Potentiality from first principles

In his gigantic post on existential inertia (section 7.13), Joe Schmid gives an argument against the act-potency distinction which can be roughly summarized as follows: the act-potency distinction commits us to pluralism about being, but pluralism about being is false, therefore we should reject the act-potency distinction. I had not heard of pluralism about being before, but Schmid quotes Merricks (a recent critic of the view) giving the following general idea:

Monism about being (monism for short) says that everything enjoys the same way of being. So monism implies, for example, that if there are pure sets and if there are mountains, then pure sets exist in just the way that mountains do. Monism can be contrasted with pluralism about being (pluralism for short). Pluralism says that some entities enjoy one way of being but others enjoy another way, or other ways, of being.[1]

Schmid’s ultimate target is the Thomistic view of existence as articulated by Aquinas in On Being and Essence, but he gets at it through the more general act-potency distinction. On the Thomistic view, God and creatures exist in different ways, since in creatures there is a real distinction between essence and being (essence is in potency to the act of being), whereas in God there is no such distinction—he is being itself, pure act. While this is true so far as it goes, it is not obvious that the “ways of being” in the Thomistic sense corresponds to the “ways of being” in Merrick’s sense, and my aim here is to show that it in fact does not.

Schmid’s argument

Now, I’ll admit that the overall structure of Schmid’s argument is somewhat confusing to me. He actually takes aim at something he calls act-potency pluralism—the view that there are two ways of being (in Merrick’s sense), namely being-in-potency and being-in-act. His discussion ends once he establishes the conclusion that we should not accept act-potency pluralism, but he doesn’t show that the Thomistic view of existence commits us to such a view in the first place.[2] As best I can tell, he assumes that the Thomistic view of existence requires the act-potency distinction and that the act-potency distinction entails (or is equivalent to) act-potency pluralism.[3] Thus, an argument against act-potency pluralism amounts to an argument against the Thomistic view of existence. On the contrary, I do not think the act-potency distinction commits us to act-potency pluralism, and that therefore the conclusion of Schmid’s argument is irrelevant to the Thomistic view of existence.

In order to show this, we must clarify what pluralism means when it speaks of “different ways of being”. If we look at premises (18) and (19) in Schmid’s argument, we can see what he takes to be the salient feature of pluralism that makes it problematic:

18. Act-potency pluralists should not accept generic existence.

19. If act-potency pluralists do not accept generic existence, then they cannot—by their own lights—state that everything is thus and so.

To “generically exist” means “to exist, or to have being, or to be something”.[4] This is the only sort of existence the monist thinks there is, while the pluralist could in principle accept that there is this sort of existence in addition to others. Nevertheless, as Schmid points out when defending (18), the motivations for pluralism militate against the acceptance of generic existence. This is all the more true of the Thomistic understanding of act and potency, since we take this distinction to be exhaustive.

This brings us to (19). If there is no generic being alongside being-in-potency and being-in-act, then there is no X such that we can say simply that everything is X, since everything a thing is is in virtue of its being, and there is no way of being that applies to everything. At best we can say something is(-in-potency) X or is(-in-act) X. As Schmid’s argument goes on to say, this is a problem for act-potency pluralism because it is the position about what everything simply is, and is therefore self-undercutting (see his post for more detail).

This implication of pluralism helps us to better understand what it means for there to be different “ways of being” in Merrick’s sense. If being is divided in this way, then there is no sense in which a thing simply is anything because there is no sense in which a thing can simply be anything—it must either be-in-potency or be-in-act.

The thing is, this isn’t how Thomists understand the act-potency distinction. For us, act just is the generic being, in virtue of which everything exists. Potency, or more specifically potentiality (a difference we’ll unpack later), is not strictly speaking being or a way of being, but something else. But, we may wonder, what could it be if not some other way of being? There are two answers we can give, which we’ll introduce each separately before showing that they are in fact equivalent.

Potentiality as relative nonbeing

The first account arises in response to the problem of the one and the many, motivated by Parmenides’ against multiplicity.[5] In order for one thing to be distinct from another, it needs to have some differentiating feature that the other doesn’t. On the one hand, it can’t be being, since this is something that both things already have. On the other, it can’t be nonbeing, since nonbeing is nothing and if nothing differentiates the things then they are not differentiated. For Parmenides, the conclusion we should draw from this is something we could call “Parmenidean collapse”—that there is no multiplicity of being, all is one and any diversity in being is merely an illusion.

