1 The Binary World: Kalām Atomism versus Falsafa Hylomorphism

I begin with an overview of central differences between classical kalām atomism and the form-matter ontology of the Peripatetic falāsifa, including Avicenna. A central problem for our purposes concerns the status of composite entities. In the kalām view, a composite is any entity that is not an indivisible atom. Here it is of central importance that however one might distinguish between simple and complex entities in classical AshÊ¿arite kalām, complex entities, strictly speaking, do not exist. Though the point often remains implicit, as Ayman Shihadeh has shown, classical AshÊ¿arites would bite the ontological bullet of kalām indivisibilism, explicitly affirming the fictive nature of our linguistic references to complex entities. They hold, for example, that the expression ‘this human’ refers only fictively or figuratively to a singular entity or unity (i.e., anything beyond a collection of atoms).¹⁰ I return to the details of this view shortly.

The Peripatetic tradition of falsafa, by contrast, presents a view of the cosmos as populated by various kinds of bodies (ajsām) or “substances”, from the four elements to living beings. Their analysis of the world is grounded in Aristotelian hylomorphism, which views material substances as composites of form and matter.¹¹ We can label this view as a kind of ‘substance pluralism,’ since there are distinct kinds of material substances, i.e. ontologically basic wholes or entities.¹² The view conflicts with ‘substance monism,’ which can be attributed to Aristotle’s Presocratic predecessors (e.g., Thales, Anaximenes).¹³ The monists hold that there is one underlying substance, e.g. water or air, from which all perceptible things are produced. On this view, at least as RāzÄ« understands it, Aristotle’s predecessors see natural things as produced solely through accidental change that occurs to an underlying substance (for example, water undergoes condensation or rarefaction to produce other elements and natural entities). RāzÄ« refers to these thinkers as the “deniers of generation and corruption,” because the primary, underlying substance in their ontology persists through change, that is, true substances do not come to be or perish. By contrast, Aristotle asserts not only the existence of multiple kinds or species of substances but also that individual substances come to be and perish. To be sure, Aristotle introduces form and matter in the Physics to ontologically ground his distinction between the two fundamental kinds of change in natural reality: substantial change (i.e. generation and corruption) and accidental change (in quality, quantity, and place). For example, the birth or death of Socrates is a case of substantial change since an independent entity simply comes to be or perishes. By contrast, in cases of accidental change, say, when Socrates turns from pale to tanned, a substance is not generated or corrupted (a pale man did not die, nor was a tanned man born). Rather, Socrates remains Socrates through the process of change; he simply acquires or loses an attribute. As RāzÄ« will argue, this distinction threatens to collapse, as substantial change itself requires underlying matter and can thus be assimilated to accidental change. To shore up this distinction, Aristotle advances a more robust hylomorphism grounded in the notions of potentiality and actuality.¹⁴ Form makes that which potentially exists exist in actuality.¹⁵ In contrast to a substrate that exists prior or posterior to the acquisition of an accident, the matter of a hylomorphic substance exists only potentially without its form. Here, the hard distinction in Aristotelian ontology between natural things, particularly living things as paradigmatic substances, and artifacts (or, more broadly, empirical compounds) will be problematized by RāzÄ«.¹⁶

Let us now return to relevant details of kalām ontology. Classical AshÊ¿arite ontology can be viewed as a kind of substance monism, since atoms, though numerically multiple, are the only true independent entities.¹⁷ As noted, bodies are viewed as being composite without any ontological unity or cohesion. To be sure, according to the AshÊ¿arites, ‘body’ (jism, jirm) is a technical term that signifies two atoms that are directly contiguous (i.e. without any intervening space).¹⁸ A body cannot be said to be space-occupying in itself. According to one interpretation, beyond being contiguous, a body is two atoms that possess, individually, the attribute of adhesion or combination (taʾlÄ«f, ijtimāʿ). Such accidents of composition inhere in each atom of the pair, a point based on the view that accidents cannot inhere in more than a single atom. The view follows from the point that body is not space-occupying and thus not a true substrate for attribute-inherence. As the AshÊ¿arites stress, even if the ordinary usage of the Arabic term jism refers to the continuous, corporeal objects that we perceive, the strict sense or the “ontologically designative” use (ḥaqÄ«qātan) refers to two adjacent or adjoined atoms.¹⁹ Given the above principles, the implication is that ‘body’ refers to a fictional entity that reduces to two atoms and to accidents that inhere in the atoms individually. As noted, this indeed is what some will explicitly hold.²⁰ As one author states, ‘body’ is used in a similar way as ‘long’ is used to describe atoms placed length-wise.²¹ In short, ‘body’, in its technical or non-technical usage, fails to refer to any real object beyond the underlying, discrete ontology of atoms and accidents.

In contrast to the AshÊ¿arites, the MuÊ¿tazilites held that composition is a single accident that inheres in two atoms.²² Moreover, the Basrians develop a more complex ontology, in which, for example, the attribute of ‘life’ gives a certain unity to a complex entity, which possesses the requisite ‘structure’ (binya).²³ Though this view attempts to account for structure and complexes by binding atoms in certain relations, the MuÊ¿tazilites still consider atoms as the fundamental entities and view animate or inanimate complexes as “essentially a conglomerate.”²⁴

Before turning to Rāzī’s analysis of Aristotle’s view of substance and motion conducted in Part I of Book II of the Mabāḥith and the Mulakhkhaá¹£, which concerns “accidents” (aÊ¿rāḍ), I conclude this section with a note on Part II of Book II on substances and, specifically, the “substantivity of bodies” (tajawhur al-ajsām).²⁵ The analysis moves from “body” (al-jism), i.e. absolute or qualityless body in chapter one (al-bāb al-awwal) of Part II, to “simple body” (al-jism al-baṣīt) in chapter two, which concerns the four elements and celestial matter.²⁶ In the following, I argue that RāzÄ« distinguishes between the analysis of “body”, which I label Body 1, from the analysis of higher-order, complex bodies, beginning with the simple, sensible elements in chapter two, which I label Body 2. The upshot of this distinction is that RāzÄ« limits the question of atomism versus hylomorphism to a question focusing on the constituents of Body 1. More precisely, RāzÄ« sees the question as regarding the (in)divisiblity of Body 1, i.e. physical magnitude or three-dimensional extension. As argued below, RāzÄ« thinks that an analysis of Body 2 can effectively be conducted independently of the question of the indivisibility of Body 1. In other words, one need not first assert a position on the ultimate indivisibility of physical magnitude to analyze what kinds of sensible bodies exist and what their natures and properties are. This, I propose, provides a constructive philosophical framework for what kalām writers, like TaftāzānÄ«, characterize as Rāzī’s tawaqquf or reserved judgment on the problem. Here, a few points can be highlighted regarding Avicenna’s approach.

First, in Avicenna’s system, Body 1 corresponds to body as a composite of corporeal form (al-ṣūra al-jismiyya) and prime matter, and is one of the five genera of substances (i.e., body, form, matter, soul, and intellect).²⁷ We can call this Absolute Body, which concerns the ultimate divisibility of physical reality.²⁸ What is of significance is that Avicenna’s Absolute Body is a form-matter substance that is ontologically prior to the simple elements and other corporeal substances.²⁹ In other words, Avicenna’s system adds a form-matter layer that is rather unique with respect to earlier Peripatetics and interpreters of Aristotle.³⁰ Here, two points can be highlighted with respect to RāzÄ«.

First, despite certain misinterpretations of his approach, Rāzī’s rejection of hylomorphism in the Mabāḥith and Mulakhkhaá¹£ is rather definitive.³¹ He roundly rejects hylomorphism throughout the Mulakhkhaá¹£ and the Mabāḥith.³² This is quite clear with respect to the analysis of Body 1. For example, RāzÄ« states, “[The corporeal form is] a thing that is neither perceptible nor known immediately (ḍarÅ«ratan), for that which is known and perceived immediately are these magnitudes and dimensions. […] No proof has been set out, in my view, that body is composed of matter and form, so of course we do not assert it.”³³ For epistemic and ontological reasons, RāzÄ« reiterates his rejection of form and matter as constituents of Body 1 throughout Part II.³⁴ In the present context, it is worth highlighting that a central point underscored in Rāzī’s critique of Avicenna’s theory is that it concerns the fundamental divisibility of corporeal reality and not simply the ordinary or perceptible division of empirical things.³⁵ As for Body 2, we will see that RāzÄ« finds the hylomorphic analysis of composite, sensible substances problematic as well. His ontological reasons for rejecting the latter are explored below.

