伊本·西纳在他的多部著作中提出了关于上帝存在的论证，主要见于《治疗论·形而上学》（Al-Ilāhiyāt al-Shifā‘）、《解脱书》（Al-Najāh）和《指示与告诫》（Al-Ishārāt wa al-tanbīhāt）。这些著作也涉及许多其他主题，如逻辑学、数学和自然哲学。伊本·西纳在这些著作中运用了一系列不同的论证：其中一些从运动（受亚里士多德启发）出发论证上帝的存在，另一些从因果关系出发，还有一些从复合性出发（这是诉诸因果关系的论证的一个重要推论）。在本章中，我将简要概述伊本·西纳来自这些著作的论证，并讨论这些论证与当代无神论者与有神论者之间争论的相关性及其作用。第1章和第2章的目的，是说明为何这一特定形式的论证在这些争论中特别有价值。其意图在于为有神论者提供理智上的“弹药”，这些弹药目前在哲学宗教学的学术文献以及上个世纪有神论者与无神论者之间最著名的公开辩论中，要么发展不足，要么完全缺失。阅读伊本·西纳时，必须牢记他常常从说明性写作转向说服性写作。因此，必须小心不要将伊本·西纳的逻辑和形而上学分类（这些分类旨在作为客观的说明）与他的形式论证混为一谈。例如，《治疗论·形而上学》第6章曾被误解为一个论证，而伊本·西纳的意图很可能只是分类性的。关于这一点，丹尼尔·德·哈恩（Daniel De Haan）指出：

Ibn Sīnā presents his arguments for God’s existence in a number of his writings, mainly in Metaphysics of the Cure (Al-Ilāhiyāt al-Shifā‘), The Deliverance (Al-Najāh), and The Pointers and Reminders (Al-Ishārāt wa al-tanbīhāt). These works address many other topics as well, like logic, mathematics, and natural philosophy. Ibn Sīnā employs a range of different arguments in these writings: some of which argue for God’s existence from motion (inspired by Aristotle), others from causation, and even others from composition (which is an important corollary to arguments appealing to causation). In this chapter, I will briefly outline Ibn Sīnā’s arguments from these sources, referring to both the relevance and function of his arguments in relation to contemporary debates between atheists and theists. The aim of Chapters 1 and 2 is to demonstrate why this particular rendition of the argument is particularly valuable in these debates. The intention is to provide theists intellectual ammunition that is currently either under-developed or completely lacking in both academic literature found in Philosophy of Religion publications and the most famous public engagements between theists and atheists in the last century. When reading Ibn Sīnā, one must keep in mind that he often switches from informative to persuasive writing. Accordingly, one must be careful not to conflate Ibn Sīnā’s logical and metaphysical taxonomies (which are intended as matter-of-fact clarifications) and his formal arguments. As an example, the 6th chapter of Ibn Sīnā’s Metaphysics of the Cure has been misunderstood to be an argument, where Ibn Sīnā’s intention was most probably simply taxonomical. On this point, Daniel De Haan states:

伊本·西纳（Avicenna）在《治疗论·神学》（Ilāhiyyāt）第一卷第6至7章中对必然存在与可能存在的分析，不可能是对神存在的形式论证，因为这两章的宗旨在于为我们提供对形而上学恰当第一原理的洞见。5

…Avicenna’s analysis of necessary and possible existence in Ilāhiyyāt I.6– 7 could not be a formal demonstration for God’s existence since these chapters are oriented towards providing us with insights into the proper first principles of metaphysics.5