But surely our experience of multiplicity in being is more evident to us than the premises of this argument![6] In particular, while it is initially plausible to think that being and nonbeing are exhaustive, perhaps there is some third option that enables us to avoid the Parmenidean collapse. This third thing, whatever it is, must have some influence over being (similar to being) but without contributing any being of itself (similar to nonbeing). But if it makes no positive contribution to being (as being does) then its influence must be negative, demarcating what being is not. And if it is to avoid completely excluding being (as nonbeing does), then this demarcation cannot be absolute. Thus, this third thing must be what we could call relative nonbeing, a midpoint between absolute nonbeing and absolute being.

Relative nonbeing, then, is that whereby being is multiplied, it is the otherness by which something is other than other things. Alice is not Bob because there is a principle within Alice by virtue of which she is not Bob, Charlie, Eve, or anything else—that is, relative to all these other things, Alice is not. We call the principle of relative nonbeing potentiality, and we call the principle of being actuality (or act). Both actuality and potentiality must be real principles of things lest they not be able to play the role they do in the multiplicity of being. Furthermore, they need to be really distinct from one another because their functions are in some sense opposed to one another, with actuality providing being and potentiality limiting or constraining it.

We may worry that there is an implicit (and unjustified) assumption lurking somewhere in the background here, of the principle of identity of indiscernibles. Why do we need some feature of a being to explain its distinction from other beings? Why not simply assume that such a distinction is primitive? In an important sense, we do take it to be primitive—relative nonbeing is postulated as an ontological primitive that isn’t analyzable into any more fundamental principles. We are not suggesting that the distinction between beings is accounted for by some property they have, since the properties themselves are part of being’s diversity which we seek to explain. What we don’t do is consider distinction to be theoretically primitive, in the sense that our theory simply assumes it without looking for its basis in reality. If we are faced with two theories, one of which is able to explain a phenomenon and the other of which simply refuses to explain it, the former is preferable when our entire aim is to give an account of reality.

Potentiality as the passive capacity for act

The second account of potentiality arises from the analyses of causation and change. Suppose Alice teaches something to Bob, which is to say that she causes him to change from being uneducated to being educated. Again, we say that act (or actuality) is the principle of being, so Bob’s change would be from being actually uneducated to being actually educated.

Now, throughout this process Bob must have the capacity for being educated—if we considered him apart from his educatedness, we would find that nothing determines it one way or the other. This is in contrast to what we find in the case of a tree or a squirrel, where the nature of each excludes the possibility of them being educated. Bob must continue having the capacity for being educated once Alice has educated him, because his being educated just is the fulfillment of this capacity—take away the capacity and he stops being able to be educated! Just as a water jar retains its capacity for holding water whether it is empty or full, so too does Bob retain his capacity whether he is uneducated or educated.

Similarly, Alice must have a “productive” capacity for realizing Bob’s “receptive” capacity. If we considered her apart from Bob’s being or becoming educated, we would find that nothing determines one way or the other about Bob. After all, Alice has this capacity before and after Bob’s being educated, and it exists even if something gets in the way of Bob being able to learn (lack of attention, distraction, misunderstanding, etc.).[7]

Generalizing this example, a potency is a capacity for some act, and we can distinguish between two types. The capacity for producing some act in another is called an active potency, which today is more commonly called a power;[8] and the capacity for receiving some act into one’s being is called a passive potency, or a potentiality. So, in our example, Bob has the potential for being educated while Alice has the power to educate him—to actualize this potential of his. An important difference between potentiality and power is that the former involves an indeterminacy within the thing that has it whereas the latter does not. This is a consequence of how we arrived at each concept: a capacity is conceived in abstraction from its act, so that if the act is within something (as in the case of potentiality) then the capacity corresponds to a non-actuality (and therefore indeterminacy) within it, but if the act is outside of something (as in the case of power) then the capacity corresponds to how it actually is (and is therefore determinate).[9]

While this is all worth exploring in more detail, here we are concerned solely with what this means for potentiality. From the perspective of causation and change, potentiality is a capacity of a thing that makes it “open” to some act without determining one way or the other whether it is in fact actual in that way.

What these accounts mean for act-potency pluralism

So, then, we have two accounts of potentiality arising from two different observations about reality, and it is relatively easy to see that they are equivalent. Indeterminacy regarding some actuality (passive potency) is equivalent to making no positive contribution of being (relative nonbeing), and being contrary to Y, Z, and so on for all options other than X (relative nonbeing) is equivalent to being for X (passive potency). Thus, these two accounts just give us two complementary perspectives on the same fundamental principle, potentiality.