Second, it is notable that a central problem that RāzÄ« identifies with respect to Avicenna’s analysis of Body 1 is that it concerns the imperceptible constituents of what is otherwise “immediately” known, i.e. body as three-dimensional extension. The view that RāzÄ« endorses as his own in the Mabāhith and the Mulakhkhaá¹£ is that body is magnitude or three-dimensional extension along with immediately related properties.³⁶ However, it is significant that RāzÄ« parses the point rather sharply. That is, his own view here is not that Body 1, as extension, is the ultimate ontological subject or corporeal substance, such that it has been proven to possess no further underlying constituents (or genus and differentia).³⁷ Rather, Body 1 is posited by RāzÄ« as the basic empirical object of analysis and subject of attribution in his system.³⁸ As such, Body 1 or three-dimensional extension and space-location (taḥayyuz) will serve as the explanatorily (rather than ontologically) basic subject for his subsequent analysis of composite, empirical substances (i.e. Body 2) in chapter two of Part 2. This reading of the Mabāḥith and Mulakhkhaá¹£ helps us to understand Rāzī’s own, positive analysis of the four elements and animate beings, which rejects hylomorphism but asserts of such composite substances a certain reality, as seen below. In other words, Rāzī’s wholesale rejection of hylomorphism need not commit him to the totalizing and reductive atomism of classical kalām. The following outlines the alternative, ‘attributive’ analysis he proposes.

To be sure, postclassical works of kalām, following the framework of Rāzī’s Mabāḥith and Mulakhkhaṣ, devote a great deal of space to the analysis of a comprehensive spectrum of composite sensible substances. If they are simply atomists in the style of classical kalām, the analysis makes little sense, as all such things are fictive, extensionless objects. Moreover, it is clear that postclassical Ashʿarite authors problematize the hylomorphic analysis of composite substances, drawing primarily on Rāzī’s texts and themes, as discussed below. It seems, then, that neither the hylomorphism of falsafa nor the atomism of classical kalām provides a satisfactory frame of analysis in postclassical works. The following suggests that, drawing on Rāzī’s approach, the mutakallimūn develop conceptual distinctions, including the attributive analysis of perceptible bodies, that allow them to conduct a more systematic analysis of physical reality.

2 Rāzī on Aristotle and Hylomorphism

In this section, I focus on several critical points that RāzÄ« raises against Aristotle and hylomorphism, before turning to Rāzī’s positive views in Part 3. Part 2 and Part 3 trace a long arch of Rāzī’s arguments developed from Book I through Part I and II of Book II. RāzÄ« devotes Book I of the Mulakhkhaá¹£ and the Mabāḥith to “general things” (al-umÅ«r al-ʿāmma) and divides Book II into “accidents” (Part I) and “substances” (Part II), as noted above. As Eichner has shown, this division would provide an organizing framework for postclassical works of kalām.³⁹ The discussion begins with Rāzī’s general analysis of simple and composite essences in Book I and then turns to his analysis of Aristotle’s view of form and matter in his chapter on motion in Part I of Book II.

Rāzī’s analysis of essence in Book I spans several topics, including various chapters on essence, mereology, and unity. One way of viewing Book I is as a summary or restatement of Avicenna’s approach to primary concepts in the Metaphysics of al-Shifāʾ.⁴⁰ Alternatively, Rāzī’s discussion in this section might be understood as developing a more general analysis of primary concepts than that found in Avicenna’s works, which tends to push towards the latter’s own Aristotelian and Peripatetic analysis.⁴¹ In the latter reading, the concepts and distinctions in Book I are developed in order to raise relevant aporias and to pursue alternative readings of primary concepts. The following analysis supports the latter reading.

In his section on essence, RāzÄ« begins his analysis of essence with a discussion of what distinguishes an essence from its concomitants.⁴² He then turns in the second section of the chapter to the division of simples (al-basāʾiá¹­) and complexes (al-murakkabāt):

An essence is either composite or simple. The composite is that whose reality obtains by the composition of many things [i.e., more than one thing], and a simple is that which is other than that. It is necessary to affirm a simple essence, otherwise all essences will be composed of infinite parts in actuality. As such, there must be a simple, since for every plurality, be it finite or infinite, the unit exists in it.⁴³

The above provides a very broad definition of simple and composite essences. Thus far, RāzÄ« makes no explicit reference to any specific ontology. The definition of simple and composite essences seems to remain general and inclusive of differing ontologies, including hylomorphism and atomism. However, as RāzÄ« proceeds, the analysis becomes more and more specific, addressing views of specific kinds of universals (e.g. genus and differentia) and their relation to ontological parts (e.g. matter and form). In the immediately following sections, RāzÄ« focuses on composition (tarkÄ«b) and the relation between simples and composite essences. In the fifth section, on how simples combine to form a composite essence, he begins with the trivial example of ‘human-plus-stone,’ “from which a unified essence fails to obtain since one is not connected with the other.”⁴⁴ He then considers medicinal paste (al-maÊ¿jÅ«n), which gives rise to certain effects (āthār) through a combination of material parts and a paste-form (ṣūra maÊ¿jÅ«niyya). In the Mulakhkhaá¹£, his examples include 10 composed of 1’s, an army of individuals, and a town of houses.⁴⁵ The analysis introduces a thin notion of matter and form, the latter of which he often refers to as the composing structure (al-hayʾa al-ijtimāʿiyya).⁴⁶ I return to aspects of these sections in Part 4 below, where postclassical thinkers offer a variety of readings.⁴⁷ To this point, the analysis establishes a general baseline regarding the composition of essences, namely that the parts of a (real) composite entity must, in some way or other, be (non-circularly) interdependent.⁴⁸ As evident in the subsequent discussions of genus and differentia, RāzÄ« is well aware that this is not the robust hylomorphism of falsafa.

Two chapters later, in the seventh section, RāzÄ« considers kinds (aá¹£nāf) of composites, which introduces distinctions regarding the relation between genus and differentia, including how general and specific things combine to produce a composite essence.⁴⁹ The discussion centers on the predicables of the Isagoge tradition but concludes with a reference to how composite essences are composed of causes and, specifically, the four Aristotelian causes, including form and matter.⁵⁰ He distinguishes between “intelligible” causes or parts, like form and matter with respect to body, and “sensible” parts, like bed composed of wooden parts and an order (tartÄ«b). He states that bed is an example of a “real relational” thing (iḍāfÄ« haqÄ«qÄ«), in contrast to “the more proximate” (al-aqrab) or “the more remote” (al-abÊ¿ad), which are mentally relational. In a bed, the wooden parts are real existents (mawjÅ«dāt ḥaqÄ«qÄ«yya) but they require the existence of order, which is a part of the bed but is a relational thing. Here, the language remains too brief and general to see what precisely distinguishes a sensible composite from an ‘intelligible,’ form-matter composite, like body.⁵¹ If one stops here in Rāzī’s analysis of essence in the Mabāḥith, one would conclude that he broadly endorses hylomorphism and follows Avicenna’s analysis of composite essences.⁵² The following sections of the chapter on essence in the Mabāḥith clarify that RāzÄ« strongly opposes a robust view of hylomorphism. To be sure, already in Book I, RāzÄ« will allude to his more systematic rejection of robust hylomorphism.

Following his introduction of universals and composition, RāzÄ« focuses on specific interpretations of genus and differentia, particularly Avicenna’s view of their relation to form and matter.⁵³ In the thirteenth section, RāzÄ« discusses how genus is “constituted” by differentia, which he states is an important inquiry that deserves attention. He outlines various distinctions relevant to how the differentia (e.g. rationality), as a part of the essence, constitutes and is a cause of the existence of the genus (e.g. animality).⁵⁴ Though he does not explicitly refer to hylomorphism here, he notes that animality requires prior preparation as the recipient (qābil) of the rational soul; that is, animality exists only potentially prior to the reception of the form of rationality.⁵⁵ According to RāzÄ«, the differentia-form is the formal cause of the composite substance (i.e. the species) but part of the efficient cause of the genus-matter.⁵⁶ Here, I set aside certain details of Avicenna’s more complex hylomorphic system, as my interest concerns Rāzī’s engagement with the fundamental principles of Aristotelian hylomorphism. In the fourteenth section, RāzÄ« directly addresses the hylomorphic status of genus and differentia and raises the following question regarding the form-matter analysis of living things:

Since it has been clarified that the genus is in need of [i.e. depends on] the differentia for its existence, it is impossible for the differentia to be in need of the [genus] in virtue of the absurdity of the circularity [this would entail]. Rather, [the differentia] must be independent of [the genus], for everything that inheres (ḥāll) in a thing is dependent on the substratum (maḥall) [in which it inheres]. Therefore, it is impossible for the differentia that divides the genus and is constitutive of the species to inhere in [the genus]. According to this, there is no problem in making the rational soul the differentia of ‘animal’ [as the rational soul, on Avicenna’s view, is immaterial]. But the problem is in making the faculty of growth (quwwat al-numuww) and its like a constitutive differentia of body. The same holds for the corporeal animal soul, because these attributes are dependent on substrata, which are bodies, but the substratum is prior in existence to the inhering [attribute], and it is impossible for that which is prior in existence to a thing to be the effect of that thing. We have discerned responses to that⁵⁷ which we will mention in the chapter on the connection of matter to form. Perhaps the truth is to hold that the subject of attribution (al-mawṣūf), whether it is the cause or the effect of the attribute (al-á¹£ifa), is the genus and the attribute the differentia, but if we state that, then the distinction between the division of genus by the differentia and the division of species by the proprium becomes invalid. We will mention what our chosen position is in the chapter on the connection of matter and form.⁵⁸

Here, Rāzī introduces, in a preliminary manner, problems regarding (Avicennan) hylomorphism, which he will address more fully throughout the Mabāḥith and Mulahkhaṣ. It bears noting that Rāzī specifies the question to the relation of form and matter in material substances, excluding the case of the relation of the immaterial rational soul to human body. That is, Rāzī is interested in how hylomorphic analysis is applied to perceptible, material substances, focusing on the case of living beings, i.e. the paradigmatic substances of Aristotelian ontology. Moreover, the passage focuses on how the form-matter relation, in a robust sense, differs from a substrate-attribute relation. That is, in the case of a bed discussed earlier, the wood as genus exists prior (and posterior) to the bed-form or order, which is the differentia. Here, the wood is not “in need for its existence” on the bed-form. However, the hylomorphism that Rāzī outlines in sections thirteenth and fourteen asserts that the genus as the proper matter of a material composite is dependent (and exists only potentially) without its form, as specified by the differentia. I return to Rāzī’s remarks on his own views in the above passage shortly.

