priority which in itself is not necessary, but is part of the contingent uniformity decreed by God. It is on the basis of this uniformity that we are able to infer the habitual effect from the habitual cause and t h e habitual cause from the habitual effect – that is when the proper conditions obtain, conditions that are part of the order which is not necessary in itself, but which is decreed by t he divine will. Thus, to give but one example, we can infer the habitual cause from the habitual effect when (barring impediments) it is ascertained that there is only one habitual cause. (This is not to deny that the uniformity ordained by God does not contain necessary connections – life is a necessary condition for knowledge, for example - but these are not causal connections). The only suspension of our ability to make scientific causal inferences is when there is a disruption of the uniformity, when the miracle takes place. (God t h en does not create in us knowledge of the uniformity, creating instead knowledge of t h e miracle). In his writings, Ghazali ordinarily (when he is not discussing the metaphysics of causality) speaks of causes and effects within the created world without having to constantly remind us that these are habitual, not real, causes and effects, in the same way that he speaks of human will and human power without having to constantly tell us that this is created will and created power that has not real efficacy. For a fuller discussion of Ghazali's interpretation of Aristotelian causal theory in occasionalist terms, see t h e authors' “Ghazali and demonstrative science,” Journal of the History of Philosophy, III, 2 (1965): 183– 204, and, “Ghazali's attitude to t h e secular sciences and logic,” in Essays in Islamic Philosophy and Science, ed. G.F. Hourani (Albany, N.Y., 1975), pp. 100–11. For a discussion of t h e causal theories discussed in the 17th Discussion of the Tāhafut in the light of the discussion of causality in its sequel, the Iqtiṣād, see the author's “Al- Ghazāli's second causal argument in the 17th Discussion of his Tahāfut,” in Islamic Philosophy and Mysticism, ed. P. Morowedge (Delmar, N.Y., 1981), pp. 85–112, and “Al-Ghazāli on bodily resurrection and causality in the Tahāfut and the Iqtiṣad,” the full reference already cited in n.l.

³⁷ See, for example, Jawāhir al-Qur'ān, ed. Qabbānī, R.A. (Beirut, 1986), p. 67.Google Scholar

³⁸ The translation is based primarily on the Ankara edition to which we have been referring. Three other uncritical Cairo printings have also been consulted. Two of these are undated and published by two different commercial presses, Maktabat al-Husayn al-Tijāriyya and Matba'at Mustafā Bābā al-Halabī; the third, published by al-Jundi press, is dated, 1972, and is edited by al-Shaykh M.M. Abu al-'Alā. These versions seem to be copies of each other. Where these vary from the Ankara edition, the latter usually, but not always, includes such variants found in the manuscripts it uses in its apparatus criticus. As will be indicated, there are places in the translation where the readings in the Cairo versions have been preferred.

³⁹ Ghazali, however, uses the term usūl, “principles”, which is more concordant with kalām terminology, rather than muqaddimāt, “premises” the term usually used in logic. But the intention is the same. “Premises,” however, seems preferable here as it conveys more directly the syllogistic nature of the argument.

⁴⁰ Knowledge received directly by the senses is “necessary knowledge,” in kalām epistemology. See note below.

⁴¹ Darūratu al-'aql, literally, the “necessity of the intellect.” The darūrat al-'aql in kalām language refers to al-'ilm al-darūrī, “necessary knowledge,” which includes self-evident truths, knowledge of one's existence, of one's states and knowledge received directly by the senses, as distinct from ‘ilm nazarī, “reflective knowledge,” arrived at through inference.

⁴² This is the disjunctive syllogism, al-sabr wa al-taqsīm, to use the terminology of the kalām, used by both the mutakallimūn and the philosophers where the disjunct is exclusive and restricted to the alternative mentioned.

⁴³ Adopting the reading in the Cairo versions and recorded in the apparatus criticus of the Ankara edition, yanqalib ‘alaykum hādhā fi al-qudra fa-innahā qadīma wa al-fa'l laysa bi-qadīm. The Ankara body of the text reads: fa-hādhā yanqalib 'alaykum fi al-qudra fa-innahā qadīma qultum lam yakun al-maqdūr qadīman, “This [argument] turns against you with respect to power; for it is pre-eternal [but] you said that what is enacted by power is not pre-eternal.”.

⁴⁴ Al-Ahkām.

⁴⁵ Iqtisād, pp. 101 ff. The divine will, also an attribute additional to the divine essence, has as its function specifying one similar thing from among other similar things for the divine power to bring into existence. The issue has to do with the creation of the world at a moment in time. Temporal moments are all similar. But the divine eternal will chooses one moment for creation. By definition the will is that which chooses between similars when there is no determinant to influence this choice. Or as Ghazali puts it, “the will is nothing but an expression of an attribute whose function is to differentiate one thing from what is similar to it.” Ibid, p. 106. This definition is argued for at length in the first Discussion of the Tahāfut, where Ghazali rejects the doctrine of a pre-eternal world. See Tahāfut, pp. 31 ff. Hence the eternal power causes its object to exist at the time specified by the eternal will. It thus does not follow that because the divine attribute of power is eternal, the act it produces is coeternal with it, as the opponent suggests.

