SOURCE: Feldman, Seymour. “Platonic Themes in Gersonides's Cosmology.” In Salo Wittmayer Baron: Jubilee Volume, On the Occasion of His Eightieth Birthday, English Section, Vol. 1, pp. 383-405. New York: Columbia University Press, 1974.

[In the following essay, Feldman analyzes Gersonides's argument against the ex nihilo doctrine of creation and his defense of the idea of the incorruptibility of the universe.]

In his opening presentation of the controversy concerning creation Maimonides cites three distinct cosmological theories, the second of which he attributes to Plato and some other Greek philosophers.¹ The chief characteristic that differentiates this theory from the Biblical and Aristotelian doctrines is that it asserts the creation of the universe from some eternal matter.² Since the Platonic theory, as understood by the medievals, asserted creation of some kind, it was more acceptable to the medieval world than the eternity hypothesis of Aristotle. But the Platonic notion of eternal matter bothered them; for it was not clear whether this part of the Platonic cosmology was compatible with the Bible.

Indeed, the dominant view in the three major philosophical-religious traditions (Jewish, Christian and Muslim) was that God created the universe ex nihilo. Quite early in the development of Jewish theology the latter doctrine was advanced as the correct interpretation of the first chapter of Genesis.³ This view became the “official” view of the Church Fathers and the Muslim Mutakallimūn as well. Amongst the philosophers too this issue was of considerable interest. In the Jewish tradition, for example, Saadia Gaon attempted to defend creation ex nihilo by several philosophical arguments.⁴ Judah Halevi and Maimonides, however, were not as certain with respect to both the incompatibility of the Platonic view with Scripture and the philosophical validity of the arguments in behalf of creation ex nihilo. To them what is essential is the belief in creation; the details are secondary.⁵ There were, however, a few philosophers who adhered to the Platonic theory. Of this latter group Levi ben Gerson (Gersonides) was probably the most vigorous defender of the Platonic hypothesis.⁶

Another aspect of Plato's cosmology was also of considerable interest to the medieval philosopher. Plato held that although the universe had a temporal beginning it has no end: the world is everlasting.⁷ Aristotle rejected this position and attempted to refute it, holding that anything generable must be destructible.⁸ On this cosmological point Maimonides was again not sure, at least not as a philosopher. He believed that the universe is indestructible and that this view was the correct teaching of the Bible. But he recognized that many of his co-religionists held the contrary, and he refused to consider them as heretics. Accordingly, he was content to cite certain Biblical and Rabbinic passages that supported his view, without entering into a philosophical argument in behalf of it.⁹ This caution and tolerance, unusual for Maimonides in philosophical matters, are absent in Gersonides. For not only does Gersonides side with Plato (as well as Maimonides), but he also defends this claim with rigorous philosophical arguments. This paper is devoted to a discussion of Gersonides' arguments in behalf of these two Platonic theses: 1) the universe is created out of some kind of eternal matter; and 2) the universe is indestructible.¹⁰

I

For the purposes of this essay creation ex nihilo shall be understood as asserting the non-existence of any “material cause” or stuff, prior to creation from which God created the universe. Let us construct the following “thought-experiment” to make this notion clearer. Since ex hypothesi the universe had a temporal beginning, let us imagine the very moment at which the universe came into being. The doctrine of creation ex nihilo asserts that “prior” to this moment there existed no corporeal substance which acted as a partner in the creative act.¹¹

Now there is at least one prima facie objection to such a theory, an objection that was evident to the Greek philosophers who admitted the generation of the universe but postulated some kind of corporeal substratum as the “stuff” out of which the universe is generated. The ordinary cases of generation are all examples of something coming from something else. This is evident both in natural generation (e.g. flowers from seeds, earth and water) and the crafts (e.g. statues from marble or bronze). To deny this generalization is to run counter to our everyday experience. Generation consists in the emergence of a new form out of some pre-existent material or new individuals out of some pre-existent material and according to a definite structure or form. We don't observe the generation of a physical object from no previously existing matter. So it was quite natural for the ancient philosophers to formulate the law—ex nihilo nihil fit.¹²

Put in another way, it could be said that the doctrine of creation ex nihilo asserts a radical discontinuity in nature. For on this theory it is claimed that an incorporeal agent creates a corporeal substance without employing any material elements. Here too the Greeks seemed to have believed in a general principle according to which like produces like. For this reason Spinoza felt it necessary to ascribe to God the attribute of extension and to collapse the difference between God and the universe. But this route was not open to the medievals, for whom the incorporeality and transcendence of God were philosophical dogmas. Accordingly, Gersonides opted for a different way out—the notion of a primordial matter. In this way divine creation is brought in line with the notion of creation in the crafts and arts.¹³ To Spinoza, however, this route led to a dead-end, and he ultimately rejected the entire doctrine of creation.¹⁴

The force of these prima facie objections, however, could be blunted if the defender of creation ex nihilo is prepared to accept certain discontinuities in and exceptions to the general course of nature. For example, it could be said that creation ex nihilo is simply a miracle, perhaps the greatest miracle of all. Or, put in a different, more philosophical way, if God is omnipotent he could create the world directly, i.e. without any secondary, or instrumental causes.¹⁵ Indeed, creation ex nihilo is, on this view, the clearest example of divine omnipotence. An omnipotent God, it is believed, can bridge the gap between non-being and being; thus creation ex nihilo is a manifestation of the infinite power of God. This conception of nature and divine omnipotence was, however, not acceptable to Gersonides. In certain contexts he was prepared to introduce certain restrictions upon the scope of natural laws, as in his distinction between particular cases of generation, which are subject to the laws of Aristotle's physics, and the creation of the whole of nature, which is not subsumable in toto under Aristotle's theory of natural generation. But this kind of distinction he introduces quite sparingly. In general, nature is a rational system; indeed, even its creation is in some sense “rational,” i.e. it does not violate good sense. This rationalism is so pervasive in Gersonides that it leads him to maintain that the doctrine of creation ex nihilo is not merely implausible, as the previous arguments contend, but actually absurd. It is not, therefore, a question of assessing the relative merits of alternative hypotheses, as it is for Maimonides; rather, one of these hypotheses is radically defective. This claim, therefore, entails that creation ex nihilo is not a genuine possibility; hence it does not fall within the scope of divine omnipotence. God can do only that which is logically possible.¹⁶ To demonstrate the absurdity of the ex nihilo doctrine Gersonides marshalls several philosophical-scientific arguments.