In the corresponding section of the Mulakhkhaá¹£, RāzÄ« focuses on a remarkable point. After stating that Avicenna’s view is that the differentia is necessarily a cause of the species-part (ḥiṣṣat al-nawÊ¿) of the genus, he states, “According to me, that is not necessary, because an animal or plant body remains [in existence] posterior to the absence of the attributes [e.g., life] in virtue of which that body came to be an animal or plant. If those things were the causes of the existence of that body, that would not be the case due to the impossibility of the persistence of the effect with the absence of its cause.”⁵⁹ Here, Rāzī’s objection corresponds, in effect, to what modern scholars have called a fundamental problem of hylomorphism in Aristotle.⁶⁰ That is, RāzÄ« considers the case of the material body or the corpse of a living substance after its death or corruption.⁶¹ In Rāzī’s view, as far as we can see, the material constituents of animals and plants exist prior and posterior to their individual existence or life. Rāzī’s point is that the analysis of animals and plants need not radically differ from the analysis of ordinary objects, like bed, in which material parts are independent of form. The dependency relation posited in Aristotelian hylomorphism is peculiar in this regard; it is unclear how the composition and dependency relation of form and matter corresponds to our empirical knowledge of things, a problem that RāzÄ« emphasizes in Part II of Book II.⁶² Avicenna’s own theory, which has been called a “radical formalism”, is undoubtedly Rāzī’s immediate target.⁶³ However, a similar problem will inform his objections to Aristotle, to which I turn shortly below.

Two points can be highlighted from our discussion of Rāzī’s chapter on essence thus far. His analysis, from section one to fourteen, illustrates how his book on general concepts in the Mabāḥith and Mulakhkhaṣ is meant to function: he moves from general definitions of concepts and examples (including essence, genus/differentia and form/matter) to specific claims and disagreements, particularly those advanced by Avicenna and other falāsifa. Rāzī points us to the Avicennan view of the relation between the genus and the differentia of an essence, which, according to Rāzī, involves a robust (and problematic) hylomorphism. Note that, in the above passage from the fourteenth section, he refers to his fuller analysis of hylomorphism in Part II of Book II. That is, the above, general analysis serves to introduce central concepts and problems and anticipates fuller discussions in subsequent parts of the Mabāḥith and Mulakhkhaṣ.

Second, Rāzī does not simply set out received views and problems, but introduces his own voice and “chosen view” on topics. In the conclusion to the above passage of section fourteen, Rāzī states, “Perhaps the truth is to hold that the subject of attribution (al-mawṣūf), whether it is the cause or the effect of the attribute (al-ṣifa), is the genus and the attribute the differentia.” As we will see, this anticipates Rāzī’s anti-hylomorphic, attributive view, which he develops and applies throughout the Mabāḥith and Mulahkhaṣ. It is evident that Rāzī’s own view is not an interpretation of Avicenna but is meant as a foundational departure. In the above passage, Rāzī introduces a fundamental problem regarding the form-matter relation but then refers to a solution, which he takes to be his own chosen view. Even this much strongly suggests that Rāzī is not simply expounding or following Avicennan or Aristotelian views and ontology. But more on this shortly.

As noted, Rāzī refers to a following discussion on form and matter in the fourteenth section. It seems that Rāzī means his better known discussion of Avicenna’s hylomorphic analysis of Absolute Body in Part II of Book II, which assesses various aspects of robust form-matter dependency. However, Rāzī provides important discussions of form and matter in earlier sections of the Mabāḥith and Mulakhkhaṣ, particularly in Part I of Book II on accidents, which to my knowledge has not been addressed in scholarship. The discussion is of particular significance as Rāzī directly addresses Aristotle and hylomorphism in the context of physics.

It has been thought that RāzÄ« is unaware that form and matter was introduced by Aristotle. As the following shows, RāzÄ« not only takes Aristotle to be the progenitor of hylomorphism but seems to be aware of its original context in physics and its relation to Presocratic thinkers. In his extensive chapter on motion in Part 1 of Book II, RāzÄ« raises the question of the relation of change to substantial forms in the Mulakhkhaá¹£.⁶⁴ The problem of “substantial change” (al-ḥaraka al-jawhariyya) becomes a prominent topic of debate in metaphysics and theology in postclassical Islamic thought, and RāzÄ« plays a central role in defining the problem.⁶⁵ However, this should not obscure the fact that the question RāzÄ« begins with centers on Aristotle’s critical distinction between substantial change (or generation and corruption) and accidental change. In a section of his chapter on motion in the Mulakhkhaá¹£, RāzÄ« states the following:

[The impossibility of motion applying to] substance according to our view is clear because substance is that which is self-subsisting (al-qāʾim bi-l-dhāt) and motion only applies to attributes (ṣifāt), as we have specified. However, on Aristotle’s view, since he affirms substantial forms (al-ṣūra al-jawhariyya), evidence for this is necessary and clarification is needed with respect to two points (maqāmayn). [A] The first point is that it is possible for substantial forms to cease to exist with their matter with respect to the [four] elements and the dispute is with those who deny generation and corruption. The [Aristotelian] philosophers (ḥukamāʾ) permit this [i.e. the corruption of substantial forms] only with respect to the four elements but not with respect to the celestial orbs or planets.

[B] Know that since the elements are four, the kinds of generation are twelve, except that if we affirm [only] three of them—namely, earth becoming water, water becoming air, and air becoming fire—the aim has been obtained with respect to all [i.e., all four substances], because this much entails the claim that matter is common to all [four]. They aim to prove that earth turns into fire on the evidence of the mechanics (ahl al-ḥiyal). […] Know that such [evidence] concerns particular things, which do not exclude various possibilities, so this does not provide a demonstrative result. We have mentioned those possible alternatives in the Mabāḥith. Indeed, what is established there is that bodies are equivalent with respect to corporeality and differ with respect to attributes [i.e., they do not differ with respect to form]. […] This method [based on the commonality of body] is strong and there are subtle inquiries regarding it that we have mentioned in the Nihāyat al-ʿuqūl, but those do not accord with the principles of the falāsifa, because they deny generation and corruption.

[C] The second point (al-maqām al-thānÄ«) concerns the point that it is not possible for the generation of these forms to occur in a gradual manner. This job is directed at them, not at us, and they have sought to prove this by three ways.⁶⁶

Rāzī begins with his own view on substance and body. As is clear from the above and related discussions, what he means by substance being “self-subsisting” is that it is not composed of the two further substantial components of corporeal form and matter. That is, substance is a substrate and is a subject only to attributes. That is, it is not a subject to forms in the sense of Aristotelian matter as a potential, indeterminate underlying thing. Moreover, he affirms what I have proposed above regarding the status of Body 1, namely, that it is the basic ontological or conceptual layer and the subject of sensible properties, including those belonging to the four elements (more on this below).

More significantly, RāzÄ« directly contrasts his own view with that of Aristotle. That is, his position not only departs from “principles of the falāsifa”, as he states in the passage, but from Aristotle’s position on substantial forms and change. Regarding the passage, my focus will be on [A] in the following discussion. However, two general points that RāzÄ« raises in the following paragraphs of the passage need noting. In [B], RāzÄ« highlights the demonstrative or epistemic status of the Aristotelian claim regarding the nature of the four elements. He asks how we know whether the change from one element to another (say, from water to air) involves substantial change rather than accidental change.⁶⁷ Again, RāzÄ« invokes his own solution to the problem, which suggests the latter reading: substances can be viewed as composed of corporeality or Body 1 plus attributes (heat, cold, etc.), rather than matter plus substantial form.⁶⁸ He will have significantly more to say about this, as indicated in his reference to the lengthier discussion in the Mabāḥith and Nihāyat al-Ê¿uqÅ«l. In [C], RāzÄ« refers to the question of the generation and corruption of substantial forms. According to Avicenna’s interpretation (which RāzÄ« takes to be a correct reading of Aristotle’s view), substantial forms do not undergo generation and corruption gradually but ‘all at once,’ in contrast to change in accidents (i.e. alteration).⁶⁹ In Part 3, I return to Rāzī’s own view of change elucidated in a text immediately following the above passage, which involves a radical relational view of the perception of accidents and accidental change.

