Monday, October 17, 2022

Aquinas's Second Way: An Analysis

My first love was philosophy of religion. When I started writing this blog, 90% of the articles were about this topic. Over the years, as my professional interests have moved into the domain of applied ethics, particularly the ethics of technology, my focus on this blog has shifted in sync. But I still read about the philosophy of religion when I get a chance, and I like to return to the topic when I can. So that's what I am going to do in this article.

In recent years, there has been a resurgence of interest in 'medieval' or early renaissance philosophy of religion, in particular in classical metaphysical proofs of God's existence such as those found in the work of Thomas Aquinas. I think this may be because of the indefatigable advocacy of someone I whose views I will discuss later in this article. But I could be wrong. Either way, I think classical arguments for the existence of God do deserve some reconsideration. Medieval, scholastic philosophy often gets short shrift from modern thinkers, probably because it is so deeply interwoven with unquestioned religious commitments, but there is more to it than meets the eye. It is very technical and sophisticated, and once you get to grips with its unique vocabulary and style there is argumentative weight to it. Certainly, I have grown to appreciate it more as I have read more of it.

In this vein, I would like to consider one of Thomas Aquinas's famous proofs (or 'ways' of proving) God's existence: the Second Way. This is the argument for the existence of God that stems from the need for a sustaining cause of all efficiently caused things. I'll review the argument in two main stages. First, I'll look at a standard translation of the Second Way, analyse the argument that emerges from this bit of text and offer some sceptical observations. Second, I'll consider an attempt to provide a sophisticated interpretation of this argument -- one more grounded in Aquinas's metaphysical presumptions -- from Edward Feser. I will also offer some sceptical observations about this sophisticated interpretation.

Everything I write comes from a sceptical perspective. Longtime readers will know that I am not a religious believer and I don't try to conceal this fact. Nevertheless, I am not necessarily interested in 'refuting' the Second Way. Like most arguments of its type, I think it is essentially inconclusive because it requires too many, unproven metaphysical assumptions to work. But I also think that a full 'proof' of a naturalistic or nontheistic worldview requires similar assumptions. Establishing worldviews is hard, and I don't think I or anyone else is capable of definitively doing so. We can only offer more or less plausible arguments for different possibilities. I think the Second Way is intriguing, but it does not convince me that I should abandon my naturalistic predispositions.

Let me try to explain why I think this.


1. A First Pass on the Second Way

Here's a standard translation of Aquinas's Second Way:

"The second way is from the nature of the efficient cause. In the world of sense we find there is an order of efficient causes. There is no case known (neither is it, indeed, possible) in which a thing is found to be the efficient cause of itself; for so it would be prior to itself, which is impossible. Now in efficient causes it is not possible to go on to infinity, because in all efficient causes following in order, the first is the cause of the intermediate cause, and the intermediate is the cause of the ultimate cause, whether the intermediate cause be several, or only one. Now to take away the cause is to take away the effect. Therefore, if there be no first cause among efficient causes, there will be no ultimate, nor any intermediate cause. But if in efficient causes it is possible to go on to infinity, there will be no first efficient cause, neither will there be an ultimate effect, nor any intermediate efficient causes; all of which is plainly false. Therefore it is necessary to admit a first efficient cause, to which everyone gives the name of God."


It is impossible to understand this argument without having some idea of what Aquinas means by 'efficient cause'. The concept is inherited from Aristotle, who distinguished between four kinds of cause: material, formal, final and efficient. The standard gloss on efficient causation is that it is the type of Aristotelian causation that corresponds most readily to our 'everyday' or 'commonsense' understanding of causation. If I say that X is the efficient cause of Y, I am saying something like 'X brings it about that Y' or 'X is responsible for the changes that make it the case that Y'.