The first set of arguments concerns the concept of a vacuum. Aristotle had pointed out that the belief in creation ex nihilo implies the prior existence of a vacuum. This can be shown as follows. This doctrine asserts that besides God there was nothing before creation; yet when the world is created it occupies a definite spatial region, since it is believed to be finite. Now, that which is empty of any body but could be the place of a body is the void. Hence, the doctrine of creation ex nihilo implies the antemundane existence of the void. But, according to Aristotle and most of the medieval philosophers, the existence of a void is impossible.¹⁷ And so creation ex nihilo is absurd.¹⁸ The absurdity of this theory can be exhibited in a different way. Since all the parts of a vacuum are homogeneous, no particular part is more fitting to be the locus of the world than any other part. God's initial creative act, therefore, would spread throughout the whole infinitive void, resulting in an infinite universe. But for Aristotle and the Aristotelian medieval tradition the universe is finite.¹⁹ Finally, even if it is granted that one particular region of the vacuum is occupied by the universe, there would still exist a void surrounding the universe. For, as before, no particular part of the remaining empty space is more fitting to be the locus for another body than any other part. Thus, either no other universe would be created, and a void would exist extra mundum, or every single point in this infinite space would be occupied, and the universe would be infinite. Since the latter consequence has already been excluded, the former consequence would ensue. Thus, the creation ex nihilo doctrine is committed not only to the existence of a void prior to creation but also to the existence of a vacuum extra mundum.²⁰

The argument positing a vacuum extra mundum requires some comment. Although Aristotle believed in a finite but unbounded universe, he denied the existence of a void beyond the universe. His main reason was that he wanted to preserve the uniqueness of this universe: if a vacuum did exist extra mundum then the existence of other worlds is more than just a possibility, since the vacuum is that which is empty of body but could be the place of a body. On independent grounds, however, Aristotle argues for the impossibility of a plurality of worlds, and then concludes that there is no vaccum beyond our universe.²¹ Thus, for Aristotle, beyond the universe there is absolute, or sheer, nothingness. (Possibly the incorporeal gods reside there, but qua incorporeal they do not occupy space.)

This point was the occasion of some discussion even amongst Aristotle's ancient commentators. Amongst the medieval Aristotelians who accepted creation ex nihilo it was even more of a problem. Thomas Aquinas, for example, is quite alive to these arguments about the vacuum; but he rejects them. He maintains that creation ex nihilo does not commit one either to the existence of a vacuum ante mundum or extra mundum. Prior to creation there was no empty space, as there would be according to Aristotle (and Gersonides); and even after creation there is no empty space surrounding the universe.²² The dimensions of the universe are created simultaneously with the universe, and there is no need to speak of them as potentially existing prior to creation.²³ This too was the answer of Gersonides' Jewish critics, Crescas and Abravanel.²⁴ On the other hand, Aristotle's definition of the void is so worded that a vacuum is implied if creation is admitted; for if the void is defined as that which is empty of body but could be the place of a body, then it is natural to conclude that if the world was created ex nihilo a void preceded creation. And since the universe is finite, a vacuum on this hypothesis would still exist extra mundum; for, as Aristotle has shown, there cannot be any other worlds. Indeed, Crescas is prepared to admit, at least as philosophically unobjectionable, the existence of a vacuum extra mundum, which confirms Gersonides' contention that creation ex nihilo is committed to the existence of a vacuum at least extra mundum.²⁵ Gersonides, following Aristotle and Averroes, maintains that although the universe is finite it is not surrounded by a vacuum, even though this is a notion that is difficult to understand. “Beyond” the universe there is just nothing.²⁶

The notion of a vacuum extra mundum is philosophically related to the first of many objections that Gersonides considers against his version of the Platonic theory. Let us assume that the universe has been created from some primordial matter. Has this matter been thoroughly exhausted or is there a surplus?²⁷ Already in Plato and Aristotle we find the view that there is no other matter besides our universe.²⁸ Origen, however, raises the following objection to this view: if all the primordial matter has been incorporated into the universe, isn't this just chance? Moreover, isn't it just chance that the primordial stuff was of the right amount so that enough matter was available for the creation of the universe? But to introduce chance into the creative act is to diminish God's omnipotence. Hence, creation must have been ex nihilo.²⁹

These questions vex Gersonides, and he is not sure how to answer them. He agrees with the view of Origen. whom he does not mention, that it is not proper to subject God's creative act to chance: it cannot be mere accident that the primordial stuff was just the right amount as our present universe. But instead of concluding that the universe was created ex nihilo, as did Origen, he cautiously suggests that there is some surplus matter.³⁰ As we shall see, however, this surplus matter does not constitute another universe, or cosmos, as Aristotle feared it would; but it exists. Accordingly, in creation some of the primordial matter was fashioned into a cosmos, whereas the rest of it remains distributed throughout space.³¹

The second major philosophical argument levelled against the defender of creation ex nihilo is the argument from possibility. Briefly stated, the argument is as follows: prior to creation the world was possible, i.e. it could have been created. Now this possibility, or potentiality for creation, must inhere in some substratum; but matter is the bearer of all potentialities. Hence, prior to creation there must have existed something serving as the material matrix from which the universe was created and as the recipient of God's creative act.³²

The notion of a bearer of possibilities sounds strange to modern philosophical ears. But consider the following examples. We can say that a block of marble is potentially a statue of Hercules; analogously, we can say that an embryo is potentially a child. When the statue has been made and the child is born, each potentiality has been actualized. In both cases the potentiality, or possibility, is true of some material substrate, e.g. marble or a fertilized ovum. In this sense the medievals said that the possibility inheres in the material substratum. For Aristotle, matter is in general the bearer of potentialities, especially in the context of change and generation.³³