In the passage, RāzÄ« highlights two problems that specifically address Aristotle. In [A], RāzÄ« asks how we know that elemental change involves substantial and not simply accidental change. Importantly, RāzÄ« identifies Aristotle’s interlocutors as the “deniers of generation and corruption.”⁷⁰ Who, precisely, are they? As suggested in Part 1, RāzÄ« is referring here to the Presocratic philosophers and particularly the substance monists, whose ontology Aristotle opposes in Physics 1 with reference to form and matter.⁷¹ In the Mabāḥith, RāzÄ« explicitly identifies the relevant interlocutors as substance or element monists.⁷² Elsewhere, RāzÄ« identifies by name a variety of Presocratic philosophers, including monists such as Thales, Anaximenes, and Heraclitus.⁷³ I will set aside the details of the discussion for a future study.⁷⁴ However, that RāzÄ« addresses Aristotle’s theory of form and matter in the context of the Presocratic monists is significant for two reasons.

First, RāzÄ« highlights a critical philosophical problem that the Aristotelian hylomorphist must address. That is, Aristotle’s monist predecessors believe that all phenomena can be explained with reference to a singular, underlying substance (e.g., fire, earth, water, etc.). The varying attributes and composite entities that we observe in the world are obtained by change (e.g. condensation or rarefaction) that occurs in some primary substance, which itself cannot undergo generation and corruption. That is, the Presocratic monists deny that primary substance, whatever it may be, undergoes generation and corruption. Rather, generation and corruption applies only to the attributes or properties that we observe and is a result of change that occurs to the primary substance itself. As such, RāzÄ« believes that the onus is on Aristotle to prove that when one element changes into another, the change in fact involves substantial change (i.e. generation and corruption, or motion in substantial form) and not simply a change in accident (i.e. alteration). The Presocratic monists, as noted, explain all such change with reference to a single underlying substance.⁷⁵

Rāzī’s discussion is part of a larger analysis of the distinction between substantial change and accidental change. The central point that emerges from Rāzī’s discussion of motion and substantial form is not that he locates a fatal flaw in Aristotelian hylomorphism—for, as he knows, there is more to say on the matter. Rather, the analysis shows that RāzÄ« is almost certainly aware of the original context of form and matter in Aristotle’s Physics and the basic intuitions underlying their introduction.⁷⁶ Moreover, he directly contrasts his own view with the Aristotelian view at a fundamental level. In Part 3, I return to Rāzī’s discussion of the Aristotelian view of the elements, and specifically the ‘fundamentality’ of the four elements, in contrast to his own attributive view.

I turn now to Rāzī’s preliminary analysis of substance and accidents in the introductory chapter to Book II. The discussion precedes his analysis of motion in Part I; the analysis follows and elaborates on the general concepts and problems discussed above in Book I. In the fifteenth section of the introduction, on whether one thing can be both substance and accident, RāzÄ« states that the problem applies especially to substantial forms. He notes that the established view attributed to Avicenna is that the substantial form cannot be a substance in one respect and an accident in another. He then states that opponents of the former view rely on three arguments to show that a substantial form can be viewed as a substance and as an accident. After raising objections to the three arguments, he states that the following is a “better” argument for the latter position. The discussion is clearly aporetic but points us to important aspects of Rāzī’s positive approach.⁷⁷ He states:

[A] If a thing inheres in a thing, then that inhering thing (al-ḥāll) has [one] aspect (iÊ¿tibār) insofar as it is in that substrate (maḥall) and [another] aspect insofar as it is in that collection (majmÅ«Ê¿) [i.e. the inhering thing plus substrate]. For example, heat has an aspect insofar as it in the ‘hot thing’ (al-ḥārr) and another aspect insofar as it is in a body. As for the aspect of heat being in the hot thing, it is clear that that does not entail accidentality [i.e. of heat to the hot thing] since heat is a part of the hot thing, whereas it is a condition of [being an] accident that it is not a part of the subject (mawḍūʿ). As for the aspect of [heat] being in the substrate, then it is either rationally conceivable that there is a substrate that can be constituted by that which inheres in it or not [i.e. such a thing is not rationally conceivable].⁷⁸ The first is false for two reasons. […]

[B] What further supports and clarifies that is that we see that if animality (ḥaywāniyya) ceases to exist, then its corporeality [i.e. its matter or body] does not cease to exist without that animality. Also, they hold that animality is composed of genus and differentia, the genus being corporeality and the differentia the animal form (al-ṣūra al-ḥaywāniyya). And they state that the animal form is constitutive of that body. This is false because that animal power (al-quwwa al-ḥaywāniyya) inheres in that body in the manner that an accident inheres in its substrate, so that if that body is in need of the animal form, then the dependency circles unto itself [i.e. is circular]. Thus, how is it rationally conceivable to state that that body is constituted by that animal form? It is clear from what we have said that the constitution of any substrate by a thing that inheres in it is impossible.⁷⁹

The above passage recalls the discussion in section fourteen of the chapter on essence in Book I, discussed above. Here, RāzÄ« provides an argument (the details of which I have set aside) against form-matter dependency, where form constitutes, in some way or other, its substrate.⁸⁰ Rāzī’s own view is, again, expressed in [B] as a substrate-plus-accident view. That is, whatever the attribute might be, the substrate or substance is a “self-subsisting” thing. Here, RāzÄ« provides two notable examples. The first concerns the relation of hot and body and, the second, the nature of animate powers. In the former, the heat in the ‘hot thing’ is not accidental to the composite ‘body-plus-heat’. This, as we will see, is the model RāzÄ« will use to interpret the four elements, not as fundamental substances but as empirical bodies. In Rāzī’s eyes, a substance-plus-accident can be viewed as producing a new substance. What distinguishes it from arbitrary composites or accidental clusters of matter (e.g. human-plus-stone) seems to be their relation to, and manifestation of, specific, empirical properties. This view, as RāzÄ« points out elsewhere, violates a central tenet of the falāsifa’s hylomorphism, since a true substance cannot be constituted of a substance and accident; rather, all true composite substances are constituted of form and matter.⁸¹

The second example he discusses is the case of a more complex object, namely animal, which he analyzes as ‘body-plus-animal-power.’ Here, animality is not a substantial form but an attribute of some kind that inheres in body. As he states, “animal power (al-quwwa al-ḥaywāniyya) inheres in that body in the manner that an accident inheres in its substrate.” Notably, RāzÄ« sees attributes as encompassing a broad category. As we will see, RāzÄ« views powers as empirical phenomena, more akin to contemporary views of ‘capacities’ or ‘dispositions.’ That is, RāzÄ« neither denies their existence outright nor does he equate powers with forms or natures, as do earlier Peripatetics, including Avicenna.⁸² The above, then, provides the basic elements of Rāzī’s attributive analysis of empirical things. Though I return to details below, it can be noted that at the end of the passage on motion and substantial form in the Mulakhkhaá¹£ discussed above, RāzÄ« explicitly provides his own view of the relation of substance to attributes and capacities, including living beings.⁸³ What is significant is that Rāzī’s alternative contrasts with earlier approaches, which first set out a specific, fundamental ontology (particularly, at the level of Body 1) on the basis of which the subsequent analysis of physical reality is conducted. In the following, RāzÄ« will not be providing an alternative insofar as it is a competing, fundamental ontology to the various views he addresses above, from the Greek monists and hylomorphists to kalām atomists. Rather, from a general, attributive view, RāzÄ« develops a conceptual and empiricist frame of inquiry for analyzing higher-order aspects of physical reality. I return to further details of Rāzī’s empiricism or indirect realism in the conclusion to Part 3.

In the Mulakhkhaá¹£, RāzÄ« provides a more detailed analysis of his position on the relation between substrate and attribute, including cases with multiple and interrelated attributes.⁸⁴ A focused analysis and application of these distinctions will more precisely reveal the structure of Rāzī’s attributive system. However, in the following, I focus on the more accessible and concrete examples of substance-plus-attribute things in Part II of Book II. In doing so, I focus on the epistemic themes guiding his inquiry, rather than the details of specific arguments, which cannot be fully addressed here.

Regarding Rāzī’s attributive analysis, it can be noted that in Book I of the Mulakhkhaá¹£, regarding the nature of composition in external reality, RāzÄ« states: “The correct [view] is that there is no meaning to the predicability of a thing but its being an attribute (á¹£ifa) of a [subject].”⁸⁵ The point is meant to oppose the Avicennan view of the differentia as an essential, constitutive part of an essence and the nature of per se predication.⁸⁶ I raise the point for the following discussion in order to contextualize Rāzī’s attributive view. The intuition behind his attributive and substance-plus-accident view, I propose, is best understood as inspired by kalām and the tradition of Arabic grammar.⁸⁷ That is, its intuitions lie in the substrate-plus-attribute (or maḥall-plus-á¹£ifa) framework of the latter traditions without adopting the reductive analysis of kalām atomism. Moreover, Rāzī’s engagement with a broad spectrum of substances and empirical phenomena, within a more systematic and complex logical system, allows him to develop a more comprehensive physics.

Tracking the longer arch of Rāzī’s argument, I turn now to his analysis of substances in Part II of Book II, focusing on his approach to the four elements of Aristotelian cosmology and more complex substances (i.e. Body 2).