其他研究伊本·西纳（Ibn Sīnā）的学者，例如法兹勒·拉赫曼（Fazlur Rahman）6

Other scholars of Ibn Sīnā, such as Fazlur Rahman6

和 Parvis Morewedge7

and Parvis Morewedge7

将这种分类本身视为一种论证。8

regard this kind of classification as an argument in and of itself.8

在《拯救》一书中，这位波斯哲学家伊本·西纳陈述了一段可能是论证神之存在最为全面而又简明的文字：毫无疑问，存在是有的。凡存在之物，要么是偶然的，要么是必然的。如果是必然的，那么对必然存在的探求便已完成。如果是偶然的，我将说明，这一偶然的存在最终必将回归于一个必然的存在。在此之前，我将先提出若干前提（以证明这一论点）。这些前提之中，有一条是：对于偶然存在而言，因果链无限回溯是不可能的。原因在于，所有偶然的存在，要么同时全部存在，要么并非同时存在。如果它们不是同时存在，而是无穷地一个先于一个，那么我们将在本书的另一节讨论这一问题。如果它们同时存在，而整个集合之中并无必然的存在，那么经过详尽的分析（无论其是有限还是无限），这一集合［的偶然之物］要么由偶然之物构成，要么由必然之物构成。如果集合之中有一个必然的存在，而集合中的其他一切都是偶然的，那么这一必然的存在将凌驾于所有偶然的存在之上。如果集合只由偶然的存在构成，那么它就需要一个赋予存在者。这一赋予者，或者在偶然存在的集合之内，或者在集合之外。如果它在集合之内，那么这一点已经阐述过了。或者，它是一个偶然的存在，并且它是该集合的原因，而该集合的原因也就延伸为其全部组成部分的原因。如果它在集合之内［或即是该集合本身］——尽管这是不可能的——这一论证思路仍可能成立，因此从某一角度看，人们便已达成了此结论的目标（即）。 因为这将证明必然存在者的存在。这是因为任何自足的事物（即所讨论的那个集合）都是必然存在者。而必然存在者不可能就是这个集合；同样，也不可能有偶然存在者存在于该集合之外，因为按定义，该集合正是所有偶然存在者的总和。因此，给予存在者（存在的提供者）必定在该集合之外，并且其本质是必然的。由此可见，偶然存在者的集合最终导向了对一个位于集合之外的必然存在者的需要，以便对其作出解释。并不是每一个偶然的结果都有一个偶然的原因，从而陷入无穷倒推。9

In what is perhaps the most comprehensive, yet concise passage written by the Persian philosopher demonstrating God’s existence, we find Ibn Sīnā state in The Deliverance: There is no doubt that there is existence. Everything that exists is either contingent or necessary. If it is necessary, then the pursuit of the necessary existence is complete. If it is contingent, I will make clear that this contingent existence will ultimately return back to a necessary existence. Before this, I will present premises (to prove this thesis). Among these premises is [the assertion] that it is impossible for an infinite regress of causes to account for a contingent existence. That is because all contingent existences either all exist at once or do not exist at once. If they do not all exist at once and they are infinite, one preceding another, we will deal with this matter in another section of this book. If they are all together and there is no necessary existence in the set, [we find that] upon exhaustive analysis [and whether or not it is finite or infinite] the set [of contingent things] can either be composed of contingent things or necessary things. If, in the set, there is a necessary existence and all other things in the set are contingent it will be such that the necessary existence overrides all of the contingent existences. If the set is composed of only contingent existences, then it requires that which supplies existence. This can either be in the set of contingent existences or outside of the set. If it is inside the set, then this has already been elaborated upon. Or, [it could be] that it is a contingent existence such that it is the cause of the set, and the cause of the set is by extension the cause of all of its constituent parts. If it is inside the set [or is the set itself] – even though this is impossible – this line of argumentation could still be valid, so from one perspective then one will have fulfilled the object in making this conclusion (i.e. because this would have proven the existence of the necessary existence). This is because anything that is self-sufficient (i.e. the set in question) is a necessary existence. And the necessary existence cannot be this; it is also impossible that a contingent existence can exist outside the set, as the set is by definition a collation of all contingent existences. Therefore, it (the supplier of existence) must be outside of it as well as necessary in essence. Thus, the set of contingent existences has culminated in the need for a necessary existence outside of the set in order to explain it. It is not the case that every contingent effect has a contingent cause as a matter of infinite regress.9

Ibn Sīnā 的论述第一阶段，是将“存在”确立为最根本、超验且普遍的分析范畴。他提出的第一个重要命题是：对于一个偶然存在者而言，其因果链条不可能无限回溯。 Ibn Sīnā 在此运用了一种类似集合论的推理方式，将具有某种特定描述的所有成员归入同一集合，然后对该集合加以分析。关于“可能”或“偶然”存在者这一点，Ibn Sīnā 指出：

The first stage of Ibn Sīnā’s discourse is to establish ‘existence’ as the most foundational, transcendental, and universal category of analysis. The first major postulation which Ibn Sīnā makes is to state that ‘it is not possible for there to be an infinite regress of causes for a contingent existence’.10 Ibn Sīnā employs a set theory type of reasoning for this, gathering together all members of a specific description in one set and analysing that set thereafter. On this point regarding ‘possible’ or ‘contingent’ existences, Ibn Sīnā states:

如果它们全部聚集在一起，而这个集合中不存在必然存在者，那么经过详尽的分析之后[无论它是有限还是无限]，这个集合[即偶然事物的集合]要么由偶然事物构成，要么由必然事物构成。11

If they are all together and there is no necessary existence in the set, [we find that] upon exhaustive analysis [and whether or not it is finite or infinite], the set [of contingent things] can either be composed of contingent things or necessary things.11

简单来说,一类事物的集合会排除其他类别的事物。例如,一套椅子的集合排除桌子;同样,一套钢笔的集合排除铅笔。归根结底,这意味着偶然存在的集合排除必然存在,反之亦然。伊本·西纳指出,如果某种必然存在能够存在于不可能存在的集合之中(尽管这在逻辑上是不可能的),那么对于任何想要否认这种必然存在的人来说,这将是一个违反直觉的主张。伊本·西纳进一步指出,正如必然存在不可能在仅由偶然存在组成的集合中出现一样,偶然存在也不可能不被所有偶然存在的集合所包含。在此基础上,伊本·西纳得出结论:"偶然存在的集合最终需要一个位于该集合之外的必然存在来对其加以解释"。在进一步总结这一论证时,赫伯特·戴维森提到,伊本·西纳的论证指出:"可能存在的事物可以追溯到一个自身即为必然的存在……存在某种自身即为可能存在的事物;因此存在某种可以自身追溯到一个必然存在者的事物"。13

Simply put, a set of one type of things excludes other types. For example, a set of chairs excludes tables. Likewise, a set of pens excludes pencils. Ultimately, this means that a set of contingent existences excludes necessary existences and vice versa. Ibn Sīnā states that if somehow a necessary existence can exist in the set of impossible existence (though this is logically impossible), this would be a counterintuitive proposition to make for anyone who aims to deny such a necessary existence. Ibn Sīnā goes on to suggest that just as it is as impossible for a necessary existence to be represented in a set of only contingent existences, it is also impossible for a contingent existence not to be represented by the set of all contingent existences. With this groundwork put in place, Ibn Sīnā concludes that ‘the set of contingent existences has culminated in the need for a necessary existence outside of the set in order to explain it’.12 Summarising the argument further, Herbert Davidson mentions Ibn Sīnā’s argument as pointing that ‘possibly existent beings are traceable to a necessary existence (by virtue of itself)…something exists which is a possible existence by virtue of itself; therefore something exists which is traceable to a necessary being by virtue of itself’.13

尽管如此，或许这是开始理解形而上学的各种划分，以及伊本·西那如何对存在进行分类的好地方，因为这实际上为他的论证奠定了基础。关于这一问题，伊本·西那在《治疗论·形而上学》第六章开篇便对“必然存在”的含义作出界定，他说：

Notwithstanding this, it is perhaps a good place to start understanding the metaphysical compartmentalisations and how Ibn Sīnā categorises existence, as this effectively lays the groundwork for his argument. Regarding this matter, Ibn Sīnā starts the 6th chapter of Metaphysics of the Cure by defining what constitutes ‘necessary existence’ by stating:

所以我们说，必然存在是无因的，而“或然存在”是有因的；必然存在在所有可能的方式和角度上都是必然的。它的存在不可能是其他任何事物存在的结果。如果是那样，就仿佛那被假定的必然存在与其所依赖的事物在存在上是相等的，并且彼此相互依存。而必然存在的存在由多种事物共同造成，在理性上也是不可能的。必然存在在理性上也不可能是与非必然之物具有任何共同方面的“实在”。因此，要被归类为“必然”，必然存在就不能被附加、构成、变化、分割，也不能是构成其自身存在的多个特定因素之一。14

So we say that the necessary existence is uncaused, whereas a ‘contingent existence’ is caused, and that the necessary existence is necessary in all possible ways and perspectives conceivable. Its existence cannot be a result of the existence of anything else. If that were so, it would be as if each of those two things (i.e. the supposed necessary existence in question and the thing which it results from) are equal in terms of existence and interdependent of each other. And it is not cogently possible for the necessary existence’s existence to be as a result of many things. The necessary existence cannot cogently be ‘the reality’ that has any common aspect of it (with that which is not necessary). Thus, in order to be classified as ‘necessary’, the necessary existence cannot be added upon, constructed, mutable, divisible, or one of multiple contributors to its own existence which is specific to it.14