⁴⁶ Al-maqdūrāt, the plural of al-maqdūr, a key term throughout Ghazali's discussion. In some contexts it means that which is enactable by power, in others, that which actually has been enacted by power. The context sometimes requires its translation as “the object of power.” We have translated the term in all three ways – depending on context.

⁴⁷ The Cairo texts add, allatī lā nihāya lahā, “which are endless/infinite”. The Ankara edition does not include this in its apparatus.

⁴⁸ Sāni' al-'ālam. The Cairo versions give Sāni' kull al-'ālam (a variant given in the Ankara edition), “maker of the whole world”.

⁴⁹ Iqtisād, p. 17, where the argument is given in its shortest form: to suppose infinite rotations in the past means that at the present an infinite has terminated, for Ghazali, a contradiction. The argument resting on the contradiction of affirming infinities that are unequal is discussed in greater detail in Tahāfut, pp. 31–3 and repeated in the discussion of the divine will in Iqtisād, pp. 104–5. The issue between Ghazali and the philosophers who believed in an eternal world is whether past events, that no longer exist, form an actual, not merely potential infinite. For Ghazali (but not Avicenna) these events, since they have existed do form an actual infinite. See Tahāfut, pp. 33–4. The story is different with future events. These, for Ghazali, do not form an actual infinite – they may indefinitely increase but they never form an actual infinite.

⁵⁰ What Ghazali seems to mean is that should the possible be realized in existence, it is realized only through the act of divine power, not that whatever is possible is always realized through the act of the divine power. The eternal will, for example, chooses only one moment of time, among other possible moments, for creation to take place, the divine power enacting the world only at that moment chosen by the will.

⁵¹ Idhā lam yamtani' al-ta'addud fi al-maqdūr.

⁵² Reading fa-unzur as in the Cairo versions. The Ankara edition gives fal-nanzur, “Let us reflect”.

⁵³ Reading wa lā ta'rif dhālika illā idhā ‘arafta ma'nā al-muhāl wa al-mumkin wa hassalta haqiqatahumā, as given by the three Cairo versions and as an alternative reading in the Ankara edition. The main text of the Ankara edition reads, wa lā na'rifu dhālika illā ba'da an 'urifa [possibly ‘urrifa] ma'nā al-muhāl wa al-mumkin wa nuhassilu haqīqatahumā, “we will only know that after the meaning of the impossible and the possible are known [/made known] and we attain their true natures”.

⁵⁴ Any temporal event which the eternal will chooses to happen, must happen. The event, however, is the consequence of the will's eternal choice, not the necessary consequence of the divine essence or nature, which would render the divine act compulsory. Hence Ghazali's insistence on the Ash'arite dogma that the eternal attributes are “additional” to the divine essence.

⁵⁵ Reading wa nujarrid al-nazar ilā dhāt al-'ālam, as given by the Cairo versions and reconstructed from variants in the Ankara edition. The Ankara text reads, wa mujarrad al-nazar ilā dhāt al-'ilm, “reflection being on the essence of knowledge”.

⁵⁶ Wa hazz al-ma'nā minhu darūrī. The sentence is highly idiomatic, its nuance difficult to capture in a translation.

⁵⁷ Ithbāt maqdūr bayna qādirayn, literally, “affirming of power between two possessors of power.”

⁵⁸ This is an expression of the Mu'tazilite doctrine that a just God does not demand of His servants that which is beyond their capacity, mā fawqa al-tāqa.

⁵⁹ The term when applied to divine action would mean “invention / creation” ex nihilo.

⁶⁰ Lā qudra li-Allāhi 'alayhā bi nafy wa lā ījād. Presumably the nafy here is after the act has already proceeded from the creaturely agent, not that God cannot prevent the creature from acting. The position here seems to be — but this is not certain — that the act being discussed is by definition creaturely, not divine; hence it would be a contradiction for God to perform such an act.

⁶¹ Adlā'ahā, literally, “their sides”.

⁶² I.e. free from angularity, not that these circular figures have straight angles.

⁶³ I.e. in containing maximum area. The text reads: inna aqraba al-ashkāl al-qalīlati al-adlā' ilā al-mustadīra fi al-ihtiwā'.

⁶⁴ Al-ard wa al-samāwāt, as in the Cairo versions. The Ankara edition gives only al-samāwāt.

⁶⁵ Wa inn furidat al-ra'da murāda li al-murta'id wa matlūba lahu. There is an ambiguity here: Is this tremor willed and sought by the person with the tremor or for him? The former makes the much better sense and as such what Ghazali seems to mean is that even if a person wills to have a tremor and seeks it, this will not produce it. Only power would affect this.

