[proof for the actual being of the Unmoved Mover]

Since there were three kinds of substance, two of them natural and one unmovable, regarding the latter we must assert that it is necessary that there should be an eternal unmovable substance. For substances are the first of existing things, and if they are all destructible, all things are destructible. But it is impossible that movement should either come into being or cease to be; for it must always have existed. Nor can time come into being or cease to be; for there could not be a before and an after if time did not exist. Movement also is continuous, then, in the sense in which time is; for time is either the same thing as movement or an attribute of movement. And there is no continuous movement except movement in place, and of this only that which is circular is continuous.

But if there is something which is capable of moving things or acting on them, but is not actually doing so, there will not be movement; for that which has a capacity need not exercise it. Nothing, then, is gained even if we suppose eternal substances, as the believers in the Forms do, unless there is to be in them some principle which can cause movement; and even this is not enough, nor is another substance besides the Forms enough; for if it does not act, there will be no movement. Further, even if it acts, this will not be enough, if its substance is potentiality; for there will not be eternal movement; for that which is potentially may possibly not be. There must, then, be such a principle, whose very substance is actuality. Further, then, these substances must be without matter; for they must be eternal, at least if anything else is eternal. Therefore they must be actuality.

[burhān al-ṣiddīqīn, version 1]

Remark. Every existent, if you consider it in itself, not considering anything else, is either such that [267] existence is necessary for it in itself, or it is not. If its existence is necessary, then it is the Truth in Himself, the Necessarily Existent in itself: the Self-Subsistent.

If it is not necessary, then it cannot be said that it is impossible in itself after it has been supposed as existent. Rather, if a condition were attached in respect of its essence, such as the condition of the non-existence of its cause, it would become impossible; or [if a condition were attached in respect of its essence] such as the condition of the existence of its cause, it would become necessary. If no condition is attached to it—neither the occurrence of a cause nor its non-existence—then a third thing is left over for it in respect of its essence, namely contingency. And it is, is respect of its essence, something that is neither necessary nor impossible. Thus every existent is either necessarily existent in itself or contingently existent in itself.

Pointer. In the case when [something] has contingency in itself it does not become existent through itself. For its existence through itself is no more appropriate than its non-existence, inasmuch as it is contingent. If either of them becomes more appropriate, it is due to the presence of something or its absence. Thus the existence of every contingent existent is from something other than itself.

Remark. If this constitutes an infinite series, then each single unit of the series is contingent in itself. The totality depends upon it. Thus it not necessary either, but becomes necessary through another. [268] Let us supply this with an explanation.

Explanation. Every unit in the whole totality is caused, and so [the totality] requires a cause external to its units. For either: (a) It requires no cause at all, so that it would be necessary and not contingent; but how is this feasible, it being in fact necessitated simply by its individuals? (b) Or it requires a cause which is the individuals all together, so that it would be caused by itself, for this totality and “the whole (kull)” are one and the same. As for “each (kull)” in the sense of “each unit (kull wāḥid),” the totality is not necessitated by it. (c) Or it requires a cause which is one of the units; but none of the units is more appropriate than any other for that, since every one of them is an effect, whereas its cause would be more appropriate for this. (d) Or it requires a cause external to all its units. And that is the only remaining possibility.

Pointer. No cause of a totality is anything amongst its units. For it is firstly a cause for the units, then for the totality. If this were not the case, then let the units not be in need of [the cause]. In that case the totality, since it is brought to completion through its units, would not need [the cause either]. Admittedly, there may be something which is a cause for one of the units and not another, but then it would not absolutely be a cause for the totality.

[269] Pointer. For any totality arranged from causes and effects in succession, involving an uncaused cause, [the latter] would lie at the extreme (fa-hiya ṭaraf), since if it were in the middle it would be an effect.

Pointer. It has become clear that every series composed of causes and effects—whether finite or infinite—if there is nothing but what is caused in it, it needs a cause external to it. However, it is doubtless connected to [that cause] as an extreme. It has also become clear that if there is in [the series] something that is not caused, it is an extreme and a limit. Thus every series terminates in the Necessary Existent in itself.

[burhān al-ṣiddīqīn, version 2]

Undoubtedly there is existence, and all existence is either necessary or possible. [567] If it is necessary, then the existence of the necessary is true, which is the conclusion sought. If it is possible, then we will show that the existence of the contingent terminates in the Necessary Existent. Before that, however, we will advance some premises.

These include that at one and the same time, there cannot be for everything that is contingent [in] itself an infinite number of causes¹⁰ that are themselves contingent. This is because all of them either exist simultaneously, or do not. If they do not exist simultaneously but rather one after another, there is no infinite at one and the same time—but let us defer discussion of this for now. As for their existing all together with no necessary existent among them, then either the totality, insofar as it is that totality (whether it is finite or infinite), exists necessarily through itself or contingently in itself. If on the one hand the totality exists necessarily through itself, but each one of its members is something contingent, then what exists necessarily is constituted (yakūnu mutaqawwiman) through things that exist contingently, which is absurd. If on the other hand the totality is something existing contingently in itself, then the totality needs for existence [568] something that provides existence, which will be either external or internal to the totality.

If it is something internal to it, then one of its units is a necessary existent, but each one of them exists contingently, so this is a contradiction. Or it is something existing contingently and so is a cause of the totality’s existence, but a cause of the totality is primarily a cause of the existence of its parts, of which it is one. Thus it would be a cause of its own existence, which is impossible, though if it were correct, it would in a way be the very conclusion that is sought; for anything that is sufficient to make itself exist is a necessary existent. Still, it was [assumed] not to be a necessary existent, so this is a contradiction.

The remaining option is that [what gives existence to the totality] is external to it, but it cannot be a contingent cause, since we included every cause existing contingently in this totality. Therefore, [the cause] is external to it and it also is a necessary existent through itself. Thus, things existing contingently terminate at a necessary existent cause, in which case not every contingent [effect] will have simultaneously with it a contingent cause, and so an infinite number of causes existing at a single time is impossible.

[specification argument]

Having established the temporal origination of the world and shown that its existence commences, it follows that the temporally originated (al-ḥādith) is that whose existence and annihilation are both possible. Yet at every moment that [the originated thing] does in fact occur, it would be possible for it to have happened some moments earlier; and it is contingent whether its existence may have been delayed beyond that moment by some hours. When possible existence occurs instead of continued non-existence, which is also made possible, the mind judges as self-evident that there must be a specifying factor (mukhaṣṣiṣ) that specifies the occurrence [of the temporally originated]. This (may God provide you with guidance) is necessarily clear and calls for neither deep investigation nor further inquiry.

Since it is clear that the temporally originated taken as a whole requires a specifying factor, that specifying factor must be either (a) an agent that necessitates the occurrence of the temporally originated thing, in manner of the cause that is necessarily productive of its effect, or (b) some nature, as was held by the naturalists, or (c) a freely choosing agent.

(a) It is wrong to say that it follows the pattern of causes, since a cause necessarily causes its effect simultaneously. If the specifying factor were assumed to be a cause, it would have to be either eternal or temporally originated. [29] If it were eternal, it would necessarily cause the world to exist eternally as well, leading to the doctrine of the eternity of the world; but we have already set out the proofs for its temporal origination. If [the specifying factor] were temporally originated, on the other hand, it would require another specifying factor, yielding a regress in the argument as to how the determining factor is determined.

(b) Those who claim that the specifying factor is a nature face the same problem. For nature, according to those who affirm it, necessitates its effects as soon as any impediments are removed. If nature were eternal, that would imply the eternity of the world. But if it were temporally originated, there would have to be [another] specifying factor. This consideration suffices to refute them, though perhaps we will refute the naturalists again later on, God willing.

(c) Thus, if it is false that the specifying factor for the temporally originated thing is a necessitating cause or a nature that by itself makes it exist without choosing to do so, then from this it follows conclusively that the specifying factor for temporally originated things is an agent that chooses to produce them, specifying their occurrence with certain attributes and certain times.

[mereological problems]

[The philosophers may] say: the conclusive demonstration for the impossibility of infinite causes is to say: each one of the individual causes is in itself either possible or necessary. If it is necessary, then it needs no cause, but if it is possible, then the whole is characterized by possibility. Every possible thing needs a cause additional to itself. Thus the whole needs an extrinsic cause.

We say: the expressions “the possible” and “the necessary” are ambiguous, unless by “the necessary” is meant that whose existence has no cause and by “the possible” that whose existence does have a cause. So if this is what is meant, let us turn back to what was said: each one is possible in the sense that it has a cause additional to itself, but the whole is not possible in the sense that has no cause additional to itself, extrinsic to it, [rather it is necessary]. (If on the other hand the expression “the possible” means something else than what we have said, then it is not comprehensible.)

If it is said, this would lead to the consequence that the necessary existent would be constituted through that which is possible of existence, which is absurd, we say: [83] if by “the necessary” and “the possible” you mean what we have mentioned, then this is question begging. We do not admit that it is impossible. It’s like saying, “it is impossible for the eternal to be constituted through temporally originated things,” even though according to them time is eternal and the individual celestial movements are temporally originated, each having a beginning but their totality having no beginning. Thus that which has no beginning would be constituted through things that do have beginnings, and whatever is true of those that have beginnings would be applicable to the individual units but not true of the totality. In just the same way, it might be said about each individual unit that it has a cause, even though the totality has no cause. Not everything that is true of the individual units is true of the totality. For it would be true of each individual that it is one, that it is some [portion], and that it is a part, but it would not be true of the totality. Any given place on earth is lit by the sun during the day and becomes dark at night, and each [of these events] is temporally originated after it was not the case, in other words, it has a beginning. But the totality, according to them, is something that has no beginning. Hence it has become clear that whoever allows the possibility of temporal events with no beginning (namely the forms of the four elements and of changing things) is unable to deny an infinity of causes. Because of this difficulty, it results that they have no way of managing to affirm the First Principle.

[kalām argument for God’s existence]

[24.6] The existence of God, the exalted and sanctified, is demonstrated as follows: “the origination of everything originated has a cause (sabab); the world is originated; it follows that it has a cause.” […]

[everything originated has a cause; preponderation]

[25.6] We have included in it two principles, which our opponent might deny. We say to him: “which of the two principles do you dispute?” If he says: “I dispute your statement that everything originated has a cause; how do you know this?” We say: “This principle must be affirmed; for it is immediately evident and necessary according to reason (awwalī ḍarūrī fī al-ʿaql).” Someone who is not moved by it is, perhaps, not moved because it is unclear to him what we intend by the term “originated” and the term “cause.” Once he understands them, his reason will necessarily agree that everything originated has a cause: by “originated” we mean that which was non-existent and then became existent. We say then: “was its existence before it existed impossible (muḥāl) or contingent (mumkin)?” It is false that it was impossible, since what is impossible never exists. If it was contingent, then we mean by “contingent” simply that which can exist and can fail to exist. However, it was not yet existent,¹¹ because its existence is not necessitated through itself. For if its existence came to exist¹² through itself, it would be necessary, not contingent. Rather, its existence required that something to preponderate it over non-existence, so that it might go from non-existence into existence. Thus its non-existence continues insofar as there is nothing to preponderate existence over non-existence: as long as there [26] exists no preponderating factor, it will not exist. By “cause” we intend nothing other than the preponderating factor (al-murajjiḥ). The upshot is that something that is continuously not existing will not go to existence from non-existence until something is established that preponderates the side (jānib) of existence over continued non-existence.

Once the meanings of these terms are fixed in the mind, reason necessarily assents [to this principle]. This is how that principle is made clear: by verifying the terms “originated” and “cause,” not by offering a proof for it.

[on Avicenna’s burhān al-ṣiddīqīn]

Speculative scholars have thoroughly discussed those problems from various angles, but most of them strayed from the straight path, like someone who tries [12] to prove His existence (that is, the existence of the Eternal) on the basis of motion. For even if this is an obvious way, and one that is sufficient to reach the objective, it is a long way to go, and requires the establishment of premises that may be dispensed with by the one who travels the correct way. I don’t deny that many useful things lie in the speculation concerning motion, but I would say that one can dispense with it when it comes to this problem as such. Imām al-Ghazālī, the “proof of Islam”—may God be pleased with him—in his book Iqtiṣād fī al-iʿtiqād devoted some ten pages to proving the Eternal. Upon my life, he may be excused for doing so, because this book of his was written in accordance with the methods of kalām, even if what he says there is superior to other works of kalām. Apart from al-Ghazālī many others have written countless pages on these problems, as is well known amongst the scholars (ʿulamāʾ), and there’s no need to get into all that here.