在《指示与提醒》中,伊本·西纳首先对可能的(或称“偶然的”)、不可能的(muḥāl,如方形的圆)以及必然的进行了分类。15 与此相近,在他的《形而上学》中,伊本·西纳直接转入对必然存在的模态特性的论述:至于必然存在,它没有原因。这是因为,如果它有原因,那么它的存在就是由于那个原因。凡是本质上自身被考虑、其存在并非自身必然的东西,都不可能是“必然存在”……因此已经清楚,必然存在是无原因的。由此推论,一个事物不可能既是自身必然存在,又是因他物而必然存在……此外,凡是本质上是偶然存在的东西,其存在或不存在都必然由某个原因而来。这是因为,如果这个[偶然的]事物可以被确定,那么它的存在就可以与虚无(或即其不存在)区分开来;如果它不存在,那么它的不存在也可以与其存在区分开来。因此,对各种可能性的穷尽式陈述表明:它的存在或不存在要么是由于他物所致,要么不是。如果是由于他物,那么那个他物就是它的原因。如果不是由于他物,那么它自身就是必然存在。16

In The Pointers and Reminders, Ibn Sīnā starts off with a classification of that which is possible or ‘contingent’, muḥāl or ‘impossible’ (like a square circle), and necessary.15 On a similar note, in his Metaphysics Ibn Sīnā moves on to speaking directly of the modal specialities of the necessary existence: As for the necessary existence, it has no cause. This is because, if it had a cause, its existence would be because of it. Anything which is considered essentially by itself and its existence is not necessary in itself cannot be a ‘necessary existence’…so it has been made clear that a necessary existence is uncaused. As a corollary to this, it is not possible that a thing is a necessary existence in and of itself and also because of another thing…Moreover, [regarding] anything that is a contingent existence in essence, both its existence or non-existence would come about because of a cause. This is because if the [contingent] thing can be located, then its existence can be discerned from nothingness (or lack of its existence), and if it is not in existence, then its non-existence would be discernible from its existence. So, an exhaustive rendering of the options would suggest that its existence or lack of existence would either be caused because of something else or not. If it is from something else, then that something else is the cause of it. And if it is not from something else, then it is itself the necessary existence.16

在《指示与警语》中，伊本·西纳谈到一个由无穷多个偶然事物构成的“jumlah”（大致可译为序列或集合），并以归谬法论证：这样的集合不可能是必然存在者。将此应用于新无神论者提出的论证，这一部分论证是反驳无限多重宇宙论题最有力、最有用的证明之一。该论题常被用来替代必然存在者或“真主”，这一点将在后续章节中详细展开。在《形而上学》第八篇第1至3章中，伊本·西纳最明确地给出了一个形式化论证。值得注意的是，伊本·西纳实际上以这一论证结束全书，前面的章节在很多方面都在为此点做铺垫。在本节中，伊本·西纳以归谬法论证无限回溯的不可能性，并以下文所强调的两种新颖方式进行：第一，论证所有“中间因”不可能具有相同的模态地位；第二，运用一种复合论证。正如将要解释的，这类论证或许最适合应对那些预设无限多重宇宙的无神论诘问。在设想出所有现存事物的集合（即前述的jumlah）之后，伊本·西纳提出：[“终极因”]不能是因果集合中的最后一个，也不能处于中间。因为集合中的中间因只能产生一个结果；此外，该结果本身并不产生任何东西。而在集合[现存事物]两端的三类事物中，每一个都具有特定的模态地位。因此，集合末端之物的特定模态地位在于：它按定义不是任何事物的原因。17

In The Pointers and Reminders, Ibn Sīnā talks about a jumlah (roughly translated to a sequence or even a set) consisting of an infinite number of contingent things, arguing ad absurdum that such a set cannot be the necessary existence. Applying this to arguments posited by New Atheists, this part of the argument is one of the most potent and useful proofs in arguing against an infinite multiverse thesis. This thesis is often employed as a substitute for a necessary existence or ‘God’, a point that will be expanded on in following chapters. In the Metaphysics, Ibn Sīnā is most explicit in making a formal argument in maqālah (section) 8, chapters 1-3. It merits consideration that Ibn Sīnā effectively ends his book with this argument, and in many ways earlier chapters seem to build up to this point. In this section, Ibn Sīnā argues ad absurdum for the impossibility of an infinite regress, doing so in two novel ways highlighted below: first, arguing for the impossibility of all ‘middle causes’ having the same modal status, and second, employing an argument from composition. As will be explained, these types of arguments are perhaps best placed to deal with atheistic interrogations which stipulate an infinite multiverse. After conceptualising a set of all existent things (the aforementioned jumlah), Ibn Sīnā posits that: [The] ‘ultimate cause’ cannot be the last in the set of causes, nor can it be in the middle. That is because the middle cause in the set can only cause one effect. In addition, the effect does not cause anything. And each thing of the three types at each extremity in the set [of existent things] has a specialised modal status. Thus, the specialised modal status of the thing at the end of the set is that it is not a cause for anything [by definition].17