Now there were two main ways whereby the defenders of creation ex nihilo attempted to meet this argument. One could simply reject the notion of possibility: either the concept of possibility is eliminated entirely or it is interpreted in a “subjective” manner. In either case, however, there are no potentialities for which a material substratum is necessary.³⁴ Or, one could interpret the possibility in question as simply the power of God to create the wolrd. As Thomas Aquinas puts it, it is not logically repugnant to ascribe this power to God. Thomas' interpretation rests upon a distinction between a passive potentiality for becoming a particular thing (e.g. the potentiality of a block of marble becoming a statue of Hercules) and the logical possibility that an agent can perform some act.³⁵ It is potency in the latter sense that is relevant in this context, and this kind of potency does not require a material substratum.

These various solutions are not acceptable to Maimonides, Averroes and Gersonides. They all believe in “objective possibilities,” i.e. to say that x is possible is to say something true about the universe, even though x is as of yet an unrealized event.³⁶ Nor will it do to ascribe the possibility only to the agent: for there are here two distinct situations. It is one thing to say that a particular agent can do x; it is another thing to say that x can be produced.³⁷ Now, according to Averroes and Gersonides, the latter sense is primary: an agent can do something if that act is possible. Indeed, Thomas himself admits that omnipotence is defined as the ability to do whatever is possible. Hence, the ability of an agent to do some act implies that the act is a genuine possibility, and this has to be determined independently of the agent. This means that the possibility of creation is a feature that cannot be attributed to the agent alone, but must be attributed to some substratum in which it inheres, i.e. matter. (To say that this matter is itself created is to open the door to an infinite regress. Accordingly, let us keep the door closed ab initio.) Indeed, if God's creative act emanates from His overflowing goodness and grace, as all creationists believe, then, Gersonides contends, there must be something that is the recipient of this divine influx. In short, making something necessarily involves three components: an agent, or maker, some material that is the subject of the act of making, and finally the product. The advocate of creation ex nihilo dispenses with the second of these factors; the critic of this doctrine argues that this factor is a necessary element in any kind of productive activity, including the creation of the universe.³⁸

Nevertheless, Gersonides' doctrine is not free from difficulties. In accepting the Platonic notion of an eternal primordial matter Gersonides must come to grips with Aristotle's criticism of one aspect of this doctrine—the problem of the disorderly motion in the “receptacle”, or primordial matter.³⁹ Whatever the correct interpretation of this notion, Aristotle construed it as implying that the primitive elements were in irregular motion. He claims this notion is absurd. For if these elements are in motion, this motion would have to satisfy some set of dynamical laws; but lawfulness implies order and form, and hence a cosmos would already be in existence.⁴⁰ Gersonides accepts this argument. But whereas Aristotle uses it against Plato's entire theory of creation, Gersonides restricts it to Plato's notion of disorderly motion. For Gersonides the primordial matter is neither in motion nor at rest, since neither state applies to it. The primordial matter is without form (the primeval chaos), and as such not capable of either motion or rest; for only a body with form is capable of these states.⁴¹ Moreover, since this matter is formless, it does not constitute a cosmos, and so our universe is unique.

In evading Aristotle's criticism of Plato's notion of disorderly motion Gersonides seems to violate, however, one of Aristotle's fundamental principles—there is no matter without form.⁴² Gersonides claims that this principle is not universally true. Firstly, as a matter of empirical fact there is even now a kind of matter that is formless: the matter between the celestial spheres, which has no form whatever and hence possesses no natural motion.⁴³ Indeed, this matter is the pristine remnant of the original primordial matter, whose other portions have been fashioned into the heavenly and earthly spheres.⁴⁴

Secondly, it does no follow from Aristotle's own system that matter must always be “informed.” To be sure this rule is true for sublunar matter, where earthly changes consist in transformations amongst contrary states or from the privation of a particular form to the possession of this form. But, as Aristotle insists, the heavenly bodies do not exhibit this kind of change, since they are not subject to privations or contrary states. Indeed, the heavenly bodies consist of a different element, the aether.⁴⁵ This difference suggested to some of the medievals that the general form-matter distinction is not applicable to the heavens, which can be considered to be simple.⁴⁶ This point is reinforced when we consider that for Aristotle forms can exist without matter, i.e. the unmoved movers, or in the medieval framework the Separate Intelligences. Now if forms can exist separate from matter, why not the converse?⁴⁷ Indeed, if it is appreciated that the heavenly bodies are simple it could be said that they are just matter.⁴⁸ Accordingly, even in Aristotle's own system the form-matter analysis is not universally applicable.

One standard objection to the primordial matter remains to be disposed of. If this matter is alleged to be eternal, then it would be equal in rank to God; indeed, it would be a god since it is eternal. But this is polytheism.⁴⁹ This objection, Gersonides replies, can be easily dissipated. The term “eternity” does not mean or imply divinity. Even for Aristotle this is so, since the world may be eternal, but it is not considered divine or a god. God is divine not because He is eternal, but because He is capable of performing certain kinds of acts, and in fact does perform them. The primordial matter doesn't act at all (since it has no form), and consequently represents the lowest degree of reality, despite its eternity. Indeed, for Gersonides, matter is a source of evil.⁵⁰

II

The second Platonic thesis that Gersonides defends is the incorruptibility of the universe. Here his task will be especially difficult, since he will have to counter the claims not only of the orthodox who maintain that this thesis puts restrictions upon God's infinite power, but also the arguments of Aristotle who attempts to prove the absurdity of the notion of a generated but incorruptible universe.

