Is God a Substance? Avicenna on Essence, Being, and the Categories

Abstract

Avicenna scholars unanimously agree that Avicenna takes the position that God is not classifiable according to the Aristotelian scheme of the ten categories. However, Avicenna scholars are in little agreement about precisely why God evades categorial classification. Scholars report numerous and, at times, mutually inconsistent arguments purportedly made by Avicenna. In this study, I argue that Avicenna has only one argument as to why God is not in the category of substance—the Essence-Being Distinction Argument—and that he makes this argument consistently throughout his major philosophical encyclopediae. Having clarified this argument, two consequences follow. First, we can study this argument to learn not only about God in Avicenna’s philosophical theology but also about the nature and structure of the categories in Avicenna’s metaphysics. I argue that Avicenna’s Essence-Being Distinction Argument reveals a route by which one may arrive at a real distinction between essence and being from a philosophical, rather than a theological premise. Second, Avicenna’s Essence-Being Distinction Argument, along with the prevalence and consistency thereof, suggests that God has an essence for Avicenna and that texts, wherein Avicenna denies that God has an essence, are exceptional and should not govern our broader interpretation of Avicenna’s philosophical theology.

1. Introduction: We All Know That God Is Not a Substance, but Why Not?

Even before Greek philosophy found its way into the Arabic language in the eighth century, it was pretty well agreed upon by many that the divine is something outside of the classificatory scheme of Aristotle’s ten categories. Notwithstanding the significant view of Aristotle, whose primary mover belongs in the category of substance (Aristotle 1924, 1072a24-26; see Houser 2010, pp. 120–21)1, the consensus that the divine is beyond the categories of being (comprising substances and accidents), has its start in the Republic with Plato’s famous “Good beyond being” (ouk ousias ontos tou agathou, all’eti epekeina tēs ousias…) (Plato 1903, 509b) and carries through the Neoplatonic tradition, both in Plotinus’s claim that Aristotle’s Categories does not consider divine, intelligible being (Plotinus 2018, p. 649)2 and Ammonius’s (1888, 1991) argument that the definition of substance does not apply to the divine because the divine transcends the categories. Ammonius says that some are confused about whether the definition given to (primary) substance, being “neither said of some subject nor in a subject” (oute kath’ hupokeimenou tinos legetai oute en hupkeimenō), can be said of the divine. Ammonius vehemently denies it. “It is entirely unthinkable to have this definition in mind for the divine substance; for, ‘not being in a subject’3 is opposed to ‘being in a subject’ and [so] it fits those things that have a relation to those things that are in a subject. However, the divine is entirely unrelated to and transcends everything, both those that are in a subject and those that are not in a subject but are a subject for accidents” (37, 4–10)4. The divine, according to Ammonius, cannot be called a substance because the definition “being not in a subject” does not apply to it, and “being not in a subject” does not apply to it because “being not in a subject” implies “as opposed to being in a subject”. This is a problem because the distinction between “being in a subject” and “not being in a subject” is a distinction that divides being into the categories of accidents and substance. If God, therefore, is not in a subject, then God is in the category of substance along with you, me, cats, dogs, and even the hideous blobfish. However, as Ammonius argues, God cannot be just one substance among many other substances—the divine transcends the categories.

Fast forward four-hundred years to the time of the great Persian philosopher Ibn Sīnā (Latinized as “Avicenna”), and we find that this tradition beginning in Plato extends into the Islamicate world in the tenth century when Avicenna argues that God is not a substance in all of his major philosophical encyclopediae. This is not a surprise. That God transcends the categories and, so, is not a substance is a view found in Avicenna’s predecessor al-Fārābī (1969, pp. 104–5; see Druart 1987, p. 93).5 Further, this is a position that is well documented in the secondary literature on Avicenna. Muhammed Legenhausen dedicated a whole study to the topic (Legenhausen 2007). If you turn to Legenhausen’s study to discover why God is not a substance (as well you should), you will find that God is not a substance because “it is the whatness [i.e., the “quiddity” or “essence”] that determines whether something is a substance or not, and because WW [wajib al-wujūd, i.e., the Necessary Being, God] has no whatness, it cannot be considered that Zayd and God are two instances of the more general concept of substance” (126). Later on, in the same article, Legenhausen summarizes his findings, saying that “…for substances, existence and whatness are distinct. For WW [i.e., God], however, there is no whatness at all, unless its existence is taken to be its whatness, and hence, WW is not a substance” (128). Further on, “Since WW does not have any such whatness, it cannot be considered a substance” (130). Legenhausen has given us at least two different reasons for thinking that God is not a substance.

1:

The “No Essence Argument”

(1)

If x is a substance, then x has an essence by virtue of which x is a substance.

(2)

God does not have an essence.

(3)

Therefore, God is not a substance.

2:

The “Essence-Being Distinction Argument”

(1)

If x is a substance, then x’s essence is distinct from x’s being.

(2)

God’s essence is not distinct from His being.

(3)

Therefore, God is not a substance.

Now, perhaps Avicenna does give us two different arguments for why God is not a substance. There is nothing wrong with having many arguments to prove the same conclusion. However, there is something wrong if the premises of those different arguments contradict each other, as is the case with the No Essence Argument and the Essence-Being Distinction Argument. For if the Essence-Being Distinction Argument is sound, then God does have an essence and His essence is being. It would follow, then that the second premise of the No Essence Argument must be false. Likewise, if one maintains that the No Essence Argument is sound, then God does not have an essence, and so the second premise of the Essence-Being Distinction argument would be false. So, Avicenna cannot use both of these arguments to support the conclusion that God is not a substance because their respective second premises are mutually contradictory.