To make this more concrete, let's consider an example. I am currently sitting at my desk writing this article on my computer. How was this state of affairs caused to be? What sequence or chain of causation brings about this state of affairs? Obviously, there is a sequence of prior events that made this possible: I woke up this morning, put on my clothes, had breakfast, walked from the kitchen to my desk, sat down, opened my word processor and started typing. This chain of events constitutes one set of efficient causes of my being at my desk. In addition to this, there is also a set of concurrent physical and natural forces that sustains my presence at the desk: my desire to write this article, the chemical, electrical and nuclear forces that bind together the atoms that make up the chair and my body (and the computer), the force of gravity that stops both myself and the chair from floating into the ceiling. These concurrent forces also represent a chain of efficient causes, though one that may not always be appreciated (as we'll see, this idea becomes crucial to Feser's interpretation of Aquinas's argument).

Hopefully that makes sense. With that clarification out of the way, we can turn to consider how best to reconstruct the argument in the quoted passage. As is typical of Aquinas, it is a dense network of premises and conclusions, consisting of both a primary argument and some sub-arguments. Here's my reconstruction of the primary argument:

  • (1) The world of the senses presents to us an order of efficient causes (i.e. a set of events, objects, states of affairs that have been efficiently caused to be in their present state/form).
  • (2) No member of this order of efficient causes (the one presented to our senses) can be the cause of itself, i.e. every member is caused to be by some other member that is 'prior' to itself.
  • (3) The chain of efficient causes cannot continue to infinity.
  • (4) Therefore, there must be a first member of the order of efficient causes that sustains this order in existence.
  • (5) The first member of the order of efficient causes must be God (or is most likely to be the God of classical theism)
  • (6) Therefore, God exists.

There is also then a subargument offered in support of (3) that works something like this:

  • (3.1) If there was an infinity of efficient causes, then there would be no first member of the order of efficient causes.
  • (3.2) If there were no first cause in the order of efficient causes, then there could be no intermediate or final member of the order (because the first causes the intermediate which, in turn, causes the final)
  • (3.3) But clearly there are intermediate and final members of the order of efficient causes (the evidence of our senses is clear)
  • (3) Therefore, the chain of efficient causes cannot continue to infinity.

I've mapped this argument, using MindMup argument visualisation software, below. Apologies for the change in numbering - this is an automatic feature of the software that I do not know how to change. The logic is still the same. 

What do we make of this argument? First, note how the argument focuses on the 'world of our senses', i.e. the empirical reality we experience on a day-to-day basis. This might be important. Aquinas is not making claims about causation beyond the physical world or about the impossibility of infinite sets of abstract objects (numbers) or events (thoughts in the mind of God). His argument is focused on the impossibility of infinite chains of causation in the world which we interact with every day (note, Feser, who I discuss below, seems to disagree with this limitation on the argument). Second, unlike other cosmological arguments -- William Lane Craig's Kalam argument being the obvious example -- Aquinas's argument is not, primarily or necessarily, about the impossibility of an infinite past sequence of events; it is about the impossibility of infinite chains of causation. This might be confusing because Aquinas uses temporal-sounding language (at least in this translation) to refer to 'prior' causal states, but this is really to denote causal priority and not temporal priority.

With those clarifications out of the way, let's consider the merits of the argument itself. Some of the premises seem uncontroversial. In its current form, I have no qualms about premise (1) though I might change my opinion on this in a moment when I consider Feser's interpretation. Based on my simplistic reconstruction there is nothing controversial about it: we do see chains of efficient causation in the everyday world. Similarly, I have no qualms about premise (3.1) or (3.3). The former is an obvious implication of what it means to have an infinite set of 'prior' causes and the latter, again, seems like an uncontroversial inference from the data presented to our senses.

