CHAPTER 2

Aquinas’s First Way

1 Introduction

Some of the most influential arguments for the God of classical theism derive from the existence
of change. They trace their intellectual heritage at least back to Aristotle, whose classic statement
of the ‘argument from motion’ or ‘argument from change’ is found in book 8 of his Physics and
book 12 of his Metaphysics. The basic form or pattern of reasoning—from the reality of change to
the existence of an unmoved mover, unactualized actualizer, or purely actual source of all
change—has been reformulated, refined, and expanded by many philosophers within the
Aristotelian tradition.1 These arguments from change, moreover, are not mere historical artifacts.
Instead, they’ve enjoyed something of a revival of interest among philosophers of religion as of
late.2
Most notable for our purposes is medieval philosopher and theologian Thomas Aquinas’s
treatment of the argument from change in his famous First Way, found in both the Summa
Theologiae (I, q2, a3) and Summa Contra Gentiles (I, ch. 13). We focus on Aquinas’s First Way—
or, more accurately, McNabb and DeVito’s (2020) contemporary formulation thereof (informed
by the writings of Aquinas)—because it serves as the clear historical precedent to Feser’s first and
most prominent argument for God’s existence: the Aristotelian proof. Defenders of both arguments
aim to demonstrate the existence of an unactualized actualizer that is the purely actual source of
all change. By understanding the problems afflicting Aquinas’s First Way, we can set the stage for
parallel problems that will afflict Feser’s Aristotelian proof.3
Perhaps the most outspoken and talented critic of arguments for God’s existence (both
traditional and contemporary) is philosopher Graham Oppy. Oppy has criticized both Aquinas’s
First Way and Feser’s Aristotelian proof, voicing objections to the former in Oppy (2006) and the

1
Maimonides, for instance, articulates and defends a version of the argument in The Guide of the Perplexed (Book 2,
ch. 1).
2
See especially Feser (2017), Oppy (Forthcoming), McNabb and DeVito (2020), Davies (2016), McRae and McRae
(2016), and Geenen and Hunt (2016) for some of the most recent treatments. The wide-ranging atheological works of
Oppy (2006) and Sobel (2003) provide influential treatments, with various aspects of the debate blossoming in later
years. Oderberg (2010), for instance, treats the Aristotelian causal principle that whatever changes is caused to change
by another; existential inertia—something which has recently been employed against Thomistic cosmological
arguments—is treated in Beaudoin (2007), Feser (2011), Audi (2019), Oppy (Forthcoming), Schmid (2021;
Forthcoming), and company; and so on.
3
Henceforth, we will speak simply of ‘Aquinas’s First Way’ or ‘the First Way’, but it should be understood that we
mean a contemporary formulation thereof (namely, that in McNabb and Devito (2020)). To be sure, there are
interpretations of Aquinas’s First Way that differ from those in McNabb and DeVito’s paper (cf. Wippel (2000, chs.
XI and XII) and Joseph Owens’ chapters on the first way in Catan (1980, chs. 6-9)). But these are not our concern in
this chapter; we are concerned only with the formulation of McNabb and DeVito (whose reading of the First Way is
deeply similar to Oppy’s (2006), Davies’ (2016), Oderberg’s (2010), and co.).
latter in Oppy (Forthcoming). McNabb and DeVito (2020) have recently defended both arguments
from Oppy’s criticisms, finding the latter to be wanting on various fronts. We’ll return to their
defense of the Aristotelian proof in a later chapter. For this chapter, we will focus on just one of
Oppy’s criticisms of the Third Way and, more importantly, McNabb and DeVito’s rejoinder(s)
thereof. In doing so, we will level new criticisms of our own toward Aquinas’s First Way. These
criticisms will also set the stage for our detailed critical appraisal of Feser’s Aristotelian proof in
the subsequent chapters.

2 Validity

Oppy’s reconstruction of Aquinas’s First Way proceeds as follows:

1. Some things are in a process of change.

2. Whatever is in a process of change is being changed by something else.
3. An infinite regress of changers, each changed by another, is impossible.
4. (Hence) There is a first cause of change, not itself in a process of change. (Oppy 2006, p.
103)

Oppy argues that this reconstruction is invalid, as there is nothing within the premises that warrants
the inference to a unique first cause of all change which is itself not in any process of change (i.e.,
which is utterly changeless). As McNabb and DeVito point out, however, there seem to be
inadequacies in Oppy’s reconstruction in light of Aquinas’s writings and line of reasoning. They
write:

Oppy is right… that the Summa Theologica lacks a phrase which explicitly addresses that
the first cause must itself be changeless, [but] Aquinas clearly implies it. For Aquinas,
‘whatever is in motion must be put in by another,’ and since postulating an infinite amount
of instrumental causes does little to help us understand the ultimate grounding of
hierarchical causation (see Aquinas’ stick example), what are we left with? Given that there
is motion, we seem to be left with the need for postulating a cause that has not been put
into motion by another. (2020, p. 2)

Following Thomist scholar Brian Davies (as well as Aquinas’s exposition of the argument in the
Summa Contra Gentiles), McNabb and DeVito proffer an alternative, more faithful presentation
of Aquinas’s First Way4:

1. Some things are in a process of change.

4
McNabb and DeVito (2020, p. 3) quote the full syllogized version of Davies’s formulation found in (Davies 2016,
p. 37). The premise that McNabb and DeVito add to Oppy’s formulation derives from Aquinas himself: “This mover
is itself either moved or not moved. If it is not, we have reached our conclusion—namely, that we must posit some
unmoved mover” (Summa Contra Gentiles, 1. 13.3, trl. Pegis).
 2. Whatever is in a process of change is being changed by something else.
3. What moves something else is either moved or not moved.
4. An infinite regress of changers, each changed by another, is impossible.
5. (Hence) There is a first cause of change, not itself in a process of change. (McNabb and
DeVito 2020, p. 3)

