Introduction

Anthony Ruffus and Jon McGinnis investigate whether two puzzles concerning divine freedom and human freedom in the thought of Avicenna can be solved (Ruffus & McGinnis, 2015). The first puzzle, which I intend to discuss, concerns the compatibility of Avicenna’s views regarding the divine nature and God’s creation of the world. Avicenna holds that since God has existed from all eternity and is immutable and impassible, he cannot come to have an attribute or feature that he has not had from all eternity. He also claims that complete causes are with their effects. The puzzle arises when we consider God’s creating this world.Footnote 1 If (as Avicenna maintains) God is immutable and impassible, then his attributes associated with his creating this world, such as power, will, and knowledge, are unchanging. Given that cause and effect are contemporaneous, it follows that God must have been creating the world from all eternity. But this, as al-Ghazali, Fakhr al-Din al-Razi, and Aquinas have pointed out, seems to amount to the claim that God’s creative act is no different from a matter of natural necessity (i.e. God’s not being different from a natural force). This is a threat to divine freedom, another important divine property, because God would then have no option/choice or will concerning creating the world. Ruffus and McGinnis, following medieval philosopher Naṣīr al-Dīn Ṭūsī, argue that this puzzle can be solved in such a way that can consistently affirm choice-based divine freedom in Avicenna's metaphysics. If Ṭūsī and thus Ruffus and McGinnis are right, Avicenna seems to offer a unique position regarding divine freedom: God does not have a choice-based freedom on what to create but on whether to create.

It is a traditional theistic view that God, as an omnibenevolent, omniscient, omnipotent, and omnirational being, is a perfectly free being. When, however, it is asked why this theistic God who needs nothing creates,Footnote 2 the notion of perfect divine freedom becomes puzzling. As Norman Kretzmann has noted, we can find two major incompatible lines of explanation in the theistic traditions (i.e. Judaism, Christianity, and Islam): the necessitarian line and the contingentist line (Kretzmann, 1991a, 1991b, p. 208). The former holds that God creates because he has a necessary productive nature (i.e. he is necessarily self-diffusive). This is based on “the Dionysian principle” held by Plotinus, Augustine, Avicenna, Anselm, and Aquinas: “Goodness is by its very nature diffusive of itself and (thereby) of being” (Kretzmann, 1991a, 1991b, p. 217).Footnote 3 It is dubious whether Augustine, Anselm (Rogers, 2022, p. 280), and Aquinas (Kretzmann, 1991a, 1991b, p. 222–223) hold the following implication of the principle but it seems to be a necessary consequence of the principle that God could not refrain from creating anything at all.

The proponents of the contingentist line, however, rejects the Dionysian principle. Al-Ghazali and again arguably Aquinas, for instance, hold that since the power to do or choose otherwise is a necessary component of libertarian freedom, God cannot be free if he is necessarily self-diffusive and thus cannot refrain from creating anything at all. Even though there might be some reasons to create, those reasons cannot be coercive for God to create something. This entails that he was perfectly free on whether to create. The proponents of this view might also be divided into two groups: those who hold that God could refrain from creating anything at all just as he could have created another world rather than this world, and those who hold that though God could not refrain from creating anything at all, he could have freely created another world rather than this world. The latter view defends the necessitarian line on whether to create while maintaining the contingentist line on what to create. There is a recent discussion between these libertarian positions (Johnston, 2019 and O’Connor, 2022).

It seems then that there are three different positions on divine freedom in the theistic tradition: (a) God has a choice-based free will neither on whether to create nor on what to create, (b) he has a choice-based free will both on whether to create and on what to create, and (c) he has a choice-based free will on what to create but not on whether to create. Naṣīr al-Dīn Ṭūsī, however, surprisingly suggests a unique position by his interpretation of Avicenna’s account of action and will: Avicenna’s God might have a choice-based free will on whether to create (but not on what to create) (Avicenna, 1957–1960).Footnote 4 He argues that though Avicenna is a proponent of the necessitarian line, his position does not exclude a choice-based free will for God on whether to create. This appears to be surprising because Avicenna’s strong accounts of divine simplicity and emanation seem to be incompatible with a choice-based free will on whether to create. The advantage of his interpretation, Ṭūsī seems to think, is that if it is true, the choice-based free will threatens neither of those accounts.Footnote 5 I argue that given that God is omnirational and thus always is acting for reasons, his interpretation which seems to offer no argument for how Avicenna’s God would have reasons for not creating anything at all fails. Before presenting my argument, however, I shall explain his motivation to offer this fourth position in following section. The third section concerns the interpretation that Ṭūsī, Ruffus, and McGinnis offer to solve the puzzle. In the fourth section, by referring to divine omnirationality, I will offer some objections to their interpretation and then consider some possible responses to those objections. I will conclude that their interpretation ultimately falls short of showing that Avicenna’s simple God can have a choice-based free will on whether to create.

The puzzle: avicenna’s simple god willing to create

The notion of divine freedom becomes puzzling in Avicenna’s metaphysics if one wants to reconcile it with his accounts of divine simplicity and emanation. Thomas Morris has suggested that the doctrine of divine simplicity should be understood as a doctrine that denies “any sort of metaphysical complexity whatsoever” for God (Morris, 1991, p. 113). Though he explains what this denial involves, R. T. Mullins provides a more helpful explanation regarding both the medieval and the contemporary literatures:

  1. (1)

    “God cannot have any spatial or temporal parts.