We can now return to the question of whether this distinction between actuality and potentiality entails (or is equivalent to) a pluralism about being in Merrick’s sense. Do our reasons for positing potentiality as a real principle of things motivate a denial of generic existence? Or does potentiality amount to some second kind of existence alongside actuality? Hopefully our discussion up until this point will have made it clear that the answer to both of these questions is no. Both act and potentiality are real and distinct principles of things, but there is only one “generic existence”, and that’s act—when we say things are thus and so we mean that they are actual. Potentiality is not another way of being (in Merrick’s sense), it is relative nonbeing and of itself is indifferent to what is actually the case.

This remains true even when what we affirm a potentiality to be thus and so, because potentiality only ever exists in virtue of actuality, either through composition with some act (its actualization) or as the second potentiality of some act. Composition is what obtains when potentiality comes to limit or circumscribe act, as we needed to posit in order to avoid the Parmenidean collapse. In this case, act supplies being while potentiality limits it to this being rather than that being. It is in virtue of the actuality, then, that the potentiality can be operative as a principle of limitation since without it there would be no positive being to limit. Second potentialities are the more intuitive way in which potentials are said to exist. These occur when the actualization of a potentiality leaves some residual indeterminacy, thereby constituting a further potentiality. Once Alice actualizes her potential to be able to read (by learning how), this results in the further potentiality to actually learn from this or that book.[10] Or, using other Aristotelian categories, Bob exists by virtue of his substantial form actualizing his underlying prime matter, which results in further potentials for having different heights, skills, knowledge, and so on. It’s second potentialities which accounts for things being in potency to various things, ie. to have the unactualized potential for being something. A second potential can exist without being actualized because it is grounded in a lower-level actuality, on account of which we say that the potential is thus and so.

So, then, I don’t think Schmid’s argument holds much weight against the Thomistic view of existence or the act-potency distinction more generally. Even if we accept premises (18) and (19), we should reject the implicit assumption that the act-potency distinction commits us to act-potency pluralism. In Merrick’s sense, the Thomist thinks there is only one way of being, namely by act. Potentiality is not another way of being, but relative nonbeing and the capacity for act.


[1] Trenton Merricks, “The Only Way to Be.”, 593, quoted by Schmid.

[2] He does note that his argument is “applied to act-potency pluralism, but it applies mutatis mutandis to the De Ente argument—simply replace ‘a-existence’ and ‘p-existence’ with ‘d-existence’ and ‘c-existence’ (representing ‘divine existence’ and ‘creaturely existence’, respectively).” However, he doesn’t explain why such a translation would leave the justifications for the relevant premises unaffected. After all, c-existence is not the same as p-existence.

[3] In footnote 25, he suggests that act-potency pluralism encapsulates the first of the 24 Thomistic Theses. However, this thesis does not say that act and potency are two ways of being, but rather that everything which exists is either purely actual things or an act-potency composite. As we shall see, this does not require that act and potency be two ways of being in Merrick’s sense.

[4] Merricks, 599, quoted by Schmid.

[5] For a very detailed discussion of Aquinas’s account in response to Parmenides, as well as the discussion of the notion of relative nonbeing, see John Wippel’s The Metaphysical Thought of Thomas Aquinas.

[6] Modern readers will recognize this as a Moorean shift, but it is also a consequence of the methodology outlined by Aristotle in Physics I. We do not have direct access to the principles of things, but only the composites of which they are principles. Thus, we are justified in believing some purported system of principles only insofar as it accounts for those things which are more manifest to us. Parmenides’ argument is rejected, therefore, because it proposes a system of principles that fails to account for the manifest multiplicity of being.

[7] These two kinds of capacity correspond to the two senses of potentia discussed by Aquinas in De Potentia I Q1.

[8] We should note that Aquinas tends to use the word “power” as equivalent to “potency”, so that for him there are active powers and passive powers. In modern times, we tend to think of powers as exclusively active.

[9] We connect non-actuality with indeterminacy because if something is neither actually X nor actually ~X, then it is indeterminate between X and ~X.

[10] Of course, there is also an active potency at play here, but we are interested in the passive potency which is realized when Alice gains knowledge from her reading.