3 Rāzī on the Four Elements: Empirical versus Hylomorphic Substances

In Part 1 above, we noted that RāzÄ« begins with the analysis of body as extension (Body 1) in his part on substances, i.e. Part II of Book II of the Mabāḥith and the Mulakhkhaá¹£. He then turns to discussing more complex substances (Body 2), beginning with the four elements and celestial matter in chapter two of Part II. As a preliminary to that latter chapter, in the thirteenth discussion of chapter one, RāzÄ« discusses the question of “affirming natural forms” (ithbāt al-á¹£uwar al-á¹­abÄ«Ê¿iyya), by which he means the species-forms of the four elements.⁸⁸ He begins by stating: “After sharing in corporeality, some bodies are receptive of different shapes with ease, which is moisture, while others are receptive to them with resistance, which is dryness, […] and we have indicated in the chapter on powers that differences in accidents are necessarily in virtue of differences of forms that are located in those bodies.”⁸⁹ As is clear, RāzÄ« begins with Avicenna’s form-matter view of the four elements, but as we will see, he will quickly problematize it.

Rāzī’s cross-reference to his discussion of powers refers to his preceding analysis in Part I of Book II. Two points may be highlighted here without getting into the details of that discussion.⁹⁰ First, he summarizes Avicenna’s fourfold division of powers, the centrality of which to Avicenna’s natural philosophy has recently come to scholarly attention.⁹¹ Second, RāzÄ« notes that according to Avicenna’s doctrine (madhhab), inanimate powers include those that are constitutive of their substrate, such as the “form, say, of fire-ness and so forth,” which RāzÄ« says is called ‘nature’ (al-á¹­abÄ«Ê¿a).⁹² While some identify sensible qualities with forms, Avicenna argues that “it is necessary that these qualities are concomitants and attributes of constitutive forms, and these forms necessitate by nature these qualities.”⁹³ RāzÄ« then states that his own view is that powers are those “that inhere in a simple body, like fire-ness, and those that are composite, like nutritive and medicinal mixtures.”⁹⁴ As noted, for RāzÄ«, powers are empirical objects that exhibit certain dispositions and capacities. As we will see, he rejects the notion of powers as natures and forms, which is central to Avicenna’s approach.⁹⁵ Let us return to his analysis of the elements in Part II.

In the Mabāḥith, RāzÄ« begins the corresponding section by stating that it is customary to discuss the point that specific attributes of bodies (e.g. their quality, place, and shape) are due to powers attributed to those bodies. RāzÄ« refers here to the arguments that Avicenna develops in his De Caelo et Mundo.⁹⁶ Notably, he then states that he is interested in a different question: “The philosophers have concurred that the powers that are the principles (mabādiʾ) of that which [inheres] in bodies in terms of quality, place, and shape, are forms rather than accidents, but they have not provided a proof for that, so it is necessary for us to investigate it.”⁹⁷ As Cerami shows, Avicenna begins “abruptly” and assumes form and matter in his analysis of elements in De Caelo et Mundo. Rāzī’s point is notable when read in the context of the attributive view discussed above. He claims that the ontology of form and matter is taken for granted in the analysis of the elements. Moreover, the overlooked alternative—i.e., forms rather than accidents—is one that would accord nicely with his attributive, anti-hylomorphic approach. At the end of the above chapter, RāzÄ« makes explicit this very point. In the following passage, which TaftāzānÄ« and others will later quote almost verbatim, RāzÄ« affirms his own preference for an attributive approach, in contrast to the hylomorphic analysis:

Know that that which has been obtained by proof is the grounding (istinād) of these attributes, like place, quality, and others, in powers that exist in a body, which are conserved in themselves (maḥfÅ«zat al-dhawāt) and which return the body to these qualities upon the cessation of constraints and obstacles. As for whether those things are causes for the existence of corporeality, so that they correspond to constitutive forms (á¹£uwar), or are not, so that they correspond to attributes (aÊ¿rāḍ), is something that has not been established by demonstration. What is closer [to the truth] according to us is not to make these causes of corporeality and to not count them as forms but rather as attributes.⁹⁸

I return shortly to Rāzī’s more detailed statements on the attributes of the four elements. In the above, it is notable that Rāzī does not deny powers; rather they are perceptibly evident. This supports the proposed reading that Rāzī views powers as empirical capacities in certain substances or kinds of bodies that give rise to certain attributes or effects. A parallel example he considered in Book I was medicinal paste and its effects. The question that Rāzī considers is whether such capacities must be viewed as caused by, or grounded in, constitutive forms. Here he means specifically the species forms (al-ṣuwar al-nawʿiyya) of the elements, as clarified below. He states that there is no demonstrative proof for the elemental forms. His own view, as indicated by “according to us,” is that the relevant properties are attributes. Given the previous discussion, it is reasonable to think that these attributes inhere in a simple body in the manner that accidents inhere in substances. Rāzī will have more to say about this, but first I turn to earlier debates regarding the nature of the elements.

There is a long history of debate regarding the relation between substances and accidents in the ancient commentarial tradition, and particularly concerning the status of differentiae as constitutive of hylomorphic compounds.⁹⁹ The immediate context for the present discussion is not a dispute between Avicenna and earlier Peripatetics, but rather a dispute between Galen’s and Aristotle’s views of the elements. As Abraham Stone aptly states the problem:

First, there is the distinction between (1) [the elemental substances] and (2) [the primary qualities (hot, cold, moist, and dry)]—if, indeed, there is any such distinction to be made. This issue particularly exercises Galen, because Hippocratic texts sometimes speak of bodies as complexions of “powers” (δυνάμεις)—among which are the primary qualities—without mentioning any species of body in which they inhere, and this was taken literally by some of Galen’s opponents. For Aristotle, however, the elements are fire, air, water, and earth, which he describes as species of body and of substance, while hot, cold, moist and dry are species of quality, which as such can never be found without substance. Thus, Galen assures us that the Hippocratic way of speaking merely involves naming the elements by the qualities which they have in the extreme; there cannot, strictly speaking, ever be qualities without underlying bodies.¹⁰⁰

As Stone clarifies, Galen’s view was taken by his opponents to be in direct conflict with Aristotle’s view that accidents cannot constitute a substance as parts and that substances, including the elements, cannot simply be a collection of accidents in a body.¹⁰¹ Notably, Galen sees his own view as an interpretation of Aristotle’s view of the elements and not as an alternative to it.¹⁰² Galen believes, like the Peripatetics, that the four elements are the fundamental constituents of living and non-living things.¹⁰³ By contrast, RāzÄ« rejects the view that such properties are attributes or propria caused by substantial forms. Indeed, he rejects the notion that the four elements are the fundamental kinds of simple bodies in the sublunary world. The above reading is supported by the fact that RāzÄ« distinguishes his attributive analysis not just from Avicenna, but from Aristotle and Galen. In Sharḥ Ê¿UyÅ«n al-ḥikma, RāzÄ« states,

As for the view taken by Aristotle and Galen, and the majority of the falāsifa and physicians, it is that the four elements are not generated from [simpler substances] other than them, and what is other than [the elements], in terms of the bodies of this world, is generated from them.¹⁰⁴

RāzÄ« then discusses two distinct methods of establishing that the four bodies are indeed usá¹­uquṣṣāt, i.e. ontologically basic elements of the sublunary world.¹⁰⁵ The first method is the method of the physicians, which he labels the method of “composition and decomposition” (á¹­arÄ«q al-tarkÄ«b wa-l-taḥlÄ«l). The second method is that of “Aristotle and his followers.” Rāzī’s analysis of the two methods is beyond the scope of the present study. What is of significance is, first, that RāzÄ« knows that the cosmology of falsafa, which holds that the four elements are the fundamental substances of the sublunary world, is the view of Aristotle and his followers.¹⁰⁶ This is notable, because RāzÄ« does not refer to Aristotle and Galen in this context in either the Mulakhkhaá¹£ or the Mabāḥith.