This formulation does avoid Oppy’s invalidity criticism, but only under specific interpretations of
the conclusion and premise (3)—interpretations that McNabb and DeVito neither clarify nor make
explicit. The conclusion follows logically only if both of the following hold: (i) the conclusion
expresses that there is at least one first cause of at least some changes, not itself in a process of
change, and (ii) premise (3) expresses that whatever moves something else is either moved or not
moved in any way whatsoever. Allow us to explain.
Let’s begin with (i). At first glance, it seems that McNabb and DeVito want to infer a
single, unique first cause of change as such. Prima facie, this is why they say that there is a first
cause of change, not merely one instance of change or one connected chain/series of changes.5
This, however, is not entailed by the premises. The denial of an infinite regress of chains of change
ordered hierarchically or per se (captured in premise (4)) only says that for each per se chain of
changes, that particular chain is not infinite.6 In other words, for each chain of per se changes C,
there is at least one first member T of C which causes the relevant changes within C but is not
caused to change in the relevant respect.
But this claim is silent on whether there is some one first member for all per se chains of
change. In other words, it doesn’t follow from the abovementioned claim that there is a single first
member T for all chains of per se changes. Just as it doesn’t follow that there’s a single, unique
counselor for all students from the fact that for each such student, he or she has a counselor, it

5
A few notes. First, we use ‘chain’ and ‘series’ interchangeably (to refer, roughly, to a connected order of movers and
things moved in respect of a particular causal power or property (heat, say, or spatial location)). Second, we do not
claim that McNabb and DeVito do in fact wish to derive this, since justifying such a claim would require access to
their private intentions—something we don’t have. It should be noted, however, that they are responding to Oppy’s
criticism here—and Oppy here explicitly raises the difficulty of uniqueness: “There is nothing in the premises of this
argument that justifies drawing the conclusion that there is a unique first cause of change that is not itself in a process
of change” (2006, p. 103). Moreover, they also consider an objection to the First Way on p. 4 that attacks the notion
of universal divine causation of all changes that occur. Tellingly, they do not respond to this objection by arguing that
this was not established by the First Way, or that they did not intend to establish this; instead, they seek other avenues
of response.
6
Chains of changes ordered or subordinated per se (as opposed to per accidens) are ones wherein the relevant causal
power or property (e.g., heat, local motion, or what have you) is wholly (concurrently) derivative in all secondary
(non-first, non-fundamental) members of the chain. Such secondary members have the relevant causal power or
property only derivatively and instrumentally; they do not possess it of themselves but merely transmit the causal
power or property bestowed to them. They possess the causality of the series only inasmuch as (or only insofar as)
they are granted it from the first member. Thus, MacDonald writes that Aquinas “characterizes the distinction as being
between causal series in which the posterior causes exercise their causal power solely in virtue of the power of a prior
cause [or causes] (those ordered per se) and those in which the posterior causes exercise their own proper causal power
(those ordered per accidens)” (1991, p. 141). Per se chains are also called hierarchical chains, while per accidens
chains are also called linear chains. Defenders of the First Way (including McNabb and DeVito 2020, p. 3) are explicit
that Aquinas is concerned with chains of change subordinated per se or hierarchically.
likewise doesn’t follow that there’s a first cause of all chains of changes from the fact that for each
such chain, it has a first cause. Premise (4)—in conjunction with the other premises—could not
warrant the inference to a unique source of all change, but only the inference that each distinct
chain of changes has at minimum one first member of that particular chain.7 This is a far cry from
a unique, purely actual source of all change—a predominant description of God within the classical
theistic tradition.8
But even if we set aside the abovementioned quantifier shift problem, it still doesn’t follow
that there must be a single, unique source of the change of a given per se chain of changes.9 For
even if a given chain cannot be infinite, it still doesn’t follow that it has a single terminus. It may
terminate in two beings, each of which imparts causal or change-related power to the rest of the
chain. In this case, we avoid an infinite per se chain of causes/changes, but it is still nevertheless
not the case that there is a single, unique source or terminus of a particular chain. So, even if the
quantifier shift problem weren’t present, uniqueness still wouldn’t follow.10
What follows from the above, then, is that the conclusion follows only if interpreted as (i)
says. But the argument’s validity also requires premise (3) to be interpreted as (ii) says (i.e.,
whatever moves something else is either moved or not moved in any way whatsoever). For only
then can we infer that the first mover of a given per se chain of changes is not itself in any process
of change.
Thus, provided that we understand the conclusion according to (i) and premise (3)
according to (ii), the argument is valid. But here’s the rub: understood according to (ii), premise
(3) is implausible, and understood according to (i), the conclusion now lends little to no support
whatsoever to classical theism. We shall justify each of these claims in turn.
Let’s consider the first of these claims. Recall the claim: understood according to (ii)—
such that whatever moves something else is either moved or not moved in any way whatsoever—