According to Aristotle and Gersonides natural destruction or corruption is the result of certain kinds of change: either the acquisition of a new form (substantial change) or the assumption of new qualities (alteration). In each case we have a passage from opposite states, either from the privation to the possession of a form or the passage from one contrary quality to another.⁵¹ The heavens, however, do not possess contrary capacities such that they are subject to qualitative change. Nor do they ever assume a different form from the one they currently possess; that is, they do not suffer substantial change. Indeed, the only kind of change experienced by the heavenly bodies is locomotion; but this kind of change does not imply any novelty in the structure of these bodies.⁵² Thus, for Aristotle, the heavens are eternal.⁵³ To Gersonides this argument establishes the thesis that the heavenly bodies are indestructible, which, as we shall see, is a weaker claim than the eternity of the heavens. If the heavenly bodies are not subject to change other than locomotion, then they are at least indestructible. Now if the heavens are indestructible, so must be the universe in toto; for if the earth could be destroyed while the heavens remain intact, the latter will have no object upon which to bestow its benefits, and hence would be gratuitous, which has been shown to be false.⁵⁴

Now it might be objected that if the universe cannot be destroyed, God's power is thereby limited. Divine omnipotence is not only manifested, according to this view, in creation ex nihilo but also in God's capacity to destroy the universe if He wishes. Indeed, although Maimonides suggested that the universe is everlasting, he did say that God, if He wishes, could destroy it.⁵⁵ For him the indestructibility of the universe is a doctrine taught by Scripture, i.e. God has revealed to us that He does not wish to destroy the universe.⁵⁶ Gersonides' claim is stronger: it asserts the impossibility of destruction. This impossibility, however, does not diminish God's power. For what is impossible cannot be done, and the destruction of the universe, like creation ex nihilo, is an impossibility.

Let us suppose that God could destroy the universe. The question can be raised, what motive could He have for doing so? If someone makes something he destroys it either because the product is unsatisfactory in some respect, or out of a fit of anger, or from vengeance. Neither of these possibilities can be attributed to God. Firstly, God created a world that is as perfect as it can be. Indeed, to say that he could destroy the world and make a better one is to diminish God's power. For why didn't He make the better one in the first place, especially since God is good?⁵⁷ In the second place, God is not subject to anger or revenge; hence He could not destroy the universe as the result of these passions. Thus, there is no reason that God could have for annihilating the universe. The impossibility of this event resides in its incompatibility with the notion of a perfect and good God.⁵⁸ In this respect, it sould be emphasized, Gersonides' claim is even stronger than that made by Plato, who makes the everlastingness of the universe dependent upon divine will.⁵⁹

The second stage of Gersonides' argument, however, is more difficult. He must now show that Aristotle's cosmological theorem—whatever is generated is destructible, and conversely—is false.⁶⁰ In the defense of this theorem Aristotle marshalls several arguments, of which two are especially crucial: 1) a metaphysical analysis showing that the concept of a generated but indestructible world is absurd; 2) a logical analysis of the terms “generated” and “destructible” showing that they are “coincident,” or coextensive.

The first argument is presented in several different versions, all claiming to show the absurdity inherent in the Platonic conception of the universe. Now ex hypothesi such a universe exhibits two contrary capacities in so far as it is a created universe: the capacity for existence and the capacity for non-existence. Now, when a substance can sustain, or have, contrary properties, these capacities must be finite, or limited, in duration; otherwise these capacities would be simultaneously realized. Let us consider first the case when the universe ex hypothesi has not yet been created. Now in this situation we have the two contrary possibilities of existence and non-existence obtaining throughout all of infinite time prior to creation. But since in infinite time all possibilities are realized, the capacity for existence would be realized prior to creation, which is absurd.⁶¹ Suppose further that the universe has been created. As a created substance it is obviously something that could have not existed, i.e. it has the capacity for non-existence. Now if it is alleged that the universe endures ad infinitum, this capacity for non-existence will be realized sooner or later, for capacities are future-directed. But this is absurd, since ex hypothesi the universe is indestructible.⁶² Put differently, contingency is an inherited property: if the universe is created, i.e. contingent, there is always the possibility of its destruction.

In rebuttal Gersonides replies to the first case by pointing out that Aristotle's argument is an ignoratio elenchi. The theory of creation does not assert that creation takes place in time; rather, it asserts that time was created together with the universe. Thus, Aristotle's first objection is irrelevant. Indeed, all the stock arguments of this type are equally beside the point, e.g. why did God create the universe at this moment and not at another? Or, since no moment is more propitious than another, the universe has not been created at all but is eternal.⁶³ All these arguments assume what the theory of creation definitely denies: the ante-mundane existence of time.⁶⁴

The second case, however, is more difficult to dispose of. It would seem that Aristotle is right in saying that contingency is an inherited property, that a substance which was once non-existent should be so again in the future. Nevertheless, Gersonides contends that this argument rests upon a confusion between the concepts of contingency and of corruption. A substance that can and cannot exist is, to be sure, a contingent substance. It would be obviously absurd to say that a substance of this sort is eternal, for, as Aristotle says, an eternal substance is a necessary existent, and conversely.⁶⁵ But an incorruptible substance is not necessarily an eternal substance; analogously, a contingent substance is not necessarily a corruptible substance. For Gersonides, then, these concepts are not identical, as they are for Aristotle. Indeed, Aristotle often uses these concepts interchangeably throughout his criticism of Plato, without realizing that these equations have to be defended.⁶⁶ Gersonides agrees that on the hypothesis of creation, the universe is contingent; but this means merely that it has been generated, that it was once non-existent. The “capacity for non-existence” has been realized, so to speak, in the non-eternity of the universe. The world's contingency is then expressed simply in the fact of its creation. But to admit this is not to imply, at least without further argument, that the world cannot endure infinitely a parte post. Indeed, this is the crucial point: whereas for Aristotle eternity and infinite duration are identical notions, for Gersonides they are different. The former is equivalent to infinite duration a parte ante and a parte post. To Gersonides, however, infinite duration is also a feature of a substance that is incorruptible, i.e. a substance that exists infinitely only a parte post. The contingency of the universe (i.e. the possibility of its non-existence) derives from the divine act of creation which, Gersonides stresses, is free. That is, to say that the world is contingent is to say that God was free in producing it. The universe is contingent not in the sense of being capable of destruction, as Aristotle maintains, but simply because it is not eternal. Eternity and contingency, then, are incompatible features of a substance, and Aristotle is right in saying that it is impossible for a contingent substance to be eternal; but this is not the issue. What is in question is the possibility of a generated and incorruptible substance, which is, to be sure, contingent. Aristotle's arguments, however, have so far not established its impossibility.⁶⁷