  2. (2)

    God cannot have any intrinsic accidental properties.

  3. (3)

    There cannot be any real distinction between one essential property and another in God’s nature.

  4. (4)

    There cannot be a real distinction between essence and existence in God.” (Mullins, 2013, p. 184)

According to both Avicenna and Aquinas, there is another aspect of simplicity: God is eternal pure act. Since God is impassible and immutable, God cannot move from potential to actual. His act is identical to himself. This divine activity can be called God’s knowing, willing, loving, and so on (Mann, 1988, p. 194 and Mullins, 2013, p. 189).Footnote 6

Avicenna (Ruffus & McGinnis, 2015, p. 182) and Anselm (Mullins, 2013, p. 185) have also offered another aspect of simplicity: since any distinct concepts imply multiplicity, there cannot be distinct concepts in a simple being’s mind.Footnote 7 Regarding Avicenna’s account of divine simplicity, then, we can add two more principles:

  1. (5)

    God is eternal pure act.

  2. (6)

    God cannot have distinct concepts in his mind.

I shall call this a strong account of divine simplicity (SDS, henceforth) and assume, as the interpretation I am objecting maintains, that weakening any of those principles (such as suggesting either that a simple God can have some extrinsic accidental properties or that he can have distinct concepts) would be unacceptable according to Avicenna’s metaphysics.

If, as Avicenna maintains, SDS is true, God, as being immutable and impassible, cannot come to possess an attribute or feature that he has not had from eternity. This is true of his knowledge, will, power, and the like. Avicenna, unlike the theologians, considers divine self-knowledge, which is identical to God’s essence, rather than divine will primary concerning his creative actions:

“If you make a sincere search, you will find nothing to say except that the representation of the universal order in the [divine] preknowledge together with its necessary and appropriate time is the thing from which that order, with its organization and details, flows in an intelligible manner. All this is providence.” (Avicenna, 2014, p. 147)

So, the cause of all existents is God’s act of knowing himself (Avicenna, 2005, p. 291–292). Divine will, however, is not excluded in this picture: since divine will is identical to divine knowledge (Avicenna, 2005, p. 295) and since causes and their effects coexist, he willed to create the world from all eternity (Ruffus & McGinnis, 2015, p. 176).Footnote 8 We may then formalize Avicenna’s argument for reconciling divine simplicity and creation as follows:

  1. (1)

    “God is immutable and impassible. (Premise).

  2. (2)

    God cannot come to have an attribute or feature that he has not had from all eternity. (From (1).

  3. (3)

    God has will. (Premise).

  4. (4)

    God cannot have some feature or act of will that he has not had from all eternity. (From (1–3).

  5. (5)

    God wills to create this world. (Premise).

  6. (6)

    Cause and effect contemporaneous. (Premise).

Therefore,

  1. (7)

    God must have been willing to create the world from all eternity. (From 4–6)” (My Dissertation, 2023, p. 50).

Regarding the divine creative action, Avicenna endorses the rule of one: “From the One, only one proceeds” (lā yaṣdur ʿan al-wāḥid illā l-wāḥid). This philosophical dictum means that God is the direct cause of only one effect and the indirect cause of everything else” (Amin, 2020, p. 124). Given his view that God is the perfect being and that from the perfect only the perfect can issue, he is the perfect cause of the perfect world (Ormsby, 1984, p. 186). We can then add the following premises to the argument:

  1. (8)

    “God is the One who is the perfect. (Premise)

  2. (9)

    From the perfect only perfect may issue. (Premise)

  3. (10)

    God must have been willing to create the perfect world from all eternity. (From 7–9)” (My Dissertation, 2023, p. 51).

Consider this argument along with his account of divine simplicity. If, as the doctrine of divine simplicity maintains, divine properties are identical to both each other and God, his having the attribute omnibenevolence means his being goodness himself. And since God has neither desires (Griffel, 2009, p. 226), nor choices (Ruffus & McGinnis, 2015, p. 181–182), or purposes and reasons (Avicenna, 2014, p. 145 and 147) for something other than himself, his creative action is his willing his goodness. So, given his endorsement of neo-Platonist emanationism, we reach the conclusion that all existents necessarily emanate from God, the Good because of his action of knowing and willing himself.

As Ruffus and McGinnis have noted, both Mu’tazilites and orthodox theologians such as al-Ghazālī and Fakhr al-Dīn al-Rāzī can object that Avicenna’s deity does not freely create the world. Avicenna’s God seems to be unable to refrain from actualizing this world, to actualize another world, or to refrain from creating anything at all. God cannot not will and he cannot choose not to act (Ormsby, 1984, p. 191). The first conclusion of the above argument, (7), then should be understood as follows: God must have necessarily created the world from all eternity. Avicenna thus adopts what we can call the necessitarian divine freedom (NDF). According to this account, unlike the libertarian account of divine freedom, the power to choose or do otherwise is neither a necessary component of freedom nor of volition in the divine case. As opposed to Avicenna’s NDF, the mutakallimūn (Muslim theologians) argue that freedom should be understood along libertarian lines. Consider the principle of freedom:

(POF) If an agent S is free with respect to a given action, then she is able to do otherwise. She has the power to choose from alternatives without being necessitated externally or internally with respect to that action.

Now consider the following argument.