Second, RāzÄ« knows that the Aristotelian view aims not simply to establish that there are four elements, i.e. that four kinds of elements are empirically known. Rather, the Aristotelians must prove that these four elements are ontologically basic kinds or species of bodies, from which all other composite substances are formed. He knows, moreover, that Aristotelian cosmology—including the notion of a singular cosmos and its form—hinges on this doctrine.¹⁰⁷ In Sharḥ Ê¿UyÅ«n al-ḥikma, RāzÄ« does not clarify his own approach, but he does emphasize that the arguments are inductive and that we cannot prove that the elements are the fundamental constituents of things. Regarding the physicians’ method, he notes, for example, that we do not observe gold as composed of the four elements, or that on exposure to heat, it decomposes into the four elements, and the same can be said about pearl.¹⁰⁸ Here, RāzÄ« refers us to his work on medicine, al-Ṭibb al-kabÄ«r, where he claims to have discussed this method extensively.¹⁰⁹

Let us now turn to Rāzī’s discussion of the above methods in the Mulakhkhaṣ and Mabāḥith. In these texts, Rāzī provides a more detailed discussion of the two methods. In the former work, in his discussion on whether “composites are produced from these four,” Rāzī focuses on two methods: the physicians’ method of composition and decomposition, and a “second method.” The second method is a summary of Avicenna’s argument in the Physics of al-Shifāʾ, a longer version of which Rāzī also examines in the Mabāḥith. In the Mulakhkhaṣ, Rāzī concludes with the following rejoinder:

The truth regarding this topic is that specialized induction (al-istiqrāʾ al-Ê¿urfÄ«) indicates that all complete generated things (al-mutawalladāt al-tāmma) only come to be upon the obtaining of these four. As for whether they [i.e. all four elements] are together the parts of those composite things or only some [of the four elements] as opposed to others are [their parts], and whether composites from [elements] other than these [four] are rationally possible, that is something that has not been established in any way by evidence.¹¹⁰

There are several significant points in Rāzī’s passage that require attention. First, he states that the truth is that inductive methods show that composites are generated from the four elements. By this, RāzÄ« means that empirical evidence establishes the contingent truth that observable composites are generated from the empirically known substances. In the Mabāḥith, RāzÄ« states that he prefers the inductive and empirical approach of the physicians over the deductive method of Avicenna, to which I return shortly.¹¹¹ Moreover, RāzÄ« opposes two claims regarding the nature of the four elements. The first is that we cannot empirically establish the precise combination of elements in actual composite things. This point might be conceded by the Aristotelians. Rāzī’s second point, however, opposes Aristotelian cosmology at a fundamental level, for he states that the evidence does not prove that it is impossible for generated things in the sublunary world to be composed of elements other than the four. That is, RāzÄ« here parses finely between the four elements conceived as contingent, empirical objects of analysis, and the four fundamental elements of Aristotelian ontology. As we will see, corresponding to this distinction, RāzÄ« distinguishes between empirical methods and the deductive or ‘rational’ methods that aim to support the Aristotelian position.

In the Mabāḥith, Rāzī’s corresponding section is titled “On the elementality of these four” (fÄ« usá¹­uquṣṣiyyat hādhihi al-arbaÊ¿a).¹¹² Here he addresses three approaches to proving the Aristotelian view.¹¹³ The first is the “method of the physicians” and the second is Avicenna’s approach in the Shifāʾ.¹¹⁴ RāzÄ« restates Avicenna’s argument in al-Kawn wa-l-fasād (Generation and Corruption)¹¹⁵ and then raises a variety of objections to it, including methodological points.¹¹⁶ Rāzī’s engagement with Avicenna’s arguments regarding the primary qualities is extensive. He again refers to and builds upon his preceding discussion in Part I of Book II of the Mabāḥith and Mulakhkhaá¹£, specifically on the nature of primary sensible qualities (al-kayfiyyāt al-maḥsÅ«sa al-Å«lā).¹¹⁷ An analysis of the content is well beyond the scope of this discussion but the following methodological points can be highlighted.

Regarding the correct combination of primary qualities in the Aristotelian view (i.e., hot-dry, hot-wet, cold-dry, and cold-wet), RāzÄ« asks whether the premises on which the theory is based rest solely on inductive knowledge or whether they are proven by ‘rational’ argument. To this, he quotes Avicenna as responding in the following manner: “Our affirming of the existence of the four elements is not based entirely on pure [rational] division but, rather, on a division that is followed by existence [in external reality], and nothing is more clear than that which the intellect indicates and existence then proves.”¹¹⁸ RāzÄ« then proceeds to raise further problems regarding Avicenna’s claim that rational division is supported by what ‘exists,’ i.e. what is known by observation (mushāhada). Here he sharply divides between accidental or attributive qualities and natural qualities, e.g. attributive heat (al-ḥarāra al-Ê¿araḍiyya) and natural heat (al-ḥarāra al-á¹­abÄ«Ê¿iyya). Avicenna’s argument aims to show the necessary connection between natural heat, which is the necessary property of the substance fire, and natural dryness. Against this, RāzÄ« argues that we may observe instances of the ‘attributes’ of heat and dry together, but we do not observe the ‘nature’ of heat as having the quality of dryness. Indeed, we do not observe any of these qualities in the sense required, i.e. as natures, and “if we do not observe that, then there remains only the rational division on its own, and you admit that that does not prove existence [i.e. in external reality].”¹¹⁹ Again, RāzÄ« sharply distinguishes between empirical knowledge and the essentialist claims about the nature and qualities of hylomorphic substances.¹²⁰

In his final analysis of Avicenna’s argument, RāzÄ« states that despite the prevalent adoption of Avicenna’s arguments in this context, they are weak and fail to go beyond what is proven by induction. He advises that it is therefore better to rely on the first method, i.e. the method of the physicians. In other words, since Avicenna’s argument falls short of its claim to be a rational demonstrative proof, one is better served by relying on the established empirical method, which is “closer to obtaining [results] (al-taḥṣīl) and precision (al-ḍabá¹­), and farther from conflation and confusion.”¹²¹ Rāzī’s own view of the elements seems to draw heavily on medicine, alchemy and the empirical sciences.¹²² I return to his analysis of alchemical compounds below.

Following the above discussion of the three methods, Rāzī concludes by connecting the discussion of elemental natures with what he sees as the overarching problem with Aristotelian cosmology:

In my view, the truth regarding this topic (al-bāb) is that whoever attempts to expound an [exhaustive] limit (bayān al-ḥaá¹£r) for these elements by means of a rational division (taqsÄ«m Ê¿aqlÄ«) has attempted something that is not possible for one to fulfill. Rather, when people investigated [bodies] by means of composition and decomposition, they found that the composition of generated things begins with these four and their decomposition terminates with these four. Furthermore, they did not find these four as generated by the composition of other, [simpler] bodies, nor do they decompose into other, [simpler] bodies, and so they naturally believed that the elements (al-usá¹­uquṣṣāt) are [none other than] these four, not because a rational proof (ḥujja Ê¿aqliyya) was established that it is not rationally possible for there to exist an element other than these [four]. Rather, [they did so] because the evidence proves only these, and this is similar to when we judge that the orbs are nine, not because a rational proof has been established that it is rationally impossible for there to exist a tenth orb. Rather, it is because observation (raá¹£d) has only happened upon these nine. This is the truth regarding this topic.¹²³

RāzÄ« draws a parallel between how Aristotelians view the microcosmic status of the four elements and the macrocosmic question of the nature and limits of the cosmos. He emphasizes the allegedly ‘rational’ nature of the arguments, which contrasts with empirical methods, including induction and the method of composition and decomposition. Rational methods aim to achieve an exhaustive understanding of the natures of things and their essential properties.¹²⁴ In various places in the Mulakhkhaá¹£, RāzÄ« indicates that the division between rational arguments and empirical evidence has a parallel in the division between ‘demonstrative’ proofs for why such and such is the case (limmÄ«), and non-demonstrative arguments for the fact that such and such is the case (innÄ«).¹²⁵ Moreover, RāzÄ« provides a similar story with regard to the ancient discipline of astronomy and the principles of later natural philosophers (al-á¹­abÄ«Ê¿iyyÅ«n al-mutaʾakhkhirÅ«n). On the basis of inductive knowledge and transmitted records, ancient astronomers came to believe that the heavens are of unchanging nature; then, later natural philosophers sought to ‘contrive’ demonstrative proofs in support of this belief.¹²⁶ It can be noted that in such discussions of the Mabāḥith and the Mulakhkhaá¹£, RāzÄ« does not refer to AshÊ¿arite theological views, although the connections are rather clear. As we will see, the connections become more explicit in postclassical works of kalām. In this way, Rāzī’s philosophical views are assimilated into the later, expanded version of kalām. Before turning to these postclassical works, I conclude with points concerning Rāzī’s positive philosophical agenda and the long arch of his argument.

Rāzī’s attributive approach can be viewed as part of a larger anti-essentialist philosophical project set against Aristotelian approaches in falsafa, and perhaps more generally, the Greek tradition of natural philosophy. In the logic of the Mulakhkhaá¹£ and elsewhere, RāzÄ« develops a systematic critique of central concepts in Aristotelian logic and epistemology, which takes a remarkably critical stance to the role of essential kinds and necessary properties in definitions and deductive inferences. The above analysis points to connections between his criticism of hylomorphism, conducted in the substantive philosophical analysis of Book II, and his anti-essentialist epistemology developed in logic. As we have seen, RāzÄ« repeatedly points out that the form and matter of things—as required in the Aristotelian analysis of substances—are epistemically inaccessible. It seems clear, to my mind, that RāzÄ« intends his anti-essentialist program in logic to prepare for the analysis of substantive philosophical views, and particularly for his own approach which emphasizes an empiricist method over the essentialist approach of the Aristotelians. But gestures toward empiricism do not yet amount to a positive alternative.¹²⁷

4 Postclassical kalām and Rāzī’s Alternative Ontology

The above discussion outlines key ideas and passages in the Mabāḥith and Mulakhkhaṣ that inform later kalām approaches to the analysis of sensible reality. Postclassical thinkers draw on Rāzī’s frame of inquiry in order to formulate novel views of complex substances. In the above, I have focused on Rāzī’s analysis of the sublunary world, and especially his systematic distinction between the four elemental substances, considered as fundamental parts of the sublunary world, and the elemental simples, considered as empirical entities. I turn now to the postclassical thinkers’ appropriation of Rāzī’s views, with a corresponding focus on their analysis of complex essences.