The certain truth concerning the establishment of the Eternal is to prove it through existence, this being the most general of things. For, if there were no Eternal in existence there would have been no existent in existence whatsoever. For existence is exhaustively divided into the originating and the eternal, that is, that whose existence does have a beginning and that whose existence does not. If there were nothing eternal in existence then nothing at all would be originated, since it is against the nature of the originated to exist through itself. Thus that which is existent through itself is necessary of existence, and the necessary in itself cannot conceivably have a beginning. From these statements results a demonstrative syllogism, which speculative scholars call the conjunctive conditional, so that it may be easily understood by a beginner who is not yet able to perceive intelligible true realities. It says: if there is any existent thing in existence, it necessarily follows that there must be something eternal in existence; this is a certain principle that cannot conceivably be doubted by anyone. Then one says: existence is known without any doubt (qaṭʿan), and this is a second principle that is certain like the first one was. From the two [13] foregoing principles one necessarily concludes to the existence of an eternal existent. This is the proof of the eternal using the method of existence. One cannot conceive of any further elucidation beyond this, either one that is more concise or one that provides verification. […]

[the ṣūfis need no argument]

[13.12] Those who are endowed with penetrating vision which penetrates the veils of the hidden, and with angelic sovereignty, have no doubt concerning the existence of an entity (maʿnā) from which existence has proceeded in the most perfect of ways. He is that which is said to be “outside the veils,” when they call Him “Allah,” the exalted, in Arabic.

[only the necessary can be a cause for existence]

We say: no quiddity that is contingently existent can ever be the cause of necessitation, unless [114] it [merely] prepares, or mediates, or does something else like what contingent existents do.

[Let] A be something contingent, and let A be the efficient cause of B’s existence. It is then known that B is contingently existent. But nothing that is contingently existent exists without its existence’s becoming necessary, so B has become necessarily existent. A, however, is not necessarily existent. Rather it is contingently existent from one perspective and necessarily existent from another: contingency of existence belongs to it through itself, whereas necessity of existence is something it acquires. So nothing but A is the cause of the necessity of B’s existence, but A is contingently existent. Therefore, the contingently existent essence will become an efficient cause of necessity of existence. But this is absurd! Thus [the cause for necessitation] cannot be a quiddity that is contingently existent.

There are however some inquiries and doubts that arise concerning this proof, including the following: [115] A might become the cause of the necessity of B’s existence only insofar as it is necessary, just as fire is the cause of igniting wood insofar as it is hot; for the other attributes of fire have nothing to do with igniting. There is no disputing this example.

Response: it is the heat that is the cause of igniting, not the essence of fire. Granted, heat can only exist in a subject, such as fire, so that igniting is related to the fire insofar it is the bearer of the efficient cause, but not insofar as it is [itself] the efficient cause. If the essence of fire were the efficient cause, then all its attributes would be involved in igniting, especially the essential attributes or those necessary concomitants of which the essence cannot be deprived. We do say that it is only insofar as it is necessary that the essence of A necessitates B. But when we say “insofar as it is necessary,” necessity would be the condition of A’s being a cause, not the cause itself. There is a difference between [116] the condition through which the cause becomes cause and the cause itself, like the cause itself of B’s necessity. This [cause itself] would need to be the essence of A, whatever conditions are applied. Furthermore, this condition—that is, A’s being considered as having necessity, which belongs to it due to something else—does not negate its being considered as having contingency, which it has due to itself. How can anything negate necessary attributes? Thus A’s essence, which is contingently existent, would on the condition of its necessity be the cause of B’s necessity. So contingency would enter into the completion of necessity and the bestowal of existence. How could this be otherwise, given that it is one of the necessary concomitants of the efficient cause, which enters into the completing of A’s essence? But how can [contingency enter into the completion of] something that A necessitates? If on the other hand being considered as contingent were negated from A’s essence as soon as it exists necessarily, then the proof is obviously unsound, since after all this way of considering it belongs to it due to itself, and cannot in any way be negated from it.

Someone may raise a doubt, saying: A’s necessity is the cause of B’s necessity. Nevertheless A’s necessity must exist in a subject, and its subject is A. In the same way, heat is the cause of igniting because it must exist in a subject. If A’s necessity is the cause of B’s necessity, [117] and A’s essence entails contingency, still the contingency that is the necessary concomitant of the subject of A’s necessity need not enter into the completion of [B’s] necessity, [so no absurdity follows].

Response: in fact A’s necessity is nothing that exists in concrete individuals. It is merely an item of intellectual consideration, something that exists only conceptually in the soul and is non-existent in concrete individuals. How then can it be a cause for an object that exists in concrete individuals? For it is not like the heat of fire, since the heat of fire does exist in concrete individuals. Furthermore, the igniting that occurs due to the heat is not anything existent either. Rather it is something privative (you will have a detailed exposition of this argument after this section). Moreover, if A’s necessity—which they imagine to be the cause of B’s necessity—were existent in concrete individuals, then the contingency of A’s essence, which is a subject, would [still] enter into the completion of necessity. For the efficient cause that requires some matter for its existence can perform its act only through the participation of the matter. [118] The matter of A’s necessity is A’s essence. So A’s essence would participate in the completion of necessity, and [therefore] its concomitants, which are contingency and privation, would participate in it as well—which is absurd.

Thus, it has become clear that all essences and quiddities emanate in accordance with an order and through an organized chain from the essence of the highest, first, true Principle alone—may He be exalted!

[on the transfer of contingency from part to whole]

One may object to this, saying that the mistake comes in when they say, “everything contingently existent (kull mumkin al-wujūd),” where by “everything (kull)” they mean the whole (al-jumla). The expression “everything” refers to each single item within the whole, but what applies to the whole (ḥukm al-jumla) is different from what applies to each single item (wāḥid wāḥid). For the whole is a multiplicity, and is either finite or infinite in number, whereas one cannot rightly say about each single item (kull wāḥid) that it is a multiplicity. How then, can you take the whole [series of contingent beings] in place of each single [contingent thing], inferring for it that which applies to the single item?

We say: what applies to the whole need not apply to the single [item] insofar as the former is the whole while the other is a single item, since the whole and the single item do differ in terms of [being] one or many. Yet they do not differ in terms of their nature and quiddity, since the quiddity of the whole and the quiddity of the single item within the whole are one and the same in respect of nature and existence. For [24] the natural place of every single [drop] of water, which is cold and wet, is limited by the outer surface of the sphere of earth and the inner surface of the sphere of air (regardless whether these are real or imaginary). The same applies to water as a whole: it does not differ in nature, location, being a cause, or being an effect. Likewise, the whole that results out of the contingent existents is just like every single one of them in respect of the contingency of existence. The fact that the contingently existent requires the Necessary Existent for its existence applies equally to every single contingent and to the whole. […]

[against an infinite regress of causes]

[26.16] If every cause has a [further] cause, every effect will have an infinite number of causes that are simultaneous with it in existence, and there will be no first cause for them. They say: that whose number is infinite does not exist or come into existence. For if the first effect proximate to us does not exist until its cause exists, and its cause does not exist until the cause of its cause exists, and the cause of its cause of its cause, and the cause of its cause of it cause of its cause, and so on indefinitely, so that without the precedence of a first [cause] that has no prior [cause], the existence of the second will never follow. Hence, the existence of the first cause is made known by the existence of the last effect that is most proximate to us, to which we made reference. Just as causes terminate at the first cause, likewise effects terminate at the last effect, since [the former] is simultaneous with [the latter] in respect of existence, being neither posterior nor prior to it. Therefore [the chain of] causes and effects does not go to infinity.

[proof from the specification of bodies]

The Master Abū al-Ḥusayn [al-Baṣrī]—may God have mercy upon him—proved the reality of the Originator in a different way. He said: we observe bodies that have in common that they are bodies, but differ in other respects, for instance that some are earth, others fire, water, or air. For this [difference] they either require something (amr) or they require nothing. If they require nothing then none of them deserve more to be earth and others water, rather than the other way around. Hence, they must require something [for their differences]. Next, this “something” must be either reducible to the body, for instance “being a body” or “having volume.” This however entails that either all these bodies would need to be water, or all would need to be earth, or fire, and would entail that they would all be one and the same body having this form. Hence, they need some other thing that is not the body. This other thing either has some specification and connection to bodies, or it does not. If it does have a connection to [the body] by way of inhering in it or by way of adjacency (bi-al-ḥulūl fīhi aw bi-al-mujāwira), then, if there is only one [item connected or adjacent to the body], then the same problem follows as before: that all bodies will have one and the same form. If however [the body] has [a connection to] many different [items], then the account we give (kalām) concerning the bodies’ need for those things is the same as the account we would give concerning their need for these forms [they have]. If on the other hand [the specifying factor] is adjacent to them, it is either a body or a substance. Why then is what is adjacent to fire that which necessitated its being fire, whereas that which implies being air is not adjacent to it? The same goes for water and air. Furthermore, the account given concerning the separation of that adjacent [factor] from these bodies, or [the bodies’] mutual separation, would be the same as the account to be given as to why these bodies need their forms. If, however, this [factor] has no connection to these bodies—whether by inhering in them or through adjacency—it must be either a necessitating [cause] or one that chooses freely. If it is necessitating it is either one or more than one. If it is one, why did it necessitate some parts to be fire rather than [156] air? How can it necessitate opposite forms, while itself being one and the same thing? Yet if it is more than one, then if it can necessitate these forms while having no connection to these bodies, it would result that all would be one and the same body all having this [same] form, since it is not more appropriate that it necessitate some of them to be fire and others to be air. So it remains only that it is a freely choosing [agent].

[kalām argument updated with the notions of preponderation and specification]

One proves it [by saying]: bodies are originated and everything originated must have an originator; hence, there is an originator of bodies. If someone says: why do you claim that everything originated must have an originator? The reply is: one may say that it is necessarily known that whatever had not been, and then came to be even though it could have failed to do so, must be due to something. Given the necessary knowledge of this, and a further proof that that thing must be freely choosing, knowledge of the originator results.

Besides which, we can mention a way of showing [the premise], by saying: whatever is originated must either have originated even though its failure to originate instead of its origination was possible, or have originated with its origination being necessary. If it originated and its origination was necessary, [154] then it cannot be that its origination was necessary at some times but not others; in that case, the necessity of its origination would have been eternal or it would not have been necessary at all. If then it originated even though it was possible that it did not originate, then it cannot be that it was more appropriate for it to originate at some times rather than others; hence [again] its origination would be eternal, or would not occur at all. Moreover, it would not deserve to originate more than not to originate; as if it can just as well originate or not, then its origination will not be preponderated over its non-origination unless this is due to some preponderating factor. Indeed, it cannot originate unless there is preponderating factor for its origination, since it must be specified along with being originated. Thus it is established that there must be something that originates the originated. Nothing that is originated escapes the [above] division, so, given that bodies are originated, there must be an originator for them. One may confirm this way [of showing the premise] by saying: if our actions such as building, writing, goldsmithing, and so on cannot be originated without being due to something or to an agent, then how can this world, with all that is in it, be originated without being due to something or to an agent?

[analogy between human and divine cases]

Our Masters—may God have mercy on them—also proved the reality of the originator of the world, saying: the world is originated and must have an originator, on analogy with our voluntarily refraining [from an action] (taṣarrufinā). This inference (al-istidlāl) requires a basis for the comparison (aṣl), a derived case (farʿ), a judgment (ḥukm), and the cause (ʿilla) of the judgment. The basis is our voluntarily refraining; the derived case is the world; the judgment is that an originator is required; and the cause is the origination. Concerning the basis—that is, our voluntary refraining—we have already established this when we established movement and rest. As for establishing the judgment in respect of the basis—that is, that we are needed for voluntary refraining—this is proven by the fact that it is necessitated to occur only due to our motivations or deterrents (dawāʿīnā wa-ṣawārifinā). We say this because, whenever we have a motivation to do something and are able to do it, it must occur; but when we have a deterrent from that action and are able not to do so, it does not occur. Indeed, we say “whatever happens this way occurs through us, because if it occurred due to some other agent, or by itself, then it could occur even while we would prefer that it didn’t occur.” Since this is impossible, we understand that it occurs through us.

[against the burhān al-ṣiddīqīn: preponderation without any reason]

This proof is the one that been passed down by Islamic scholars; we have already explained how it proves the existence of the Originator in the beginning of the chapter. However this proof does not on its own prove the Necessary Existent; this is why its proponent only mentions [in the conclusion] that a cause must be either present or absent [determining, respectively, the existence or non-existence of the world]. Furthermore, the proof does not rule out that the origination of the contingent is its own cause, or its absence [i.e. the absence of a cause that would prevent it from existing].¹³ And there is another point, namely that the view taken in their school requires them to allow that something contingently originating can originate not on the basis of anything. For, they say: nature provides well-designed acts that correspond to what benefits humans, and it organizes [these] acts giving priority to whatever should come earlier and postponing whatever should come later—due to nothing additional to the essence of nature, and with no deliberation, thought, or knowledge concerning the priority and posteriority, and even though its essence is the same in relation to the prior and the posterior.

[against the burhān al-ṣiddīqīn: contingency does not transfer from part to whole]

One may object to the proof in another way, different from those already considered, on the basis of a view taken in their own school. For, they say: every single motion of the sphere has a beginning, yet their aggregate (majmūʿ) has no beginning. Thus they have made a distinction between the aggregate and its units, even though the aggregate is identical to the units. So, one may say to them: if this is so, according to you, how can you exclude that the whole series composed out of causes and effects is different from its units, such that each of them would be caused, while the whole is not caused; or such that each of them would be contingently existent, while the whole is necessarily existent, just as each of them is finite while the whole is infinite?