  1. (P1)

    Given (POF), if God is free with respect to creating, then he is able to refrain from creating anything at all or to create another world instead of this world without being necessitated externally or internally. (Premise)

  2. (P2)

    God, as a perfect being, is free with respect to creating. (Premise)

  3. (P3)

    God is able to refrain from creating anything at all or to create another world instead of this world without being necessitated externally or internally. (From P1 and P2)

  4. (P4)

    But if (7) is true, God is neither able to refrain from creating anything at all nor to create another world instead of this world. (Premise)

Therefore,

  1. (P5)

    (7) is false. (From P3, P4)

If this is the case, Avicenna’s deity can be free only if he is able to refrain from creating anything at all or to create another world rather than this world.

One quick objection here the Avicennian might arise is that (POF) is not true. For, there are at least two conceptions of free will: libertarian and compatibilist. Unlike the former, compatibilist free will, as a Frankfurtian might insist, does not require that an agent must be able to choose or do otherwise for an action to be free (Frankfurt, 1969). If this is the case, Avicenna`s deity could still be free provided that he has a compatibilist free will. It thus appears that (POF) begs the question and does not show that Avicenna’s deity who can have compatibilist free will cannot be free with respect to creating.

This objection, I think, would be plausible only if the Avicennian’s attempt is not to show that both Avicenna and the theologians’ positions are similar in the following respect: they both can accommodate choice regarding God’s creative choices on whether to create. Yet, as noted above, since this is exactly what Tūsī is attempting to do, this objection cannot go so far. The Avicennian, like al-Tūsī, however does not need to endorse this compatibilist version. She might hold what Joseph Keim Campbell calls strong compatibilism. According to Campbell, strong compatibilism contends the following:

  1. 1.

    “Alternative possibilities of action are necessary for both free will and moral responsibility.

  2. 2.

    Both free will and moral responsibility are compatible with causal determinism.” (Campbell, 1997, p. 319)

It seems then that by referring to Frankfurt’s weak compatibilism, the Avicennian cannot show that both Avicenna and the theologians’ positions are similar concerning God’s options on whether to create. For, weak compatibilism denies (1) which is assumed by Tūsī.Footnote 9 He argues that an agent who can do otherwise with respect to an action is a “capable” agent. And if doing and not doing is due to his will, it means that he is acting (or refraining from doing) voluntarily. If it is not due to his will, he is compelled with respect to that action (Tūsī, 1992, p. 25). Yet necessity is compatible with choice. A voluntary agent has two attributes, namely power and will. Power makes both doing an action and refraining from doing that action possible for the agent. When will, however, is added to power, either doing or refraining becomes preponderant. If doing is the preponderant option, then the occurrence of the act is necessary and the occurrence of refraining is impossible (and vice versa if refraining is the preponderant option) (Tūsī, 1992, p. 25–26). If the voluntary agent wills to do something, it is then impossible that he can refrain from doing it. It seems then that if an agent is a voluntary agent, it is necessary that he has power to do an action and refraining from doing that action (i.e. it is necessary that there are alternative possibilities of action). But when he wills to do or refrain from doing it, the preponderant option becomes necessary for the agent. By accepting (1) and (2),Footnote 10 thus, Tūsī might claim that God, as a voluntary agent, had the power to choose creating this world or refraining from creating it but given his eternal knowledge and will, creating this world was the preponderant option from all eternity. He then seems to be able to show that both Avicenna and the theologians hold a choice-based free will for God on whether to create even though they endorse different accounts of the choice-based free will.

The theologians, however, might rightly claim that since strong compatibilism cannot consistently hold both (1) and (2), it cannot be a choice-based view of free will. For example, if it is really open to an agent to choose or do otherwise with respect to an action, she cannot be determined externally or internally to do the given action. What the theologian understands from the ‘could have done otherwise’ condition (i.e. from (1)) is different from Tūsī’s version. If that is right, Tūsī is wrong in arguing that both Avicenna and the theologians hold a choice-based free will for God on whether to create. But suppose that Tūsī is right, and we should hold the strong compatibilist version of ‘could have done otherwise’ condition.Footnote 11 We shall then see whether this poses a threat to divine simplicity.

Ruffus and McGinnis’ willful understanding

Ruffus and McGinnis argue that Tūsī’s interpretation of Avicenna’s account of action can show that God is free concerning whether to create. They first present Avicenna’s account of volition. According to Avicenna, volitional actions are those that do not originate from an external agent but from oneself. The criteria for intellectual volition which is applicable to humans, Intellects (i.e. angels), and God is as follows:

“volition must (1) proceed from the agent’s recognition (ma‘rifa) or understanding (‘ilm) that the agent is the cause of the action, (2) the agent is not impeded in its action and (3) the agent consents (raḍiya, yarḍa, riḍan) to the action.” (Ruffus & McGinnis, 2015, p. 172)