The other premises, however, strike me as being more controversial. Take premise (2) for starters. As with all theistic arguments, Aquinas has to do some special pleading here. He has to argue that nothing can be the efficient cause of its own existence, *except God*. What justifies this exception? Theists always have some argument for this. Aquinas's version seems to be that God's essence is to exist and so He can be the sustaining cause of his own existence (we'll say more about this below), but this type of argument never seems wholly persuasive to me. Why is it only God that has this special property or power? Could there not be some utterly simply metaphysical substance -- perhaps the simplest most basic form of matter or energy -- that has this power but lacks the other attributes associated with God (mind, creative agency, omniscience etc). To me, this seems like a plausible metaphysical starting point, although, as with God, I don't think there is any decisive evidence in favour of it. Furthermore, even if you reject that, there are some possible counterexamples to (2). I am not personally a believer in libertarian free will -- roughly the view that humans can be the originating cause of their own actions -- but if I were then I might wonder whether every act of free agency violates premise (2). There is certainly an argument to be made to that effect. It is central to the libertarian view, after all, that there are no necessary or sufficient prior efficient causes of your act of free agency.

Moving on to premise (3), this is probably the most philosophically contentious premise. This may be why Aquinas dedicates a sub-argument to it. Like many atheists/naturalists, I don't find this premise and its supporting sub-argument (particularly premise 3.2) to be persuasive. There is supposedly a puzzle (or contradiction or absurdity) inherent in the idea of an infinite series of 'prior' causes. The puzzle is often framed as a question: If the series continues indefinitely in the prior direction (be that causal or temporal) then how did we ever end up in the present? To get to the present event/object/state of affairs, there must be a finite prior series that enables us to get from there to here. But I don't see the puzzle. Assuming that nothing (other than God) can be a cause of itself (and, as should be clear, from my comments on premise 2, I don't assume that), then all that matters for causation purposes is that for every identifiable member of the series there is some 'prior' cause. Continuing this chain infinitely in the prior direction then makes sense because it ensures that every member has a prior cause. You could even use this observation to flip premise (3) on its head and say that if the present is to have a causal explanation then there *must* be an infinite series of efficient causes.

There is nothing new in this observation. Hume made it years ago and, from what I have read of it, Aquinas appears to have accepted it when it comes to historical series of events. He did not, for example, think that you could prove that the universe had a beginning by philosophical argument alone.

"St. Thomas did not believe that it can be proved philosophically that the world was not created from eternity: he admits the abstract possibility of the world's creation from eternity and this cannot be admitted without the possibility of a beginningless series being admitted at the same time" 
(Coplestone, History of Philosophy Volume 2, p. 341)


Nevertheless, he seems to have thought that there was something different about concurrent efficient causes. I'm not sure why. I appreciate that some people find the idea of infinite series of prior causes puzzling but, to my mind, they are no more puzzling than the idea that there is some uncaused first cause that brings the series into existence out of nothing. When we get to such abstruse metaphysical realms, it is all a little puzzling and counterintuitive.

Finally, moving on to premise (5), it is pretty obvious from my formulation, and from the quoted passage stating the Second Way, that this argument is under-defended. Nothing in the prior premises explains why the first cause must be God, or have the attributes of God as classically conceived. To be fair, Aquinas has plenty of reasons for thinking that the first cause ought to have those properties, but it takes reams of complex argumentation to support that point of view. The Second Way, in and of itself, can only get you to the view that there must be an uncaused first cause of the series of efficient causes. Whether that first cause is God, is another matter. I would add that the God of classical theism is quite an abstract and, dare I say it, bloodless character. He is perfectly simple, timeless, immutable and impassible. He is a long way from the more personable and interventionist God that most religious believers (Christians included) worship.

These objections to the original formulation have been added to the argument map, below.

2. Feser's Interpretation of the Second Way

As noted at the outset, the first run through of Aquinas's argument is based on my own interpretation of the text. In recent years, the foremost apologetical interpreter and defender of Aquinas's argument is, probably, Edward Feser. Indeed, Feser has somewhat single-handedly rehabilitated classical metaphysics and the associated proofs of God's existence, and made a powerful case for them to be included in the armamentarium of the theistic apologist. To be clear, interest in classical arguments and metaphysics has never entirely waned in academia or in religious institutions (perhaps most notably the Catholic church). But I do think Feser is responsible for 'popularising' these arguments for a new generation of religious defenders.