The story of how Avicenna argues that God is not in the category of substance does not stop there, however. Legenhausen also gives us another argument for why God is not a substance. Explaining that, in the famously treacherous middle books of Aristotle’s Metaphysics, Aristotle considers whether substance is form, matter, or the combination thereof, Legenhausen concludes that “[e]xistence, however, is not form or matter or any combination thereof. Hence, existence is not a substance. However, that whose existence is necessary, WW, is existence itself, and hence, according to Avicenna, existence itself is God. Therefore, God is not a substance” (126). So, let us add a third argument to our list:

3:

The Aristotelian Disjunctive Proof

(1)

If x is a substance, then x is either form, matter, or a combination thereof.

(2)

Being is not a form.

(3)

Being is not matter.

(4)

Being is not a combination of form and matter.

(5)

Therefore, being is not a substance.

(6)

God is being.

(7)

Therefore, God is not a substance.

So, in one article on the topic of whether God is a substance in Avicenna, we find three different reasons why God is not a substance—two of which are mutually inconsistent. Turning to other Avicenna scholarship enforces this sense that Avicenna offers many arguments for God’s not being a substance: the No Essence Argument is reported in Bertolacci (2006, p. 455; 2022, p. 258),6 Adamson (2013, p. 173),7 Houser (2011, pp. 366–67),8 and Morewedge (1972–1973, p. 54).9 The Essence-Being Distinction Argument is reported in Acar (2005, p. 105)10 and Shehadi (1982, p. 85).11 Indeed, we can find a fourth argument in an article by John Rosheger (Rosheger 2002, p. 235):12

4:

The “Being is not a Genus Argument”

(1)

If x is a substance, then x has an essence that is contained in a genus.

(2)

God’s essence, being, cannot be contained in a genus.

(3)

Therefore, God is not a substance.

In all, in the secondary literature, we find a lot of reasons to think that God is not a substance: either because he has no essence, or because his essence is being, and so either being is not a genus, being is not a substance, or anything in a category has a distinction between essence and being. Moreover, we find the answer to the question, “is God a substance?”, tied up with one of the thorniest issues in Avicenna scholarship: the question of whether God has an essence (see Gilson 1960; Flynn 1974; Judy 1976; Macierowski 1988; Wisnovsky 2013, p. 202).

I do not intend to solve the issue of whether Avicenna’s God has an essence here. Indeed, I doubt it could be solved in a mere article. Rather, I want to focus on the question of whether God is a substance and how Avicenna goes about proving this. I argue that Avicenna does not offer us four different proofs that God is not a substance. Rather, Avicenna consistently offers one argument: what I have called the Essence-Being Distinction Argument. That Avicenna gives only one argument is significant for two reasons: first, it indicates to us just how important Avicenna’s doctrine concerning the distinction between essence and being is to understanding his view of the categories, and second, between the mutually contradictory No Essence Argument and the Essence-Being Distinction Argument, that Avicenna favors the latter over the former gives us some indication that it is more correct to say that, for Avicenna, God does have an essence and His essence is being. In the next section, I shall present the major arguments for God’s not being a substance and show that Avicenna consistently presents one argument: the Essence-Being Distinction Argument. In the third section, I shall offer some suggestions as to why the scholarship has been confused about the character of Avicenna’s argument. Fourth and finally, I shall discuss the implications of my argument, namely, that Avicenna’s argument against God’s being a substance reveals a lot about how Avicenna thinks about the categories and the distinction between essence and being.

2. Avicenna’s Argument That God Is Not a Substance: Three Texts

There are three major texts that I shall consider in this study: Avicenna’s arguments in the metaphysical parts of the Shifā’, the Ishārāt, and the Dānesh Nāmeh. I will begin with the Ishārāt and the Dānesh Nāmeh texts and then move to the Shifā’, despite the fact that the Shifā’ text is the best known of Avicenna’s arguments for God’s not being a substance. As I shall argue here, I think that the Shifā’ text is the source of the confusion in the secondary literature regarding just what Avicenna’s argument for God’s not being a substance is. Though it is the most famous argument, it is also the most crabbed version of the argument, lacking an express conclusion and relying on other texts to complete the argument, as we shall see below. So, for methodological purposes, I shall let the doctrine found in the Ishārāt and the Dānesh Nāmeh lead the way—for theirs is the doctrine that prevails. For the moment, let us not be concerned with correctly mapping out and explaining the argument. Instead, in this section, my only goal is to show that these three texts all present the same, singular argument by showing that they all address the issue of whether God is a substance by discussing the same themes.

With those preliminaries established, let us begin with Avicenna’s Ishārāt. Avicenna writes:

[T1 Avicenna, al-Ishārāt w-l-tanbīhāt: Ilāhīyāt ch. 25 (Avicenna 1983)]

It is often thought that the maʿna13 of “a being not in a subject” is common to the First and to everything else [with] the generality of a genus such that [the First] falls under the genus of substance. But this is an error. For “a being not in a subject”—which is like a description for substance—does not mean “a being in act (mawjūd b-l-fʿil) being (wujūdan) not in a subject”, such that from the fact that one knows that Zayd is a substance in himself, one knows from this that he is a being in act absolutely, let alone the manner (kayfiyya) of this being. But, the meaning that is predicated of substance like a description and in which particular substances participate in potency (just as one participates in a genus) is that it is an essence and reality14 (ḥaqīqa) whose being [wujūdhā] is only [ever] not in a subject. And this predication is [said] of Zayd and ʿAmr by virtue of themselves, not by virtue of a cause. But concerning being “a being in act”, which is a part of being “a being in act not in a subject”, this may belong to it from a cause, so how is it composed from it and from an additional maʿna? So, it is not fitting to predicate of the Necessary Being at all that which can be predicated of Zayd like a genus because He does not possess an essence to which this rule (ḥukm) attaches; but rather necessary being belongs to him just as essence belongs to the rest. And it is known that, since “a being not in a subject” is not said of the well-known categories like a genus, it does not become a genus for a thing by the addition of a negative meaning. For since “a being” is not among the constitutive elements (muqawwimat) of the essence, but rather is among its concomitants (luwāzimhā), it does not become, since it is not in a subject, a part of the constitution [of the essence], for then it would become a constitutive element, save for becoming a genus for accidents which are beings in a subject by the addition of the maʿna of a relation to it.