Regarding the divine creative action, they underline the importance of the distinction between intention and volition in Avicenna’s system. Avicenna claims that God, unlike us, cannot act intentionally or act for the sake of something other than himself. So, God’s creating the world is not for the sake of the good of the world (Ruffus & McGinnis, 2015, p. 174). As Ruffus and McGinnis have emphasized, the reason why Avicenna holds this view is that God’s intentional action would jeopardize divine simplicity (Ruffus & McGinnis, 2015, p. 175). But how? According to Avicenna’s metaphysics, simplicity and necessity are co-implied (Amin, 2020, p. 125). God thus as a necessary being must be simple. Now we know that there are both multiplied and composed things in the universe. But then how can a simple God be the creator of the universe? Avicenna thinks that God cannot be the direct cause of such things because then he would no longer be simple and thus necessary. Direct creation of both multiplied and composed things implies a multiplicity in God himself, for it suggests that God might have in himself a knowledge that causes an intention that in turn causes his creative action.Footnote 12 Since causes and effects are distinct, God’s having an intention causing him to act would jeopardize divine simplicity and thus his necessary existence. To reconcile divine simplicity with divine creative action, Avicenna offers his famous principle of emanation: the rule of one. Thus, while God is the direct cause of the First Intellect which is immaterial, it is the First Intellect that is the direct cause of multiplied and composed things.

Avicenna contends that emanation is neither a natural action nor an intentional one; rather, it is a volitional action. As Ruffus and McGinnis have pointed out, most Muslim theologians have seen choice as a necessary component of volition. Avicenna allows for choice in two different senses (Ruffus and McGinnis call those nuances of choice “choiceo” (with respect to options) and “choiceg” (with respect to the good)). Nonetheless, he denies that choice is a component of the volition involved in God’s creative action.Footnote 13

One might still claim that if Avicenna is right, God creates as a matter of natural necessity; for he creates in the sole way that his nature requires. Al-Ghazali, for instance, claimed that the God that has been creating from all eternity cannot be a free creator of the world in three respects: “with respect to the agent, with respect to the act, and with respect to a relationship common to act and agent” (Al-Ghazali, 2000, p. 55). For now, let’s put the last two respects aside in order to avoid redundancy and focus on the first one. Avicenna’s God, al-Ghazali claims, cannot be the agent of what he wills. He writes: “‘Agent’ is an expression [referring] to one from whom the act proceeds, together with the will to act by way of choice and the knowledge of what is willed” (Al-Ghazali, 2000, p. 56) and adds that the agent’s existence, knowledge, power, and will are a condition for her being an agent (Al-Ghazali, 2000, p. 62). Al-Ghazali argues that Avicenna’s God can be neither a willer, nor a chooser, nor a knower of what he wills (Al-Ghazali, 2000, p. 55). He cannot be a willer because the action (the effect) proceeds from him (the cause) as a necessary consequence but not as an intentional action. He could not have refrained from creating the world. He neither can be a chooser, for if an agent does not have will and choice with respect to an action, by simply being its cause, she cannot be the agent of that action. Action can only be attributed in a real sense to an agent if that agent brings it about through her will. And if God is not the willer of an action, he cannot be the knower of what he wills either. Since action necessarily entails will and since will necessarily entails knowledge, Avicenna’s God who has neither will nor choice, al-Ghazali argues, can be neither the agent nor the maker of the world in the real sense (Al-Ghazali, 2000, p. 56–59).

Ruffus and McGinnis argue that Naṣīr al-Dīn Ṭūsī’s interpretation might help to show that unlike what al-Ghazali or other theologians assume, Avicenna’s deity does not create of necessity. According to his interpretation of Avicenna’s account of divine simplicity and creative action, though it is true that God, as an eternal agent, cannot produce a temporal effect but only an eternal effect, “it is, however, false that such an agent cannot choose absolutely” (Ruffus & McGinnis, 2015, 183). For it is still open to him to choose either to create eternally or to refrain from creating anything at all. If we accept Ṭūsī’s interpretation, they argue, divine simplicity is safeguarded:

“for creating or not creating do not correspond to two distinct concepts in the divine mind. Not creating simply is the negation or absence of the idea of creating. In other words, the options for the divinity are not between distinct types of action, which would require distinct concepts, as, for example, choosing between either reading a book or going to a movie. Instead the only thing that Necessary Existing knows, and to which It consents in either the case of creating or not creating, is the essential goodness that is the very divine being itself.” (Ruffus & McGinnis, 2015, p. 184)

Thus, Ṭūsī’s interpretation might help an Avicennian to show that (P4) is false because even if (8) is true, God had options for his creative actions, namely creating the world and not creating anything at all. According to Ruffus and McGinnis, Ṭūsī thus shows that Avicenna’s account of God’s having options for his creative action (namely to create eternally or not to create anything at all) is no different from that of the mutakallimūn. As noted in the previous section, I don’t think that it is no different from that of the theologians. But let’s put that aside and see whether God’s omnirationality is a threat to Ṭūsī’s interpretation and thus to divine simplicity and freedom on Avicenna’s account.