[an infinite series cannot be taken as a whole]

As for the further premise that needs to be established, namely that if causes proceeded to infinity, then there would be a whole that encompasses an infinite number of units, each of which is caused: what is the proof for this premise? For an opponent may say: whatever is infinite cannot be described as a “whole.” “Whole,” “every,” and “all” (jumla, kull, jamīʿ) are among the accidents and concomitants of finite things. As will be obvious to you, here the meaning of “every” is not “every one (kull wāḥid).” Rather it means the aggregate (majmūʿ) of units, with none of them lying beyond it. According to this viewpoint there is no cause beyond the aggregate of causes, so causes taken all together (kull al-ʿilal) have no existence. But it is rather the case that, whichever aggregate you take, there will be a further cause beyond it, and it will proceed like this to infinity.

[249] This is just like what you say about the rotations of the sphere. According to you, they have no beginning or inception, their past number being infinite. Yet every single rotation does have a beginning and inception. Still, you do not say that because every single [rotation] has an inception and a beginning, the whole [collection of rotations] would have an inception and a beginning. Likewise, on your view human souls that are separated from bodies after death are infinite in number, while beforehand there was time at which they did not exist. Every single [soul] has a temporal beginning for its existence, that is, it is preceded by the time in which it did not exist. Nevertheless, you do not say that, since every single [soul] was preceded by the time at which it did not exist, the same goes for the whole, on the grounds that the whole is the aggregate of these units.

Nor are you forced to say this, precisely because “every” and “whole” do not apply here. Rather, whichever number of them you take, there would be something further beyond it. It would never come to a limit beyond which there is nothing further, so that this limited thing could be “every” and “whole.” If you however conceptualized here an “every” that is the aggregate of these units, so that none of them would remain outside it, then you would necessarily be forced to say that the whole has temporal beginning—that is, that it was preceded by a time at which it did not exist. For if time preceded every single unit out of ten, then whole ten is necessarily preceded by a time at which it did not exist. This does not follow [with an infinite series], though, because there is no “every” nor any whole here at all.

[al-Khayyām’s argument: only the necessary can be a cause]

There are people who claim that this proof [i.e. burhān al-ṣiddīqīn] does not require rejecting circularity or stopping the infinite regress. They said: this is because, if there is a necessary existent among things, then the desired conclusion is achieved. If on the other hand there is no necessary existent among them, then all [things] are contingently existent; but the existence of a contingent existent cannot be traced back to [another] contingent existent, and this for two reasons. First, if the contingent were to produce the existence of something else, then its essence would be involved in that production, since the “being existent (al-mawjūdiyya)” of the producer in involved in its “bringing-into-existence (mūjidiyyatihi).” But the essence of the contingent, as such, is contingently existent. Thus if the contingent were to produce the existence of something else, its contingency would be a part of its being a producer. Yet contingency cannot be a part of being a producer, given that insofar as something is contingent it is not necessary, whereas insofar as it is productive it is necessary. One and the same thing cannot be both contingent and necessary in one and the same respect.

[do we need to show that the world is contingent?]

Know that some people believe that, for establishing the reality of the Necessary Existent, one also needs to establish that the world is contingent. This is not so. Rather we may be satisfied by what has been mentioned above: if among existents there is a necessary existent, the goal [of the proof] has already been achieved. If on the other hand nothing among them is a necessary existent, so that everything is contingent, nonetheless the contingent must be traced back to the necessary, and there will in any case be among existents the necessary existent. Then, when we thereafter undertake to enumerate the attributes of the Necessary Existent, it becomes evident at that point that the world, given the substances and accidents in it, is no necessary existent but rather among the effects of its existence.

[argument from motion]

As for the natural philosophers, they proved [God’s existence] on the basis of motion, on the basis that it cannot but terminate in unmoved movers, and that anything that is possibly true of the unmoved mover must be the case eternally. And whatever is like this must be a necessary existent.

[preponderation argument]

The first way of proving the existence of the Necessary Existent is by way of the contingency of essences. We say: it cannot be doubted that the true realities and quiddities [of things] are existent. In the case of each existent, its true reality is either susceptible (qābila) to non-existence, or not. If its true reality is not susceptible to non-existence due to what it is (li-mā hiya hiya), then this kind of existent is the necessary existent through itself, and this is the conclusion sought. If on the other hand its true reality is susceptible to non-existence, we say: whenever the true reality of an existent is susceptible to non-existence, its true reality relates equally to existence and non-existence. The existence of such a thing is not preponderated over its non-existence, unless through a preponderating principle (murajjiḥ), and this preponderating principle cannot but be existent. Furthermore, if this preponderating principle is contingent, then one may apply the same reasoning to it, so that either a circle or an infinite regress will follow, and both are absurd. Therefore it must terminate in the necessary existent through itself.

[four ways of proving the Creator]

Know that one may infer the existence of the Artificer either through contingency or through origination. Both may apply either to essences or to attributes, so that there are four ways.

[1. contingency of essences: a version of burhān al-ṣiddīqīn]

The first way is “the contingency of essences”. We say: there is no doubt that the existent exists. If this existent is necessary through itself then this is our goal. If it is contingent, then it must have a producer (muʾaththir). If that producer is necessary, then this is our goal. If it [too] is contingent, then it [too] has a producer. If that [second] producer is the effect of the [first] producer, then it follows that each of them requires the other, so that both must require themselves, which is absurd. If however the [second producer] is something other [than the effect of the first producer], then either an infinite regress follows, or it terminates in [43] the necessary. The infinite regress is ruled out, because this aggregate [of an infinity of contingent units] requires every single unit. But every single one of these¹⁴ is contingent, and whatever requires [something] contingent is a fortiori contingent. Therefore the aggregate is contingent and has a producer. It producer is either itself—which is absurd, since the producer is prior to the effect in rank, and it is absurd that something be prior to itself—or it is a part that is internal to [the aggregate]. But this is also absurd, for that which produces the aggregate also produces every single unit belonging to this aggregate. So if we supposed that the producer of the aggregate is one of its units, it would follow that this unit produces itself, which is absurd. Or [this unit] might produce that which produces it, but this would be circular, which we have rejected. Or [finally], the producer of that aggregate may be something outside it. But we know that whatever is outside the whole [aggregate] of contingents is not contingent, but rather necessary. In which case it follows that everything contingent through itself must terminate in an existent that is necessary existent through itself, which is what we wanted to show. Thus it has been established that there must be a necessary existent that is necessary existent through itself, which is the conclusion sought. Concerning the proper characteristics of the necessary through itself, we have mentioned that it must be unique (fard) and immune to division. But every body, and everything that subsists in a body, is composite and divisible. So it has been established that the necessary existent through itself is an existent that is distinct from these bodies [in the world] and from the attributes that subsist in bodies, which is the conclusion sought.¹⁵

[2. origination of bodies: the classic kalām argument]

The second way is to infer the existence of the Necessary Existent from the origination of bodies. We say: bodies are originated and everything originated has an originator. The knowledge of this is necessary, as we have shown. So all bodies have an originator. That originator cannot be a body or anything corporeal, otherwise it would have to originate itself, which is absurd. Nevertheless, one might still ask here, why can’t the originator of bodies be contingent in itself? So to reject the vicious circle and the infinite regress, we need the former proof.

[3. contingency of attributes: the specification argument]

The third way is an inference on the basis of the contingency of attributes. We say: we have already proven that all bodies are equal in respect of their complete quiddity (tamām al-māhiyya). This being so, the specification (ikhtiṣāṣ) of the body of the celestial sphere by which it turns out to be a celestial sphere, and the specification of the body of the earth by which it turns out to be earth, is something merely possible. So it cannot be without a specifier. If this specifier is a body it requires itself for its composition and arrangement, which is absurd. If however it is not a body, this is the conclusion sought.

[4. origination of attributes: the design argument]

The fourth way is an inference based on the origination of attributes. It focuses on the “signs of the horizons and ourselves,” as the exalted put it: “We will show them our signs in the horizons and within themselves” (Q. 41:53). The clearest way [to put it] is for us to say: semen is a body that is homoiomerous in respect of form. It is either homoiomerous by virtue of itself (fī nafs al-amr) or not. If the former is the case, then we say: that which produces the natures of organs and their shapes cannot be nature. For a single nature would yield a spherical shape, so that animals would have to be born spherical in shape, according to a single, simple nature, but this is ridiculous. But if the latter is the case, then each of those parts must have spherical shape, and it would follow that animals [consist] of [44] spherical shapes added one to another, and this [too] is ridiculous. Hence it has been established the Creator of bodies of animals is not nature, but a freely choosing agent. But then, to establish that it is necessarily existent through itself, we [still] need the first way we have mentioned above.

[simultaneity of the infinite series of causes]

If someone says: there is no beginning for the movement of the spheres, according to us, but rather a motion before each motion. Since this is possible, why can’t there be a cause before each cause without¹⁶ end? […] [42.13] The response is to make clear the difference between cause and effects, on the one hand, and the motions of the spheres on the other, namely that whenever something is a cause of the existence of something else it must be existent simultaneously with the existence of the effect. The reason for this is that the existence of the effect is together with either the existence of the cause or its non-existence. But the non-existence of the cause cannot be together with [the effect], since the non-existence of the cause is not the cause of the existence of the effect, so that the existence of the cause must be the cause of the existence of the effect. This being the case, the cause must be existent simultaneously with the existence of the effect. Having confirmed this judgment, we say: if we supposed the existence of an infinite number of causes and effects, they would have to occur at one and the same moment (dufʿatan wāḥidatan), so that contingency of existence and the need for a producer can be predicated of them [all]. As for the motions of spheres taken as a whole, they are never existent [all together]. To the contrary, no two parts of [this series] can exist at the same moment. Hence one cannot predicate of them contingency and the requirement for a limit, given that absolute negation and pure non-existence cannot be described with existing (al-wujūdiyya) attributes. Thus the difference between causes and motions is evident.

[causes and effects must be simultaneous]

Let it not be said: why can’t one say that a cause, while it is [still] existent, necessitates the existence of the effect after the cause perishes? On this assumption, the existence of the effect would be caused by that cause that existed before it. For we say: it is false to say that this cause necessitates at this hour the occurrence of the effect the next day. For, given that it is true that this cause does at this hour truly necessitate that effect, the necessitation of the effect either (a) means the origination of that effect or (b) means something else, from which the origination of that effect derives. (a) In the former case, the existence of the effect must originate at this hour, not afterward. For, given that it is true of that cause that it necessitates [119] that effect at this hour; and this necessitation means the origination of that effect, so long as there exists such necessitation at this hour; that effect must occur at this hour. Saying that the effect has not occurred at this hour but at some other hour contradicts the first statement. (b) As for saying that the necessitation of that effect means something different from that effect, and that effect derives from it, this is false. For it must be true of that different item that at this hour it necessitates that effect at another hour. Its necessitation of that effect would therefore be additional to its essence, so that an infinite regress of necessitations would follow, but all this is absurd. So what we mentioned above has been established: an effect cannot but exist together with the complete producer (al-muʾaththir al-tāmm).

[there is a necessary existent that causes all bodies]

[245r21] It has been established from these points of view, [245v] some of which were confirmed already with the others to be confirmed later, that every body is contingent. You have already learned that both options [that is, existence and non-existence] relate equally to everything contingent. You have also learned that everything like this has a producer. Hence, every body has a producer. Nothing that is the producer of any body is a body. Otherwise it would depend on and be producing itself. Nor is it anything corporeal, since otherwise a vicious circle would follow. Therefore, it has been established that the world has a principle that is neither a body nor anything corporeal. […]

[246r20] Section two: that the Governor (al-mudabbir) [of the world] is a necessary existent. If the Governor of the world is a necessary existent, then this is the conclusion sought. If not, it is contingent and requires a producer. Its producer is simultaneous with it, as you have learned in the chapter [246v] on causes: the existence of the producer must be simultaneous with the effect. Then, if that producer requires another producer, it yields either an infinite regress or a vicious circle, and both are absurd. Or [the series] terminates in the existent that is necessary existent through itself, which is the conclusion sought.

[the need for preponderation is self-evident]

Those who assert the former [that is, that the need for preponderation is self-evident] argue for their position by saying: we observe that sound-minded people agree that, whenever they perceive with the senses that something is originated, they seek after a reason for it; when they hear a sound of a man, they are impelled to know that the man is there; when they see the origination of a building, they are certain that there exists a builder. Actually we may go further, saying that this kind of knowledge occurs in the souls of children who have yet to reach full development of the intellect. For, whenever a child has a place or position which it selects as being at his disposal, and then finds some food in that place which he has not put there, or finds that something is absent after having put it there, it cries out, “Who has taken it?” “Who has put it here?” This indicates that the inborn nature (fiṭra) of that child is aware that the contingent cannot be without a preponderating factor, or the originated be without an originator. Since this knowledge is seated in the instinct (gharīza) of that child’s soul, we understand that it is the strongest of self-evident ideas. [75] In fact, we say that this kind of knowledge is seated in the souls of animals as well. For if an animal hears a sound of a snake, it flees. It flees simply because its awareness of the sound of a snake implies its awareness of the snake’s existence. This indicates that inborn nature and the soul make a transition from the effect to the producer, in the souls of children and even animals.