7
We will address an objection from Occam’s Razor in favor of uniqueness in Section 6.2.
8
And it doesn’t seem mistaken to say that this is what Aquinas took his argument to establish. As Michael Augros
points out, “When Aquinas at the end of the First Way reaches a first mover, which is moved by no other, he concludes
‘et hoc omnes intelligunt Deum.’ Who are ‘omnes’? Chiefly the learned, among both Christians and non-Christians.
The learned among the Christians know that God alone is the first source of motion and change, the initiator of all
changeable things, and that he does not change” (2007, p. 99). And again: “Now the First Way proves the existence
of an unmoved mover, which is a principle of change in all things, … giving them their agency and receiving its
agency from none (or at least not by means of a change)” (2007, p. 100). And David Twetten: “When ‘motion’ is
taken in this universal sense [i.e., the actualization of potential], then, the subsequent proof will conclude to a first
mover that is unmoved in the sense of having no potency for further actuality. … [T]he proof concludes to a mover
‘unmoved’ in the sense of ‘not further reduced or reducible from potency into act’” (1996, p. 268). And Aquinas
himself in the Summa Contra Gentiles: “Therefore it is necessary to suppose that there is some primary unmovable
mover” (primum movens immobile). We will argue in the subsequent section that the inference to the unmoved
mover’s being immovable or unchangeable in all respects is a non-sequitur on many fronts. But it’s important to note
that this is precisely what many defenders and commentators of the First Way take it to establish.
9
Recall that we are not accusing McNabb and DeVito of the quantifier shift fallacy. Rather, we are temporarily
operating on the assumption that (i) is false, i.e., under the assumption that McNabb and DeVito’s conclusion expresses
a single source of all change that is itself unchanged in all respects.
10
It’s worth noting that the criticisms we raise don’t solely apply to the reconstruction of McNabb and DeVito. For
instance, they equally apply to Oderberg (2010, p. 43), whose articulation proceeds: (i) Everything that is changing is
being changed by something else; (ii) But the series of changers and things changing cannot be infinitely long; (iii)
Therefore, there must be a first cause of all change, i.e., God.
premise (3) seems implausible. The justification on behalf of premise (3) seems simple: an infinite
per se chain of changes is impossible; hence, each such chain has at least one first member. Call
one such first member T. Now, if T were moved, then T would no longer be the first member, as
there would be some prior mover of T. But we’ve supposed, ex hypothesi, that T is the first
member. So, T is not moved—simpliciter, full stop, in any respect whatsoever. So, premise (3) is
true.
But the simplicity of this justification is misleading. For the finitude of per se chains only
justifies an inference to at least one first member T of a given series which is not being moved in
respect of the causal power or property of that particular series (heat, say, or spatial motion). This
is because not all changes are situated within a single per se series. For instance, when the stone is
moved by the stick, in turn moved by the hand, in turn moved by the mind (which is the source of
the motive causal power in this particular series), the causal power or property of this series—the
power of (initiating or maintaining or bringing-about-in-something-else) spatial motion—is
unrelated to the causal power or property of the series of the water heated by the pot, in turn heated
by the stove, in turn heated by the fire (which is the source of the causal property of this particular
series, viz. heat). These are completely different per se chains of change with completely different
causal powers/properties that secondary members derive from the primary or fundamental
members. Moreover, each of these per se chains is finite and ends in a member with the underived
or ‘built-in’ causal power/property of the series. But neither of the first members in each chain (the
mind and the fire, respectively) are unmoved in all respects whatsoever. And this doesn’t detract
from their status as first member of their respective chains. For their status as first member of their
respective chains would only be compromised if they derived the relevant causal power or property
(motion or heat) from without. This is irrelevant to whether they are moved in other respects that
have nothing to do with their status as the underived causal source of the motion/change of the
series. For instance, suppose that the water-pot-stove-fire apparatus is small and portable, and
suppose that someone carries the whole apparatus across the kitchen. In this case, the fire is
unmoved in respect of the causal power of the series for which it serves as terminus, but it is not
unmoved in all respects whatsoever. For it is moved or actualized in a different respect, to wit,
spatial motion.
It follows, then, that the ‘simple justification’ we discussed earlier in favor of premise (3)
is, indeed, too simple. For it doesn’t follow that the first member’s status as first (in a given per se
chain) is compromised by the mere fact that it is moved. For it could easily be moved in respects
that are irrelevant to the causal power or property of the series in question. And this is perfectly
compatible with the first member’s having ‘built-in’/non-instrumental/underived/unmoved power
in respect of the causal power of that series.
We will have cause to return to these issues in the subsequent section, since—as we will
argue—they undergird a host of new non-sequiturs afflicting the First Way. The key takeaway for
now is that the first member of a given per se chain need not be unmoved in all respects
whatsoever.11 And this, in turn, disallows the inference to a first mover that is unmoved in all
respects whatsoever (based solely on the First Way, that is). In principle, it could easily be the case
(for all the First Way shows—and for all McNabb and DeVito say on its behalf) that all per se
chains of change are finite, but that nevertheless each first member (of each such per se chain) is
moved (or even possibly moved) in some other respect—a respect that locates them as non-first in
a different per se chain.
Now let’s consider the second claim we leveled earlier. Recall the claim: understood
according to (i)—such that there is at least one first cause of at least some changes, not itself in a
process of change—the conclusion now lends little to no support whatsoever to classical theism.
First, as we hope to have shown above, we have not arrived at something ‘not itself in a process
of change’. Rather, we have only arrived at something that is not (presently) moved in respect of
the causal power or property of the series for which it serves as terminus. So, we’ve only concluded
to at least one first cause of at least some changes. This is a far cry from a unique, purely actual
source of all change. It is this latter description that is (more or less) properly deemed the God of
classical theism. The former description seems to get us only to a world populated by disparate,
mundane unmoved movers (fire, minds, etc.), each of which has the built-in power to cause
changes in their respective series of changes. Classical theism seems very far off indeed.
But perhaps this can be redressed. Perhaps, in other words, we can avoid or redress the
quantifier shift problem that initially led us to add the provisos ‘at least one first member’ and ‘at
least some changes’. And perhaps we can also infer that the unique first member must be purely
actual (thereby disallowing the possibility that the first member of one per se chain is non-first in
a different per se chain—for if it were non-first in any per se chain of changes, then it would be
moved in some respect; but a purely actual being cannot be moved in any respect). These are
potential paths forward. Let’s explore them.

3 A Path Forward?

Some in the Thomistic tradition have, indeed, sought ways to address (or circumvent, as the case
may be) the foregoing problems. One way is to concede that the argument only gets to at least one
thing that causes the changes in some per se chain of changes without itself thereby changing in
the relevant respect (i.e., in respect of the causal power or property of the series), but then to argue
that: (i) any such unchanging terminus of any given per se chain of changes would be purely actual
(i.e., utterly unchangeable in all respects); (ii) there can only be one purely actual being in
principle; and so (iii) each such member that stands at the terminus of each such per se chain of