In this text, Avicenna addresses a problem: it seems true to say that both God and created substances are beings that do not need to inhere in a subject in order to be. Avicenna explains that “a being not in a subject” is like a description—something short of a definition15, what I shall call a “quasi-definition” (see Gilson 1974; Aavani 2009)—of substance and that it seems entirely reasonable to say that God is not in a subject. After all, if one denies that God is not in a subject, it is tantamount to saying that God is a being in a subject like an accident, and that is absurd. So, why not say that God is a being not in a subject? Furthermore, if “a being not in a subject” is a description of a substance, then why not say that God is a substance?

The goal of the text is, of course, to show why one cannot say that God is a substance, and Avicenna accomplishes this by way of a consideration of the quasi-definition of substance. First, and perhaps a bit perplexingly, Avicenna explains that when we use the phrase “a being not in a subject” as a quasi-definition of substance, we must be cognizant of the fact that “a being” is, as all good Aristotelians know, ambiguous.16 For “a being” could signify that something actually exists here and now. However, if “a being” is taken in this sense, then the quasi-definition of substance would be “an actual being existing not in a subject”. Avicenna says this is not how we should understand the “a being” in the phrase “a being not in a subject”. Otherwise, by knowing that some x is a substance, one could infer that x is a being in act—i.e., that x actually exists here and now. As Avicenna makes clear, this is not an inference one can make: from the knowledge that Zayd is a substance, one knows nothing about whether Zayd exists at all or how he exists. This is because Zayd’s being in act is not something that pertains to him by virtue of himself but rather by virtue of a cause. However, Zayd’s being a substance is something that pertains to him by virtue of himself. Being a substance is an essential fact about Zayd, and since “a being not in a subject” is a quasi-definition of substance, it would follow likewise that “being not in a subject” is said essentially of Zayd (see Benevich 2017, pp. 257–58). Thus, in order for “a being not in a subject” to be said of Zayd essentially, the “being” in “a being not in a subject” cannot signify that Zayd is a being in act. To call Zayd “a being not in a subject” is to specify what kind of being he is—not to specify that he actually exists here and now. This is why Avicenna clarifies that the proper quasi-definition of substance is “an essence and reality whose being is only [ever] not in a subject”. Again, when this quasi-definition is predicated of Zayd, it says nothing of whether Zayd actually exists but specifies how Zayd exists when/if he exists—namely, not in a subject. Therefore, when we say that a substance is “a being not in a subject”, by “a being” we mean “an essence and reality”, and by “not in a subject” we mean “whose existence is only [ever] not in a subject”. This is the quasi-definition of substance, properly disambiguated and articulated.

How does this excursus on the quasi-definition of substance help us to solve the question of whether God is a substance? Avicenna tells us that nothing that is predicated as a genus of Zayd can be predicated of God. Now, Zayd is in the genus of substance, and the quasi-definition of substance is “an essence and reality whose existence is only [ever] not in a subject”. Thus, since substance is the most general genus under which a thing can fall, and the quasi-definition of substance is “an essence and reality whose existence is only [ever] not in a subject”, it follows that “an essence and reality whose existence is only [ever] not in a subject” is predicated of Zayd as a genus. However, this cannot be predicated of God: “God is an essence and reality whose existence is only [ever] not in a subject” is false. Why is this proposition false? Avicenna explains that the predicate in this proposition entails a certain composition between, on the one hand, the essence and reality and, on the other, its existence. Avicenna says that the part of the quasi-definition of substance “whose existence is only [ever] not in a subject” is a judgement or rule (ḥukm) which attaches (yulazm) to the essence. The reason why this quasi-definition does not apply to God is that, in the case of God, existence (wujūd) is not something that attaches to Him but is in fact His very essence—wujūd is for God what essence is for everything else. So, Avicenna understands “whose existence is only [ever] not in a subject” to be like an add-on to “an essence or reality”.

In sum, we find five distinct themes in this argument: (1) the problem that, since God is not in a subject, he must be in the genus of substance; (2) the need to disambiguate the term “a being” in the quasi-definition of the genus of substance such that “being in act” is not included in the quasi-definition; (3) the correct quasi-definition of the genus of substance includes essence as a quasi-genus term and a “rule” regarding how the essence exists when or if it exists as the quasi-differentia term; (4) the way that the quasi-differentia term “a being not in a subject” is predicated of substances; and finally, (5) the idea that God’s essence is being.

When we look at Avicenna’s Persian encyclopedia the Dānesh Nāmeh, we find a very similar argument, raised in the same context (coincidentally, in the same chapter number as the Ishārāt), and following the same steps. I have inserted into the below text bracketed step numbers to indicate where T2 follows the steps of T1.

[T2 Avicenna, Dānesh Nāmeh ch. 25 (Avicenna 1955, 2004)]17

That the Necessary Being is neither substance nor accident.

Substance is that which, when it exists, the essence (ḥaqīqat) consists in the fact that its being (wujūd) is not in a subject—which does not mean that it already has a realized (ḥāṣl) being not in a subject [(2)]. Following this, you do not doubt that the body is a substance, but you can ask whether such a body which isa substance exists or not, and then ask yourself whether its being is or is not in a subject. So, substance is that which has an essence, e.g., corporeity, spirituality, humanity, and horseness. This essence has this condition: until its being (ʿinniyyat) is found not in a subject, you cannot know whether it has being (ʿinniyyat) or not [(3)]. Everything that is such has an essence other than being. So, everything whose essence is not other than being is not a substance. As for the accidental, it is clear that the Necessary Being is not in something. Since the being of the Necessary Being is not [said] univocally and generically [i.e., as a genus of] with the being of other things, the “being not in a subject” which pertains to Him is not found in the sense of a genus with the “being not in a subject” which pertains to a human and to things other than human—and this is because being is applied by analogy and not by univocity or genericity and because “that which is not in a subject” is not [said] analogically [(4)]. So, “being not in a subject” is not the genus of things except in the sense that we have said, [namely that] substance is the genus of all of the things which are substances. So, the Necessary Being is not a substance and, in short, He is not made a part of any of the categories because the being (wujūd) of all of the categories is accidental and added to the essence, being outside of the essence, [whereas] the being of the Necessary Being is His very essence [(5)]. So, from everything we have said, it has become clear that the Necessary Being does not have a genus and, as a result, does not have a difference and consequently does not have a definition. It has become clear that He is neither in a receptacle nor in a subject; so, He does not have a contrary. It has become clear that He does not have a species; so, he has neither equal nor like. It has become clear that He does not have a cause; so, He has neither change nor divisibility.