Aristotle's second major argument against the possibility of a generated but incorruptible world consists in a linguistic analysis of the relevant concepts.⁶⁸ Let us consider the terms “always existing” (i.e. eternal) and “never existing” to be ontological extremes, or contraries. The contradictory of the former is “that which is capable of not existing at some time,” whereas the contradictory of the latter is “that which can exist at some time.” Now if the latter two predicates are (non-simultaneously) applicable to something, then that thing will have the capacities of existence and non-existence, i.e. contingency, and hence will be the intermediate entity between that which always exists and that which never exists.⁶⁹ Aristotle then identifies this intermediate entity, or term, as the generable and the destructible.⁷⁰ From this identification he then concludes that “generable” and “corruptible” are convertible predicates.⁷¹

Gersonides maintains that this argument is not valid. Aristotle is wrong in identifying the intermediate term with the generable and the destructible. For any generated substance is something that can and cannot exist (non-simultaneously), and, therefore is, contingent. Whether this substance is destructible or indestructible is irrelevant. And to assume that every contingent is corruptible is to beg the question, as we have already seen. Consequently, Gersonides' hypothesis of a generated but indestructible universe describes a world that is also intermediate between a world that is eternal and a world that cannot exist, just as much as the notion of a generated and destructible universe. Gersonides suggests that Aristotle should have begun his analysis with the terms “existence” and “non-existence”, which are the most general and basic ontological categories. Then he would have seen that under the former notion we have the sub-classes of eternal and of contingent beings; and of the latter category the generated and incorruptible is an instance, just as much as the generated and corruptible.⁷²

If we now try to form a general picture of Gersonides' cosmology we see that it asserts the existence of a universe that endures infinitely in one direction only, i.e. a parte post. But for Aristotle the notion of a “one-sided” infinite is absurd, and at this point in this dialectic an argument crops up that gives Gersonides considerable difficulty. It would seem that a magnitude, e.g. a length, that is infinite in one direction, i.e. at one of its termini, is “smaller” than, or a part of, a magnitude that is infinite in both directions. But for most medievals, and most likely for Aristotle too, infinite magnitudes are all equal. Hence, an infinite part of an infinite whole would qua infinite be equal to the whole; however, this is absurd, since according to common sense the whole is greater than any of its parts. Consequently, the notion of a universe that endures infinitely a parte post only is absurd.⁷³

This argument can be spatially represented as follows: Let line A represent the duration of a universe that is temporally infinite in both directions, i.e. Aristotle's conception of an eternally existing universe; let line X represent the duration of a universe that is temporally infinite in one direction only, a parte post. …

The argument contends that since X is shorter than A it is a part of A. But ex hypothesi X is infinite and qua infinite it is equal to A. Thus, the part is equal to the whole, which is absurd.

To dissipate the force of this objection Gersonides levels three arguments against it. Firstly, the objection is equally valid against the eternity hypothesis. Consider any instant on A; it demarcates two segments each infinite but “smaller” than the whole. That is, the present delineates two parts of time, each infinite, yet as parts presumably smaller than the whole of time. Accordingly, whatever conviction the argument carries is equally pertinent to Aristotle's own theory. Secondly, although from a strict numerical, or quantitative, point of view X is no greater than A, it is possible to consider X “smaller” in so far as it is terminated at one side. In so far as X has a terminus it is a “part” of A, although the term “part” has to be construed in a loose sense.⁷⁴

Finally, Aristotle's argument is again an example of an ignoratio elenchi: it assumes that time is infinite a parte ante as well as a parte post, and that, for the Platonist, the universe exists in only one part of this infinite dimension. But this is not the case. It is not alleged that the universe was created in time, which is infinite in the past and then continues ad infinitum. Rather, it is claimed that time had a beginning: there is no infinite time a parte ante. Aristotle's argument is, therefore, irrelevant.⁷⁵

CONCLUSION

Gersonides' cosmogony turns out to be one of the more unique and and bold theories of creation in the Middle Ages. He defends the doctrine of the primordial matter vigorously, which brings him into conflict not only with his more traditional co-religionists for whom creation ex nihilo had become a dogma, but also with Aristotle for whom Plato's cosmology was utterly defective. In his criticism of the traditional dogma of creation ex nihilo Gersonides forges out a path that Spinoza was to pursue even farther. For both thinkers the notion of creation ex nihilo is too radical a break in nature. Similarly, in advocating the Platonic thesis of the indestructibility of the universe Gersonides defends it in a strong form going far beyond the more modest statements of Maimonides or even Plato. Again, Gersonides points in the direction of Spinoza, for whom also the destructibility of the universe is utterly impossible. Spinoza took the steps that Gersonides was unwilling to take—the complete denial of creation and the elimination of genuine contingency in nature. But in his version of Plato's cosmology Gersonides approaches and prepares the way for Spinoza's radical break with the medieval world.

Notes

1. Moses Maimonides, The Guide of the Perplexed, trans. S. Pines (Chicago, 1963), II, 13.

2. The interpretation of the Timaeus is notoriously difficult. But irrespective of the correct interpretation of Timaeus, 48 e-53 c, the medievals interpreted “the receptacle” to be some kind of eternal and formless matter out of and within which the universe is alleged to have been created. Presumably this interpretation was based upon Aristotle's identification of the receptacle with matter (Physics, 209 b, 10). For a critical discussion of Aristotle's interpretation see H. Cherniss, Aristotle's Criticism of Plato and the Academy (Baltimore, 1944), chapter 2, especially pp. 112 ff.