为了说明这一论证,设想一个从P到P10的线性序列。伊本·西那只是在陈述:根据逻辑上的必然,P5在该序列中只能是P6的原因,而不能是位于它之前之物的原因。P10不是该序列中任何事物的原因。他继续这一推理,指出:在该序列的另一端(即其起点),原因(依其定义)是该序列中所有其他事物的原因。处于该序列中间的事物所具有的特殊模态状态在于:它必须由一物所引起,同时又是一物的原因。18

To illustrate this argument, conceptualise a linear set from P to P10. Ibn Sīnā is simply stating that by logical necessity P5 can only be a cause for P6 in the set, and not for those things which come before it. P10 is not the cause for anything in this set. He continues this line of reasoning by stating: At the other extremity of the set (i.e. at the beginning of it) the cause (by definition) is the cause for all other things in that set. The specialised modal status of the thing which exists at the middle of the set is that it must be caused by one thing and be the cause for one thing.18

如果这一系列是有限且线性的,那么P5必然由P4所引起,并且必然是P6的原因;即便P1至P10之间的中间环节是多个,情况依然如此,正如伊本·西那所指出的:无论中间的事物是一个还是多个,情况都是如此。若中间的事物是多个,那么它们要么是有限的,要么是无限的:若是有限的,则这一系列位于两个端点之间(即第一因与最终的结果)。因此,位于这两个端点之间的每一个事物都将具有特定的地位(这与该系列的有限性相对应)。19

If the set is finite and linear, P5 must be caused by P4 and must be the cause of P6; this is the case even if the elements between P1-10 are multiple, for as Ibn Sīnā suggests: That is [the case] regardless of whether the thing(s) at the middle is one thing or more than one thing. If it is more than one thing(s), then it is either finite or infinite: if it is finite, then the set is between two extremities (i.e. the first cause and the final effect). Therefore, each of the things in between these two extremities will have a specialised status (which corresponds to the finitude of the set).19

伊本·西那在《指示与提醒》和《解脱》两书中，通过考察一个有限的循环序列，对这一无限序列的论证作了进一步的扩展。赫伯特·戴维森在其关于中世纪卡拉姆（经院神学）的著作《永恒的证明》中讨论了这一点：……一个自足的循环回归，仅凭适用于它本身的论证即可证明其荒谬。在循环回归 x y z 中，x 将是 z 的远因，而 z 将是 x 的直接原因。于是 x 将是它自身的远因，伊本·西那认为这是荒谬的。同样，x 也将是它自身的远果，这同样荒谬。这一点还可以用稍有不同的方式再次说明：x 的存在将依赖于某个在存在上后于它的事物 z。但是“当某物的存在依赖于另一个在本质上后于第一者的事物的存在时，第一者的存在是不可能的。”因此，一个自足的循环因果回归不可能存在。

Ibn Sīnā expands this argument of an infinite set in both The Pointers and The Deliverance by considering a circular finite set. Herbert Davidson addresses this point in his work on medieval Kalām (scholastic theology) entitled as Proofs for Eternity: …a self-contained circular regress is shown to be absurd by an argument applying only to it. In the circular regress x y z, x would be a distant cause of z, and z would be the immediate cause of x. x would consequently be a distant cause of itself, which Avicenna regards as absurd. By the same token, x would be a distant effect of itself, which is equally absurd. And the point can be made again in a slightly different way, as follows: x would be dependent for its existence upon something-z-whose existence is posterior to it. But “when the existence of something depends upon the existence of something else that is essentially posterior to the first, the existence of the first is impossible.” A self contained circular regress of causes cannot, therefore, exist.