[answering objections to the need for a preponderating factor]

Those who assert that the world is originated are more numerous than those who assert that it is eternal. Though they do outnumber [their opponents], they are forced to say that He—may He be exalted—has become the agent for the world after He was not an agent for it. They agree that this agency arose without any reason (laysa bi-sabab), and then an instance (maʿnā) of origination and arising occurs for no reason in this case. If however the impossibility of this [view] were a matter of necessary knowledge [as claimed by the opponents, who insist on a need for a preponderating factor], then the sound-minded would not go along with this.

Someone who is fleeing from a beast and encounters two paths that are equivalent in all respects chooses one over the other, without any preponderating factor. The same applies to someone who chooses between two equivalent cups of water to drink from: he chooses one over the other without any preponderating factor. Clearly there are many examples of this kind. So, most people agree that preponderation may occur without any preponderating factor (al-rujḥān lā li-murajjiḥ). If this premise were self-evidently known to be impossible, then a large faction of sound-minded people would not endorse the claim. […]

[79.18] When there occurs to our minds the fact that one is half of two, and then the statement that the contingent requires the preponderating factor, we find that the judgment of the mind concerning the first proposition is more evident than its judgment over the second one. The variance in strength between two judgments indicates that the rejection of the principle of preponderation (al-marjūḥiyya) may occur, which in turn indicates that this principle of preponderation is a matter of belief, not certainty (ẓanniyya lā yaqīniyya). […]

[119.8] We do not concede that there is no preponderating factor here [in the case of choosing a path to flee from a wild beast]. [That there is such a factor] is shown in two ways. First, because his movement is in one place rather than in the other. And the second reason: the person’s volition is for one of two movements rather than the other. Furthermore, we do not say that this volition occurred due to another volition in his heart.¹⁷ Otherwise an infinite regress would follow. Instead, this volition in his heart originated due to higher reasons whose details are not revealed to us.

[Abū al-Ḥusayn al-Baṣrī’s argument for preponderation and its rejection]

I have seen that Abū al-Ḥusayn Muḥammad b. ʿAlī al-Baṣrī, among the most insightful Muʿtazilites, proved the truth of this premise in the book he called Taṣaffuḥ. He said: the contingent is that which is indifferent on both sides. If preponderation occurred without a preponderating factor, the preponderation would need to occur along with the occurrence of indifference (al-istiwāʾ). This, however, is a co-occurrence of contradictories, which is absurd. To which one might object: the contradiction does not follow, because the contingent is that whose quiddity does not require the preponderation of one side over the other. The contradictory of this proposition would be that this quiddity does require the preponderation of one side over the other. But to assume the [proposition] “its true reality requires no preponderation,” and then that preponderation occurs neither through itself not through anything else, entails no contradiction.

[contingent things need an external preponderating factor]

The contingent is that to which existence and non-existence relate equally. Such a thing cannot enter into existence unless its existence is preponderated over its non-existence. That preponderation (al-rujḥān) must be an attribute of something else, which preceded the existence [of the contingent thing]. The existence [of the contingent] cannot be the subject of inherence for that preponderation, because if that preponderation were an attribute of its existence, it would be posterior to its existence. We have however shown that it must be prior to the existence [of the contingent], so a vicious circle would follow, which is absurd. Therefore, that preponderation must be an attribute of something else and follow from its existence. That thing is the producer (al-muʾaththir). So it has been established that everything contingent requires a producer.

[objection to preponderation, and response]

The contingent thing’s requirement for a producer will be simultaneous with either [the contingent thing’s] existence or its non-existence, but both options are false, so the talk of a requirement is false. The distinction [between the two options] is obvious. The reason the requirement cannot be simultaneous with the existence [107] is that it would imply bringing [something already] existent into existence, which is absurd. But the requirement cannot be simultaneous with the non-existence either, because the producer is that which has an effect, whereas non-existence is pure negation. So saying that production happens simultaneously with the effect’s being pure non-existence is absurd. […]

[113.9] Response: why can’t one say that the producer’s production of existence occurs simultaneously with existence, not before it or after it? The opponent says: this would entail bringing into existence [something already] existent, which is absurd. To which the response is: it is indeed absurd to bring into existence an existent that was already existent before it was brought into existence. However, bringing into existence an existent that has not yet been existent before being brought into existence, but occurs only at the moment it is brought into existence—why do you say that this is absurd?

[non-existence requires no preponderation, so neither does existence]

The possibility of existence is connected to two sides, that is, existence and non-existence. So if the possibility of existence entailed that existence requires a producer, then the possibility of non-existence would [likewise] entail that non-existence requires a producer. But it is absurd that non-existence should require a producer, since non-existence is sheer negation and pure privation, so it absurd to make it an effect or something produced. Also, given that non-existence endures from eternity forever and that which remains cannot be traced back to a producer while it remains. Thus it has been established that, if the possibility of existence made existence require a producer, then the possibility of non-existence would also make non-existence require a producer; but it has been established that it would be absurd for non-existence to require a producer. It follows that it is also absurd that existence require a producer.

[response]

[115.9] There is no dispute as to whether contingency is connected with two sides, existence and non-existence. Nevertheless, the preponderation of existence is due to the existence of whatever produces existence, whereas the preponderation of non-existence is due to the non-existence of whatever produces existence. This is the well-known thesis that the non-existence of a cause is the cause of non-existence.

[argument against contingency]

The first proof is: we say either that existence is identical to quiddity or that existence is distinct from quiddity. Either way, one cannot speak about contingency. (a) As for the option that existence is identical to quiddity, we say: [in this case] contingency is inconceivable. For, whenever something is described as being contingent with respect to existence and non-existence, it may sometimes be described with existence and sometimes with non-existence. Yet existence cannot remain while there is non-existence! Thus it has been established that if existence is assumed to be identical to quiddity, one cannot judge the quiddity to be contingent with respect to existence and non-existence. (b) As for the option that existence is distinct from quiddity, we say: in this case, that which is described with contingency is either (b1) the quiddity, (b2) existence, (b3) or the fact that the quiddity is described with existence. All three options are false. (b1) Quiddity cannot be that which is described with contingency, because if we say “black can be black and can fail to be black”, we would mean that black could be judged to be not black. This is absurd, since it entails that while it is black it can be not black, which is co-occurrence of contradictories, and that is absurd. (b2) Nor can [203] it be existence that is described with contingency, because this would amount to saying that existence can become not-existence, and this is obviously false. (b3) Nor can it be quiddity’s being described with existence that is described with contingency, because the points we mentioned in the case of the quiddity and existence also apply to quiddity’s being described with existence. Thus it has been established that speaking of the contingency of existence is unintelligible, regardless whether we say that existence is identical to the quiddity or distinct from the quiddity. […]

[response: intuition of contingency]

[205.18] By “contingency” we mean that something can keep on being as it has been before, or also not continue to be as it has been. Once our understanding of “contingency” is clarified [in this way], the aforementioned doubts will vanish. For we know that it is not impossible for a sitting man to remain sitting, nor is it impossible that his sitting ends. This knowledge of this is necessary (ḍarūrī). Having understood this, we say: whenever something originates after eternal non-existence, [206] there is necessary knowledge that it could have remained in that original state of non-existence, and that this non-existence can be replaced with existence. So long as both options are indifferent, neither preponderates over the other without a preponderating [factor].

[everything that exists exists necessarily]

Eleventh chapter: So long as the contingent does not become necessary, it does not come to exist.

Demonstration: the contingent [taken] together with its cause is either in the same state as it was [taken] without its cause or not. The first is false, since if this were so, then the cause would not be a cause, and this is a contradiction. If however the state [of the contingent] is different from that previous state, and without the cause it is in a position of indifference [to existence and non-existence], then [taken] together with the cause it leaves the position of indifference and one side becomes more appropriate for it. We say then: the preponderated side (al-marjūḥ) cannot possibly occur, since when it is indifferent and not preponderated it cannot possibly occur. However, at the point when [that side] becomes preponderated,¹⁸ thereby not being any stronger in respect of the impossibility of occurrence, [the impossibility of its occurrence] is even more appropriate.¹⁹ Whenever it is impossible for the preponderated (al-marjūḥ) side to occur, it is necessary that the preponderating (al-rājiḥ) side occur, since it would be absurd to go outside the two sides of the contradiction.

If someone says: the contingent fluctuates between existence and non-existence, not between necessity and impossibility. So how can you posit necessity as prior to existence? We say: the contingent has two kinds of necessity: one occurs to it after its existence. As you have learned, this is the fact that something is necessarily existent (wājib al-wujūd) given the condition of its existence (bi-sharṭ wujūdihi). The other is prior to its existence. As has been shown, this is the fact that so long as it does not leave the position of indifference and does not enter the position of necessity, it is impossible that existence should occur to it. However, given that existence and non-existence are two terms of necessity and impossibility, one cannot but say that the true reality fluctuates between existence and non-existence, not between necessity and impossibility.

[traditional kalām argument]

A second way for putting forward this proof is to say: bodies are originated, and everything originated requires an originator, regardless whether that originator is possible or necessary. Most masters of theology (kalām) relied on this way [of arguing]. Then they have two ways [to support the proof]:

First, those who say there is necessary knowledge that the originated requires a producer. They say: an indication of this is the fact that whenever one sees a building, whether large or small, one must surely know that it has a builder and artificer, such that if someone allows that that building originated without an agent or builder, everyone thinks he’s crazy. So we know that this premise is just self-evident.

Second, those who say that this premise is susceptible of proof. These are most of the Muʿtazilite masters, such as Abū ʿAlī and Abū Hāshim [al-Jubbāʾī]. Their approach is to establish that the servant [of God] brings his own acts into existence. Then they establish that our acts depend only upon us, given that they originated after having been non-existent. So it is evident [129] that origination is the reason (ʿilla) for the requirement of an originator. Since the world is originated, it necessarily requires an agent.

However there are complex questions that arise about their argument.

First: we do not concede that anyone among us originates acts himself. (We will explore this, God willing, [when discussing] the problem of the creation of acts.) But why can’t one say that our acts originate along with our intentions and motives, not through our power and motivation but accidentally (ʿalā sabīl al-ittifāq), without any producer? If they say that accidental origination is absurd, then they should have mentioned this from the very beginning when speaking about the origination of the world, so that the origination of the world would indicate the existence of an agent without any need for this argument [of theirs].

[improved version of the traditional kalām argument]

The first way is an inference on the basis of the origination of essences. We say: (a) the world is originated; (b) everything originated has an originator; therefore the world has an originator.

The proof that (a) the world is originated has been already presented.

The proof that (b) everything originated has an originator is that (b1) everything originated is contingently existent, and (b2) everything that is contingently existent requires an existent producer for its existence. From which it follows that everything originated requires an existing producer.

The proof that (b1) everything originated is contingently existent is that the true reality of the originated is either susceptible to non-existence or not. If not, it is never non-existent and is [400] eternally existent [which is absurd given that it was stipulated to be originated]. If however it is susceptible to non-existence as well as to existence, then it is contingently existent, since this is precisely what we mean by “the contingent.”

The proof that (b2) everything contingent requires an existing producer [is in two steps]. As for its requirement of a producer, the reason for this is that it is susceptible to both non-existence and existence; if there were no [additional] factor (amr), then the one side would not preponderate over the other or vice-versa. There is necessary knowledge of this. As for the fact that that producer must be existent, the reason for this is that there is no difference between a negated producer and the negation of a producer. Hence, claiming that a negated producer suffices amounts to claiming that there is no requirement for the producer. Yet we have already rejected this. Therefore it has been established that contingents require an existing producer.

[proof from specification]

The exposition of the inference on the basis of the contingency of attributes is that bodies have corporeality in common, yet differ in attributes and directions (jihāt). Each of them is specified, insofar as it is specified, either not due to something—but then the contingent would occur for no reason—or due to something. Which would be either those bodies [themselves]—but this is absurd, given that equivalence in corporeality entails equivalence in all those [specific] features—or something else. The latter might be (a) something that inheres in them [sc. the bodies]. But this absurd, since (a1) if it is their necessary concomitant, then its concomitance is (a1a) either due to corporeality, but this just takes us back to what is shared in common; or (a1b) it is due to something additional, and then the argument concerning [this additional factor] is the same as the first argument [about the specific attributes of bodies], yielding either a circle, an infinite regress, or what was just mentioned about what is shared in common. (a2) If however it is not a necessary concomitant, then no necessary concomitant [such as the specific character of a certain body] will occur because of it. (b) Alternatively [the distinct specifying factor] is their [sc. the bodies’] subject of inherence. But this is absurd, given what you have learned about the impossibility of corporeality inhering in a subject. (c) Or it neither inheres in [bodies] nor is their subject of inherence. In the latter case, if it is a body or something that specifies [body] by inhering, or by being a subject of inherence, then we are just back to the [same] set of options. If however this is not the case, then it is either more appropriate for some bodies, rather than others, to receive a given effect from that separate producer—but then the same set of options can be applied to this “appropriateness”—or not. Therefore, that separate [specifying cause] is not necessitating. Otherwise it would be no more appropriate for some bodies, rather than others, to be specified by being receptive of a given effect from that separate [cause]; nor would [the given effect] be more appropriate than any other. Thus, it has been established that bodies and their attributes require a freely choosing agent.