3. Midrash Bereshit Rabbah, I, 9; Second Maccabees 7:28; Philo Judaeus, De Opificio mundi, VII; H. Wolfson, Philo (Cambridge, Mass., 1948), I, 300-310.

4. Saadia Gaon, The Book of Beliefs and Opinions, trans. S. Rosenblatt (New Haven, 1948), pp. 46-70.

5. Judah Halevi, The Kuzari, trans. H. Hirschfeld (New York, 1946), pp. 47-48; Maimonides, Guide, II, 25. Maimonides' attitude is particularly interesting. Although he opts for the ex nihilo view, he is prepared to accept the Platonic theory if the latter can be demonstratively proven. He suggests that the literal text of the Bible could be interpreted to fit the Platonic theory, if the latter were proven true.

6. Abraham ibn Ezra suggests the Platonic theory in his commentary on Genesis 1:1. This interpretation of ibn Ezra (cf. Guttmann, Philosophies of Judaism [New York, 1969], p. 120) was rejected by Krochmal (cf. N. Krochmal, Moreh Nebukei ha-Zeman, ed. S. Rawidowicz [London 1961], pp. 306, 324).

7. Plato, Timaeus, 31 b.

8. Aristotle, On the Heavens, I, 9-12.

9. Maimonides, Guide, II, 27-30.

10. Gersonides, Milhamot Hashem (The Wars of the Lord), (Leipzig, 1866). Hereafter abbreviated as MH.

11. Naturally, since on this view time is created with the universe, terms like “prior” or “before” are not really proper. But we may use these terms so long as we take them figuratively. Cf. H. Wolfson, “The Kalam Problem of Non-existence and Saadia's Second Theory of Creation,” JQR, XXXVI (1945-46), 371-391; idem. “The Meaning of ex nihilo in the Church Fathers, Arabic and Jewish Philosophy, and St. Thomas”, Medieval Studies in Honor of Jeremiah Ford (Cambridge, Mass., 1948), pp. 355-367.

12. Aristotle, Metaphysics, 1032 b, 30; Gersonides, MH, pp. 364-366.

13. Gersonides, MH, p. 364.

14. The problem that is fatal to any creation theory is how an incorporeal agent can act upon corporeal substance. For Spinoza there could be no such causal action; hence he rejected any kind of creation, voluntaristic or emanationist. Spinoza, Ethics, I, first six propositions. Cf. H. Wolfson, The Philosophy of Spinoza (New York, 1958), vol. I, chapter 4.

15. This is an application of the Principle of Parsimony. Cf. Saadia Gaon, The Book of Beliefs and Opinions, p. 69. In his defense of the doctrine of creation ex nihilo Thomas Aquinas explicitly connects this thesis with the notion of divine omnipotence (Thomas Aquinas, Summa Contra Gentiles, II, c. 20-21).

16. Gersonides implicitly accepts the definition of divine omnipotence explicitly stated by Maimonides: God can do whatever is logically possible. Cf. Maimonides, Guide, III, 15.

17. Aristotle, Physics, IV, 6-9; A. Koyré, “Le vide et l'espace infini au XIV siècle,” Archives d'Histoire Doctrinale et Littéraire du Moyen Âge, XXIV (1949), 45-91.

18. Aristotle, On the Heavens, 302 a, 1-9; Averroes, Tahafut al-Tahafut, trans. S. van der Bergh (London, 1954), vol. I, paragraph 89; Gersonides, MH, pp. 364-365.

19. Gersonides, MH, pp. 364-365; Aristotle, Physics, 203 b, 27-29. These latter two arguments make use of a version of the Principle of Sufficient Reason: if there is a homogeneous vacuum there is no way to account for the creation of the universe. Cf. Leibniz, Letters to Clarke, III, paragraph 6, reprinted in Leibniz Selections, ed. P. Wiener (New York, 1951), p. 224.

20. Gersonides, MH, op. cit. This was the view of the Stoics. Cf. Philo, De aeternitate mundi, 102.

21. Aristotle, On the Heavens, I, 9, especially, 279 a, 12-20.

22. Thomas, Summa Theologiae, I, q. 46, a. 1, ad 4; idem, In De Caelo et Mundo, Liber I, Lectio XXI; Liber III, Lectio VIII.

23. Thomas Aquinas, II Sententiarum, d. 1, q. 1, a. 5, ad 4.

24. Hasdai Crescas, Or Adonai III, 1, 5; Isaac Abravanel, Mif'alot Elohim VI, 3. Cf. I. Efros, The Problem of Space in Jewish Medieval Philosophy (New York, 1917), pp. 85-86. (The latter has been translated into Hebrew [Tel Aviv, 1965]. In the Hebrew version the pages are 62-63.)

25. Some of the Latin Scholastics appreciated the force of this objection. William of Burleigh, for example, realized that creation ex nihilo seems to imply the existence of a vacuum extra mundum. Cf. A. Koyré, “Le vide et l'espace infini au XIV siècle,” pp. 75 ff.

26. Gersonides, MH, p. 386; H. Wolfson, Crescas' Critique of Aristotle (Cambridge, Mass., 1929), pp. 421-422. In his supercommentary on Averroes' Middle Commentary on Aristotle's On the Heavens Gersonides points out that the hypothesis of a plurality of universes, which for Aristotle implied a vaccum extra mundum and is false, does not imply a vacuum. For between these universes there is absolute privation, or non-being …, which excludes the possibility of any other bodies. Nevertheless, Gersonides rejects this hypothesis on different grounds. As we shall see, Gersonides believes that our universe does not exhaust all the available matter, even though this universe is unique. But, as his main argument suggests, he wants to deny also the existence of a vacuum extra mundum on any hypothesis concerning the number of universes. Cf. Gersonides, Supercommentary on Averroes' Middle Commentary on Aristotle's On the Heavens (Parma manuscript, 805), klal 8; MH, pp. 378-380.