11
Other commentators on the First Way explicitly recognize this, too. Kerr (2012, p. 546), for instance, recognizes
that different per se chains with different causal properties can and do terminate in different first members (members
with the causal power or property of that series in an underived manner). He gives the example of the mind’s being
first (and thus unmoved) in respect of the causal power of the mind-hand-stick-stone series but nevertheless moved
(and hence secondary) in respect of other properties and powers in other ordered series of change. And Aquinas
himself: “Therefore, the man of himself is the primary mover, and he moves the stone through several intermediaries”
(In libros Physicorum 8.9). As we will argue in the subsequent section, this actually undergirds a number of new non-
sequiturs that afflict the First Way.
changes is one and the same unique purely actual being (from which all change derives).12 Let’s
call this dialectical avenue ‘Path’. Below, we offer new criticisms of Path.
The first problem with Path lies in the distinction between y’s being unchanged in the
relevant respect and y’s being unchangeable in the relevant respect. Allow us to unpack this by
means of an example.
Consider one such chain of changes (one we considered earlier): at time t, the noodles are
heated by the water, in turn heated by the pot, in turn heated by the stove, in turn heated by the
fire. Suppose further that this is a per se chain of changes, and that such chains must terminate.
Because per se chains are concurrent chains of hierarchical dependence (at a given, single
moment), all we could conclude here is that there’s something with the power to make something
else hot at time t without actually deriving that power (to make other things hot) at t. This says
nothing about other causal powers such an entity may have; it says nothing about whether it can (in
some possible world) derive this causal power but simply does not in fact (i.e., in actuality) derive
it; it says nothing about whether the entity has the causal power non-derivatively at t but fails to
have it non-derivatively at some t’ distinct from t; and so on. It also says nothing about this entity’s
being fully, purely actual. For the entity could have potencies that simply have nothing to do with
the relevant chain of changes in question, or potentials which are simply not right now reducing
from potency to act, or what have you.
More generally, then, the first problem for Path is that it simply fails to deliver that the
being is unchangeable or unmovable even in respect of the causal power of the chain for which it
serves as the primary member. Although the fire may have the ability at t to impart heat to the
series without actually having heat imparted to itself (and thereby move others to be heated without
itself being moved in such a respect), it wouldn’t follow that it is metaphysically impossible for
heat to be imparted to the fire at t (and even if that did follow, it also wouldn’t follow that the same
is true for times distinct from t).13 Hence, although the fire is unmoved at t (in respect of the causal
power of the series, viz. heat), it is not (thereby) unmovable at t. This is a problem for Path, since
Path aims to infer something that is unchangeable or unmovable in the relevant respect, not merely
in fact unchanged or in fact unmoved.

12
Broadly, this is the approach Feser takes in his Aristotelian proof—he focuses on a single per se chain of
actualizations of potential, argues that the terminus of this chain must be purely actual, and from there argues that
there could only be one purely actual being. And from here he arrives at a unique source of all change.
13
Objection: The fire, in order to be first mover of the chain in respect of property/power P, must possess P of itself.
But then it possesses P by its very nature, and hence it is not the sort of thing which could possibly lack P or derive P
from without. Response: First, it doesn’t follow from S’s possessing P of itself that P is essential to S. For all this
objection shows, P may very well be a non-essential property or power of S which S has at t but which S does not
derive from without at t (and hence has it ‘of itself’ in the relevant sense, i.e., the sense required in order for S to serve
as the underived source of P for the relevant chain(s) for which S serves as terminus). (Just consider: the Earth holds
other things aloft and is thereby ‘first’ in a whole host of per se causal (or actualization-of-potential-related) chains.
But it’s still possible for the Earth to be held aloft if, say, a giant planet was placed ‘underneath’ (or ‘above’, depending
on how you look at it!) Earth.) Second, even if (contrary to what we’ve argued) S’s having P of itself entailed that P
is essential to S, this seems perfectly compatible with S’s having P non-derivatively in some worlds in which it exists
but having P derivatively in all the other worlds in which S exists. In this case, it is not possible for S to lack P; but,
nevertheless, S derives P in some worlds but not others.
 But let’s suppose (contrary to what we’ve argued) that the First Way could establish that
the first member of a given per se chain of changes has the relevant causal power in an
unchangeable or unmovable manner—and not only at t (i.e., not only when the member is causing
the series), but also at every other time at which the first member exists. Even then, this still
wouldn’t get to a purely actual being (i.e., one that’s unchangeable in all respects), since this would
only entail that the terminus of the given per se chain of changes is unchangeable in respect of the
causal power or property of that series (heat, say, or some other causal power or feature). But this
is perfectly compatible with such a being having other potentials (unrelated to the causal power of
the series for which the being stands as terminus) that are simply not presently being actualized or
are not required to be actualized for the being to serve as the terminus of the per se chain in
question. This is the second problem, then, for Path.
One response that some Thomistic thinkers might level at this juncture is the principle
agere sequitur esse, “according to which what a thing does reflects what it is. If the first cause of
things exists in a purely actual way, how could it act in a less than purely actual way?” (Feser
2017, p. 185). Elsewhere Feser articulates the principle as “the way a thing operates reflects its
mode of existing” (Feser 2020).
As Feser thus explicates the principle, it would seem to be something like: if S exists F-
wise (i.e., in an F way or manner), then S acts F-wise. Under the assumption (which we’ve
challenged) that the terminus T of a given per se chain is at least unchangeable in respect of the
causal power of the series in question, we could use the principle agere sequitur esse to infer that
T would be unchangeable in every respect (and hence purely actual full stop). For if T existed in a
changeable manner, then T would act in a changeable manner. But at least one of T’s acts (the
causal power or feature of the per se chain in question) is not changeable (ex hypothesi). So, T
exists in an unchangeable way. And—by a second application of agere sequitur esse—every one
of T’s acts would thereby be unchangeable, meaning T is purely actual (full stop).
What to make of this riposte? First, it presupposes that the First Way has established that
T has at least one unchangeable act. But as we’ve already seen, the First Way has not established
this. Second, the principle agere sequitur esse seems false, as it faces a number of
counterexamples. Consider, for instance, the fact that nothing has control over whether it exists
(for it would already have to exist in order to have such control). So, everything exists in a manner
that is outside of its control. But it’s simply false that everything thereby acts in a manner that is
outside of its control. Another counterexample is that while God exists in a necessary way, it is
simply false that God chooses in a necessary way; under classical theism, God is free to create or
not, to perform a particular miracle or not, and so on. It is not metaphysically necessary that God
choose to create or perform a particular miracle.14 Overall, then, the appeal to agere sequitur esse
is of no help here.
The third problem for Path runs as follows. Suppose that we granted that the first and
second problems I’ve raised fail, and that we actually can establish that T is purely actual full stop.