[design argument, drawing on an analogy given by Abū Bakr al-Rāzī]

To their question, “why can’t a natural begetting power be a producer of the origination of these bodies?” We say: we know necessarily that the natural powers existing in our organs have neither complete wisdom nor complete power. And we know necessarily that whatever is devoid of knowledge and power cannot bring into existence an arrangement such as this, that encompasses these wondrous benefits. Whatever one might mention in confirmation of this premise would be a matter of clarifying things that are already clear, and whatever the opponent mentions to refute these things would be a matter of rejecting the obvious.

[Abū Bakr] Muḥammad b. Zakariyyāʾ al-Rāzī has written a treatise on proving the existence of the wise God on the basis of the human body. At the beginning of this treatise he said: whoever sees a jug and considers how it is composed, sees its head like a wide funnel, sees its structure that is well balanced between narrowness and width, sees its handle that has a specific shape, and then understands that its [224] wide head is good because one can pour water into it, and understands that the rest of it has a good structure, which falls right in the middle between width and narrowness, because water will flow out of it with a moderate power, and understands that its handle is good for taking²⁰ it in the hand. Thus whoever has a mind devoid of all sorts of defect, and pure, will be certain that this jug, composed as it is out of these parts that are good for these benefits, has not come to be on its own nor created itself. Moreover it has not come to be due to nature, which is devoid of awareness and perception. Rather he is certain that a powerful, knowledgeable agent knew that what is beneficial in light of certain specific purposes with this jug would be fully attained only through the realization of these three parts, in accordance with their specific attributes. […] [225.14] After Muḥammad b. Zakariyyāʾ provided this nice discussion to prepare the way for his main purpose, he started to explain the effects of the wisdom of the Merciful on the creation of human body. He mentioned something of the wondrous compositions in it and the forms that correspond to wisdom and advantage (maṣlaḥa), and then said: the sound mind is aware that these wondrous and astonishing things in the composition of the body can come into being only from a wise and powerful [agent] who created the structure [of the body] through His power, and its wisely-chosen features through His wisdom.

You should know that this proof mentioned here by Muḥammad b. Zakariyyāʾ is one that is excellent and complete. Furthermore, it is evident to the sound mind that the contrived arguments that are mentioned to show that these wondrous effects that occur in the creation of the human body can proceed from nature alone, which is bereft of knowledge and power, are unconvincing and unworthy, like [226] when the disk of the sun is hidden by dust.

[popular belief establishes theism]

Ancient histories indicate that people of this world were like this [that is, believers in God] from all eternity and at all times, and not a single one of them denied the existence of God, the exalted. This being established, we say: it is among the things known by necessity that the mind of the people of the East and West together, over a duration of more or less seven thousand years, outstrips the mind of a single obscure person. Thus, if some problem occurs to a single person, or he doubts the existence of God the exalted, then he must decide that this doubt or problem is due to the inadequacy of his mind and his lack of understanding, not because of the non-existence of that which is sought [i.e. God]. For the sound mind bears witness that the minds of all people over these long periods are more perfect than the mind of this one [person]. This is a strong and powerful way to establish the knowledge of the existence of the wise God, so long as someone concedes that his mind is weaker than that of everyone.

[no atheists in foxholes]

Those who like to raise doubts and pose problems exert themselves to put forth false imaginations and set up fallacious doubts about God, the providential. Yet whenever some misfortune befalls them or something dreadful occurs, they find submission and apparent obedience to a God of the world in their sound minds and hearts. They ask Him to free them from suffering and to deliver them from their trial. The intuition in this state is like things known necessarily by induction and consideration. Then, after they are freed from suffering, they often return to raising doubts and inventing fanciful worries.

[wager argument]

It is more appropriate (awlā) to take precautions in all things than to take no precautions. So we say: acknowledging the reality of God, who chooses to assign [duties], is more on the side of taking precautions (aqrab ilā al-iḥtiyāṭ) than denying Him. Therefore acceding to this position is more appropriate. The proof that this is the better precaution is determined by saying: either this world has a God or not. If not, then acknowledging His reality [does no harm]. If however it has a God, then rejecting Him] is harmful.²¹ Thus it is established that admitting a God of the world is a better precaution.

Furthermore, we say that the God of the world is either a voluntary agent or not. If He is not a voluntary agent, acknowledging the reality of a voluntary agent does no harm. If however He is voluntary, then rejecting Him is harmful. Therefore, acknowledging the reality of a voluntary agent is further from harm and a better precaution.

[slap argument]

Some people of sound mind used to say: even one slap to the face of a young man shows that there is a God for this world. […] [273.10] When this young man feels that slap, he immediately cries out and asks, “Who hit me? Who slapped my face?” If people gathered around him and said, “The slap just happened by itself, with no agent,” he would not accept this response, and what they say would have no impact on him. This shows that the sound intellect and initial, inborn nature (awwal al-fiṭra) bear witness that no act can be without an agent, and nothing originated without an originator. This premise being evident, we say: if the sound mind finds it implausible that that slap would originate without an agent, how can the origination of all that has arisen in the world of the [heavenly] spheres and the elements be intellectually grasped as having no originator or agent?

[part-whole inference]

Every single contingent is contingent, thus the whole [of them] is contingent; not because we base the predication of the whole on the predication of every single unit, and solely on this basis claim that the predicate of the unit can be applied to the whole, but because the whole is an effect (maʿlūl) of the units. When causes are contingent, the effect is contingent all the more.

[proving necessary existence]

There are two approaches to establishing necessary existence (al-wujūd al-wājibī). One way is to prove its existence and then establish its unity. The other way is to establish that the necessary existent must be one, and then to establish that bodies and their features are many, so that none of them is necessary; hence one is left with the option that they are contingent and require a preponderating factor, which is either necessary or leads back to a necessary preponderating factor.

[specification argument]

Among the ways [to prove God’s existence] is establishing that the world of bodies is contingent. It is obvious that there is composition in the bodies that fall under species—regardless whether or not the inquirer agrees that there is prime matter that is simpler than bodies, and whether or not he agrees that there are forms, which is [a teaching] held only by the Peripatetics. For he cannot but agree that there exist features (hayʾāt) that are additional to bodies, through which the bodies are distinguished and individuated. [He must also agree] that these features are not necessarily existent through themselves—since otherwise they would not require a subject (mawḍūʿ)—and that no body entails them through itself (li-dhātihi)—since otherwise they would all be alike. Also, bodies require something that somehow distinguishes them: either form, as the Peripatetics claim; or, as everyone else says, the whole range of distinguishing features, of whatever sort. On any doctrine it follows that in their multiplicity, bodies are not necessarily existent, given their requirement for some distinguishing factor (mumayyiz); rather they are contingent. If there were no distinguishing factor, then there could be only one entity. Whatever determines multiplicity for the entities that fall under multiplicity is that which determines their existence, given that if there were no multiplicity, not a single entity among them would be possible. If there were no determining factor for multiplicity, the single items would not be determined. If then bodies and their features are contingently existent, and none of them is a preponderating factor for any other (since otherwise it would follow that something is prior to that which is prior to it, and [hence] prior to itself, which is absurd), and their contingency thus requires a preponderating factor. This cannot but lead to the necessary existent.

[argument from motion]

A second approach argues on the basis of motions. You have learned that nothing that is in motion necessitates motion through itself. Rather, it requires a mover. It has been shown that the motions of the bodies of the spheres [389] are due to a soul, not to nature, while it is soul that enacts motion. There must be a goal for this. Once it is shown that its goal is nothing below them [sc. in the sublunary world], nor a state that would belong to some of them relative to others, and given that [their motions] are due neither to the bodies below or above them, nor to the souls below or above them, as has [also] been shown, it remains only that they are due to something incorporeal that has no connection to matter, so that its existence is necessary. This is the intended [conclusion]. But if [the incorporeal mover] is contingent, it needs a preponderating factor, and this leads to the necessary existent in itself, which is what was sought.

Now at first glance, there is nothing to “preponderate” this approach over others. Yet sound inborn nature (al-fiṭra), after having investigated other methods, chooses this one over the other approaches of the Peripatetics, since it has an admixture of intuition (shawban ḥadsiyyan). It is the one on which Aristotle relied. For perishing things are obviously contingent, and the things brought about in the elements by the heavens do not exist self-sufficiently; for heavenly bodies govern (qāhira) elementals. Stars are the noblest of heavenly things, while the sun is the most outstanding of heavenly [bodies] in evidence and governance (aẓhar wa-aqhar). No false supposition will befall someone who inquires into [this issue], unless he is mistaken about the heavenly [bodies]. This approach involves denying that [the heavens] are the utmost goal, and posits something more perfect beyond them. It is that which moves [the heavens], not by enacting [the motion] or causing alteration, but by [being an object of love] and supplying light. Thus motions eliminate this false supposition [that heaven is the utmost goal].

[preponderation and inclination]

Another approach, which belongs specifically to this book: there is no doubt that the inclination (mayl) of a motion in our everyday experience (ʿindanā), such as that of an arrow or a top, decreases bit by bit, as is perceived with the top [as its spinning slows down], among other cases. It is not the case that some of the inclination that has accumulated perishes, while some of it remains. You learned that this is false in the chapter on intensity and weakness: when there is a decrease the whole [inclination] perishes, and a lesser one originates [in its place]. There is a preponderating factor for this, which cannot be the nature of an arrow, for example, since [the nature] would be incompatible [with any other degree of inclination]. Nor is [the preponderating factor] the initial inclination, since it does not remain once the second [inclination] exists; nor can the [initial] inclination necessitate [the second one] along with itself, since the same would need to apply to the second and third [inclinations], and so on, so that all inclinations would be gathered together at the same moment and perish together, which is absurd. Moreover, the subsequent [inclinations] would need to become of stronger intensity, and not of weaker intensity. Nor is the preponderating factor of subsequent inclinations the agent, since he is no longer bringing them about.²² If, after loosing [the arrow, the archer] does not want these different [inclinations] to occur, [the arrow] will not obey him. So for every one of these different [inclinations] there is a need for some external preponderator. This is its [real] mover, and the person only fancies himself to be its mover. Nor is its mover the air, since air is subject to compulsion by burning and division. Nor is it any other body, since it would never end the way that the weakness of the inclination leads to an end. Thus it remains only that [the preponderating factor] is something separate; if this is necessary, our intended conclusion is reached. If it is contingent, then this leads to something necessary through itself.

[proof of the Light of lights]

If an incorporeal light (al-nūr) is dependent in its quiddity, its need is not directed towards the lifeless dusky (al-ghāsiq) substance, which is in no way fitting to give existence to that which is nobler and more perfect than it. How could the dusky substance emanate light? Thus, the realization of the incorporeal light depends on a self-subsistent light. Furthermore, these self-subsistent lights ordered in ranks cannot form an infinite regress. For you know by demonstration that an ordered simultaneous series must be finite. Therefore the self-subsistent and accidental lights, the barriers (al-barāzikh), and the features of each (hayʾātihā) must terminate at a light beyond which there is no light. This is the Light of lights (nūr al-anwār), the All-Encompassing Light, the Eternal Light, the Holy Light, the All-Highest Almighty Light, the Dominating Light. It is absolutely independent (al-nūr al-qahhār), since there is nothing beyond it.

[preponderation and the Light of lights]

Bodies participate in corporeality and differ in the luminosity²³ that occurs to bodies. The luminosity of bodies is their being evident (ẓuhūr). Insofar as accidental light subsists through something else, and does not exist through itself, it does not show itself evidently (laysa ẓāhir al-dalāla). If however it were subsistent through itself, it would be light through itself. Our rational souls, by contrast, are evident through themselves and are subsisting lights. But we have already shown that they are originated, so there must be a preponderating factor. It is not bodies that bring [our souls] into existence, since nothing brings into existence that which is nobler than it. Hence, their preponderating factor too is an immaterial light. If it is the Necessary Existent, this is the intended conclusion. If not, then it leads to the necessary existent through itself, the Living, the Subsisting. [62] The subsisting soul provides an indication of the Living, the Subsisting, but the Subsisting is evident. It is the Light of lights, separate from bodies and their attachments. It is veiled because of the intensity of its evidence.

[principle of sufficient reason]

If every posited existent is contingent, and they are infinite in number, then either they follow upon each other (mutaʿāqiba) or they are all together. If they follow upon each other, then one can hardly posit any of them to exist by taking it to exist singly [without another]. It is impossible that it be the end of the sequence of existence. For, so long as we do not posit as necessary the existence of whatever brings it into existence, it will not exist. The same applies to whatever brings it into existence in relation to whatever brings it into existence, and so on. If its existence depends on the existence of whatever is before it, where that other thing [20] also has as a condition something else that is prior to it, and so on to infinity, its existence is impossible. We can see this from the case where someone says, “I will give you a coin only after having given you [another] coin first, and so on to infinity.” Then there is no way for him to give any coin at all.