Of the five themes found in T1, we can find all of them in T2. (1) is not explicitly stated in T2 as it is in T1, but you can see it in the background, so to speak, as the motivation for Avicenna’s emphasizing that “being not in a subject” is not said of both God and created substances. Regarding (2) in T2, Avicenna makes the same point as before, namely that although “being” or “existence” figures into the quasi-definition of substance, it does not signify actual existence or being in act. Whereas Avicenna uses the phrase “a being in act” (mawjūd b-l-fʿil) in T1, in T2, Avicenna uses the phrases “realized existence” (wujūd ḥaṣl) and “being” (ʿinniyyat). Regarding (3), wherein T1 Avicenna discusses a “judgment” or “rule” (ḥukm) that attaches to the essence of a substance, in T2 Avicenna discusses a condition that applies the essence of a substance, namely that one does not know whether the essence of a substance exists until one discovers that essence instantiated in a subject (see Taylor 2022, pp. 458–64). So, the condition under which an essence exists when or if it exists (that is, existing as a subject) is a characteristic feature of substance. Of some essence, one might ask whether it exists (i.e., whether there is such a thing as x), and then, if one discovers that x does indeed exist, one might ask whether it exists not in a subject—that is, whether, when it exists, it exists as a subject or in a subject.18 After first discovering that some × exists, we can then enquire into the condition of its existence and, depending on what we find, determine whether x is a substance or accident. Regarding (4), Avicenna makes the same point as in T1 that “existing not in a subject” is not predicated of God, though here in T2 we obtain more rationale for that judgment. In T1, Avicenna denies that “existing not in a subject” is predicated of God because God does not have the sort of essence to which that rule of existing attaches. However, in T2, Avicenna tells us that “existing not in a subject” is not predicated of God, because no condition of existing is predicated of God and creatures univocally. However, “existing not in a subject” is said univocally of all substances. This leads to theme (5) in T2 which is the same as T1: Avicenna affirms that God’s essence simply is existence. Now, though there are differences between the texts, it is well worth noting that all of the themes are present and accounted for. Moreover, they are even presented in the same order as T1. The differences between T1 and T2 are not doctrinal, but for the most part, are verbal. The argument in T2 does not diverge from that of T1—rather, it expands upon that of T1.

Aside from these modifications to the five themes, we also acquire some new information in T2. Whereas the scope of T1 concerns only whether God is a substance, the scope of T2 is broader in that God is found to be neither a substance nor an accident. Avicenna’s conclusion in T2 goes one step beyond that of T1 in that T2 denies that God is within the scope of the categories at all because the scope of categorial being is defined by the composition of essence with existence. The essences in the categories are conditioned essences—substances are essences whose existence is not in a subject and accidents are essences whose existence is in a subject. For the things in the categories, their essences are committed to a relation to a subject and they are defined by that relation: substances are essences that are not in a subject because they are subjects and accidents are essences that are in subjects. Though this idea is implicit in the reasoning of T1, here in T2, it is explicit.

Having seen that T1 and T2 argue to the conclusion that God is not a substance by the same line of reasoning, we are now prepared to evaluate Avicenna’s argument in the metaphysics of the Shifā’. The Shifā’ text, in my view, is the source of the proliferation of arguments in the secondary literature that I catalogued in the Introduction. In order to fully understand the Shifā’ text, there is some context that needs to be established. First of all, Avicenna’s argument that God is not a substance in the Shifā’ concludes a lengthy chapter concerning God’s attributes. In the metaphysics of the Shifā’ 8.4, Avicenna first proves that the Necessary Being is rightly called “the First” because He is the cause of all things. Moreover, the First is also perfectly simple in that he is not something having necessary existence, but rather is necessary existence itself. This is significant because if the First was a thing having necessary existence, then necessary existence would not be the First’s essence, but would instead attach to the First’s essence. However, Avicenna tells us near the beginning of the chapter that “the First does not have an essence other than being” (la māhiyyat lahu ghair al-inniyyat) (Avicenna 2005, p. 274). Thus, the First’s essence is being itself simply and absolutely, whereas everything else has a quiddity to which being merely attaches. After having proved this point, there is a shift in Avicenna’s language: rather than say that the First does not have an essence other than being, Avicenna surprisingly says that the First simply has no essence.19 “Hence, everything that has a quiddity is caused. The rest of the things, other than the Necessary Being, have quiddities. And it is these quiddities that in themselves are possible in being, and being only attaches to them from without (min khārij). The First, hence, has no quiddity. Those things possessing quiddities have being emanate on them from Him” (Avicenna 2005, p. 276) (some modifications to Marmura’s translation). From the fact that the First has no essence, Avicenna deduces that the First has no genus, no differentia, no definition, and no demonstration of Him (because He has no cause).