God’s omnirationality

God, in the classical theistic traditions, is considered omnibenevolent, omniscient, omnipotent, and perfectly free. Alexander Pruss has pointed out that omnirationality is an underexplored divine attribute. I agree and will now argue that God’s omnirationality is an obstacle to the interpretation of Tūsī that Ruffus and McGinnis draw upon in their attempt to reconcile Avicenna’s accounts of divine simplicity and freedom. Pruss defines omnirationality as follows:

“God is omnirational: whenever he does anything, he does it for all and only the unexcluded reasons that favor the action, and he always acts for reasons.” (Pruss, 2013, p. 1)Footnote 14

As he has pointed out, we should note that those unexcluded reasons are good reasons that have rational force.Footnote 15 Omnirational God, thus, cannot act for reasons that are not rationally eligible.Footnote 16 Suppose that A, B, and C are three alternatives and X, Y, and Z are three considerations. If A is a better alternative than B and C with respect to X, and if B is a better alternative than A and C with respect to Y, and if C is a better alternative than A and B with respect to Z, then A, B, and C are all rationally eligible. But suppose that A, B, and C are three alternatives, and they can be comparable only with respect to consideration X that favors A instead of B and C. Further, there are no good reasons or incommensurable values that favor B and C. In that case, a rational agent is supposed to choose A, for rational agents are supposed to do the best they can do (Levi, 2004, p. 159).Footnote 17

Now suppose that A stands for creating the world from all eternity, B stands for not creating anything at all, and C stands for creating another world instead of our world. One might claim that if A is a better alternative than B and C with respect to God’s omnibenevolence and omnirationality, God can choose neither B nor C. Considering all and only unexcluded reasons that favor either A, B, or C, the good reasons that favor B or C might be overwhelmed by the reasons that favor A. This does not imply that God is not free in his decision; it shows only that neither B nor C is a genuine alternative that can motivate an omnibenevolent and omnirational being. Alternatively, it might be the case that A, B, and C are all rationally eligible, all optimal, and have incommensurable values. In such a case, if there are some good reasons that favor B or C, God would be justified in choosing either B or C instead of A (even if there are similar reasons that favor A over B or C). For instance, Alexander Pruss has argued that there might be some positive reasons that uniquely favor B (no creation) such as maximal simplicity, uniformly maximal excellence, and all beings achieving the ultimate telos of the universe (Pruss, 2016).Footnote 18

Now, if Ṭūsī’s interpretation is correct, then it seems that there were only two options that were open to God, namely A and B. Consider the following argument:

  1. (T1)

    If there are at least two options that are open to God with respect to his creative action, then he is free in his action.

  2. (T2)

    Only A and B were the options open to God with respect to his creative action.

  3. (C1)

    God is free in his action.

  4. (T3)

    If A and B do not correspond to two distinct concepts, God’s simplicity is safeguarded.

  5. (T4)

    A and B do not correspond to two distinct concepts.

  6. (C2)

    God’s simplicity is safeguarded.

We can thus conclude that God’s simplicity is compatible with his freedom.

I think (T1) is plausible, for it underlines the importance of the principle of alternative possibilities. But I think that both (T2) and (T4) have flaws. Let’s first consider (T2). Though it seems to be compatible with (T1), one might claim that it is hard to see why C was not open to God as an option, especially considering the possibility of God’s having distinct reasons for creating another world (or different possible worlds). Ṭūsī’s rejection of the possibility of C is incompatible with our intuitions that God could create different worlds. This appears to be a problem for his interpretation, for if C is an option open to God, then since C is not simply the negation of A (unlike the supposed relation between A and B), the combination of A, B, and C correspond to distinct concepts in the divine mind. God’s simplicity, thus, cannot be safeguarded according to Avicenna’s account. Further, if C is rationally eligible or favorable with respect to some values, then there must be an explanation for why an omnirational and omnibenevolent God would refrain from willing C. Since God always acts and refrains from acting for rational or moral reasons, God’s not willing C implies that there are some reasons that make C either rationally ineligible for God or, unlike A and B, unfavorable.

Ṭūsī however has a reason to reject the possibility of option C. His interpretation assumes both Avicenna’s strong theory of divine simplicity and account of emanation. These two theories together contend that it is impossible to be open to God to create a multiplicity of distinct universes.Footnote 19 If A and C stand for distinct possible worlds, and if we suppose that God creates A but not C (in other words, his knowledge causes A but not C), this means that God knows those distinct accounts/concepts. But this is a threat to divine simplicity because it assumes that there is a multiplicity in divine mind. And regarding the emanation account, as noted above, the rule of one holds that from the One, only one proceeds. Therefore, unless there are two distinct divine selves from one of which A emanates and from the other C does not emanate, multiple worlds seem impossible.

One might still see some flaws with Ṭūsī’s reason to reject the possibility of C. Notice that though he holds strong compatibilist view of free will on whether to create in order to secure divine freedom, he does not consider the same account sufficient enough to secure divine freedom on what to create. He seems to suggest that we should hold a weak compatibilist view of free will on what to create, the view that does not require alternatives for an action to be free. This might be a problem for a theist who would want to have a unified account of free will in the divine case. It might also seem surprising: one can expect that given Avicenna’s emanationist view, a strong compatibilist view of free will (which is a choice-based free will) is supposed to be held on what to create but not on whether to create. For, according to Avicenna, God creates because he knows and wills himself. Since creative impetus is a necessary component of divine knowledge, will, and goodness which are all identical to God, God inevitably creates something or other. So, if a choice-based free will is to be attributed to God at all, it appears that it can be held only on what to create but not on whether to create. Further, referring to the abovementioned quote from McGinnis and Ruffus,Footnote 20 one might object that if God wills himself either in the case of creating or not creating and if the only thing that Necessary Existing knows is his being itself, this should also be true on what to create (i.e. in the case of actualizing C).