In the chapter on essence in Book I discussed above, Rāzī devotes several chapters to the analysis of the relations that obtain between simples and complexes. Of particular concern to him is the question of whether various kinds of composites can be viewed as possessing a unity, and how simples constitute composites. In section five discussed above (titled “On the manner of the composition (ijtimāʿ) of the composite essence by simples”), Rāzī assesses the dependency relations that hold between the parts of an essence and the essence as a whole. More specifically, he is interested in assessing whether simple parts can form more than accidental composites. As noted, he begins with the example of a rock lying beside a person, the two of which form a composite without being, as he puts it, a “unified reality” (ḥaqīqa muttaḥida). He then provides other examples of composite unities:

As for the composition (takawwun) of ten of the units it contains, [the composition] of paste (maÊ¿jÅ«n) through the composition (ijtimāʿ) of medical ingredients (al-adwiya), the [composition] of an army of individuals, and the [composition] of a village of houses, they are [all] due to a unifying structure (al-hayʾa al-ijtimāʿiyya), which is one of the parts of the composite. It is the formal part (al-juzʾ al-ṣūrÄ«) on which the rest [of the parts] depend.¹⁴⁰

As we will see, in postclassical kalām, these and additional examples of composites that Rāzī provides elsewhere become the paradigmatic examples. Most important will be the example of paste and the concept of the ‘composing’ or unifying structure, to which I return shortly. The above examples of composite unities are striking in several respects. First, they are not compelling examples of ontological unities from the perspective of either kalām or falsafa. Indeed, Rāzī’s examples depart from the more obvious examples of composite unities in falsafa and Aristotelian philosophy, where the paradigmatic cases are the substances of living species (biological kinds), the four elements, and the celestial intellects. Rāzī’s examples of artifacts and numbers are not ontological unities for the falāsifa.

The above passage raises questions regarding what RāzÄ« means by identifying the unifying structure as the “formal part” and the components or ingredients as the material part. As I have argued elsewhere, the unifying structure is a mind-dependent concept that identifies ‘phenomenal’ properties.¹⁴¹ There is, however, more to be said about it, for RāzÄ« addresses a variety of relations between material parts and form in the Mabāḥith and the Mulakhkhaá¹£, which attempt to explain, not a deeper ontological unity, but kinds of empirical unity.¹⁴² For example, a paste brings together ingredients that are otherwise independent, not by virtue of a substantial form, but by virtue of applied processes, such as mixing and cooking, that produce the paste. RāzÄ« responds to an objector, who denies the need for dependency to obtain between the parts of paste:

If it is stated: Is it not the case that the [medicinal] paste is formed by a composition of parts, each of which is independent of another. We state: This is not the case. Rather, the collection of those parts is like a single part of the entity, which is the material part. As for the other part, which is the paste-form (ṣūra majÅ«niyya) that is the principle of the effects manifested by it, it is dependent on the first part [i.e., the material part].¹⁴³

Recall that the analysis up to this point remains general, neither endorsing a robust hylomorphism nor problematizing it. Rāzī’s notion of the unifying structure seems to be a general term that avoids the ontological implications of Aristotelian substantial forms.¹⁴⁴ I will now turn to how postclassical thinkers put it to use.

In his commentary on al-Ījī’s (d. 1355) Mawāqif, al-SharÄ«f al-JurjānÄ« discusses essences in the second chapter of the book on general concepts.¹⁴⁵ The discussion begins with the example of the non-unity of ‘human-and-stone’ and proceeds to analyze the examples of composite unities, including those that RāzÄ« provides above, namely, the medicinal paste, the army, and number.¹⁴⁶ JurjānÄ« begins by stating that the composition of a (real) composite essence (al-māhiyya al-murakkaba) requires its parts to be interdependent, “for if each (of the parts) were independent of the others, a single essence would not obtain (as a real unity).”¹⁴⁷ That is, if there were no dependency relations between the parts of a thing, a true unity or essence would not obtain—a point underscored by RāzÄ« in the above text. Here, JurjānÄ« considers an objection to the examples of the paste and the army, since both consist of units or individuals that seem to be independent of the other parts. He then examines an initial response: “The formal part in both (which is the unifying structure [al-hayʾa al-ijtimāʿiyya] that is attributed to each individual or the units as a whole) is dependent on the material [part].”¹⁴⁸ Again, the point is precisely that made by RāzÄ« above. Still, kalām thinkers will raise problems regarding how precisely to understand the distinction, which remains rather vague. Addressing Ījī’s dissatisfaction with the answer, JurjānÄ« raises an obvious objection, stating that this response is weak because a ‘mental’ or ‘conceptual structure’ (al-hayʾa al-iÊ¿tibāriyya) can be applied to ‘human-and-stone’ as well, but this does not make the latter a real or mind-independent unity. JurjānÄ« then comments on Ījī’s more “apt” response:

It is more apt to say in response: As for the paste, it is necessary that a mixture obtains in it (that is, a species form that applies to the mixture), which follows the qualities (and effects produced from it) and which (that is, that mixture, in the sense of form, is part of the paste) is dependent on the parts (by virtue of the [form’s] inhering in the parts).¹⁴⁹

The passage clearly follows Rāzī’s mereological analysis, though it is unclear how Jurjānī views (robust) hylomorphism. Here, Jurjānī directly quotes the Mabāḥith and expands on the implications of Rāzī’s view, highlighting a critical ontological point developed by Rāzī:

What we have mentioned is supported by the view of Imām RāzÄ« in the Mabāḥith al-mashriqiyya: “As for the other part, which is the paste-form and which is the principle of the effect produced from it, it is dependent on the other part, which is the collection of the individual parts.” According to this, there is no problem. However, if ‘mixture’ is interpreted in its real sense and is made a part of the paste that is dependent on the rest of the parts, it follows that the substance, which is the paste, is composed of a substance and an accident. Some [thinkers] have permitted this, affirming the composition of the bed from a substance [i.e. substantial matter], which is the wooden parts, and an accident which is the specified ordering and the structure entailed by it (al-tartÄ«b al-makhṣūṣ wa-l-hayʾa al-murattaba Ê¿alayhi). He [RāzÄ«] states: What is impossible is the composition of a substance from an accident that inheres in it, since the latter is posterior to the [substance], and is thus not a part of [the substance], but not its composition from another substance and an accident that inheres in the other substance, because what is then implied is the posteriority of one part to another part. Indeed, it is impossible for an accident to be a part predicated of a substance.¹⁵⁰

Here JurjānÄ« introduces the possibility of an ontology in which substance is not constituted by form and matter, but of substance plus accident. As noted above, substances cannot be constituted by accidents. JurjānÄ« refers to some thinkers who admit this alternative position as valid. Below, we turn to thinkers who expressly endorse this view and the examples they offer of substances viewed as substance-plus-accident composites. It should be noted that JurjānÄ« seems to assimilate the case of medicinal paste to that of a bed, i.e. an artifact, which requires wooden parts and a specific ordering (tartÄ«b) or structure (hayʾa) and construction, which are accidents. As discussed earlier, the latter point is highlighted by RāzÄ« in the Mabāḥith, where he notes that wood is the “real existent part” of the bed and the order is a relational thing but a necessary component that cannot exist independently of the parts.¹⁵¹ RāzÄ« views bed as a real, relational composite. He is well aware that this is not an example of robust form-matter analysis; however, although the more astute readers will recognize the distinction, it is not clear whether all postclassical thinkers distinguished between a robust Aristotelian form-matter analysis and Rāzī’s critical points regarding substances. As we will see, kalām thinkers will suggest distinctions that can capture varieties of composite essences.

The Razian framework informs the later thinkers’ approach to a broad scope of problems concerning essences, including the questions of the unity (waḥda), individuation (tashakhkhuṣ), and causing of essences (jaʿl al-māhiyya). Rāzī seems to be the first to introduce the latter problem in a major work of kalām. However, he assumes his reader to be aware of the term, suggesting that it was already in use prior to him. It seems to be a reframing of the old kalām question of the ‘thingness’ of the non-existent essence (shayʾiyyat al-maʿdūm). Following Rāzī, later kalām works devote a section of their chapter on essence to this question.

In a chapter titled “Whether essences are caused (majÊ¿Å«la) or not”, JurjānÄ« considers various positions on whether simple and composite essences are caused by some causer (bi-jaÊ¿l jāʿil).¹⁵² He first considers composite entities possessing physical magnitude or extension that are constituted of parts with distinct essences, like ‘one tree,’ which he refers to as “one in virtue of composition (al-wāḥid bi-l-ijtimāʿ).”¹⁵³ This kind of composite unity is contrasted with unity that obtains through contact or ‘contiguity’ (ittiṣāl), e.g. two bodies that are immediately next to each other. The unity of the tree, or the composite entity that is one in virtue of composition, is significant in the context of the above discussion regarding substance. JurjānÄ« comments regarding the tree: “For it is composed of extended parts that differ with respect to their essence, in contrast to a simple body like water on the view of the atom [i.e., according to those who endorse atomism], since the parts of [water], although they exist in actuality and as being composed, are similar in essence.”¹⁵⁴ Jurjānī’s reference to atomism here is rather puzzling, since one who holds to the atomism of kalām effaces any ontological difference between a composite like tree and an alleged simple like water. As discussed above, all composites are equally fictional. Here, water elements are not real simples according to kalām atomism; water does not have an essence above and beyond the atoms and accidents that constitute it, and is just as much a perceptual construct as tree. JurjānÄ« seems to be alluding to some distinction available to the mutakallim between homogenous and heterogenous composition of bodies.