Finally, we reach Avicenna’s argument that God is not a substance. Now, a standard way to think of this argument is as a continuation of the preceding enumeration of attributes that do not apply to God because God lacks an essence. So, in this view, God is not a substance for the same reason that he does not have a genus, differentia, definition, or demonstration, namely, because He does not have an essence. So, at the end of the day, Avicenna’s argument that God is not a substance is the final negative attribute in a long list of negative attributes. There seems to be a good reason to read the argument this way. The argument that God is not a substance is presented as a response to an objection to Avicenna’s earlier claim that God does not have a genus. For if God is a substance, then God is in the genus (i.e., category) of substance, and then God has a genus. This is why, I think, it is very natural to read the argument that God is not a substance in the Metaphysics of the Shifā’ 8.4. as a No Essence Argument.

However, do the details of the argument support that reading? Let us look at the text to find out.

[T3 Avicenna, al-Shifā’: Ilāhīyyāt 8.4 (Avicenna 2005, pp. 277–78)]

Someone may say, ‘‘Although you have avoided assigning the name ‘substance’ to the First, you do not avoid assigning Him its meaning. This is because He is a being not in a subject; and this is the meaning of substance, which you have rendered a genus” [(1)]. We answer: this is not the meaning of the substance we have made a genus. Rather, the meaning of [the latter] is that it is the thing having an established quiddity whose being is not in a subject—for example, a body and a soul [(3)]. The proof that it would not be a genus at all if this is not intended by “substance” is that the thing referred to by the expression “a being” (mawjūd) does not require its being generic. The negation that follows it does not add [anything] to it above and beyond being (wujūd), except the relation of distinctiveness. This [latter] meaning does not include any realized thing after being (wujūd), nor is it a meaning of something in itself, but it is only in terms of relation. Hence, the being (mawjūd) is not in a subject. It is only the affirmative meaning in it that can belong to some entity [that] is the being. What comes after it is a negative, relative thing, extraneous to the identity (huwiyya) belonging to the thing. Taken in this way, this meaning would not be a genus. This you have learned in a perfected way in the Logic. You have also learned in the Logic that, if we say, for example, “all A”, we mean everything that is described as “A”, even if it has a reality other than A-ness. Hence, in defining substance, our statement “a being not in a subject” means that it is the thing of which it is said “a being not in a subject” in that “the being not in a subject” is predicated of it, and has in itself a quiddity—as, for example, man, stone, and tree [(4)]. It is in this way that substance has to be conceived in order to be a genus. The proof that there is a difference between the two, and that genus is one of them but not the other, is that you would say of some human individual whose being (wujūd) is unknown that he is necessarily [someone] whose being (wujūd) would [consist] in his not being in a subject. But you do not say that he necessarily is a being now not in a subject [(2)]. It seems as though we went into great detail in explaining this when we discussed [it] in the Logic.20

There are several things to note about this text. First, all but one of the themes that we found in T1 and T2 are present in this text, though very much out of order. The only theme that is missing is (5): that God’s essence is His being. However, all of the other themes are here in this text. Now, this is not to say that there is nothing new going on in this text. Indeed, there surely is. At the center of this text lies a rather unique discussion of why “a being not in a subject” is an inadequate quasi-definition of substance. Avicenna makes the case here that “a being not in a subject” could not be the quasi-definition of the genus of substance because the only affirmative meaning in that phrase is “a being”, and “a being” cannot be a genus. Avicenna’s idea seems to be this: if you make “a being not in a subject” the quasi-definition of the genus of substance, then the quasi-genus term of the quasi-definition would be “a being” and the quasi-differentia term would be “not in a subject”. But, Avicenna says, “a being” is not a genus. Therefore, “a being not in a subject” cannot be the quasi-definition of substance.

It is tempting here to say that Avicenna’s mention of the fact that “a being” is not a genus gives evidence that the argument of T3 is the Being is Not a Genus Argument as outlined above. However, this is not the case. Notice that the conclusion of the Being is not a Genus Argument from Rosheger above is that God is not a substance. The argument about “a being” not being a genus in T3 does not aim to prove that God is not a substance. Rather, it concludes that “a being not in a subject” is not the quasi-definition of the genus of substance. Avicenna does not make mention of the fact that “a being” cannot be a genus in order to show that God is not in a genus (and is, therefore, not a substance). Rather, it is to show that “a being not in a subject” is not the correct quasi-definition of substance.21

Another big difference between T3 and T1-2 is that theme (4) of T3 is notably muted. Whereas T1 and T2 land firmly on the claim that “being not in a subject” is something like a rule or condition that applies to the essence, (4) in T3 requires some unpacking. Avicenna explains that the “being not in a subject” that is in the quasi-definition of substance should be interpreted as “the thing called ‘a being not in a subject’”. This is because every universal affirmative, like “every A”, should be taken as “everything that is an A”. Thus, to say that every substance is a being not in a subject is equivalent to saying that every substance is a thing that is a being not in a subject. As Avicenna makes clear earlier on in the metaphysics of the Shifā’, to call something a “thing” is to say that it is of a determinate sort, nature, or essence. Thus, substances are things of a determinate sort, like humans or horses, to whom belongs the attribute “being not in a subject”.

Regarding theme (2), Avicenna refers us to a point he made in the Logic, but also that he made in T1 and T2, with which we are already familiar: that by knowing that some x is a substance, one knows that x’s existence is not in a subject, though one may doubt whether x exists at this very moment. Of course, Avicenna does not tell us exactly where he discussed this, and the Logic of the Shifā’ is rather long. I cannot say with certainty that this is the text Avicenna has in mind, but it is a plausible candidate:

[T4 Avicenna, al-Shifā’: Maqūlāt 3.1 (Avicenna 1959)]

Consider some person like Zayd when he is absent from you; or a species of substance with the ability of lapsing from the world (if in your view its lapsing is possible); or a species whose being (wujūd) one doubts: then you know that its whatness, when it is an existent among individuals, is not in a subject. And, you know that this notion [that is, not being in a subject] is the primary constitutive element (muqawwim) for its essence, since you know that it is a substance [(3)]. But you do not know whether it is “a being not in a subject” among individuals in actuality—rather, perhaps in your view it [will be] nonexistent afterwards [(2)]. Indeed, actual being (wujūd b-l-fi‘l) in an individual not in a subject is not a constitutive element for the whatness of Zayd or for a thing that is a substance. Rather, it is an item (ʿamr) that attends as an attachment to the being (mawjūd), which is not true for the whatness of things, as you have learned.