接着讨论有限集合之后，伊本·西纳转而论述无限回溯的反驳，他说道：如果另一方面这一集合是无限的，那么这样的两端（第一因与最终结果）便不会实现，集合中的所有事物都将具有处于两个集合之间的那些事物的［同等］地位（即它们都不是第一因或最终结果）。21

Having dealt with a finite set, Ibn Sīnā moves on to address the infinite regress objection, stating: If the set, on the other hand, is infinite, then such extremities (the first cause and final effect) will not have been realised, and all things in the set will have the [equivalent] status of the things between the two sets (i.e. they will not be first causes or final effects).21

根据这一释义,以及回到我们P1至P10的例子,在无限的集合中,P1、P5和P10都是没有意义的。所有的P在模态规定上都是相同的,它们不可能被视为"起始者"、"中间点"或"最终结果"。伊本·西那就这一点进一步阐述道:[关于]这个无限集合,无论你对它增加还是减少,其"无限"的状态都保持不变。在这种[无限集合]的情形下,原因的集合不可能存在而没有一个本身无因、自发起源的原因。因此,集合中的所有事物都将具有中间事物的地位(即既不是最终结果也不是第一因),而这在逻辑上是不可能的。22

On this rendering – and returning back to our P1-P10 example – P1, P5, and P10 are meaningless in an infinite set. All Ps have the same modal specification, and they cannot possibly be considered ‘initiators’, ‘middle points’, or ‘final effects’. Ibn Sīnā elaborates on this point by stating: [Regarding] this infinite set, if you add or subtract from it, its status as ‘infinite’ will remain the same. In this case [of the infinite set], it is not possible for the set of causes to exist without having a cause that is itself uncaused and originating. Accordingly, all of the things in the set will have the status of the middle things in the set (i.e. being neither final effect or first cause), and this is logically impossible.22

伊本·西那与亚里士多德一样，是永恒论者。换言之，他确实相信宇宙是前永恒的。这一令人惊讶的事实使他与许多反对此类观点的伊斯兰思想家（最著名的是安萨里）处于对立面。尽管相信宇宙永恒，伊本·西那和亚里士多德一样，明确论证了因果链无限回溯的不可能性。23 除了中间因的论证之外，伊本·西那还以如下方式提出了一个关于复合的论证：因此我们说，必然存在不能被视为一种“复合构造”，即其中有某种“是什么”的属性[归于它]，而这种“是什么”本身就是必然性的性质……在这样一种“是什么”之中存在实际的实在，在逻辑上是不可能的，因为如果它有某种实在，并且该实在与必然存在相区别，那就表明某种非必然的东西造成了某种必然的东西。这将表明所讨论的事物实际上并非必然的。24

Ibn Sīnā, like Aristotle, was an eternalist. In other words, he actually believed that the universe was pre-eternal. This startling fact puts him at odds with many Islamic thinkers who opposed such notions (most notably al-Ghazālī). Despite believing in an eternal universe, Ibn Sīnā, like Aristotle, argued quite clearly for the impossibility of an infinite regress of causes.23 In addition to the argument from middle causes, Ibn Sīnā makes an argument from composition in the following manner: So we say [that] the necessary existence cannot be considered a ‘composite construction’ such that there is a certain ‘whatness’ attribute [attributed to it], and that such ‘whatness’ is itself the quality of necessity…it would be logically impossible that in such a ‘whatness’ there is an actual reality, for if it had a reality and that it was differentiated from the necessary existence then it would indicate that something unnecessary had brought about something necessary. And this would indicate that the thing in question is not, in fact, necessary.24

论证由组合而来的论点认为，任何被组合的事物都是被产生的，因此是依赖性的。如果某物之内有一种品质，使其成为“必然”并能够从中被抽取，同时又具有差异性的特征，那么这将表明该实体需要这种品质。因此，这足以证明该实体的偶然性。采纳伊本·西纳的前提：任何由其部分而存在（而非由其自身存在）的复合实体都是复合的，而任何复合体都是依赖性的，那么存在无限多偶然事物的设想就变得无法成立。赫伯特·戴维森如此概括这一论证： 因此，依照这一另一种观点，“无论该群体是有限还是无限”，它都需要一个能够持续“赋予其存在”的因素。伊本·西纳假定，这一因素必定要么（β1）在该群体之内，要么（β1）在其之外。然而，假定整个群体（β1）最终由其自身的某个成员所维持，就等同于假定该成员是其自身的原因。因为，作为一个群体存在的原因，“首先”就是作为各个成员的原因；而既然所假定的原因本身即是该群体的成员之一，它就必然是其自身的原因。但是，我们已经假定，作为群体的成员之一，该原因是可能存在的；而可能存在者，恰恰是不因其自身而存在者。因此，它不可能是它所属的那个整体的原因。26