Anyway, in his book Aquinas, Feser argues that many lay interpreters of Aquinas get his arguments wrong because they are not sufficiently well-versed in the rich metaphysical schema that Aquinas presupposes for those arguments. He may be right about this. I certainly have some famous 'sceptical' books on my shelves that give Aquinas an unduly short shrift (Michael Martin's Atheism: A Philosophical Justification is an obvious example of this). I want to avoid making the same mistakes so I will try to learn from what Feser has to say. I also want to be clear that although I am referring to this as Feser's interpretation of the argument, I don't think it is unique to him nor do I believe that he would he claim that it is. For instance, Frederick Coplestone offers a similar (albeit much shorter) interpretation of the Second Way in his History of Philosophy.

Two important conceptual distinctions undergird Feser's interpretation of the argument. The first is the distinction between causal series per accidens and per se. I have essentially distinguished between these already but not using this terminology. A causal series per accidens is organised linearly through time: an event at T1 causes an event at T2 and so on. A causal series per accidens is organised 'hierarchically' as a set of concurrently sustaining causes. Think back to my example of sitting at the desk. As mentioned, there is a set of sustaining causal forces that ensures that I stay seated at my desk writing this article: the mental desire to do so, the chemical, electrical and nuclear forces binding the atoms together in my body, chair and computer, the force of gravity holding us down to the Earth's surface, and so on. These forces are arranged in a causal 'stack' that operates concurrently, not in a linear sequence.

Why is this conceptual distinction important? Because Feser argues that Aquinas's argument is only about causal series per se and not per accidens. This is something that other interpreters agree upon, e.g. Coplestone argues that Aquinas's argument is about 'vertical' causation and not 'horizontal' causation, and uses this terminology in a manner similar to the per accidens and per se terminology. This is an important correction but, for what it is worth, I think that for anyone not well-grounded in the broader Thomistic metaphysics the assumption that the argument might be referring to linear causation as well is understandable. My reading of the literature on the Aristotelian notion of 'efficient causation' suggests that it can be understood to cover linear causation.

In any event, you may still be wondering: why does limiting the argument to per se causal chains matter? The answer is that there is something unique about per se causation: each member of the set of per se causes must exist at the present moment in order for the causal chain to give rise to the relevant effect. This is different from linear causation where prior members might pass in and out of existence without affecting the existence of the present effect. Here's how Feser puts it, referring to an example of causation involving a hand moving a staff, which is taken directly from Aquinas:

"Contrast this with a causal series ordered per se or 'essentially'. Aquinas's example from the First Way of the staff which is moved by the hand is a standard illustration, and we can add to the example by supposing that the staff is being used to move a stone, which is itself moving a fallen leaf. Here the motion of the leaf depends essentially on the motion of the stone, which in turn depends on the motion of the staff, which itself depends essentially in turn on the motion of the hand. If any member of higher up in the series ceases its causal activity, the causal activity of the lower members will necessarily cease as well.' 
(Feser 2009, 70)


Got that? Good. Now we can move on to the second important conceptual distinction. This is the distinction between an entity's essence and its existence. Following standard Platonic/Aristotelian metaphysics, Aquinas believes that things have essences, i.e. certain essential properties that determine their nature, character, and form. The essence of a triangle, for example, is that it has three sides, the essence of a human being is that it is a rational animal, the essence of a table is that it is a hard surface for supporting objects, usually lifted off the floor, and so on three-sided. For Aquinas, it is crucial that an entity's essence is distinct from its existence. Knowing that humans are essentially rational animals does not explain why or how they came to exist. Only one entity has the feature that its essence is to exist and, surprise surprise, that entity is God.