God causes evil actions without causing the evil in actions

On a recent episode of Unbelievable?, William Lane Craig and James White discussed whether Molinism or Calvinism provide the better approach to God’s providence in light of the reality of evil. Craig is a proponent of Molinism, which seeks to reconcile libertarian freedom with divine providence by positing a special kind of knowledge in God called middle knowledge. White, on the other hand, is a proponent of the compatibilist model of providence common among Reformed theologians, which posits that divine determinism is compatible with human freedom because the latter doesn’t require alternate choice.[1] In the course of this discussion, Craig presents an argument for thinking that the counterfactuals of creaturely freedom (CCFs) are part of God’s middle knowledge. These CCFs are propositions about what free creatures would freely choose when put in various circumstances. Both Craig and White agree that such propositions are contingently true, but they disagree on whether they are decided by God or are true logically prior to God deciding anything. Craig’s argument, which he presents here, can be formulated as follows:

  1. The CCFs are either true logically prior to the divine decree, or they are true only posterior to the divine decree.
  2. If the CCFs are true only logically posterior to the divine creative decree, then God is the author of evil.
  3. God is not the author of evil.
  4. Therefore, the CCFs are true logically prior to the divine decree.

My interest in this post is to evaluate this argument from the perspective of the Thomistic model of omni-instrumentality I have explained in previous posts (see here).

A major feature that differentiates omni-instrumentality from Molinism and compatibilism is that it doesn’t treat human choices as an atomic reality. While Molinism and compatibilism may grant that there are various psychological and other parts that make up a human choice, when it comes to explaining how God can exert control over these choices these parts do not play an explanatory role. When considering a creature’s choice of X, for instance, the Molinist focuses on the circumstances of the choice (as captured by the relevant CCFs) while the compatibilist focuses on God’s ability to make the creature choose X, both of which are factors external to the details of the choice itself. Even if we consider the underlying models of concurrence, which analyze the components of the choice, we see the same thing. The Molinist holds to the view of simultaneous concurrence, wherein God provides the indeterminate being of the choice and the creature provides the details that determine it to this or that particular choice, and so the atomicity of the details of the choice remains. The compatibilist holds that God causes the whole choice, details and all, and so there is just no need to decompose the choice any further in the account.

By contrast, the omni-instrumental view requires that we analyze the details of the choice in terms of what aspects of it originate from God and the creature. I discussed this in some detail in the third post on the view, but to summarize the point briefly: the goodness of the choice—its proper ordering toward an end truly worth pursuing—arises from God as the primary cause, whereas the evil of the choice—its failure to fully realize this goodness—arises from the creature as a limited secondary cause. Any choice is a result of one or both  of these factors coming together in some way. A purely good choice is one in which the creature introduces no privation of the goodness originating from God, whereas a choice is said to be evil precisely insofar as such a privation is present.[2] The upshot of this is that God causes evil actions without causing the evil in actions.

This conclusion helps us to see that there is an important ambiguity in Craig’s argument: when it speaks of “evil” in (2) and (3), is it speaking of evil actions or the evil in actions? This ambiguity is not relevant to Craig or White because their respective positions treat actions as atomic. But for the omni-instrumental position, it affects which of the two premises we should reject.

If what the argument has in mind is evil actions, then we should deny (3). God is the author of evil actions in the sense that he is the primary cause of these actions. This does not mean, however, that he is the cause of evil as such, because God the evil of these actions does not arise from him, but from the creature.

If, on the other hand, the argument has in mind the evil in actions, then we should deny (2). The CCFs concern the actions taken by creatures, whether good or evil. These are logically posterior to the divine decree because in causing creatures to act freely as they do, God thereby determines how they would act in various circumstances. However, as we have said, this occurs without God thereby causing the evil in the actions, and therefore (2) is false.

So, then, as far as the omni-instrumental position is concerned, Craig’s argument fails to establish its conclusion. But the exact reason why depends on how we nuance the reference to “evil” in the argument, in light of the distinction between causing evil actions and causing the evil in actions.


[1] I have discussed both views in my posts “Omni-instrumentality 4: Contrasting Views” and “God’s control and our free will.” As mentioned in the former post, the word “compatibilism” admits of multiple senses. Here we take it in its most common sense (in the context of this debate) or soft determinism, which rejects the libertarian view of human freedom in order to justify divine determinism as the mechanism of providence.

[2] There is no such thing as a purely evil action in the sense of there being no good whatsoever, since evil is the privation of good and therefore parasitic upon it. Nevertheless, there are purely evil actions in the colloquial sense of an action being so significant that there is no way for it to be morally justified by the good within it.