[refutation of al-Rāzī’s argument for the preponderation principle]

As for the first argument, [it fails] because the opponent may deny that entering into existence depends on a preceding preponderation (al-tarjīḥ). He may say that preponderation means nothing more than entering existence rather than non-existence, or vice-versa. Hence preponderation cannot precede existence. So this talk of circularity does not follow. Why would it? If preponderation did precede existence, it would be an attribute of something other than existence, so that the circle would not follow as established. If this were so, however, then that which is described with it would be the preponderating factor (al-rājiḥ), not existence itself, but this is absurd.²⁴

[an infinite series cannot be taken as a whole]

They argue: we do not concede the existence of what you are calling a “whole” (jumla) in the case of the infinite. We say: that which is called a “whole”—that which you have described as being infinite—is without doubt different from each one of the units (kull wāḥid). For every unit is finite, whereas what may be described as infinite is numbers posited in such a way that none is excluded [from the series].

[argument against the need for a cause]

As for your claim that, if something exists after not existing, its existence must be due to something else, since otherwise it would not previously have been non-existent: [we object that] if its existence were due to something else this other thing must either always be a cause, or its being a cause was originated. If it is always a cause, then the existence [22] of its effect cannot be delayed after its existence, nor can it be preceded by non-existence. If however its being a cause originated, then whatever applied to the effect applies to its cause, and so on. This leads to the fact that it is neither non-existent, nor preceded by non-existence, but that is absurd. Or, it leads to an infinite number of causes and effects, but you deny this. […]

[response]

[26.2] As for your argument that the originated does not need an originator, this would follow only if [the originated] were not traced back to an intention and volition, but only to nature and cause. But this is not so.

According to the view of the philosophers (al-falāsifa), who speak of bringing into existence through causation, [the reason why the objection fails] is that the spheres perpetually move such that their contingent positions occur successively and are constantly renewed. They seek to resemble their object of love and to attach themselves to their object of desire, yielding circular motions through an eternal volition that belongs to the souls of the bodies of the spheres. Through the intermediary of the motions, there exist effects such as mixtures, proportions, and so on down below [the spheres], as well the reception of substantial forms and human souls among their receivers. If something fails to exist, this is simply due to the non-existence of receptivity, not because efficient causality is absent; for the efficient cause is the Active Intellect, which exists together with the body of the sphere of the moon.

But the Islamic view is as follows: the origin of all that is originated and that to which it goes back, is the willing and freely choosing Creator. By eternal volition He demanded, and by everlasting will He made to arise, every one of [the originated things], at the moment when its existence was demanded by Him, as will be verified in what follows, God willing. That which brings the originated into existence is not itself originated, such that it would need [another] originator. Nor does it bring [the originated] into existence as cause or nature, such that whatever proceeds from it would have to be eternal because it is eternal.

[argument that nothing is contingent when it actually exists]

As for their claim that if causes and effects were infinite, then each of them would be contingent in respect of its essence, how will they respond if someone lays down as a condition for the contingently existent that it is non-existent, and that whatever is described as existent is necessarily existent, while the necessarily existent is not contingent? If it is said to be contingent, then only equivocally. […]

[response]

[27.16] As for laying down as a condition for the contingently existent that it is denied to exist, this is very problematic. For the contingently existent also is contingently non-existent. So, if one laid down as a condition for the contingently existent that it not be existent, one should also lay down for the contingently non-existent that it not be non-existent. Just as entering into existence would mean its coming into the necessity of existence, so entering into non-existence would mean its coming into the necessity of non-existence. This implies that the contingent would never be either existent or non-existent, which is absurd.

[the existence of the Knower through itself]

Particular existents are many in terms of classes and sub-classes. In some of those existents there is revealed a trace of being knowledgeable (dānandigī) and a sign of awareness. Clearly that among particular existents in which there does appear a sign of awareness and of being knowledgeable, does not know through itself. But the existence of the particular is through it [that is, through being knowledgeable]. For, such a trace can be found in it only when it is alive and soul is connected to it. When the connection to the soul vanishes, its life perishes, and those traces and signs of awareness are no longer to be found. Hence those particular individuals are knowledgeable through [their] souls, not through themselves (bi-nafs, na bi-khūd). Neither is the soul knowledgeable through its essence and itself. For if it were knowledgeable through itself, some trace of being knowledgeable would be found in every particular individual that possesses a soul, just as one finds the traces of life [in all of them]. Yet not everything that has a soul or is alive is knowledgeable. Rather one may find the traces and signs of soul’s being active (kunandagī) in everything that has a soul. Hence, activity does belong to soul in itself, but being knowledgeable does not [belong to soul] through its essence and through itself. Therefore, being knowledgeable through itself belongs to something other than the soul, and soul is knowledgeable through that, not through itself. There can be only one knower through itself; there are not two things each of which is a knower through its essence and itself. For, by “knower through itself (dānā-yi bi-dhāt),” we mean something for whom being knowledgeable, essence, and its existence are all one. For, if there were two things for whom being knowledgeable were existence, and being knowledgeable were essence, but whenever there are two, one is different from another, and whatever is different [226] from the knower through itself is not a knower through itself, then neither of them is the knower through itself, even though we presupposed that both of them are knowers through themselves.

[only God is a cause of existence]

Everything contingent requires a preponderating factor (murajjiḥ) through which its existence is necessary, as has been argued. Either this preponderating factor is necessary through itself or it is contingent through itself. The second is absurd, since if it were contingent through itself then it would require a cause, and then the effect that requires it would require its cause, given that if one thing requires another, any [further] thing required by the latter is also required by the former. And what requires the cause of something is not rendered necessary through that thing [but rather the cause of the thing]. Thus it remains that [whatever the contingent ultimately requires] is necessary through itself, and everything contingent is necessary through an existent that is necessary through itself.

Let it not be said: if every single contingent were necessary through the necessary existent in itself, [344] then nothing contingent could ever be a producer (muʾaththir) for anything else, given that its necessity is [in fact] through the necessary existent. For we say: if by “producer” you mean that through which the existence of something is necessary, then we concede [your] conditional proposition (al-sharṭiyya). But why do you say that the consequent is false? According to us, the producer²⁵ through which the existence of all contingents is necessary is [indeed] the necessary through itself. But if by “producer” you mean that which, after it is existent, renders [another] contingent thing necessary, then we reject [your] conditional proposition. For the contingent may become necessary through the necessarily existent in itself, [but only] after the existence of another contingent thing. Nonetheless the second [contingent thing] does not become necessary through the first. For instance, a composite quiddity becomes necessary only after the final part [of it] exists. But it is not necessary through [that part], since the existence of the composite does not become necessary through [its] formal part alone.

[can the eternal cause the temporal?]

Let it not be said: if every contingent were necessary through the necessary in itself, then every contingent would be eternal, due to the everlastingness of the cause that necessitates the necessity of its existence. For we say: we do not concede the inference that it would be eternal. This would follow only if it did not have an [additional] preparatory condition. This is because everything originated depends on condition that is itself originated—not meaning that [this condition] is a part of the cause through which its existence is necessary, but rather that [the originated effect] is necessary after [the condition], through the necessitating cause, which is necessary through itself. This amounts to saying that everything contingent that does not depend on a preparatory condition is always necessary through the necessary existent in itself; but whatever does depend on [a preparatory condition] is necessary through [the necessary existent in itself, only] after the existence of the preparatory condition.

[rejection of the kalām argument as presented by al-Rāzī]

As for the first, we do not concede that if the producer of all bodies were a body then something would produce itself. Why can’t the producer of each body be some other body, so that bodies [produce one another] in order, to infinity? And even if we admitted that the producer cannot be a body, why couldn’t it nonetheless be a corporeal power? As for [al-Rāzī’s] statement that a vicious circle would then follow, we say: we do not concede this. This would follow only if the power in question were inherent in this same body which is its effect, but why do you insist that this is so? Why couldn’t the power in question be inherent in another body, with its producer being still another power that inheres in still another body, and so on to infinity? Then too, even if we did concede this, why does [the infinite series] terminate at the necessary existent principle? A demonstration is needed for this. Anyway, the falsehood of the circle and the infinite series was already [shown] at the very beginning of establishing the Creator, so there is no need for this long argument.

[rejection of the specification argument]

As for the second, we say: we do not concede that the specification of certain bodies with certain attributes is due either to corporeality or to something extrinsic. Why can’t it be due to corporeality, but on the condition that some preparation (istiʿdād) occurs, due to certain material constituents (mawādd) as opposed to others?

[rejection of the design argument]

As for the third: why have you said that if the semen is homoiomerous, and if that which produces it were a corporeal power, then the human would be circular in shape? Why can’t the mixture of some simple [elements] with others prevent a circular shape from arising?

[taking all existing things, including God, as an aggregate]

All existents taken as a whole (jumla), insofar as they are this whole, are not necessary through themselves, since they require something else. Hence they are contingent in themselves. Their existence is from a preponderating factor that precedes them in existence. This preponderating factor is not outside [the whole of all existents], since whatever is extrinsic to it is non-existent, and the non-existent cannot be a preponderating principle for that which exists. Therefore [the preponderating factor] is among [the whole of all existents]. Either it is necessary through itself or contingent through itself. The latter is false, since if it were contingent, and if it is a cause of its cause either immediately or through some other intermediary, then it would precede both of them, which is absurd. If however it is not a cause, then other things would be independent [of it] even though we supposed that it is the preponderating factor [of everything], which is a contradiction. Therefore, it remains that [the preponderating factor] is necessary through itself. […]

[the uniqueness of the necessary existent]

[51v19] Once it is established that the necessary in itself is one, and since it was established that a regress of causes and effects is absurd—rather they must terminate at an existent that is necessary through in itself—it follows then that the principle of everything existent is one and is the necessary in itself, which is the conclusion sought. […]

[part-whole problems]

[51v25] If someone says: the preponderating factor of the whole of all existents is no single [element] of the whole, since the whole depends on more than one of [its elements], so the preponderating factor cannot be only one [element of the whole]. Also, the preponderating factor is that which necessitates the existence of the effect, even when one supposes that everything else is non-existent. No single [element] of the whole, however, necessitates the existence of the whole when everything else is supposed as being non-existent. Therefore, a single [element] cannot be the preponderating principle of [the whole]. We say: we do not concede that if the whole depends on more than one of its [elements] then [one element] cannot be its preponderating factor. This would follow only if the preponderating principle of “all” (kull wāḥid) were not one of its [elements]. If however one of its [elements] is the preponderating factor of “all”—either in a mediated way or immediately—then it is the preponderating factor of the whole.

As for his claim that the preponderating factor is that which necessitates the existence of the effect, when we suppose that everything else is non-existent, we say: we do not concede this. Rather the preponderating principle is that which, when everything else [52r] that is not its effect is supposed as not existing, necessitates the existence [that effect]. Every single [element] of the whole that is other than the necessary in itself, is its effect: if we posit the necessary in itself, alongside the non-existence of any single [element of the series], then what has been posited would be the preponderating principle alongside the non-existence of something else which is its effect, so the existence of the effect need not be necessary.

[objection to taking God as a member of the whole set of existing things, and a reply]

As for his statement: “if the producer of [the whole] were within it, then it would produce itself, since whatever produces the aggregate (al-majmūʿ) produces each of its parts,” we say: we do not concede this. For the whole that is composed out of the necessary in itself and all other existents is contingent in itself, because of its need for the parts. Yet the producer of the aggregate is within that whole and is the necessary in itself.

[response to the reply]

As for his statement regarding the second argument: “we do not concede that the producer [of the whole] cannot be within [the whole],” [64] we say: [no member of the whole can produce the whole] because by “producer” we mean a cause that is complete [i.e. sufficient] for the realization (taḥaqquq) of the aggregate. But nothing that is within the aggregate is a complete cause for its realization, since the aggregate depends on all of its parts. Someone might say in response: why can’t the complete producer be all the parts [taken together], as when the aggregate is composed of the necessary in itself and all other existents, and has a producer which is all its parts? The same would apply to all other true realities that are composed. The complete producer of the whole that is made up from its parts would be all its parts. In answer to this we would say: we know necessarily that every whole that is composed out of units, all of which are contingent in themselves, requires a cause that is outside them. That which is outside all the contingent things is necessary in itself.

[full proof taking into account simultaneity and part-whole problems]

If there is an existent in existence, then there exists an existent that is necessary in itself. The antecedent is true, therefore the consequent is true.

The conditional [proposition] is shown as follows. If that existent is necessary in itself then this is the conclusion sought. If however it is contingent in itself, then it has a cause through which its existence is necessary. If that cause is necessary in itself then this is the conclusion sought. If however it is contingent in itself then [this further cause also] has a cause. This yields either a circle or a regress, or terminates at a principle that is necessary in itself.

The circle is shown to be false in the following way. If X is an effect of Y, then X requires Y.²⁶ If [Y also] requires [X], then [X] is not its effect. Otherwise, given that X requires Y, and whatever requires that which requires something [further] requires this [further] thing as well, then X would require itself, but this is a contradiction.