Why does any of this matter? It matters, according to Feser, because the argument from the Second Way is about explaining the existence of objects, events and states of affairs. That is the type of causal explanation to which it appeals. This makes it distinct from the First Way (which I haven't discussed in this article) which is about explaining change in the world around us. Here's how Feser puts it:

"whereas the First Way is concerned to explain why things undergo change, the Second Way is intended to explain why they exist at all...just as the First Way is intended to show that no motion or change would occur here and now unless there were a first unmoved mover operating here and now, the Second Way is meant to show that nothing would even exist here and now unless there were a first uncaused case sustaining things in being here and now" 
(Feser 2009, 84)


For what it is worth, I think this interpretation of the Second Way, and of the importance of the essence-existence distinction to understanding it, may be more controversial that the preceding point about the importance of the per accidens/per se distinction. Some commentators, from what I have read, do not perceive a major conceptual distinction between the First and Second Way. Feser's interpretation requires drawing from other Thomistic texts to support this reading (as I understand it -- I am not an expert in this area). But that seems fine to me. I am happy to accept that the Second Way is, primarily, about explaining existence, not motion or change.

With that out of the way, we can now reconstruct the argument from the Second Way to acknowledge Feser's interpretive points:

  • (1*) The world of the senses presents to us a per se order of efficient causes (i.e. a set of entities that are concurrently caused to exist by causally prior entities).
  • (2*) No member of this per se order of efficient causes can be the cause of itself (because it is not the essence of any entity to exist (apart from God)).
  • (3) The chain of efficient causes cannot continue to infinity.
  • (4*) Therefore, there must be a first member of the order of efficient causes that presently sustains this order in existence.
  • (5) The first member of the order of efficient causes must be God (or is most likely to be the God of classical theism)
  • (6) Therefore, God exists.

The sub-argument is essentially unchanged, though you may want to amend the sub-premises to acknowledge that the argument is about per se orders of efficient causes and not per accidens orders. I won't do that in the text but I will do it in the diagram I share later on.

3. Evaluating the Argument

Now that we have clarified the argument, using Feser's interpretation as our guide, we can turn to its evaluation. Is this version of the argument more robust?

I'm not convinced, but before I explain why I want to acknowledge a debt. Feser has written a whole book in which he offers up-to-date formulations of classical arguments for God's existence. This book draws in particular from the work of Aristotle and Aquinas. The book is called Five Proofs for God's Existence. Joseph C. Smid and Graham Oppy have written some interesting critiques of the arguments found in this book, particularly the Aristotelian argument (which is, as I gather, similar to Aquinas's First Way). I found those critiques informative and they helped me to understand some potential problems with the reformulated version. That said, a lot of my criticisms are basically the same as they were for my initial run through the argument. I don't think there is much in this modified version that makes is more compelling than the original.

There are four particular problems I have with the reformulated version.

First, in order to find it persuasive, you have to buy into the larger metaphysical schema upon which is depends. So, for example, you have to think that entities really do have essences, that essences are distinct from existence and that existence demands an explanation of some sort. I'm not sure I buy that larger metaphysical schema. These are difficult questions but, if push came to shove, I think I would deny the idea that physical entities have essences. I lean towards nominalism instead. I think essences are just human conceptual constructions that often provide a convenient way of grouping and understanding physical entities, but also lead us astray. For example, I find the idea that humans are, essentially, rational animals problematic since there are many apparent exceptions to this. Abstract entities are a different story. I think there may be essences to mathematical objects, for instance, but I also don't think they need to be explained. (I should add that I think this problem -- the problem of buying into the larger metaphysical schema -- is perhaps a bigger problem for some other classical proofs of God's existence, e.g. the argument claiming that something must explain the movement or action we see in the world around us).