Here the glosses on Jurjānī’s commentary furnish us with some critical insights. In his glosses, the Ottoman theologian Ḥasan ÇelebÄ« al-FanārÄ« (d. 1481) makes the critical point that even the atomist is not committed to the view that all things beyond atoms and accidents are simply fictional or mental constructs.¹⁵⁵ Fanārī’s point is not an isolated remark but a new approach that some post-RāzÄ«an kalām thinkers take to the analysis of the sensible world. Here the old form-matter ontology of Aristotelianism is problematized and largely set aside.¹⁵⁶

The position that sensible composites can be viewed as possessing certain kinds of unity, which was stated above as a theoretical possibility, was pursued in practice by later kalām thinkers. Later thinkers begin to forge new ways of analyzing natural substances, like simple elements and biological composites. In this regard, Ê¿Abd al-ḤakÄ«m al-SiyālkÅ«tÄ« (d. 1656), another glossator operating in the Indian subcontinent, states in somewhat more systematic terms: “According to those who hold that indivisible atoms are homogenous, it is not necessary that the composite and simple body is of the same genus, since the accidents by which simple bodies differ are constitutive of the [complex bodies] according to them. Hence, the complex body divides into extended parts that are not homogenous like the elements [which differ in terms of being fire, earth, etc.], and simple bodies divide into parts that are homogenous.”¹⁵⁷ This is a remarkable reinterpretation of the elements and the fundamental constituents of natural reality. SiyālkÅ«tÄ« deploys the theoretical point gestured to by JurjānÄ«, namely that complex substances can be composed of substance-parts plus external attributes. As such, a tree is not a form or a vegetative soul that is a cause of the existence of the material parts of simple elements and natural mixtures; rather, some additional property (perhaps the structural property of biological processes) defines the complex unity of tree beyond the simple elemental matter that constitutes it. The details of such views require further study. Notably, SiyālkÅ«tÄ« views the above as consistent with atomism, which must differ from the classical atomism of kalām.

These discussions in postclassical kalām seem to mark a turn in the kalām tradition: mutakallimūn take seriously the analysis of complex reality. Moreover, kalām thinkers develop a frame of inquiry that departs from falsafa’s hylomorphism as well as the reductive ontology of classical kalām. It is significant that the above discussion, located in the section on general concepts in Ījī’s Mawāqif, provides the theoretical framework for the substantive analysis of sensible things that follows. Thus, as in Rāzī’s Mabāḥith and Mulakhkhaṣ, the questions in general concepts and logic anticipate substantive discussions of sensible reality in later sections. A detailed analysis of Ījī’s treatment of substances cannot be pursued here, but a few notes can be made with regard to the present discussion.

First, we turn to Ījī’s discussion of substances referred to in the above passage.¹⁵⁸ ĪjÄ« begins by rehashing the philosophers’ position on substantial forms, specifically the corporeal species forms. He describes how the contrary qualities of the elements can mix to form intermediary states—say, between absolute heat and cold—that are ‘homogenous’ with respect to each mixed, composite part. Importantly, they note that this homogeneity obtains as an objective feature of the world—i.e. ‘in itself’ (nafs al-amr) or in external reality (fÄ« al-wāqiÊ¿)—as opposed to being something solely dependent on the mind or perception. When such a stable intermediary state is obtained, ĪjÄ« states that “this homogenous quality is called ‘mixture’ (mizāj), and what is prior to that composition is called ‘combination’ (imtizāj).” That is, in line with the Aristotelian and Galenic views, true mixtures possess a reality that physical blends or combinations do not possess.¹⁵⁹ Regarding the falāsifa’s definition of mixture as obtaining necessarily by means of the contact of different elemental forms, ĪjÄ« raises two problems (ishkāl) concerning the definition and status of forms.

First, Ījī asks whether ‘contact’ is a necessary condition for such mixtures to occur. He provides the example of the sun, which can heat objects without being in contact. It is notable that in response to Ījī’s criticism, Jurjānī cites Rāzī as stating the following:

In the Mabāḥith al-mashriqiyya, [RāzÄ« states that] the correct [approach] here is to leave [rational] demonstration and rely on observation, so that one states that the matter concerns the parts of a mixture, which are no doubt in contact. We also observe that some [parts] neither affect others nor are affected by them, except by means of contiguity and contact. Hence, it is not relevant to ask why it is not rationally possible for an element to affect another [element] without contiguity and contact, because that is not what is required with regard to what we are concerned with [in this context]. Rather, the truth is that effects between them can obtain without contact, although it is rare.¹⁶⁰

The quote corresponds nearly verbatim to what RāzÄ« states in his discussion of mixture.¹⁶¹ Jurjānī’s point seems to be that the analysis concerns empirical objects and not necessary and underlying essences.

Following the above point, ĪjÄ« asserts the alternative view, drawing on Rāzī’s attributive distinction between external attributes and bodies. In his second problem related to the above definition of mixture, ĪjÄ« states: “Why did you state that there are forms beyond the qualities that are the agents? Why is it not possible that the [underlying] bodies are homogenous and the difference is in virtue of attributes rather than forms?”¹⁶² Echoing RāzÄ«, ĪjÄ« wants to admit in the analysis of sensible reality only those properties that are empirically known.

Rāzī’s discussion informs not just the central teaching texts of Sunnī kalām, but also the later kalām commentaries produced on Naṣīr al-Dīn al-Ṭūsī’s (d. 1274) Tajrīd al-ʿaqāʾid. Perhaps the most important commentary, at least for the study of physics, is that of the well-known astronomer ʿAlī al-Qūshjī (d. 1474). I turn to his discussion of general concepts, and specifically the mereological analysis of the dependency relation between the parts of composite essences, which correspond to the above chapters of the Mawāqif. Remarkably, the framework is evidently Rāzian, from the examples cited to the central concept of the unifying structure. In his discussion of the composite essence, Qūshjī states:

Composition can be either mind-dependent (iÊ¿tibāriyyan), so that there are numerous things that the mind (Ê¿aql) considers as one thing, even if it is not one in reality, and one can posit for it a name, like ‘ten’ [that is composed] of ones, and ‘army’ [composed] of individuals. In this case, the dependency of some parts on others is not entailed. If it is stated: If what is meant is a lack of dependency absolutely, the [point] is invalid because the dependency of the structural form (al-hayʾa al-ijtimāʿiyya) on the material parts is self-evidently entailed. If what is meant is the lack of dependency with regard to that which obtains between the material parts, that is not required of the real composite either, as is the case of elemental simples [composing] mineral composites for example. We state [in response]: The first is meant [i.e., the lack of dependency absolutely], as structural forms in mental composites are purely judgements of the mind, without obtaining in external reality, for there is nothing of the army in external reality except the individuals. […] The analysis of [mental composites] contrasts with real composites, since the latter possess structural forms that obtain fÄ« nafs al-amr, as is the case with a house. Indeed, it can be that a mixture is generated by the interaction of its parts, as is the case with a medicinal paste. Indeed, substantial species forms are principles of wondrous effects, as is the case with the theriac (tiryāq [an ancient antidote]). If it is stated: In each mixture, the structural form is an accident. How can it then be a part of the paste or the house, which are substances? We state: There is no impossibility in the composition of a substance from two parts, one of which is a substance and the other an accident subsisting in the substance that is its part. Rather, what is impossible is only the composition of a substance from an accident subsisting in that substance, since [the accident] would be posterior to it, whereas the part of a thing is prior to it.¹⁶³

From the above, it is clear that QÅ«shjÄ«, like his predecessors in kalām, follow the guide of RāzÄ« in pursuing alternative avenues of ontological analysis beyond simple atomism and the problematic ontology of falsafa. Following the above analysis, QÅ«shjÄ« goes on to consider the category of ‘real’ composites: “[A composite] can be real such that one essence, with real unity, obtains by means of the composition of numerous entities specified by external attributes and effects. It is necessary in such a composite that some dependency obtains between some parts and others. For if each of the parts were independent of the others, a single essence, with real unity, would not obtain, as is the case of a stone placed beside a man.” Here, QÅ«shjÄ« states that “the minds of the scholars have been confounded by the question of how an essence is composed of predicated parts, and they have differed according to four approaches.”¹⁶⁴ He proceeds to analyze the various positions in detail, including the Aristotelian view of constitutive differentia and genus. Although clearly inspired by Rāzī’s framework, over 250 years after him, kalām discourse was still evolving.

Acknowledgments

I thank Jari Kaukua, Asad Ahmed, Aun Hasan Ali and the anonymous reviewers for their invaluable comments and corrections. I would also like to thank Andrew Arlig, Calvin Normore and the participants of “The Philosophy Occasional Workshop” in the Department of Philosophy, UCLA, for their critical engagement with an earlier version of the paper.