The context of this argument makes no mention of God’s not being a substance, so we do not find (1) or (5) anywhere nearby. However, we do find themes (2) and (3) here. Avicenna makes the familiar point that the “being not in a subject” is in the quasi-definition of substance and does not signify that the substance, by definition, actually exists. This is because actual existence does not fall into the definition of substance because it is something that attaches to the essence of a substance. If it were in the definition of a substance, then it would be a constitutive element of the substance and, as Avicenna makes clear here, it is not a constitutive element. We have already seen the term “muqawwim”—“constitutive element”—in T1 above. There, Avicenna told us that “mawjūd”—“a being”—is not a constitutive element of an essence but is instead a “lawāzim”—a “concomitant” of an essence. Here, in T4, we have different vocabulary but seemingly the same idea: the distinction between “existing not in a subject” as a constitutive element of an essence and “actually existing not in a subject” as a concomitant of an essence. This is further evidence that the reasoning in T3, though notably different from T1-2, is yet connected to them by way of a line of reasoning previously laid out in the Shifā’. By distinguishing “existing not in a subject” as a constitutive element of substance from “actually existing not in a subject” as a concomitant/attachment, T4 connects the reasoning of T3 to T1.

Having reviewed these three arguments from Avicenna’s major philosophical encyclopediae, it is clear to see that Avicenna makes the same, singular argument that God is not in the genus of substance in each encyclopedia. Though there are differences in order and vocabulary, the doctrine appears to be the same throughout and follows five major themes: (1) the problem of whether God is in the genus of substance; (2) the correct quasi-definition of substance; (3) the need to disambiguate the sense of “being” that appears in the quasi-definition; (4) the way that “being not in a subject” is predicated of substances; and finally, (5) the idea that God’s essence is being. When we line them up, as in

Table 1. Stepwise Comparison of Three Iterations of Avicenna’s Essence-Being Distinction Argument.

T1: Ishārāt	T2: Dānesh Nāmeh	T3: Shifā’
(1) It is often thought that the maʿna of “a being not in a subject” is common to the First and to everything else [with] the generality of a genus such that [the First] falls under the genus of substance.	(1) MISSING	(1) An objector might say: “surely if you have refrained from assigning to the First the name ‘substance’, you do not pretend to refrain from assigning to him its meaning. This is because [He is a] being (mawjud) not in a subject, and this meaning is the substance with which you put him in a genus classifying him”.
(2) For, “a being not in a subject”—which is like a description for substance—does not mean “a being in act (b-l-fʿil) existing (wujūdan) not in a subject[.]”	(2) Substance is that which, when it exists, the essence (ḥaqīqat) consists in the fact that its being (wujūd) is not in a subject—which does not mean that it already has a realized (ḥāṣl) being not in a subject.	(2) [Y]ou would say of some individual person whose being is unknown that he is undoubtedly something whose being (wujūd) is not in a subject, and you would not say that he is undoubtedly a being [or existent (mawjūd)] now in a subject.
(3) But, the meaning that is predicated of substance like a description and in which particular substances participate in potency (just as one participates in a genus) is that it is an essence and reality (ḥaqīqa) whose being [wujūdhā] is only [ever] not in a subject.	(3) So, substance is that which has an essence: e.g., corporeity, spirituality, humanity, horseness. This essence has this condition: until its being (ʿinniyyat) is found not in a subject, you cannot know whether it has being (ʿinniyyat) or not.	(3) We answer: this is not the meaning of substance which is its genus. Rather, that meaning [which we propose] is a thing that is the possessor of a determinate essence whose being (wujūd) is not in a subject, like body or soul.
(4) So, it is not fitting to predicate of the Necessary Being at all that which can be predicated of Zayd like a genus because He does not possess an essence to which this judgment attaches[.]	(4) Since the being of the Necessary Being is not [said] univocally and generically [i.e., as a genus of] with the being of other things, the “being not in a subject” which pertains to Him is not found in the sense of a genus with the “being not in a subject” which pertains to a human and to things other than human—and this is because being is applied by analogy and not by univocity or genericity and because “that which is not in a subject” is not [said] analogically	(4) [I]n the definition of substance, the meaning of “the being (mawjūd) not in a subject” is “the thing called ‘a being not in a subject,’” on account of the fact that “a being not in a subject” is predicated of it, though it is an essence in itself, e.g., a human, a stone, or a tree. This is the way one must conceive of substance so as to be a genus.
(5) [N]ecessary being belongs to him just as essence belongs to the rest.	(5) So, the Necessary Being is not a substance and, in short, He is not made a part of any of the categories because the being (wujūd) of all of the categories is accidental and added to the essence, being outside of the essence, [whereas] the being of the Necessary Being is His very essence.	(5) MISSING

, we can see how similar these texts are.

Having shown that the same themes appear throughout T1-3, in the next section I shall explain precisely how this argument works and what it tells us about divine nature and the nature of the categories.

3. Conclusions: God, the Categories, and the Essence-Being Distinction

In the previous section, we found that Avicenna’s response to the question of whether God is in the genus of substance consistently reaches the same conclusion through an argument addressing five themes. What precisely is Avicenna’s argument throughout these texts? The conclusion is clear enough: God is not in the genus of substance. Why is this true? Avicenna’s reasoning in T2 is the clearest: “the Necessary Being is not a substance and, in short, He is not made a part of any of the categories because the being (wujūd) of all of the categories is accidental and added to the essence, being outside of the essence, [whereas] the being of the Necessary Being is His very essence”. This is what I have called the Essence-Being Distinction Argument.

• P1 If x is in the genus of substance, then x’s essence is distinct from x’s Being.
• P2 God’s essence is not distinct from God’s Being.
• C1 God is not in the genus of substance.