27. Gersonides, MH, p. 368.

28. Plato, Timaeus, 32 c; Aristotle, On the Heavens, 278 b, 4-9, 23 ff; Philo Judaeus, De aeternitate mundi, 21.

29. Origen, “Fragment on Genesis,” preserved in Eusebius' Praeparatio Evangelica VII, 335 a-b; On First Principles, trans. G. W. Butterworth (New York, 1966), Book II, chap. 1, pp. 79-80. Cf. Wolfson, “Plato's Pre-existent Matter in Patristic Philosophy,” The Classical Tradition, ed. L. Wallach (Ithaca, 1966), p. 419.

30. Gersonides, MH, p. 372.

31. Ibid., pp. 367-368. Gersonides' discussion of Origen's problems is puzzling. Neither Plato, Aristotle nor Philo are aware of Origen's problem. Of the Jewish philosophers writing in Arabic or Hebrew only Saadia, as far as I know, mentions the general question of the surplus matter; but he does not raise the issue of chance. (Cf. Saadia Gaon, Commentaire sur Sefer Yesira, trans. M. Lambert [Paris, 1891], pp. 18-19.) Nor does Averroes raise this problem in his commentary on that passage in On the Heavens where Aristotle claims that all the available matter is contained in the universe. Averroes, Long Commentary on De Caelo (Venice, 1562), ad locum.

32. Gersonides, MH, p. 365; Averroes, Tahafut al-Tahafut, paragraph 100 ff.; cf. Maimonides, Guide II, 14, the “fourth method;” Thomas Aquinas, De potentia, q. 3. a. 1, 2nd objection. In Maimonides and Thomas this argument is cited as an argument in behalf of the eternity of the universe; Gersonides uses it to show the absurdity of ex nihilo creation.

33. Aristotle, On Generation and Corruption, I, 3; Metaphysics, 1032 a, 20. Indeed, this is why for Aristotle matter cannot be created. Cf. Aristotle, On the Heavens, 302 a, 1 ff; Gersonides, Supercommentary on Averroes' Epitome of Aristotle's Physics, Book I.

34. This was al-Ghazzali's objection to the philosophers. Cf. Averroes' Tahafut al-Tahafut, paragraphs 93 and 103.

35. Thomas Aquinas, Summa Contra Gentiles II, 37, paragraph 5; De potentia, q. 3, a. 1 ad 2.

36. Averroes, Tahafut al-Tahafut, First Discussion, paragraphs 94-107.

37. Maimonides, Guide II, 14; Gersonides, MH, p. 373. Maimonides maintains, however, that in the context of creation there is no need for a substratum for the second possibility; since what is true for natural generation need not be true for generation that is non-natural.

38. Gersonides, MH, p. 373.

39. Plato, Timaeus, 30 a.

40. Aristotle, On the Heavens, 300 b, 17 ff; Gersonides, MH, pp. 362-63; cf. G. Vlastos, “Creation in the Timaeus: Is It a Fiction?”, Studies in Plato's Metaphysics, ed. R. E. Allen (New York, 1965), pp. 401-419.

41. Gersonides, MH, p. 374, reply to the 4th objection.

42. Aristotle, Physics I, 6-9; idem, On Generation and Corruption, I, 3-5; Gersonides, MH, p. 370, fifth objection.

43. In the astronomical portions of MH Gersonides introduces this matter to account for certain astronomical phenomena. Gersonides, MH, pp. 193-194.

44. Ibid., p. 367, p. 425. According to Gersonides this matter is formless not only in the topological sense but also in the deeper structural sense of Aristotle's notion of ειδος. Since this matter has no natural motion, or intrinsic activities, it has no form, or essence; for according to Aristotle form is neither identified with or most clearly manifested in the recurrent activities of a substance. Hence, this matter has no form. (MH, p. 193, pp. 315-316. Cf. P. Duhem, Le système du monde, 2nd ed. [Paris, 1954], V, 221-222.)

45. Aristotle, On the Heavens, I, 1-3; Metaphysics, 1069 b, 25-27.

46. Averroes, Tahafut al-Tahafut, Fourth Discussion, paragraph 271. Wolfson, Crescas' Critique of Aristotle, pp. 594 ff. Gersonides, MH, p. 193.

47. Gersonides, MH, p. 374, reply to fifth objection. This argument seems to be based upon Maimonides' second proof for the existence of a first unmoved mover. (Guide II, 1; cf. also the remarks of Efodi and Shem Tob ad locum.)

48. Averroes, op. cit., quoted in Wolfson, Crescas' Critique, p. 595.

49. Saadia Gaon, Book of Beliefs and Opinions, p. 48; Gersonides, MH, p. 370, third objection.

50. Gersonides, MH, p. 373. This too is a Platonic motif, which in the hands of Aristotle became a strict equation. (Cf. Aristotle, Metaphysics, 988 a, 14-15. H. Cherniss, Aristotle's Criticism of Plato and the Academy, pp. 95 ff.)

51. Aristotle, On Generation and Corruption I, 7; Gersonides, MH, p. 359.

52. Aristotle, Physics, 261 a, 23-24.

53. Aristotle, On the Heavens, 270 a, 13-35.

54. In his proofs for creation Gersonides argues that the heavenly bodies have positive influences upon the earth. (Cf. Gersonides, MH, pp. 359, 310-316).

55. Maimonides, Guide II, 27. Maimonides defends the possibility of its everlastingness by restricting Aristotle's theorem that whatever is generated is corruptible to natural phenomena and generation. The creation of the universe, however, is not a natural occurrence, or generation. Gersonides provides a different kind of argument, as we shall see.

56. This is similar to Thomas' view, according to which the world is not subject to natural corruption but only to supernatural annihilation if God so wills. Annihilation, for Thomas, is the contrary of creation: both are supernatural events. (Cf. Summa Theologiae, I, q. 104, a. 3-4; idem, Commentary on Aristotle's Physics, trans. R. Blackwell, R. Spath and W. Thirkel [London, 1963], paragraphs 986, 1147.)