For an Avicennian, like Tūsī, however, these are not primary concerns. When describing God, Ṭūsī considers simplicity as God’s primary attribute, but not others such as omnipotence, omnirationality, or freedom. In order to secure divine simplicity, he does not hesitate to give up on holding a unified account of free will in divine case or to give up God’s omnirationality regarding the explanation why C is not rationally eligible provided that God wills only himself either on A, B, or C. It seems then that if Ṭūsī shows that God has a choice-based free will on whether to create (though not on what to create) and it is compatible with divine simplicity, he will be proving that Avicenna, like the theologians, attributes to God a free element in creation. If he fails, however, in his attempt, it will follow that first, Avicenna’s God could not refrain from creating anything at all (i.e. he did not have two options, so (T2) is false) and second, Ṭūsī’s reason to reject holding a unified account of free will in the divine case and thus denying the possibility of C in order to secure divine simplicity is ungrounded. This then brings us to the question whether his interpretation shows that divine simplicity is safeguarded even if we assume that (T2) is true (i.e. God has two options namely not creating anything at all and creating).

We shall then now evaluate whether (T4) is true. Ruffus and McGinnis argue that Ṭūsī’s interpretation suggests that Avicenna’s position on A and B is no different from that of the mutakallimūn. But I think this is mistaken. It is hard to see how A and B can be considered two options if they are not related to some distinct reasons that can correspond to distinct concepts in divine mind. When the mutakallimūn, like al-Ghazali, argue for the idea that it was open to God either to create our world or to remain within himself by not creating anything at all, they mean (and as omnirationality requires) that God could have acted for distinct positive (or negative) reasons with respect to the alternatives he had, and thus he would be justified in choosing either of them. This commits them to rejecting (T4). Now consider the following:

  1. (O1)

    If A and B are two options God has, then there are distinct reasons that favor each of them.

  2. (O2)

    A and B are two options God has.

Therefore,

  1. (O3)

    There are distinct reasons that favor each of them.

  2. (O4)

    But distinct reasons correspond to distinct concepts.

  3. (O5)

    If God has two options, namely A and B, then God has distinct concepts in his mind. (from O1 to O4)

  4. (O6)

    Given the falsity of (T4) (i.e. A and B don’t correspond to distinct concepts), the argument for the compatibility of divine simplicity and freedom is not sound.

Thus, I think the mutakallimūn have not seen those options/alternatives as mere logical possibilities but as options providing reasons that God can act for. If God is omnirational, then his willing to eternally create our world, or not to create, or to create any other world requires his having unexcluded reasons to do so in each case. Now, if not to create anything at all is an option for Avicenna’s God, then the reasons for his willing to remain within himself would be different from those that he would have for willing to create. The mutakallimūn, thus, would not see B as a mere negation of A as merely declining to create our world. If this is the case, then their worry how Avicenna’s God’s creative action appears to be a matter of natural necessity remains.

The Avicennian might object that both (O3) and (O4) are problematic. (O3) seems to be the case only if the principle of omnirationality is true. McGinnis, for instance, states his worry on omnirationality as follows:

“Omnirationality is also potentially problematic, particularly if it requires a reason distinct from either the agent or even the agent’s actions. In that case, omnirationality would be incompatible with divine simplicity, and so to assume it in an argument against divine simplicity would beg the question. The fact is that much of the dispute between Avicenna and his defenders and the mutakallimūn focuses on which divine attribute one takes to be primary when describing God: simplicity (Avicenna) or omnipotence (the theologians). Also I would note that omnirationality might be incompatible with a strong form of omnipotence. If one believes that an omnipotent God can do anything even something without a reason or even opposed to reason (and there were theologians who certainly maintained as much), then, omnirationality limits what God can do, since he is limited to only action favored by reasons.” (McGinnis, 2022, Email to the author)

He further notes that reason can be either external or internal to the agent. If the former like the good of another, it must be clear that there cannot be an external cause that can cause God, the Cause of causes. If the later like acting for its own good, the agent is either simple or composite. If simple, we get Avicenna’s position. If composite, we get the theologians’ position. The whole discussion then will come down to whether the doctrine of divine simplicity is true but not to whether omnirationality is a threat to God’s simplicity.

And regarding (O4), he offers a counterexample by referring to the notion of self-reference assumed by the doctrine of divine simplicity:

“If in a state of self-reflection someone asks me, “What are you thinking about?” and I say, “Myself” and then they subsequently ask me what is the reason for my thinking of myself, a legitimate answer to the question, “Why is myself an object of my thought” is because I was thinking of myself. Now, they can go on and say, “No, why did you go from a state of not thinking of yourself to thinking of yourself.” In that case, I would presumably have some other answer, but God’s immutability precludes asking the analogues of this latter question of him. All of that to say that self-reference might provide a counterexample to O4, and self-reference goes right to the heart of Avicenna’s doctrine of simplicity.” (McGinnis, 2022, Email to the author)

If McGinnis is right, the Avicennian might provide a counterexample to (O4) and thus show that it might be false.

Let’s begin with his objection to (O3). McGinnis, I think, is right that omnirationality is incompatible with a strong form of omnipotence which assumes the possibility of divine actions not favored by any reason.Footnote 21 But since my aim is not to defend a strong form of omnipotence but to show that omnirationality is incompatible with divine simplicity, the problem that the theologians who maintain a strong form of omnipotence might encounter should not be worrisome. I argue that omnirationality is incompatible with divine simplicity even if it does not require a reason distinct from either the agent or even the agent’s actions. Suppose, as Ṭūsī’s interpretation assumes, that Avicenna’s God has two options (i.e. A and B) concerning whether to create and also suppose that omnirationality requires that God causes an intention/action for a reason that is not distinct from himself. In option A, then, he actualizes A for an internal reason such as knowing or willing his own goodness which is identical to himself. In option B, God again actualizes B for the same internal reason. In either option, God wills his own goodness, wills himself. It seems then that omnirationality does not particularly require that divine reason must be distinct from divine essence.