14
Even if one disagrees with classical theism on this point, there seems to be nothing impossible in principle about a
necessary being that acts in a non-necessary way (i.e., indeterministically).
Even then, Path still fails, since there seems to be no problem in principle for there being more
than one purely actual being.
Feser’s argument for the uniqueness of a purely actual being is as follows. For there to be
two (or more) purely actual beings, a differentiating feature must obtain between them. “But,”
writes Feser, “there could be such a differentiating feature only if a purely actual actualizer had
some unactualized potential, which, being purely actual, it does not have” (2017, p. 36).
By our lights, however, this argument is not successful. A differentiating feature could
easily be in terms of some difference in actual features between the two things. An elephant, an
amoeba, and a planet for instance are distinguished by many features other than unrealized
potentials. And while ‘having different actual features’ entails that one being does not have a
feature the other has, the mere absence of a feature does not entail potentially having that feature.
For example, humans don’t have the feature ‘being made entirely of gold’, but they are not even
potentially made entirely of gold.
Moreover, this line of argument (if successful) seems inconsistent with Trinitarian
conceptions of God, according to which there is one God in three divine persons.15 For in order
for there to be more than one divine person, there would have to be some feature that one had that
the other(s) lacked, in which case—according to Feser’s reasoning—at least one of the divine
persons must have some unactualized potential. But this runs into trouble on two fronts. First, it’s
incompatible with such a person being divine. For God is by nature purely actual, and hence
anything with the divine nature (including a divine person) would also be purely actual. Hence, if
one such divine person has an unactualized potential, it wouldn’t be divine after all. Second, if one
divine person has unactualized potential, then there exists unactualized potential within God (since
each divine person is intrinsic to God). And this is incompatible with God’s being purely actual.
Overall, then, Path fails on at least three fronts, and thus the original criticisms leveled
towards the First Way retain their teeth.16 We’ve also seen, moreover, a number of additional
problems for the First Way in its inference to the pure actuality of the unmoved mover or
unactualized actualizer. We turn next to a further non-sequitur problem for the First Way.

4 Mind the Gap

15
While this is not an objection or problem per se for the inference (since it relies on a Christian doctrine), it is
nevertheless valuable to uncover. To the extent that one finds the doctrine plausible (or even rationally defensible),
one thereby has pro tanto reason to reject (or doubt) the inference. While not a direct challenge to the inference, then,
there is value in bringing this tension to light.
16
Another avenue some take at this stage in the dialectic is to appeal to the real distinction in created, finite beings
between essence and existence (esse or the act of existence). In particular, some might propose, at this juncture, that
anything that is an admixture of act and potency is an essence-existence composite, and any essence-existence
composite derives its being from the causal activity of something in which essence and existence are identical. But
this response makes Aquinas’s First Way wholly parasitic on Aquinas’s De Ente argument for God’s existence from
essence-existence composition. (Remember that we are targeting McNabb and DeVito’s formulation of the First Way,
which is a distinct argument from the De Ente argument.) Thus, to pursue this dialectical avenue is to grant that the
First Way fails as an independent demonstration of God’s existence. And showing this is our sole purpose in this
chapter. (Note, further, that we address Feser’s rendition of the argument from essence-existence composition—the
Thomistic proof—in Chapter 10.)
In the context of cosmological arguments for God’s existence, the ‘Gap Problem’ refers to the
difficulty in bridging the gap from the being inferred by the argument (an unmoved mover, say) to
God. We think the Gap Problem is particularly poignant for the First Way, since its defenders
typically aim to demonstrate not merely God’s existence but the classical theistic God’s existence.
With this in mind, we present another new problem for the First Way: it is categorically insufficient
to establish the classical theistic God’s existence as opposed to the neo-classical theistic God’s
existence (or even the existence of a non-theistic necessary being). Allow us to explain.
Ignoring all the previously articulated problems for the First Way, we are entitled at most
to infer the existence of a purely actual source of all change. But—crucially—this is perfectly
compatible with neo-classical theism. For according to neo-classical theism, God is a
metaphysically necessary being. God cannot fail to exist, nor can he begin or cease to exist. He is
necessarily actually existent. While he has potentialities, these are not potentialities for ceasing,
beginning, or failing to exist; they are merely potentialities for accidental change (as opposed to
substantial or essential change wherein the substance itself comes into or goes out of existence).
But what this means is that the neo-classical theistic God’s essence or nature (else: his
collection of essential properties) is purely actual. For suppose that his essence (or essential
properties) has some potential for change. Then God could cease to exist. For if a substance’s
essential properties change, then the substance in question ceases to exist—essential properties are
those properties without which a substance cannot exist, and hence to lose, gain, or alter a
substance’s essential properties is to make it cease to exist. So, if God’s essence or essential
properties had potential for change, then God could cease to exist. But God—under neo-classical
theism—cannot cease to exist. Hence, God’s essence or essential properties have no potential for
change. Hence, they are unchangeable and thus purely actual. It follows, then, that the neo-classical
theistic God’s essence or nature is purely actual. The same reasoning holds true for any
metaphysically necessarily existent substance (whether it be theistic or non-theistic).
What this means is that non-classical theistic views, such as neo-classical theism, are
perfectly compatible with the existence of a purely actual source of all change. For under neo-
classical theism, God’s essence or nature could be the ultimate source of all change: God’s
accidental changes (like going from not intentionally acting to create to so acting) could easily be
explained, ultimately, in terms of God’s perfect nature (say, in terms of God’s goodness, or God’s
love, or God’s necessary and essential reason-desire pairs, or certain rational structures that are
built into the fabric of God’s being (as it were), or whatever). And the changes within created
reality are ultimately traced back to God’s creative activity—which, as we’ve just seen, could
easily be explained in terms of God’s nature or essential properties. Thus, even under neo-classical
theism, all changes can ultimately be sourced in something purely actual: God’s essential nature
(or essential properties). And similar reasoning can equally apply to the essential nature of a
foundational, ultimate necessary being in (certain) non-theistic worldviews, too.
The First Way therefore seems to face a serious problem: it is unable to establish the
existence of a being or substance that is purely actual full stop, simpliciter. For even under neo-
classical theism—on which God has various potencies for accidental change—there could be a
purely actual source of all change (to wit, God’s nature). And the same holds for non-theistic views
that embrace the existence of a foundational metaphysically necessary being. These non-classical
theistic views are perfectly compatible with all the premises of the First Way: the denial of infinite
per se chains of changes is preserved; the causal principle is preserved, since there are no
unexplained (i.e., unactualized) changes; and so on.
By our lights, this is a formidable new challenge to the First Way, one which—as far as we
are aware—is found nowhere in the extant literature. We offer it as a tool, then, for further inquiry
and fresh exploration in debates concerning arguments from change.17 Before moving to the next
section, though, we wish to address an objection to our argument in this section.
First, one might object that, on any metaphysics that features real essences, all things with
real essences cannot change their essences or essential properties and so would be purely actual in
the sense we’ve articulated. For instance, given that S is essentially human, S cannot gain or lose
S’s humanity. In that sense, S’s essence or essential properties are purely actual, with no potential
to be gained or lost.
We have two responses. First, this objection crucially ignores potentials for ceasing to exist
and potentials for being absent from reality altogether. For while S’s essence or essential
properties may not have the potential to change (i.e., become different than they in fact are), they
most definitely have potentials to cease to exist.18 S’s properties, surely, cannot float free from S,
and so once S ceases to exist, S’s essential properties likewise do.19 Thus, S’s essence and essential
properties are not, after all, purely actual—they have potentials to cease to exist and to fail to be
absent from reality altogether. (With respect to the latter potential, the same will hold for any
contingent object.)
Second, suppose the objection is right and our first response is wrong. Even still, the
objection doesn’t actually target our argument of this section. The objection simply grants that
essences or essential properties are purely actual. But in that case, our point stands: the neo-
classical theistic God’s essence or essential properties could easily be a purely actual source of all
change. To be sure, the neo-classical theistic God still changes (and thus has potentials); but the
changes are not inexplicable, brute happenings; instead, they are explained in terms of more
fundamental aspects of God—aspects that are part of God’s very nature. And since (as this