57. Although Maimonides allows for the possibility of the destruction of the world, he expresses revulsion for the view that God created a succession of worlds, presumably because this doctrine opens the door for doubt upon God's perfection and wisdom. (Maimonides, Guide II, 30; Wolfson, “The Platonic, Aristotelian and Stoic Theories of Creation in Halevi and Maimonides,” Essays in Honor of the Very Rev. Dr. J. H. Hertz [London, 1943], p. 436.) This suggests that Maimonides was moving towards Gersonides' position, according to which the destruction of the universe is inconsistent with God's nature. (Cf. Philo, De aeternitate mundi, 39-44.)

58. Gersonides, MH, pp. 359-361. Gersonides' arguments anticipate Leibniz' principles of sufficient reason and perfection, although Leibniz himself did not consistently develop their implications; for Leibniz accepted Thomas' notion of the supernatural power of annihilation. Gersonides, however, correctly saw that if God acts for a reason and if He chooses to create our universe because it is the best possible world, then it would be inconsistent with the nature of God to ascribe to Him the power to destroy this universe. There is no restraint on God's power here; just the realization that God cannot do what is impossible. (Cf. Leibniz, Monadology, paragraphs 4-6; Letters to Clarke, V, paragraphs 73-76.)

59. Plato, Timaeus, 41 a-b.

60. Aristotle, On the Heavens I, 9-12; Gersonides, MH, pp. 403-415.

61. Aristotle, On the Heavens, 283 a, 20-24.

62. Ibid., 283 a, 11-19; 283, b 13-14.

63. Aristotle, On the Heavens, 283, a 11-14; Averroes, Tahafut al-Tahafut, First Discussion, passim; Leibniz, Letters to S. Clarke, III, paragraph 6, in Selections, ed. P. Wiener, p. 224.

64. Gersonides, MH, p. 406 bottom. Cf. Thomas Aquinas, Summa Contra Gentiles, II, c. 35.

65. Aristotle, On Generation and Corruption, 337 b, 35-338; a 4.

66. Aristotle, On the Heavens, 279 b, 18 ff.; 282 a, 23 ff. It is only from 282 b, 10 ff. that Aristotle attempts to prove these equations. But, as we shall see, these proofs are guilty of a petitio principii.

67. Gersonides, MH, pp. 405-408. The notion of a generated but incorruptible universe differs, then, from the notion of an eternal but contingent universe that Thomas alleges to be a possible, i.e. non-self-contradictory, concept. (Thomas Aquinas, Summa Theologia, I, q. 46, a. 2, ad 1, where Augustine is quoted as reporting this view.) This was the view of Avicenna who maintained that the universe is contingent per se but eternal ab alio, i.e. through its eternal cause, God. (Wolfson, Crescas' Critique, pp. 680-684.) Averroes rejected this notion of Avicenna, and as far as I can tell, so did Gersonides. (MH, pp. 312-313.)

68. Aristotle, On the Heavens, 281 b, 33-282 a, 13; 282 b, 13-283, a 3.

69. Ibid., 281 b, 32-282 a, 13.

70. Ibid., 282 b, 10-12.

71. Ibid., 282 b, 9-10.

72. Gersonides, MH, pp. 410-411. Aristotle's other proofs for the convertibility of “generable” and “corruptible” are also invalid. Either they presuppose the previous argument, which has been shown to be invalid, or are invalid on other grounds. Consider, for example, this indirect proof: if it is premised that anything generable or destructible is contingent, Aristotle concludes that whatever is ungenerated or indestructible is eternal, from which it could be inferred that “generable” and “destructible” are convertible. (On the Heavens, 282 a, 23-38.) This argument, however, is invalid, since it presupposes that the only intermediate term between “eternally existing” and “never existing” is “generated and destructible.” (MH, p. 410.) Nor is Aristotle's direct proof (On the Heavens, 282 b, 9-24) any more successful, since in addition to the premise that “the generable and the destructible” is the intermediate ontological predicate, it introduces the additional premise that the contingent, e.g. the generated, must be finite in duration, which is an obvious petitio principii.

73. Gersonides, MH, pp. 405-407. This argument does not appear explicitly in Aristotle, although the denial of a one-sided infinite is stated in On the Heavens 283 a, 10-11. For a discussion of the principle that infinites must be equal see H. Wolfson, The Philosophy of Spinoza (New York, 1958), I, 286-291; S. Feldman, “Gersonides' Proofs for Creation of the Universe,” PAAJR, XXXV (1967), 131-132; N. Rescher and H. Khatchadourian, “Al Kindi's Epistle on the Finitude of the Universe,” Isis, 1965, pp. 426-433.

74. Gersonides, MH, p. 406; Wolfson, Crescas' Critique, pp. 423-424. Although Cantor's theory of transfinite numbers has opened up to us diverse realms of infinite numbers, it is possible to employ the modern theory as a means to elucidate Gersonides' point. The fundamental notion here is that of an infinite set A set K is infinite if all the members of a proper sub-set of K can be matched one-to-one with all the members of K. Accordingly, the set of integers N is infinite since a proper sub-set of N, the series of positive integers, can be matched one-to-one with all the members of N. In this sense, the set of positive integers is as large as the set of integers, although the latter contains the former. But in another sense N is “larger,” since it also contains the set of negative integers, which is also infinite; hence N encompasses two sub-sets that are each infinite. Although the terms “larger” and “smaller” are mathematically inappropriate here, one might say that N is “richer”, or more inclusive, than any of its sub-sets. To the degree that one of the sub-sets is impoverished it is thereby a “part” of the whole. In our particular case this would mean that just as the set of positive integers does not exhaust all of the set of integers, so too X, infinite time a parte post, doesn't exhaust all of A, and thus can be considered as a “part” of A.

75. Gersonides, MH, p. 406, bottom. Gersonides refutes the thesis of infinite past time in an earlier stage of his argument against Aristotle's theory. Cf. chapter 11 of MH, Book VI, part 1. (Cf. Feldman, “Gersonides' Proofs for Creation of the Universe,” pp. 125-137.)