The regress is shown to be false in the following way. If causes and effects regress to infinity, then either (a) they coincide in respect of existence, or (b) they do not coincide. [160v] (b) The second option is not acceptable, since it would imply that the contingent exists without its cause, and so would be existent through itself and not through a cause, so that the contingent in itself would be the necessary in itself, but this a contradiction. (a) Nor is the first option acceptable, since if [causes and effects] were to coincide in existence then the whole, taken as a whole, would be contingent in itself, given that it requires each one of its units. Hence [the whole] has a cause through which it is necessary. The cause of the whole must either be (a1) within it or (a2) outside it. (a1) It cannot be within it, since if the cause of the whole were within the whole, either (a1a) the whole depends on its other parts, or (a1b) not. (a1b) If it does not depend [on its other parts], then it follows that the whole exists without the other parts, and this is a contradiction. (a1a) If however it does depend [on its other parts], that [part] which we mentioned will not be a complete cause, yet we supposed that it is a complete cause, and this is a contradiction. Also, if something inside [the whole] were a complete cause, then it would either (a1a1) be the cause of other parts or (a1a2) not. (a1a2) If it is not the cause of the other parts, then the whole is necessary through it, without the other parts co-existing with it, and this is absurd. (a1a1) If however it is the cause of the other parts then its own cause is either (a1a1a) within the whole or (a1a1b) outside it. (a1a1a) If it is within it, then it is the cause of the cause of itself and it requires itself, but this is a contradiction. (a1a1b) If however [its cause] is outside the whole—and whatever is outside the whole [of contingents] is necessary in itself—then the infinite series is interrupted. The cause of the whole is therefore outside it, but we supposed it was within it, and this is a contradiction. Therefore, (a2) the whole has a cause that is outside the whole. That which is outside the whole [of contingents] is necessary in itself. Hence, causes and effects terminate at a principle that is necessary in itself. But we supposed that they form a regress, and this is a contradiction. The regress is therefore false, and there must be an existent that is necessary in itself.²⁷

[God as Light of lights]

Lights are either sensible or intelligible [140v] or a composite of both

The sensible [light] is that which is evident (ẓāhir), and whose observation does not depend on the observation of anything else. It is of two kinds. The first is that which is self-evident, but makes nothing else evident, such as the stars whose light is hidden […]. [140v4] The second is that [whose light] is stronger than the former: it is that which is self-evident and makes other things evident, such as the Sun, the Moon, and burning lights. As for intelligible light, it is that which perceives (al-mudrik) something: for instance, the faculties of outer and inner perception, intelligible and sensible perception, and the vision (al-raʾy) of something. […]

[140v14] As for the light that which both sensible and intelligible, it is that which is evident in itself (bi-nafsihi) and from itself, through its self (bi-dhātihi), from its self, and [evident] in its self and making other things evident (li-dhātihi wa-li-ghayrihi). This is the strongest light. It is God the exalted, since He is evident in Himself, as there must necessarily exist that which makes all things evident through the emanation of the lights of existence upon them.

[part-whole problems; the chair example]

We do not concede that whatever produces the whole produces each of its parts. It could be that the whole requires a producer while some of its parts do not, or occur through some other producer. If it were necessarily so that [the producer produces all the parts], then in the case of an effect some of whose parts precede others in time, such as a chair, if its complete cause existed simultaneously with a part that comes first, then the effect would have to follow with a delay after the complete cause, [which is absurd]; but if [the complete cause] were simultaneous with the part that comes last, this would imply that the effect comes earlier than the complete cause [which is also absurd].

[infinite series of infinite series]

But suppose we concede this. Still, why have you said that whatever is outside that whole is outside the whole of contingent existents? This would follow only if that whole included all contingent existents. But this is rejected, because there could exist an infinite whole, each of whose [parts] encompasses an infinite number of contingent existents.

[refutation of infinite series is not necessary]

But suppose we concede this too. Still, from the fact that whatever is outside [the whole of contingent beings] is a necessary existent, it does not follow that there can be no regress. You should prove this. [26] Rather, once one has been forced into [admitting either] a circle or regress in order to avoid the conclusion, one ought to say that inferring a circle is false, as has been shown, whereas if the inference is a regress, then this may be either false or not. But either way, the intended conclusion [i.e. the existence of God] will follow.

[al-Kātibī: there could be an infinite series of infinite series]

[111.8] They say that whatever is outside the whole (jumla) of contingents is an existent that is necessary in itself. Which is true. But why does it follow from this that whatever is outside this aggregate (al-majmūʿ) is necessary in itself? This follows only if all contingent existents are included in this series. This however is not obvious, because there could be more than one series, each of which encompasses only some contingent existents, not all of them.

[al-Kātibī: the whole containing the Necessary Existent]

[114.3] One may say: we do not concede that whatever produces the aggregate (al-majmūʿ) produces all (kull) of its parts. Why can’t [something] produce the aggregate, taken as an aggregate, yet not produce each (jamīʿ) of its parts, due to the fact that some of its parts are sufficient without a producer, or occur through a producer other than this one? Do you not see that the aggregate composed out of all existents—that is, the necessary in itself and all contingent existents—is contingent in itself? For it needs its parts, which are distinct from it, and this entails its contingency. Its cause is the necessary in itself, even though [the necessary in itself] is not a cause for itself, because it is sufficient without a cause.

[al-Kātibī: the chair example]

[114.17] If the producer of the aggregate composed out of contingent units were necessarily a producer for each of its parts, one of two things would follow. Either the effect would be prior to the complete cause, or the effect would be delayed after the complete cause. But both are absurd. The inference is explained as follows. When something is composed out of two parts, each of which is contingent, and one part is prior to the other in time—like a chair, because one of its parts, the matter, is prior to the other part, the form of chair—then if its producer were to produce each of its parts, then the producer must either exist simultaneously [115] with the prior part or not. If not, then the effect is prior to the complete cause. But if it does exist [simultaneously], then the effect is delayed after the complete cause. The absurdity of both options is evident.

[al-Kātibī: another version of the argument, with objection]

[116.22] We have taken the trouble to formulate the following argument for establishing the conclusion sought: the contingent as such (al-mumkin min ḥaythu annahu mumkin) is existent. From this follows the existence of something that exists through itself.

[117] The first premise [holds], because this contingent is existent, and the contingent as such is a part of this contingent thing. But a part of the existent is existent. Hence the contingent as such is existent.

The second premise [holds], because whenever the contingent as such is existent, it is either necessary in itself or contingent in itself. If the first is the case, then we already have the conclusion sought. But if the second is the case, then doubtless it has a cause, given that every contingent requires a cause to bring it into existence. This cause is either (a) itself, that is, the contingent as such; or (b) one of its elements (afrād); or (c) something that exists outside it and its elements. (a) The first is absurd, because an essential cause is prior to the effect, and nothing can be prior to itself. (b) The second is also absurd, because each of the elements of [the whole] requires [the whole], given that the part must require the whole.²⁸ So if one of its elements were a cause for [the whole], then that [i.e. the whole] would require that element, given that the effect must require the cause. It would follow that each of them requires the other [i.e. the whole would require the element and vice-versa], but such a circle is absurd, as you have learned, since it implies that something is prior to itself. Having refuted both first options (a) and (b), the third option (c) remains: the cause of the contingent as such is something existent outside of it and its elements. And everything that exists outside of it and its elements is necessary in itself. Thereby is established the existence of the necessary in itself, which is the conclusion sought.

But this too is inadequate. For we say: we do not concede that the contingent as such is a part of this contingent. Rather it is a merely conceptual consideration (iʿtibār ʿaqlī) that is applied to each of the contingent elements; namely that neither its existence nor its non-existence is through itself; rather either one arises for it due to a cause outside it. Clearly this intellectual [consideration] has no extramental existence or reality. This being so, it cannot be a part of something that is extramentally real and present.

[al-Ṭūsī: there must be a first in the series of causes]

[120.10] One [aspect] of this demonstration requires additional clarification. So let us put forward the following premise: we say that every ordered series of causes and effects, where the causes are complete in giving rise to their effects, must have a first among its causes. This series, being continuous after the first cause—regardless whether it is finite or infinite in the other direction—can be traced back only to that which has been posited as the first of the causes. Otherwise it would not be a complete cause. Through the occurrence of all units of the series, the whole occurs. If there were no cause among them that is first, but instead the causes ascended up the causal chain infinitely, then there would be no cause for that series or its parts, so that neither the units nor the whole could be traced back to it. For nothing outside [the series] can be its cause, because otherwise two independent causes would co-occur for one and the same thing, given that the parts taken together are an independent cause [of the whole]. Nor can any element among the units [of the whole], nor any [smaller] whole which is a part of the series, be the complete cause [of the whole]. For if it were a cause then first it would be the cause of the independent proximate cause [of the whole], which is the parts taken together. If this were the case, then that element or that [smaller] whole would be the cause of itself and of its causes, which is absurd. […]

[al-Ṭūsī: the causal relation between parts and whole]

[122.7] The complete proximate producer of the aggregate can be nothing other than the parts taken together. For if there are two things, one of which is essentially prior to the other, and if the prior cannot be disjoined from the posterior in existence or non-existence, then the prior is the complete cause (ʿilla tāmma) of the posterior. This being so, the aggregate does exist as essentially posterior to the parts taken together, while the parts taken together are prior to it, and cannot be disjoined [from the whole] in existence or non-existence. Therefore, the parts taken together are the complete cause of the aggregate. Nothing other than the parts can be the producer of the aggregate, given that there cannot be two distinct causes for one and the same effect. […]

[123.4] As has been shown, the issue he [sc. al-Kātibī] is investigating is whether the producer of an aggregate is the producer for each of its parts. Now, nothing but the parts [taken together] can be conceptualized as the producer of the aggregate taken as an aggregate, given that nothing other than the ordered units can be the producer of the ordered [series of] causes and effects. The cause of the ordering (al-tartīb) is the parts themselves, taken as causes and effects. So what we have shown on this issue is that the cause that produces the aggregate, and nothing else, must first produce the parts. […]

[al-Ṭūsī: response to al-Kātibī’s chair case]

[123.12] We said that the complete producer of the aggregate is the complete producer of all parts taken together, but we have not said that it is the complete producer of only some of its parts, so that the implication you drew would follow. The complete producer of the matter of the chair is not the complete producer of the chair. Once the complete producer of the form of the chair is added to it, then they will together become the complete producer of the chair and of its parts. In saying that the complete producer of the aggregate is the complete producer of its parts, we do not mean that it is itself an essential producer of every part [taken singly]. Rather we mean that it encompasses the complete producers of each and every part, even if the producer of a given part is distinct from the producer of each and every part and from [the producer] of the aggregate; for this does not detract from our argument. […]

[al-Ṭūsī: on al-Kātibī’s argument from contingency]

[125.22] The contingent as such is neither existent nor non-existent, nor is it a part of this contingent. There is no division of it into necessary and contingent. The contingent, qualified in this [126] way, cannot be described with anything else, since its meaning is the contingent on the condition that nothing else is [taken] together with it. Yet the contingent that is described as “existent” is taken with no qualification (qayd), not even the qualification of lacking any qualification; rather it admits of being either qualified or not qualified. It is this sort of the contingent that is part of the existent contingent.

There can be no division of existence taken with an eye to [its] contingency [again] into the necessary and the contingent, because a subject of division must be shared, but the contingent existent is not shared between the contingent and the necessary. If one takes the existent and divides it into the necessary and the contingent, then the contingent cannot belong to the division of the necessary. Rather, the contingent is the contingent existent, and requires a cause not for its being contingent (the cause of its being contingent is its own essence), but requires a cause for its being existent. Its cause may be something else that is contingent, but this other thing will not be one of its elements. Rather it will be distinct from it, and something to which existence has already occurred. So the first demonstration that contingent causes must terminate at the necessary in itself, comes down to this. […]

[al-Ṭūsī: the principle of preponderation]

[126.15] Saying that one of two equal [options] can be preponderated (tarajjuḥ) without a preponderating factor goes against that which is evident to the mind. If one allows this, it leads to the existence of something contingent for which the option of existence is equal to the option of non-existence, but without anything that would bring it into existence.

A problem arises concerning the difference between two cases. To say that in one case, there is an agent that preponderates one option over the other even though it has no claim [to be preponderated, in its own right], while in another case there is no agent, so that nothing preponderates one of the two options, does not show that there is a difference. For the agent is posited as having an equal relation to both options. The argument based on someone who flees from a hungry beast does not prove that there is no preponderating factor. It only amounts to showing that there the preponderating principle is unknown; but nothing necessarily follows for each case where the existence of [a preponderating factor] is unknown. […]

[al-Ṭūsī: refuting the infinite regress is not necessary]

[146.6] Not only we have proven the falsehood of an infinite series by arguing that its cause cannot be among the items outside and inside [the series], but we have also proven its falsehood by showing that there cannot be a cause for all [its parts], given that neither they themselves, nor what is among them, nor what is outside them, can be [their cause]. But if someone does not admit the impossibility that [the cause] be something outside them, and we confine ourselves to refuting the options that the cause is [the members of the series] themselves or among them, then insofar as they require something other than them, they require something outside them. So we have established the necessary on the basis of the requirement that the infinite series [plural] encompassing all contingents have for something outside it. For this, one does need not to refute the infinite series.

[the parts taken together are the cause of the infinite series]

The refutation of the regress calls for further investigation. For [Rāzī] establishes a producer for the aggregate of the infinite number of things, on the grounds that it requires its units. From this, however, it follows only that the aggregate has an infinite number of producers which are [its] units. From the fact that none of these units is the cause for itself, or the cause for its own causes, it does follow that it is not all by itself the cause of the aggregate. It does not follow, however, that it cannot be the cause when taken together with all other units. In fact this is true! In which case the causes of the aggregate are within it, and from this one cannot infer that the cause of the aggregate is outside it, so the sought conclusion is not reached.