17
Another Gap Problem consists in the difficulty of inferring various divine attributes from the notion of a purely
actual being. We assess many of these inferences (from Feser 2017) in Chapters 7 and 8.
18
A complication arises here concerning such potentials. We postpone discussion of this complication until Section 6
of the next chapter (Chapter 3). For now, we can simply understand the ‘potential’ in question as either a dispositional
feature of something or—and this is important—a possibility (without this possibility having to be some inherent
dispositional feature of the entity in question).
19
Of course, other people’s humanity will remain once S ceases to exist, and presumably so will the abstract universal
humanity if there is such a thing. But S’s humanity will not. We are here concerned with ‘concrete natures’ (cf. Pawl
(2020, p. 14) for an exposition of ‘concrete nature’ in the context of Christology). These are particular things that
inhere in and make up objects. (We are thus concerned with a constituent as opposed to relational ontology. This is
appropriate in the present dialectical context, since—as Vallicella (2019) and others have pointed out—classical
theism is naturally and traditionally undergirded by a constituent ontology.)
objection grants) this nature is purely actual, it follows that deriving a purely actual source of all
change is unable to establish classical theism.20

5 Summary Thus Far

Thus far, we’ve examined the criticism that the First Way is invalid. The recent formulation of
McNabb and DeVito is only valid under certain interpretations of the conclusion and premises—
interpretations that are either implausible or quite distant from classical theism. Along the way,
we’ve raised a host of new problems for Aquinas’s First Way by uncovering (at least) six non-
sequiturs in its reasoning, ones which have gone un- or underappreciated by philosophers working
on arguments from change.21 As we’ve seen, the First Way only concludes that, for each per se
chain of changes C, there is some terminator T that is unactualized (i.e., unmoved) at time t of C
in respect of the causal power or feature F of C. But the inference from this to T’s being purely
actual simpliciter is marred by at least six non-sequiturs:

1. From the fact that T is unactualized in respect F at time t, it doesn’t follow that T is
unactualized in respect F at times other than t (since there may be some T* distinct from T
that serves as the terminus of C at another time, and T may be actualized in respect F at
such times). But suppose it did follow. Still:
2. From the fact that T is unactualized in respect F at times other than t—say, at each moment
at which the relevant chain of changes C exists—it doesn’t follow that T is unactualized in
respect F at all times at which T exists (since perhaps C no longer exists, and so even if T
is the unactualized/first member of C at every time at which C exists, T may still exist (after
C has ceased to exist) and be actualized in respect F). But suppose it did follow. Still:
3. From the fact that T is unactualized in respect F at all times at which T exists, it doesn’t
follow that T is unactualizable (as a matter of metaphysical necessity) in respect F. But
suppose it did follow. Still:
4. From the fact that T is unactualizable in respect F, it doesn’t follow that T is unactualizable
in every single respect (i.e., that T is purely actual, full stop). But suppose it did follow.
Still:

20
Note that we have not said that the neo-classical theistic God is purely actual, or that a non-theistic necessary
foundational being is purely actual. We have simply argued that their essence or essential properties are purely actual.
21
There are other problems for the First Way that we haven’t leveled. One recently articulated problem is “the
challenge of explaining why a first cause in an essentially ordered series could not have been caused by things within
a non-essentially ordered causal series” (Rasmussen 2010, p. 817). Another problem: the act-potency analysis of
change upon which the First Way rests seems to presuppose that eternalism is false—for according to eternalism, all
times (and their contents) are tenselessly and equally actual. And given that there is change over time, it would be
false (under eternalism) that such changes involve a transition from potentially existent things or attributes to actually
existent ones. While this isn’t a problem in itself for the First Way, it certainly limits its dialectical force (insofar as
eternalism seems to be a rationally defensible position in philosophy of time). We return to considerations in the
philosophy of time in later chapters.
 5. From the fact that T is unactualizable in every single respect (and hence purely actual), it
doesn’t follow that T is the single source or terminus of every chain of changes (i.e., of all
change). But suppose it did follow. Still:
6. From the fact that T is a purely actual source of all change, it doesn’t follow that the God
of classical theism exists.