Notice however that omnirationality requires that there must be a reason or an aspect of a reason that can explain why agent chooses an option rather than another. On option A, the reason the agent has (i.e. knowing/willing himself) is twofold: God actualizes A rather than B because he wills himself and his essence is necessarily diffusive. Though the first conjunct does not necessarily make B impossible for God, the second conjunct does provided that the Dionysian principle (assumed by Avicenna) is true. The divine reason on this option then is sufficient to explain why God actualizes A rather than B. If option B is really open to God, there must also be a sufficient reason that can explain why God actualizes B rather than A. On option B, the reason the agent has (i.e. knowing/willing himself) might be sufficient to explain why God actualizes B (because he would still be knowing and willing himself) but does not explain why he refrains from actualizing A by acting against his own nature. Unlike on option A, there must be an additional reason on option B that can explain God’s refraining from actualizing A. Since the needed additional reason cannot be the agent himself on option B provided that the agent is self-diffusive, that reason will require multiplicity and the existence of incompatible reasons (for the needed reason will be incompatible with the reason for self-diffusion) in divine mind. This will, however, threaten divine simplicity. Thus, if I am right that there must be an additional reason on option B that God cannot have on option A, (O3) is true.

McGinnis also finds (O4) implausible because he argues that self-reference might provide a counterexample to it. He is, I think, right that God’s immutability precludes asking why God goes from a state of not thinking of himself to thinking of himself. For God necessarily and eternally knows himself. But that’s not where the problem lies. What (O4) along with (O3) shows is that though the notion of self-reference is sufficient to explain how Avicenna’s simple God can actualize A, it is not sufficient to explain the possibility of option B for his God. For, if option B is really open to God, it means that God’s power could have stayed unmanifested. This should not be a problem for a theist like al-Ghazali who defends that God could have refrained from creating anything at all but it is a problem for a divine simplicity theorist. For, since God’s power is identical to his will, his knowledge, and ultimately his substance, the real possibility of option B entails that his will, knowledge, and substance could have stayed unmanifested with respect to external manifestation. But this is entirely incompatible with Avicenna’s account of divine simplicity and emanation. Notice that even if the notion of self-reference might explain how a simple God can remain within himself, it cannot explain how a simple God who is necessarily diffusive can remain within himself. The latter must have a reason indistinct from himself to refrain from creating anything at all but it should be clear that it is impossible for him to have such reason.

(O4) along with (O3) also shows that God cannot have only a single concept provided that he has options A and B. Since B is not a mere negation of A, it is not an empty concept either. For instance, had God actualized B by refraining from creating anything at all, he would have actualized the bare world in which only God, numbers, and other uncreated things exist but not angels, human beings, and the like. As noted above, the notion of self-reference would not be sufficient to explain why Avicenna’s God had actualized the bare world (i.e. B) rather than A. The additional reason that is distinct from God and thus corresponding to a distinct concept is unavoidable in option B.

I think a possible way to escape the objection I raise might be showing that though God has distinct concepts in his mind, this cannot be a threat to his simplicity. For instance, Michelle Panchuk attempts to reconcile divine simplicity with God’s having the multiplicity of ideas and concepts in his mind. She argues that if divine knowledge is a non-propositional knowledge and if “(1) It is distinct from, but continuous with, the propositional knowledge that humans have; (2) It is richer than propositional knowledge; (3) It is more direct than propositional knowledge” (Panchuk, 2021, p. 399), then divine simplicity is safeguarded even if there is multiplicity of concepts in the divine mind. I do not think that divine knowledge can be considered only a non-propositional knowledge but Panchuk’s view, I believe, might help an Avicennian to escape the objection only if the Avicennian is willing to accept that A and B correspond to different concepts. I doubt that Avicenna himself would accept Panchuk’s view (because it would weaken his SDS) but even if the Avicennian is willing to hold that there can be multiplicity of concepts in the divine mind, another problem arises. If A and B correspond to different concepts and if, as Avicenna maintains, God’s knowledge of himself is the cause of the existence of things (i.e. God’s knowing, unlike ours, is active), it seems that it was impossible for God’s knowledge of B to be active. For it should be clear that God’s knowledge of B (or his knowledge in B) is distinct from his knowledge of A. In B, for instance, God is supposed to know the following propositions: (a) he exists only along with necessary abstract objects, numbers, and other uncreated things, (b) angels, human beings, and the like don’t exist, (c) he could have actualized A but has decided to refrain from actualizing it. Unlike B, in A, God knows that (a), (b), and (c) are false. But since God’s knowing a thing entails his causing that thing according to Avicenna’s metaphysics, God’s knowledge of B must entail his causing B. Given however (as Avicenna maintains) that God causes A eternally and does not cause B, it must follow that his knowledge of B is impossible. If that is right, B cannot be an open option to God. Further, notice that it is hard to reconcile his account of knowledge with the strong account of compatibilism which we have assumed for Ṭūsī’s interpretation. As Campbell notes, the strong account of compatibilism maintains that the ‘could have done otherwise’ condition assumes two constraints on possible worlds that are accessible to an agent:

“A1. Accessible worlds must have the same laws of nature as the actual world.