The argument from composition suggests that anything which is composed is generated and is therefore dependent. If there is a quality within something that makes it ‘necessary’ and capable of being extracted from while also being characterised with differentiation, then it would suggest that the entity in question is in need of it. As such, this would be enough proof to indicate the contingency of that entity. Taking Ibn Sīnā’s premise that any composed entity that exists by its parts (and not by virtue of itself) is composite and therefore any composite is dependent,25 the prospect of an infinite number of contingent things in existence becomes an impossible prospect to maintain. Hebert Davidson summarises this argument in the following way: Hence, on this alternative, “whether the group is finite or infinite,” it stands in need of a factor that will continually “provide it with existence.” The factor, Avicenna assumes, must be either (β1) within the group or (β1) outside it. Assuming that the whole group is (β1) ultimately maintained by one of its own members would, however, be tantamount to assuming that the member in question is a cause of itself. For to be a cause of the existence of a group is “primarily” to be the cause of the individual members; and inasmuch as the supposed cause is itself one of the members, it would perforce be a cause of itself. Yet the supposed cause has already been assumed, as one of the members of the group, to be possibly existent; and the possibly existent is precisely what does not exist by reason of itself. Therefore, it could not be the cause of the collection of which it is one member.26

这一论点与围绕“新无神论”运动及其对有神论的质疑之讨论尤为相关。彼得·阿达姆森（Peter Adamson）在其关于伊本·西拿（Ibn Sīnā）的三部分专题播客中提到，这一论证与当代有神论者和无神论者之间的讨论有关。在谈到伊本·西拿的论证中，即若整体的各个部分都是偶然的，那么整个存在的集合是必然的，这一点是难以置信的时候，阿达姆森提到了伯特兰·罗素（Bertrand Russell）著名的“组合谬误”（fallacy of composition）反驳。关于这一反驳，本书将在关于反驳的章节中作更详细的探讨。阿达姆森举了一个由许多小零件组成的大钟的例子，并指出，仅仅因为宇宙内部存在原因，并不意味着宇宙本身有一个原因。阿达姆森解释说，这种观点： “……听起来像是当代无神论者可能会说的话。伊本·西拿对此并不在意。他看到对手可能会提出这一反驳，于是像耸耸肩一样说，在这种情况下，对手其实只是把他想要的东西交给了他……毕竟，他的目的就是要证明必然存在的存在。对手实际上已经承认了这一点——只不过对手认为，这个必然存在者就是宇宙本身。这个反驳根本算不上反驳，而是一种投降。”27

This point is particularly relevant to the discussion on the New Atheism movement and its interrogations of theism. In his three-part specialised podcast on Ibn Sīnā, Peter Adamson mentions the relevance of this argument to modern day discussions between theists and atheists. When discussing the part of Ibn Sīnā’s argument in which he examines the implausibility of the entire set of existence being necessary if its parts are contingent, Adamson refers to Bertrand Russell’s famous fallacy of composition objection. The latter will be revisited in greater detail in the chapter on objections. Adamson provides the example of a big clock made up of small parts and mentions that just because there are causes within the universe, it does not necessarily mean that the universe itself has a cause. Adamson explains that this: …sounds like the sort of thing a modern-day atheist may say. Avicenna is relaxed on this point. He sees that an opponent might raise this objection and, as if shrugging his shoulders, says that in that case, the opponent would just be giving him what he wants…after all, he is out to prove that there is a necessary existence. The opponent has actually admitted that – it’s just that the opponent thinks that the necessary existence is the universe itself – the objection is no objection at all, but a capitulation.27

话虽如此,阿达姆森解释说,按照伊本·西纳的观点,宇宙(或这一整套单元)不可能是必然存在者,原因在于:如果一个必然存在者有部分,那么就需要某种东西来区分这些部分彼此。但是,按照我们刚才使用的同样推理,这些部分最终会彼此不同,于是它们就不再是必然的了——然而,一个必然存在者又怎么可能拥有偶然的部分呢?28

Having said this, Adamson explains that according to Ibn Sīnā, the universe (or the set of units) cannot be the necessary existence because: If a single necessary existent had parts, then something would need to distinguish those parts from one another. But then, by the same reasoning, we just used, the parts would wind up being different from one another, and then they would not be necessary – but how can a necessary existent have contingent parts?28