Now, P2 is theme (5) above, and P1—the crucial explanatory premise—is found in T2 exclusively, so it is clear where those claims are found in Avicenna’s texts. However, if this is all there is to the argument, then what exactly is the significance of themes (1)–(4) to the Essence-Being Distinction Argument? Theme (1) merely introduces the issue, so that is not a mystery. What is yet a mystery is how to stitch together (2)–(4) to acquire a coherent argument. I argue that themes (2)–(4) are meant to prove P1, and so all of them must contribute to answering the question: why is it the case that anything in a category must have a distinction between its essence and its being? Avicenna suggests that the right place to start answering this question is by thinking about the quasi-definition of substance and accident. A substance is a thing having a determinate essence that exists not in a subject. An accident is a thing having a determinate essence that exists in a subject. Both of these quasi-definitions include a quasi-genus term and a quasi-differentia term22: an essence and what I have called a condition of existence—i.e., a rule governing how the essence exists when/if it exists. The condition of existing not in a subject is a constitutive element of being a substance. As a result, it is univocally said of every substance. Now, to say that the condition or rule of existing not in a subject falls into the quasi-definition of substance does not signify that a substance actually exists by its very nature. For something can be a substance and not actually exist. Moreover, “existing not in a subject” is predicated of a substance per se and univocally, but “actually existing” or “being in act” is not predicated univocally, but analogically. Therefore, “being in act” or “actually existing” is not a constitutive element of a substance such that it falls into the quasi-definition of substance. Rather, “being in act” or “actually existing” is a concomitant of the essence of a substance. “Being in act” is external to and attaches to the essence of a substance. Thus, “being” is predicated of creatures in two ways: one, as a constitutive element, it signifies the condition of existence proper to the essence; two, as a concomitant and accident that attaches to the essence. “Being in act” is predicated of a substance only when the existence of that substance is realized, and when the existence of a substance is realized, the substance exists not in a subject.

In sum, Avicenna’s argument appears to be something like this:

1. If x is in the genus of substance, then x’s essence is distinct from x’s actual existence (i.e., x’s being-in-act).

1.1. For, if x is in the genus of substance, then x is a thing having a determinate essence to which belongs the condition of existing not in a subject (Theme (3)).

1.2. If x is a thing having a determinate essence to which belongs the condition of existing not in a subject, then the condition of existing not in a subject cannot entail that x actually exists.

1.2.1 For if the condition of existing not in a subject entails that x actually exists here and now, then actually existing here and now is a constitutive element of x’s essence.

1.2.2. However, actually existing here and now is not a constitutive element of x’s essence.

1.2.2.1 For if actually existing here and now is a constitutive element of x’s essence, then “actually existing here and now” would be said univocally of all substances (Theme (4)).

1.2.2.2. However, “actually existing here and now” is not said univocally of all substances.

1.2.2.3. Therefore, actually existing here and now is not a constitutive element of x’s essence (Theme (2)).

1.2.3 Therefore, the condition of existing not in a subject does not entail that x actually exists here and now.

1.3. If the condition of existing not in a subject cannot entail that x actually exists here and now, then x’s essence is distinct from x’s actual existence.

1.3.1 For there is a distinction between the essence of the substance which is constituted by the condition of existing not in a subject, and the actual existence of the essence which is not a constitutive element of the essence of the substance, but is instead something caused from outside its essence.

2. God’s essence is not distinct from God’s actual existence (Theme (5)).

3. Therefore, God is not in the genus of substance.

The hinge on which this whole argument turns is Avicenna’s distinction between “being” or “existence” as a constitutive element and as a concomitant. In one sense, “being” or “existing not in a subject” is the quasi-defining feature of substances that distinguishes substances from accidents. As such, this condition of existing belongs to substances per se and is said of every substance univocally. In another sense, being-in-act or actually existing is not a constitutive element of a substance. As such, it is not said of substances univocally. The distinction amounts to a distinction between actually existing and the rule governing how a thing exists when/if it actually exists. Since “being not in a subject” is a constitutive element of being a substance, the “being” cannot signify “being in act” or “actually existing here and now”. Since the categories of substance and the accidents are distinguished from each other according to diverse conditions of existing, and the condition of existing definitive of the categories of substance and the accidents do not entail being-in-act or actual existence, it follows that being-in-act or actual existence is outside the definitions of the categories of substance and accidents necessarily.

Further, I do not think Avicenna’s claim that “being-in-act” or “actually existing” is outside of the quasi-definition of the categories is an ad hoc contrivance of a rule. It seems to be the case that, for any x such that x is distinguished from other things with which it shares a genus by virtue of a differentia that comprises or includes a verb, the verb in the definition of x is not signified as actually or actively being carried out, but instead as a disposition to act in a certain way. For example, consider Stockfish. Stockfish is a computer program that is able to analyze chess positions and then generate one or more optimal moves. If you are so inclined, you can go online and play a game of chess against Stockfish right now. So, a definition I could give of Stockfish is “a computer program that plays chess well”. “Plays” figures into the definition of Stockfish—it is part of the differentia term that distinguishes Stockfish from other computer programs. If Stockfish were stored on a hard-drive and forgotten in a desk somewhere, never again to face the Vienna Gambit or to calculate the best response to the Caro Kann Advance Variation, it is still correct to define Stockfish as a computer program that plays chess well, even though it is not playing chess at this very moment nor, indeed, will ever play again. This same thesis about the verb “plays” as a differentia term applies to the quasi-definitions of substance and accident: because “being” or “existing” figures into the quasi-definitions of the categories of substance and accident as the quasi-differentia term by which substance and accident are distinguished, the manner of “being” or “existing” that a substance or accident has in its quasi-definition does not signify actual existence, but instead a sort of disposition to exist in a certain way. This is what I have called the “rule for existing” or the “condition of existing”. Further, just as one can truly predicate the definition “a computer program that plays chess well” of Stockfish without knowing whether Stockfish is actually playing chess right now, and indeed even if Stockfish is destined to never play chess again, so too is the definition “an essence existing (not) in a subject” predicated of substances and accidents without knowing whether that substance or accident are actually existing right now or, indeed, never destined to exist. Hence Avicenna’s insistence that “being not in a subject” or “being in a subject” is a constitutive element of substances and accidents—since “being” or “existing” shows up in the quasi-definitions of substance and accident, and whatever is in the definition of a thing is a constitutive element of it, it must be the case that “being” or “existence” in the quasi-definitions of substance and accident signifies something other than “being in act” or “actually existing here and now” because these are concomitants rather than constitutive elements of an essence.