The defender of the First Way therefore has a number of new hurdles to clear before they can
demonstrate the God of classical theism.22
In the following section, we address two final rejoinders (on behalf of the First Way) to
(some of) our criticisms thereof.

6 Rejoinders

6.1 Act’s Priority

One might respond that Aquinas himself anticipated the problems we leveled in Section 3
concerning the (purported) possibility of the unmoved mover’s having some potency. For Aquinas
himself writes:

[T]he first being must of necessity be in act, and in no way in potentiality. For although in
any single thing that passes from potentiality to actuality, the potentiality is prior in time to
the actuality; nevertheless, absolutely speaking, actuality is prior to potentiality; for
whatever is in potentiality can be reduced into actuality only by some being in actuality.
Now it has been already proved that God is the First Being. It is therefore impossible that
in God there should be any potentiality. (Summa Theologiae I, q3, a1)

The idea is as follows. Act is prior to potency; hence, anything which is absolutely first in the order
of being (such that nothing is prior to it) could not have potency. For if it had potency, some distinct
actuality would be prior to it—in which case, it wouldn’t be first in the order of being. Something
would be prior to it. We can formalize this argument like so:

1. Potentials require some prior actuality.

2. If the first being has potential, and if potentials require some prior actuality, then
something is prior to the first being.
3. Nothing is prior to the first being.
4. So, the first being has no potential.

22
It should be noted that MacDonald (1991) brings to light the problematic inference from ‘unmoved’ to ‘unmovable’,
arguing that—for this very reason—Aquinas’s First Way is parasitic on another, different argument for God’s
existence found elsewhere in Aquinas. However, we have brought to light a whole host of other problematic inferences
(or, more accurately, problematic by our lights) that MacDonald does not discuss.
What to make of this argument?
There are at least two serious problems with this rejoinder. First, the consequent of premise
(2) simply doesn’t follow from the (conjunctive) antecedent. It doesn't follow that there would be
an entire concrete object (i.e., a being) that is prior to the ‘first being’. Instead, the actuality (or
actualities) of a single being B could be prior to the potency (or potencies) of B. Hence, act would
be prior to potency, but there would be nothing prior to B. Hence, B can be a first being while
nevertheless having potency. (Recall the points defended in Section 4.)
Second, as we’ve seen, the First Way only gets us to the claim that there is at least one first
member T of chain C that is actually unmoved at time t in respect of the causal power or property
of C.23 It therefore does not establish the existence of a ‘first being’ that is the most fundamental
being—more fundamental than all else and in every respect. Instead, it only shows (at most) that
each chain has a first member T of that chain C which has nothing prior to it in respect of the
causal power of C only during the time period at which T stands as first member of C. But this is
perfectly compatible with T standing in potency in some other respect (e.g., some respect that has
nothing to do with the causal power/property of the C for which it serves as terminus). Hence, it
is perfectly compatible with there being something more fundamental than T in that particular
respect but yet not with respect to the causal power/property of C.
We conclude, then, that this rejoinder does not succeed.

6.2 Occam’s Revenge

Recall that one of our criticisms of the First Way was its inability to establish the uniqueness of
the unmoved mover. But one might respond that Occam’s Razor justifies the inference to a single
source for all chains rather than a plurality of sources, each of which serves as a terminus for their
respective chains. For the former is surely simpler than the latter and hence is to be preferred.
We have three responses. First, taking this line of response renders the First Way no longer
a metaphysical demonstration but rather an argument based on a defeasible rule of thumb. This is
significant in its own right, as the argument is often taken by both proponents and detractors to be
an attempt at strict, deductive demonstration from first principles of metaphysics. It is also
significant in its own right, as it (logically) weakens the argument.
Second—and more importantly—this objection fails to distinguish between quantitative
and qualitative (i.e., categorical) simplicity. The former refers merely to the number or quantity of
entities within one’s ontology, whereas the latter refers to the number of fundamental (irreducible)
kinds or categories within one’s ontology. The latter is typically taken (rightly so, by our lights)
to be the more important (i.e., the more theoretically virtuous).24 While the objection at hand might
succeed in showing that the ‘single source’ hypothesis is quantitatively simpler than the ‘multiple
source’ hypothesis, the latter is actually qualitatively simpler than the former. For in the latter case,
our explanatory work is done by ‘mundane’, ordinary first movers with rather mundane, ordinary

23
Assuming, of course, the truth of the First Way’s causal principle and the requirement that per se chains be finite.
24
Cf. Koons and Pickavance (2017, p. 141), Lewis (1973, p. 87), and Nolan (1997).
causal powers. For instance: the first member of a given chain of heat-related changes is just (say)
the fire; in another case of heat-related changes, the first member is the sun; in another case wherein
the relevant causal power or property of the series is (something like) ‘being held aloft’, the first
mover will simply be the Earth (or the Sun, or the gravitational influence they exert on one
another), since the Earth has the power to hold other things aloft without itself being held aloft by
something ‘beneath’ it; in another case of spatial motion (e.g., the mind-hand-stick-stone case),
the first mover is simply the human mind; and so on. In this case, there is a qualitative simplicity,
since the first movers are all mundane, ordinary, this-wordly, changeable, natural entities. This
kind of categorical uniformity and simplicity is not present in the single source hypothesis under
consideration, which posits a thoroughly extramundane, unordinary, unchangeable, supernatural
being as first mover, categorically unlike any of the mundane movers previously articulated. Thus,
far from supporting the single source hypothesis, arguably Occam’s Razor cuts against it.25
Third, it seems that the single source hypothesis isn’t, after all, even quantitatively simpler.
For the ontological commitments of the multiple source hypothesis seem to be a proper subset of
the ontological commitments of the single source hypothesis. After all, the single source
hypothesis is committed to the very same mundane movers that the multiple source hypothesis
takes as the stopping point of explanations of change. But in addition to such mundane movers,
the single source hypothesis posits an extra being appended to all such (per se) chains of change.
Thus, the single source hypothesis seems committed to all the entities posited by the multiple
source hypothesis and more. It seems, then, that the multiple source hypothesis is both qualitatively
and quantitatively simpler.
For these three reasons, the appeal to Occam’s Razor won’t be of much help to the defender
of the First Way.