A2. In accessible worlds, agents cannot have any abilities or capacities that they lack in the actual world.” (Campbell, 1997, p. 324)

In the divine case, A1 is not necessary. But if God, as theism rightly maintains, is essentially and necessarily perfect in every world, A2 must be true in the divine case. This, however, implies that God’s knowledge in A and his knowledge in B differ with respect to abilities or capacities provided that Avicenna’s account of knowledge is true. For it is the case that God’s eternal knowledge is the same in those possible worlds because he knows himself and all other existents through himself. If it is the same, God’s knowledge in A is necessarily active while it is not in B (provided that B is really an option). So, they differ with respect to abilities or capacities. But if Campbell’s principle A2 is true, either A or B is not accessible to God. Since the Avicennian cannot accept the former, she will have to admit that B is not accessible and thus not an open option to God. If I am right, the Avicennian then seems to need to weaken not only Avicenna’s strong account of divine simplicity but also his account of divine knowledge if she wants to hold that option B was open to God.Footnote 22

Another alternative for the Avicennian might be endorsing the view that concepts are neither mental representations nor abstract objects (i.e. senses) but only abilities. According to this view, “.concepts are abilities that are peculiar to cognitive agents....The concept CAT, for example, might amount to the ability to discriminate cats from non-cats and to draw certain inferences about cats.” (Margolis & Laurence, 2021). The concept about option A for God, then, might amount to the ability to discriminate A from B and to draw certain inferences about A. This view might help the Avicennian to suggest that God had a single concept regarding the options A and B: the ability to discriminate A from B (i.e. the ability to discriminate creation of this world from not creating anything at all). If the content of B is empty and if the theory that the content of a concept is determined by its relation to other concepts is false with respect to the options A and B, then God had a single concept/ability but not two. In such a case, the Avicennian might hold a modified conceptual atomism and suggest that the content of A is determined by its relation to God but not to B. And since the content of B is empty, it is not true in the divine case that there was a concept about option B that amount to the ability to discriminate B from A. So, though there were two different options namely A and B for God, they did not correspond to distinct concepts in his mind. I think, however, this is still problematic for the Avicennian. If the concept about option B were an empty concept (i.e. the mere negation of A), then it is true that B would not correspond to a distinct concept. But as I have pointed out, the content of concept B is not empty. If B is really an option and if God is wise and omnirational, then there must be an explanation how Avicenna’s God could choose B instead of A. The possible reasons that would explain God’s choosing B instead of A imply that the content of B cannot be empty.

Another alternative might be, as Sandra Menssen and Thomas Sullivan contend, to claim that the Dionysian principle that goodness is necessarily self-diffusive does not require that God, who is identical to goodness, must create something or other (Menssen & Sullivan, 1995). I find their interpretation of the principle implausible but if they are right, my objection that even if self-reference explains how a simple God can remain within himself, it cannot explain how a simple God who is necessarily diffusive can remain within himself might collapse. Assume that the Dionysian principle does not require any external divine manifestation. We can then take the principle as only requiring divine internal manifestation. And we can understand this manifestation either as nothing but God’s self-recognition of his own goodness or as intellectual emanation. While the former only implies that God is naturally and necessarily aware of his own goodness, the latter requires a necessary internal emanation. It must be clear that the former understanding of divine manifestation cannot satisfy those who hold the Dionysian principle. It should come with the latter. For example, Aquinas holds the latter as follows:

“On the classical Christian view, God the Father is always emanating an internal Word that is “light from light, true God from true God.” This means that the Good is already manifested in the most important and perfect way (albeit a way much different than contingent material reality). This can accommodate the insight that goodness is diffusive by internal necessary emanations, without requiring or involving anything contingent or external.” (Cohoe, 2022, p. 271)

If Aquinas is right, the Dionysian principle does not require that God must create something or other. The principle along with divine simplicity then does not make the counterexample based on self-reference to (O4) implausible.

There are, I think, two problems with this understanding of the Dionysian principle for the Avicennian. First, it requires an endorsement of the Christian doctrine of the Trinity. Since Avicenna, as a Muslim philosopher and a proponent of the strong account of divine simplicity, denies the doctrine of the Trinity, he has both theological and metaphysical reasons to reject the Thomistic understanding of the principle. Second, there is also a metaphysical reason for him to reject it because as Kretzmann has pointed out, it is quite dubious whether this is a correct understanding of the principle (Kretzmann, 1991a, 1991b). Given Avicenna's understanding of divine knowledge and the Dionysian principle, creative impetus must be a necessary component of divine goodness and thus of its manifestation.Footnote 23 Since the insight noted in the above quote does not capture the true meaning of the Dionysian principle according to Avicenna’s metaphysics, it is impossible for option B to be open to God.

Conclusion

I have argued that Ruffus and McGinnis have failed to show that Avicenna’s simple God can be considered free in his creative action as regards whether to create. For the interpretation they offer establishes neither that not creating is simply the negation of creating for an omnirational God nor that God did not have an option to create another world instead of this world. If (as seems plausible) either or both of these claims is in fact false, then divine simplicity as Avicenna understands this doctrine is incompatible with any choice-based account of free will.