We now have a reconstruction of Avicenna’s argument for why God is not a substance, and we have a rationale motivating that argument based on the distinction between a concomitant and a constitutive element of an essence. At the risk of beating a dead horse, as it were, I shall presume that there is yet some life in this old nag and, as a conclusion, ask and answer a few final questions: why is it so important to get Avicenna’s arguments for whether God is in the category of substance right? Aside from cleaning up some of the details in the secondary literature, what is at stake in correctly understanding this argument? What exactly have we learned from the above texts?

What have we learned from the above? First of all, we have seen that in three of his major philosophical encyclopediae, Avicenna consistently makes the same argument—the Essence-Being Distinction Argument—to prove that God is not in the genus of substance. Contrary to the literature on this topic, there is not a plethora of arguments, but one consistent argument. This is perhaps surprising given the fact that the context behind T3 strongly suggests that the reason God is not in the genus of substance is the same reason that God does not have any genus, difference, definition, or demonstration—namely, that God has no essence. However, as we saw above, this is not the reasoning that Avicenna gives. It is not His lack of essence that leads us to infer that He is not in the genus of substance, but rather it is the absence of a distinction between His essence and His being that leads us to infer that He is not in the genus of substance.

Now, I do not intend to offer a final or decisive word about whether God has an essence in Avicenna here. The goal of this study is humbler—I simply want to show that Avicenna consistently offers the same argument for why God is not in the genus of substance across major works. That being said, this humble study does suggest something about Shifā’: Ilāhīyyāt 8.4: though there is a central part of the text that states that God has no essence, the text is bookended by sections that affirm that God has an essence and that His essence is being. Perhaps when Avicenna says that “the First has no essence”, we must understand this statement to be qualified in some way like “the First has no essence like other essences” (on God’s uniqueness in Avicenna, see De Smet and Sebti 2009). This reading seems plausible and even preferable given the fact that Avicenna’s reasoning about God’s not being a substance consistently relies on the premise that God has an essence and that essence is being. I cannot entirely explain why, then, Avicenna says in Shifā’: Ilāhīyyāt 8.4 that God has no essence; all that I can say is that we should prefer the interpretation that God has an essence because God’s having an essence plays a crucial role in Avicenna’s last argument and that the broader context of Shifā’: Ilāhīyyāt 8.4 does not demand that God not have an essence. The conclusion that God does not have a genus, species, differentia, definition, or causal explanation does not rely on the claim that God has no essence as it appears in Shifā’: Ilāhīyyāt 8.4. For, as we saw above in T2, Avicenna thinks that he can derive those conclusions from the claim that God is not in the categories because God’s essence is being. However, this is a humble consequence that follows from my above analysis; I leave the bigger project of sorting out what’s going on in Shifā’: Ilāhīyyāt 8.4 and exhaustively cataloguing texts on whether God has an essence in Avicenna to someone like Bertolacci. For now, my analysis favors that God does have an essence—His essence is being.

What else have we learned? Why is it important to get this argument right? The true significance of the doctrine that God is not a substance lies not only in what it tells us about the divine nature, but also about the nature of the categories. The reason that God is not in the category of substance—indeed, in any category whatsoever—is that to be in a category requires a distinction between one’s essence and one’s being. Why? Because something is in the category of substance or one of the categories of accidents on account of the fact that it is a sort of thing (i.e., an essence) that follows a certain rule (ḥukm), or has a certain condition (ḥāl), concerning the way that the essence exists when or if it exists. Substances, when or if they exist, exist not in a subject, but as a subject. Accidents, when or if they exist, exist in a subject. To be a subject for accidents is a constitutive element (muqawwim) of a substance and to be in a subject is a constitutive element of an accident. However, this being not in a subject or being in a subject is not the same thing as being in act or actually existing here and now as or in a subject. This entails that, to be in a category, something must have an essence to which a certain condition of existing belongs essentially, and this condition of existing is distinct from actually existing or being in act. In other words, the categories are categories of essences distinguished according to diverse conditions or rules of existing. For these categorial essences, actual existence or being in act is not a constitutive element—being in act is not a part of the quasi-definition of the categories. This, of course, does not preclude that essence is always found with being in act as mutual concomitants—there are no non-actual or pre-existent essences (see Bertolacci 2012, pp. 271–72). However, essence and being in act are distinct because being in act cannot be a constitutive element.

Therefore, it follows from Avicenna’s understanding of the categories as essences distinguished according to diverse conditions of existence that God, Who does not have conditional existence but is existence absolutely, does not fall into the domain of the categories. Avicenna is ultimately in agreement with other Neoplatonic commentators on God’s being outside the domain of the categories though his reasoning is unique. Whereas other Neoplatonists understand the categories as categories of composite, sensible being (as opposed to the divine, intelligible being about which Aristotle does not intend to speak in the Categories), the composition definitive of the categories for Avicenna is not that of form and matter, but rather that of essence and being. Why Avicenna focuses on the composition of essence and being rather than form and matter as constitutive of the categories requires further study, but we have a gesture at the answer: the categories of substance and the accidents are not distinguished according to diverse ways of joining form to matter, but instead, they are distinguished according to the different ways that essences have their existence.23