7 Conclusion

We began this chapter by noting the historical and contemporary significance of arguments from
change for classical theism, and we also stressed the parallel between one such argument—
Aquinas’s First Way—and Feser’s Aristotelian proof. We then raised a variety of problems for
Aquinas’s First Way. First, its validity is preserved only in light of certain interpretations of the
conclusion and premises. But such interpretations were found either to be implausible or to lend
little to no support to classical theism. We also uncovered at least six non-sequiturs afflicting the
First Way. One of these was the categorical inability of the First Way to establish the existence of
the God of classical theism. We concluded by examining and rebutting objections to our critical
appraisal based on actuality’s priority to potency.
With this critical appraisal in hand, we can proceed on our journey through Feser’s five
arguments for the God of classical theism, with the next argument to be evaluated—the
Aristotelian proof—being an intellectual descendent of Aquinas’s First Way. As we shall see,
however, it is descent with modification and, hopefully, improvement.

25
Pun intended.
 References

Audi, P. 2019. “Existential Inertia.” Philosophic Exchange, 48(1): 1-26.

Augros, M. 2007. “Ten Objections to the Prima Via.” Peripatetikos, 6: 59-101.
Aquinas, T. In libros Physicorum.
Aquinas, T. Summa Contra Gentiles.
Aquinas, T. Summa Theologiae. Translated by Fathers of the English Dominican Province.
Online edition (2017, Kevin Knight).
Beaudoin, J. 2007. “The world’s continuance: divine conservation or existential inertia?”
International Journal for Philosophy of Religion, 61(2): 83-98.
Catan, J.R. 1980. St. Thomas Aquinas on the Existence of God: Collected Papers of Joseph
Owens. Albany: State University of New York Press.
Davies, B. 2016. Thomas Aquinas’s Summa Contra Gentiles: A Guide and Commentary.
Oxford: Oxford University Press.
Feser, E. 2011. “Existential Inertia and the Five Ways.” American Catholic Philosophical
Quarterly, 85(2): 237-267.
Feser, E. 2017. Five Proofs of the Existence of God. Ignatius Press.
Feser, E. 27 Feb, 2020. “Agere Sequitur Esse and the First Way.” Blogspot. Retrieved
from http://edwardfeser.blogspot.com/2020/02/agere-sequitur-esse-and-first-way.html.
Geenen, R. and Hunt, R. 2016. “The Prime Mover Removed: A Contemporary Critique of
Aquinas’ Prime Mover Argument.” In R. Arp (ed.) Revisiting Aquinas’ Proofs for the
Existence of God. Leiden, The Netherlands: Brill.
Grant, W.M. 2010. “Can a Libertarian Hold That Our Free Actions Are Caused by God?” Faith
and Philosophy, 27(1): 22-44.
Kerr, G. 2012. “Essentially Ordered Series Reconsidered.” American Catholic Philosophical
Quarterly, 86(4): 541-555.
Koons, R.C. and Pickavance, T.H. 2017. The Atlas of Reality: A Comprehensive Guide to
Metaphysics. Wiley-Blackwell.
Lewis, D.K. 1973. Counterfactuals. Oxford: Blackwell.
Macdonald, S. 1991. “Aquinas's Parasitic Cosmological Argument.” Medieval Philosophy &
Theology, 1: 119-155.
Maimonides. The Guide of the Perplexed.
McNabb, T. and DeVito, M. 2020. “Has Oppy Done Away with the Aristotelian Proof?” The
Heythrop Journal, 1-9.
McRae, H.T. and McRae, J. 2016. “A Motion to Reconsider: A Defense of Aquinas’ Prime
Mover Argument.” In R. Arp (ed.) Revisiting Aquinas’ Proofs for the Existence of God.
Leiden, The Netherlands: Brill.
Nolan, D. 1997. “Quantitative Parsimony.” The British Journal for the Philosophy of Science,
48(3): 329-343.
Oderberg, D. 2010. “‘Whatever is Changing is Being Changed by Something Else’: A
 Reappraisal of Premise One of the First Way.” In J. Cottingham and P. Hacker
(eds.) Mind, Method and Morality: Essays in Honour of Anthony Kenny, Oxford
University Press, 140-64.
Oppy, G. 2006. Arguing About Gods. Cambridge University Press.
Oppy, G. Forthcoming. “On Stage One of Feser’s ‘Aristotelian Proof’.” Religious Studies.
Pawl, T.J. 2020. The Incarnation. Cambridge University Press.
Rasmussen, J.L. 2010. “Cosmological Arguments from Contingency.” Philosophy Compass,
5(9): 806-819.
Schmid, J.C. 2021. “Existential inertia and the Aristotelian proof.” International
Journal for Philosophy of Religion, 89: 201-220.
Schmid, J.C. Forthcoming. “Stage One of the Aristotelian Proof: A Critical Appraisal”,
Sophia.
Sobel, J.H. 2003. Logic and Theism. Cambridge University Press.
Twetten, D. 1996. “Clearing a ‘Way’ for Aquinas: How the Proof from Motion Concludes to
God.” Proceedings of the American Catholic Philosophical Association, 70: 259-278.
Vallicella, W. 2019. “Divine Simplicity.” Stanford Encyclopedia of Philosophy.
< https://plato.stanford.edu/entries/divine-simplicity/>
Wippel, J.F. 2000. The Metaphysical Thought of Thomas Aquinas: From Finite Being to
Uncreated Being. Washington, D.C.: The Catholic University of America Press.