⁶⁶ Reading aw as given in the Cairo versions. The text reads, fa yu'addī ilā ijtimā' al-haraka wa al-sukūn ilā al-khuluwwi 'anhumā. If this reading is accepted, it would make better sense if the ilā is eliminated, whereby the sentence would translate: “The [simultaneous] combination of motion and rest would lead to being devoid of both.”.

⁶⁷ Wa al-ikhtirā' yatasāwā, literally, “invention ex nihilo is equal”.

⁶⁸ Wa inn kāna ma'ahu. In other words, the relation of the created power to the object of power is that of concomitance. It is not a causal relation as the text will further show.

⁶⁹ In the Third Discussion of the Tahāfut, the use of the term “act,” is confined to animate beings. The inanimate do not “act” and when the term is used to refer to the action of an inanimate thing like fire, this use, for Ghazali is metaphorical. Tahāfut, pp. 100–1.

⁷⁰ The divine attributes of will and knowledge are not causally connected, but connected nonetheless.

⁷¹ Ghazali here is arguing against the Mu'tazilites on their own terms. For they held that power in created things exists before the act, not only at the time of the act as the Ash'arites held. If then power exists before the object of power, is it at this preceding time, connected or not connected with the object? See Part C above for our comment on Ghazali's argument.

⁷² The Mu'tazilites would then have to say that there is a power that is not connected with anything – causally or not causally.

⁷³ Al-qādiriyya. For the Mu'tazilites God is qādir (powerful) in Himself, not through qudra, an attribute additional to the divine essence. Ghazali is presenting the argument using Mu'tazilite language, but whether one uses qudra or qādiriyya the logic of the argument remains the same.

⁷⁴ That is, “by it” at the time when his power and its relation exist. The “if” clause of this conditional sentence reads literally: “If, then, it is not a necessity of either the existence of power or its connection with the object of power for the object of power to exist by it, …”.

⁷⁵ In other words, if one hypothesizes a power that exists before the object that is caused by it, as the Mu'tazilites hold, where the object of power has not as yet come to existence, or if one supposes a created power simultaneously with an object of power (though the object is created by divine, not human power), as the Ash'arites hold, in either case there is a power without an object of power caused by it.

⁷⁶ The text reads minhumā, a printing error corrected as mahmā in the list of corrections, appended to the text (p. 269).

⁷⁷ Literally, “Your statement.”

⁷⁸ The condition here is a necessary, but not a sufficient, condition. If life is a condition for the existence of the attribute of knowledge, it would be unintelligible to speak of an existent that has no life as having knowledge. Thus it would be unintelligible to speak of an inanimate object, a stone, as having knowledge. See Tahāfut, p. 294.

⁷⁹ For a detailed used of this example, see Tahāfut, pp. 278 ff.

⁸⁰ Alternatively, “its exiting from coldness itself”.

⁸¹ Reading thumma huwa mūjib li-al-'ajz wa al-tamānu' kamā sabaqa as a variant given in the Ankara text and as given in the Cairo versions. The editors of the Ankara text have adopted the reading lā al-tamānu', “not preventiveness,” instead of wa al-tamānu' (and preventiveness). Accordingly, the sentence would then translate something as follows: “Moreover, it leads necessarily to [the attribution to God] of impotence, not [merely] preventiveness”. But if God is prevented from performing an act, then this leads to His inability to perform it. Hence, the reading adopted in the translation seems preferable. The kamā sabaqa, ‘as has been previously [seen]” does not refer directly to any explicit mention or use of the concept of “preventiveness” in the chapter, although the argument is implicit therein: if there is such a thing as an efficacious secondary cause, animate or inanimate, that is preventive of divine action, then this would mean the attribution to God of the inability to execute such an action. The reference, however, is to a previous discussion, that God is the only creator of the world (pp. 73–80). Ghazali refutes the argument that it may be possible that there are several creators, each confined to bringing about some part of the creation. Thus he argues that if, for example, one creator is confined to the creation of substances, another to the creation of accidents, this would be impossible since the existence of substances and accidents are mutually dependent. Hence one of the hypothesized creator's refusal to create accidents, for example, would prevent the other from creating substances (p. 78). Ghazali argues further in the same vein against dualism, the belief of a creator of the good and a creator of evil. The argument based on al-Tamānu' is encountered in earlier Ash'arite writings, in similar but not identical context. Al-Bāqillāni, for example, gives a version of the argument which he refers to as dalīl al-tamānu' (proof from preventiveness) to demonstrate that the celestial spheres have no causal efficacy on events in the terrestrial world. His argument in its barest essentials can perhaps be paraphrased as follows: It is either the case (a)that God cannot prevent the supposed causal action of the celestial spheres and create terrestrial events directly or (b) He can. If (a), then the supposed causal action of these celestial spheres are preventive of divine action. “This leads to His deficiency and to His being created.” This is impossible. If (b), it becomes false that these astral beings have actions and influences (batala an takūna li-hādhihi al-kaw'āl wa ta'thīrāt). Al-Baqillānī, Kitāb al-Tamhīd, ed. McCarthy, Richard J. (Beirut, 1957), pp. 50–1.Google Scholar