[perfection argument]

And from this [consideration] it becomes evident that the relationship between the prophets [on the one hand] and mixtures and elements (mawālīd wa ummahāt) [on the other hand] is like the relation between the heavens [on the one hand], and the mixtures and elements [on the other]. Whatever is closer to the Agent is nobler. Among mixtures, whatever is stronger than another, its perception, ability to discern, knowledge, wisdom, laudable character traits (akhlāq), and its esteemed qualities are greater and nobler. This order and organization are among decisive proofs and plain indications that show that these elements, their [derived] mixtures, and the heavens [must] have an Artisan who is immaterial and is void of actions. Indeed, He is an essence with all-encompassing grace and perfect wisdom. This composite [universe], which is the best of all possible structures, is derived from Him.

[fine-tuned universe argument]

The fully developed wisdom of the exalted, wise Artisan entails that the order of the elements be according to its present structure. If something touching (mumāss) the heavenly sphere [from the outside] were some body other than fire, and fire were in another element, then that other body which was adjacent to the heavenly sphere would by its contact with the heavenly sphere turn it into fire; the nature [of fire] would thereby become even more intense and the equilibrium [in the heavens] would be destroyed. Likewise, if in the center [12] of the universe there were a body other than earth and earth were in that other body, then that body, due to its being furthest from the heavenly sphere, would become cold and dense and thus would be transformed into earth. In this way the nature of earth would become even more intense and equilibrium would be destroyed. Such noble composition and delicate order could have arisen only through the governance of the wise Artisan, and the ordinance of a knowing, eternal Creator (mūjid).

[astrological argument]

An argument for the existence of the Artisan based on the signs of the Zodiac in the heavens. When we consider the order of the Zodiac, we find that [the signs] are at variance with the order of the elements, and with the order of the human body. For in the order of the Zodiac of the first heavenly sphere there is a fiery sign, which is Aries; then there is an earthy [sign], which is Taurus; then there an airy [sign], which is Pleiades; and then there a watery [sign] which is Cancer; the rest are in that same order up to Pisces. Thus it is evident that the elements are ordered in one way in the lower world, in another way in the human body, and in still another in the higher world. If this ordering were demanded by the elements, it would have been alike in each case. But this is not so. Thus it is evident that beyond nature, there is something else that has given existence to these different bodies with these different orders, as demanded by His own wisdom and will.

[proof considering all possible part-whole problems]

If there existed no existent that is a necessary existent, then all true realities (al-ḥaqāʾiq) and quiddities would be contingently existent. Everything that exists contingently requires a cause for existence that exists simultaneously with it and preponderates it to exist rather than not existing. Hence the aggregate of contingent existents requires an existent of this kind. This existent is either (a) that aggregate itself; (b) inside it; or (c) outside it.

(a) If it is the aggregate itself, then one either (a1) thereby means all units, without considering their composition (al-taʾlīf), or (a2) with the composition taken into consideration. (a1) If one means the case where the composition is not considered, then let us talk about just the units. (a1a) They are not the cause of themselves, given that the cause must be different from the effect. Otherwise it would be [406] essentially prior to itself and it would require itself and acquire existence from itself, which is obviously false. (a1b) Nor is the cause [of the elements altogether] some of these elements, given that they cannot be the cause of themselves or of their causes. For if one thing is a complete cause of another, it cannot require anything else outside it. Rather, if this thing were composed out of contingent items and some of these contingent items required something outside the thing, then it would follow that this thing would require that external [cause] as well. For it requires its part, which in turn requires the external [cause], and if one thing requires another, the former requires whatever the latter requires too. Thus the complete cause would not be complete, which is a contradiction. So if some elements of the whole were the complete cause of the whole, then some other [elements] would not require anything outside them. It would follow that their causes are their effects, and that they are their own effects. Not only is this obviously impossible, but it also implies that more than one proceeds from the one, and you have already learned that it is impossible. (a1c) Nor is the cause [of the elements altogether] anything outside them, since (a1c1) if these units are infinite, this will be false. For each of them, and every subset (kull jumla minhā), is traced back to the complete cause which is not outside the infinite series. [That cause] is prior to this unit or subset. So if the cause of the units, taken altogether, were outside them, then a cause for some of them would be simultaneous with the complete cause [for all of them], and you have already learned that this is impossible. (a1c2) If however these units are not infinite, then they must terminate at an uncaused cause, and that will be the necessary existent. If therefore the units, taken altogether, are caused, then [paradoxically] it is impossible that they have a cause, given that their cause can neither be they themselves, nor something inside them, nor something outside them, on the assumption that there is no [407] necessary existent involved in their cause. (a2) If one however means by “the aggregate” something where composition is taken into consideration, then something will be the cause of itself, which is evidently false.

(b) If that which is the cause of the aggregate of contingents is something inside it, then it is that cause either (b1) all by itself, or (b2) together with the other units. (b1) It cannot be [the cause] all by itself. Otherwise it would be the cause of itself or of its causes, [which has been disproven] by the preceding argument. (b2) Nor can it be the cause together with the rest of the parts, as this would be tantamount to saying that the cause is the aggregate taken in one of the aforementioned meanings [from option (a)]; but you have just learned that this is absurd.

(c) It remains then that the cause of the whole of contingent existents is outside them. If whatever is outside the whole of contingent items were itself contingent, then it would belong to that whole and not be outside it. Therefore, it must be the necessary existent.

[argument from the persistence of originated things]

We know that there exists an existent that has persistence (thabāt), such as the body that is the bearer of motion, the soul that moves the spheres, prime matter, the substance in the human being that grasps itself, and so on—in fact everything originated apart from motion. For the moment (ān) of its origination is distinct from that of its perishing. Some time elapses between the two moments, and this time is its persistence. The causes of persistence are simultaneous [with that which persists], since nothing can persist once that which makes it persist perishes. The aggregate of that which is contingently persistent is itself contingently persistent. Its persistence is rendered necessary by something else. Otherwise its persistence would be through itself and it would be necessary in itself, despite being contingent in itself, which is absurd. This “something else” must be necessary in itself. For if it were contingent in itself, then its persistence would be due to some cause, and then the persistence of the aggregate of contingents would be necessary through both it and its cause, and its persistence would not be [412] through it alone, as was assumed, which is a contradiction.

[overview of previous thinkers on the part-whole problem]

The imām Athīr al-Dīn al-Abharī—may God have mercy upon him—expressed doubts concerning [this proof], saying: we do not concede that whatever produces the aggregate, insofar as it is an aggregate, is that which produces every single [part] of it. It might become the producer of the aggregate, insofar as it is an aggregate, by being the producer of its final part, without being the producer of all its parts. He has illustrated this with the fact that the aggregate made up of the necessary and the contingent is contingent, given that it cannot be without what is contingent; and its producer is the necessary in itself, without its being the producer of every [part] of [the aggregate] since it cannot be the producer of itself. [Al-Abharī] answered his own objection by saying that the claim [needed for the proof] is that whatever produces the aggregate of contingents is the producer of every single [part] of [the aggregate]. But to this he answered that [the claim] is not self-evident, and stands in need of demonstration.

He has also organized the proof in another way, but again raised doubts about it. This way is as follows: the aggregate must have a complete cause (ʿilla tāmma). Its complete cause cannot be the aggregate itself, nor can it be something within it, since whichever [part] from within the aggregate [you take], the aggregate will always depend on another [part as well]; and by this reasoning the complete cause cannot be anything that belongs to the aggregate. Hence, the complete cause will be either outside [the aggregate] or composed out of what is inside it and what is outside it. Either way one must accept that there exists an existent that is the necessary existent.

[167] [Al-Abharī’s] doubt concerning this proof is as follows. If by “complete cause” one means a whole, each item of which is truly needed by the aggregate, then why can’t you say that the aggregate itself is the “complete cause” understood in this way? But if [the complete cause] is that which produces the existence of the aggregate, but only given the condition of something else, then why have you said that an effect does not depend on anything other than the complete cause, understood in this way? For support [al-Abharī] appealed to what he had already mentioned, namely that whatever is composed out of the necessary and the contingent things does have a cause, which is neither the aggregate itself nor anything outside it, but something inside it.

This is the thrust of [al-Abharī’s] doubts. I have already given an answer in my response to the doubts expressed by the imām ʿAllāma Najm al-Dīn [al-Dabīrān i.e. al-Kātibī]—may God reward the people of knowledge by making his life long and securing his duration—on [al-Rāzī’s] book al-Maʿālim: when the aggregate is composed out of units each of which is contingent, then the complete cause of this aggregate must be that which produces every single unit from which the whole is composed. Otherwise whichever [units] it does not produce would either have no producer at all—so that something contingent would require no producer, which is absurd—or it would have a producer other than that cause. This would be either necessary, which is the conclusion sought, or it would be contingent, and if it were not factored in, those units would not occur at all. Yet so long as those [units] do not occur, neither does the aggregate. Hence that complete cause will not [really] be complete, since the effect will be delayed after it, and this is a contradiction.

In the book Muntahā al-afkār, imām Athīr al-Dīn al-Abharī, after having argued for the falsehood of this demonstration, pursued [the following argument]: if contingent things regressed to infinity, the whole composed out of this regress would need to have a cause through which the aggregate is necessary, either through [this cause] itself or through something that [the cause] implies. For [the whole series] is contingent, and whatever is contingent requires a cause of this kind. [168] One knows this necessarily. That cause cannot be within the aggregate, since when so understood, the cause cannot be preceded by another cause. Otherwise the aggregate would [in fact] require the prior cause, in which case [the cause we are considering] will not be the cause [after all]. Whatever is within the whole series that is composed out of units characterized by contingency is preceded by another cause, so [anything in the series] is not a cause, so understood. Nor is [the cause of the aggregate] the aggregate itself given that it cannot be prior to itself. Therefore [the cause] is outside the aggregate. Yet whatever is outside the aggregate [of contingents] is necessary in itself. Therefore the infinite series must terminate, whereas we supposed that that it exists [as infinite], and this is a contradiction.

[argument from motion combined with specification argument]

Know that the cause of motion is not corporeality itself, because otherwise [motion] would be eternal due to the eternity of [body], so that every body would be in motion; but this is not so. Also it would follow that motions would be the same and not different, since corporeality is the same in [all] bodies; when the causes resemble one another, so do the effects. Therefore, there must be a cause of motion which is incorporeal. If it is the necessary existent, then this is the conclusion sought. But if it is contingent then it leads to the necessary existent, as has been explained more than once.

[argument from souls]

You have already learned that the rational soul originates with the origination of body. Hence it is contingent and requires a cause. Its cause cannot be a body, since nothing can give existence to something that is nobler than it. If then its cause is the necessary existent, then this is the conclusion sought. But if it is contingent, it leads to the necessary in itself, as you have learned more than once.

[part-whole problems regarding the impossibility of infinite regress]

[The part cannot be the cause of the whole], since otherwise it would be the cause of its cause and of itself, since the cause of the aggregate is the cause of every single [part] (kull wāḥid). For, if some [parts] occurred through something else, then the cause would be a cause only taken together with this other thing, and then it would not be an independent [cause]. […]

[143.9] This calls for further investigation. For, why not concede that some [parts] are not due to a cause? Why can’t whatever is beyond the first effect be the cause, and so on to infinity? For, as soon as it is realized (taḥaqqaqa), so necessarily is the aggregate, and in that case none of the consequences you mentioned would follow. Whatever is beyond the first effect, to infinity, would be the cause of those [parts], and likewise the cause of their cause, and so on indefinitely.

[argument from motion]

The natural philosophers (al-ṭabīʿiyyūn) used to prove [God’s existence] in another way which is distinctive of them. They said: the spheres do not move by themselves, given the aforementioned [arguments] that nothing can move itself. Nor are they violated and moved by nature. Hence they are spiritual and they must have a goal (ghāya). Their goal is not anything corporeal above or below them. Hence [the goal] is something non-corporeal. If it is contingent, an infinite series follows. Otherwise it is necessary. [520] This way of arguing is based on false premises whose falsehood will be shown in what follows. Moreover, it amounts to the first way [that is, Avicenna’s burhān al-ṣiddīqīn] anyway.

[short version of al-Ṭūsī’s proof]

You should know that establishing the Necessary is nearly self-evident, for inevitably, there is some existent. If it is necessary, this is the conclusion sought, but if not then it is contingent. If there is a regress then the aggregate of contingent things is contingent and must have a cause. But simply having a cause does not suffice for the existence of the contingent; rather there must be a complete cause, with which [the contingent] is rendered necessary and without which it is rendered impossible. The complete cause for the aggregate of contingents must be necessary. For if it were contingent, and if it were the complete cause of every single contingent, then it would be the cause of itself, since it belongs to the whole of contingents. But if it were the cause of some contingent things but not others, it would be a part of the complete cause for the aggregate of contingents, while at the same time being the cause of the aggregate of contingents, which implies that something is a part of itself, [which is absurd].