在以伊本·西那的论证进行牧灵劝导时，重要的是要留意对话者何时已从无神论的解释转向与自然神论或古典有神论更为相称的解释。这些逻辑论证的目的，并不是像伊本·西那在上文那样从第一原理出发，说服对话者接受经文所载神的全部属性。相反，首要目标是向持无神论立场的反对者证明：一种不承认存在之起源者的世界观是有缺陷的，并且很容易被驳倒。概括而言，伊本·西那的"证明"（Burhān）论证可归为一种源自偶然性的因果论证。该论证的要旨是：存在不能仅由偶然的实在构成，因为偶然的存在无法使自身产生。一组偶然的存在，无论有限还是无限，都不是自足的，也不是必然的，因为这样一组存在由许多组成部分构成。如果有某种东西能够区分甲部分与乙部分，那么这种东西就表明甲部分或乙部分并非必然，整体（即所描述的那个"整体"）也非必然。因此，必然存在者不可能由组成部分构成。由于同样的理由，一组偶然的存在，无论有限还是无限，也不可能是必然的存在。所以，必然存在者必须存在于偶然存在的集合之外，并最终使这一集合及其所有成员得以存在。这样一个位于偶然存在集合之外的必然存在者，必须是自足的；它无需任何东西来产生或致其存在，因为若非如此，便必然表明它是偶然的。

When using Ibn Sīnā’s arguments pastorally, it is important to note the threshold at which an interlocutor has shifted away from atheistic explanations and moved towards explanations which are somewhat more commensurate with deism or classical theism. The purpose of these logical arguments is not to convince the interlocutor of all of the attributes of God in accordance with scripture by arguing (as Ibn Sīnā does above) from first principles. Rather, the prime objective is to demonstrate to an atheistic detractor that a worldview which does not acknowledge the status of an originator of existence is deficient and can be easily repudiated. To summarise, Ibn Sīnā’s Burhān argument can be categorised as a cause-based argument from contingency. The thrust of the argument is that existence cannot contain only contingent actualities, as contingent existence cannot bring rise to itself. A collection of contingent existences –whether finite or infinite – is not self-sufficient or necessary, as such a collection is made up of many component parts. If there is anything to differentiate Part A from Part B, such a thing would indicate the non necessity of Part A or Part B, as well as the ‘whole’ which is being described. For this reason, the necessary existent cannot be composed of component parts. In addition, because of this reason, a set of contingent existences – whether finite or infinite – cannot be a necessary existence. Therefore, there is the requirement for a necessary existence to subsist outside of the set of contingent existence, which ultimately causes this set, as well as all of its members. Such a necessary existence which is outside the set of contingent existences must be self-sufficient; it requires nothing to generate or cause it, as the failure to be such would necessarily indicate its contingency.

理查德（Richard）我在引言中已经介绍过，他可能会以无限多元宇宙来取代必然存在者的命题。可以向理查德提出的问题是：这样一个多元宇宙是自足的，还是依赖的？如果理查德回答说它是依赖的，那么问题就是：它依赖的是某种依赖性的东西，还是某种独立（自足）的东西？如果是独立的，那么至少在一点上双方达成了一致，即存在一个“独立实体”。如果是依赖的，那么是否存在依赖事物的无穷倒退？如果存在，那么这种无穷倒退又依赖于什么？理查德可能会说，它不依赖于任何东西。如果是这样，理查德就是承认无限多元宇宙是独立的，这同样是一个共识点，因为理查德将同意存在某种自足的事物。真正的主要分歧在于这个独立存在者究竟是什么。有人可以主张，任何由可分离／可附加部分构成的事物都是依赖的，而宇宙正是以这种方式构成的，因此宇宙是依赖的。同样的道理也适用于无限多元宇宙。这一点将在接下来的各章中进一步阐述。

Richard, who I have described in the introduction, may reject the proposition of a necessary existence in exchange for an infinite multiverse. The question that may be posed to Richard is whether such a multi-verse is self-sufficient or dependent. If Richard replies that it is dependent, then the question is: is it dependent on something which is dependent or something which is independent (self-sufficient). If independent, then there is agreement at least on the point that there exists an ‘independent entity’. If dependent, then is there an infinite regress of dependent things? If so, what does such infinite regress depend on? Richard may say that it depends upon nothing. If this is so, Richard is admitting the infinite multiverse is independent, again a point of agreement as Richard would be agreeing to the existence of something self-sufficient. The major point of disagreement relates to what the independent being is. One can argue that anything composed of detachable/attachable parts is dependent, and since the universe is composed in this way, therefore the universe is dependent. The same thing can be said about an infinite multiverse. This will be fleshed out in the chapters to come.