Second, in order to find it persuasive, you have to think (a) that linear causal series cannot adequately explain the present existence of entities and/or (b) that we always need a per se explanation presently existing entities. Neither is obviously true. Oftentimes linear explanations do satisfy our explanatory demands. To understand why my leg is broken right now it seems sufficient to point out that someone tackled me on the football pitch with force sufficient to break the bones in my leg. These are all prior events and they explain why, now, my leg is broken. Appealing to a vertical or per se explanation seems to miss the point of the explanatory demand. Admittedly, there are some cases in which we do need to a per se explanation, and so you might reasonably argue that a puzzle remains for those cases. For instance, the fact that the candle on my desk is lighting now does not seem to be sufficiently explained by the fact that I took a match to the wick a few minutes ago. We also want to know about the concurrent causal forces that keep the flame going. But do we always need such explanations?

Joseph Schmid has written a good article that examines this issue in depth, focusing on the necessary conditions for per se explanations. Simplifying ever so slightly, Schmid argues that we only need per se explanations when the entities in question would, without intervening causal force, tend towards some opposite state. So, for example, we know that in the absence of oxygen, flames will die out. Hence we need the concurrent presence of oxygen to explain the ongoing flame. But if there is no countervailing force, then no per se explanation is needed. So, for example, an object moving in a void requires no explanation for its ongoing motion. It is a fundamental principle of physics that objects will continue in motion unless operated on by some countervailing force.

This, of course, is to focus on the need for per se explanations for particular entities within our universe. Aquinas's argument isn't about that. It is about explaining the total set of presently existing entities within the universe. Applied to this total set, Schmid's principle implies that per se explanation is only needed if we assume that the natural tendency of entities is toward destruction or annihilation. But that's a controversial assumption. It's not obvious that existence is more in need of an explanation than non-existence. And this is the assumption you need to accept in order to motivate the Second Way. The whole debate about 'why is there something rather than nothing' is a complicated one, and many people feel that existence does require an explanation, but when you ponder these deep metaphysical issues it is really not obvious why it does. After all, if you appeal to God to explain existence you are still left with the puzzle as to why he exists, rather than fails to exist. Claiming that his existence is necessary or essential to his nature doesn't eliminate this problem. It just begs the question.

Third, in order to find the argument persuasive, you have to accept that no infinite series of concurrent per se causes would suffice to explain the existence of some present entity. But, again, it's not obvious to me why we should accept this or why an infinite series is not an adequate explanation. Assume that there is an infinite set of per se causes (e.g. that for each fundamental force that explains the present state of matter and energy we can find another, yet more fundamental force, that explains that why that force is operative). Why do we need to assume that there must be a final, grounding fundamental force? If each member of the set is explained by a more fundamental, causally 'prior' member, then where is the problem? This is, admittedly, to return to the classic Humean objection. Feser claims that this won't work for the reformulated Second Way. But I'm not sure I understand why. To paraphrase the famous story about what is holding up the Earth, I don't see why 'its turtles all the way down' is an unsatisfying explanation.

Feser, I think, would argue that this objection doesn't work because, under Aquinas's metaphysical schema, the set of actually existent entities have no power to explain their own existence. God is the only being whose essence is to exist. As he puts it:

"It is causal series ordered per se that the Second concerned with, and here the need for a first cause follows from the fact that in such series all causes other than the first cause are purely instrumental, having no causal power [to explain their own existence] of their own. Extending the series back to infinity would not change that in the long as the existence of each member is distinct from its essence, it will have to be conserved in existence at each moment by a first cause in whom existence and essence are identical."

 (Feser 2009, pp 88-89)


But why should we accept this view? For starters, it is not obvious to me that existence and essence are always distinct for the entities we observe in the physical universe. Nor it is obvious to me why God is the only possible being for whom essence and existence are identical. I think primordial forms of matter and energy could have the essence of existence (I'm not sure but it doesn't strike me as an absurd view). Also, I would point out that even if there were some abstract, metaphysically primary being whose essence is to exist, and who could, therefore, explain causal series ordered per se, this being would be a long way from the personal God presupposed in most religious traditions. Indeed, such a being may be so abstract that the distinction between a worldview founded on it, and a worldview founded on naturalistic presuppositions, would be largely indecipherable.

So I'm still sceptical of the Second Way.