Happiness as a common good

Suppose you’re wandering through a desert wasteland, outside of the borders of any political community, and you come across another person. Why should you care about their well-being rather than attack them or forcibly take their possessions for yourself? The aim of this post is to briefly outline an Aristotelian approach to this question.[1]

As Aristotle explains in the first book of the Nicomachean Ethics (NE), everything we do is done ultimately in pursuit of happiness. At the outset, “happiness” is simply the label for something we don’t yet understand, but which we know enough to reason about. Pretty quickly, Aristotle establishes that (1) it is the chief or final good, something desirable for its own sake and never for the sake of something else; (2) it is self-sufficient, such that when isolated it makes life worth living and lacking in nothing; and (3) via the function argument that it is about living life in accordance with complete virtue, rather than in simply having something or being in some passive state. For the most part, Aristotle spends NE exploring the different aspects of these ideas, like how we develop the virtues, what virtues there are, the role and nature of friendship, how virtue relates to pleasure, and so on. Throughout all of this, it is clear that Aristotle sees human life as something lived in community with others, but as far as I’m aware there is no explicit discussion about this until the closing section of the book, which serves as a transition to the Politics. Then, early in the first book of the Politics, he makes the following point:

We thus see that the city exists by nature and that it is prior to the individual. For if the individual is not self-sufficient when he is isolated he will stand in the same relation to the whole as other parts do to their wholes. (Politics I.2)

His reasoning here presupposes the foundation laid in the NE. We said that humans by nature pursue happiness, and that whatever happiness is it is self-sufficient and an activity of life. So, if individual humans are not self-sufficient, then whatever happiness is it cannot be limited to an individual life alone—that is, it cannot be something which individuals have in isolation from one another. Rather, it must be that happiness is a common good rather than a private good.

Generally speaking, a common good is a good of a community. But when we say that happiness is a common good rather than a private good we have something more specific in mind. To see this, we need to distinguish between different kinds of common goods. First, there is what we could call a cumulative good, which is a sum of goods belonging to the members within the community. This is opposed to a singular good, which is the good of just one member of the community. My wealth is my singular good, but the wealth of the entire country is a cumulative good made up of the wealth of all the individuals therein. Second, there is what we could call a communal good, which is a good that belongs to the community rather than to the members of that community. This is opposed to an individual good. With communal goods, the community is a sort of quasi-individual, so that a good either belongs to the community or to one of the individuals, but not both. Public parks or government buildings are communal goods, whereas private property is an individual good. When we say that happiness is a common good, we have neither the cumulative nor the communal senses in mind. Rather, what we mean is that happiness is a communicable good, something which is shared by the members of a community without thereby being diminished. This is opposed to a private good, which is something which either cannot be shared or something that is diminished when shared. A public radio broadcast, justice, and victory in a sports match are communicable goods, whereas particular food or my own health are private goods. So, when we say that happiness is a common good, we do not deny that it is an individual good. Rather, it is something enjoyed by each of the individuals in the community without having to be carved up and apportioned to those members.[2]

Happiness must be a common good because it must be self-sufficient, and individual humans lack self-sufficiency in all sorts of ways. Let’s briefly summarize the areas in which we lack self-sufficiency. First, we depend on others for both the acquisition and maintenance of the resources we need to live a happy life, including external goods, bodily goods, and goods of the soul such as the virtues. Second, we depend on others in the proper exercise of the virtues: the complexities of life and our own limitations all make it difficult to figure out how best to act in any given situation (prudence), and even when we know what to do it is difficult to distance ourselves from our passions, in order to critically evaluate them so that we might regulate our desires (temperance) and our responses to obstacles (fortitude).[3] Such reasoning and distancing, which is required for living in accordance with the virtues, can only be done reliably in cooperation with others. Third, MacIntyre suggests that our own conception of self depends on others knowing us, and is necessary if we are to imagine possible futures for ourselves that guide our daily choices toward virtuous lives.[4] And fourth, we are vulnerable to various impairments throughout our lives, which hinder our ability to live in accordance with complete virtue. We start out impaired as children, totally dependent upon those around us to develop us into responsible practical reasoners. As adults we remain vulnerable to illness, injury, disability, old age, extortion and manipulation by others, and unfortunate or unexpected outcomes.[5]

Once we understand the ways in which individual humans lack self-sufficiency, we can draw conclusions about what a community must be like in order for its common good to constitute happiness. For one, the common life of this community must be governed by a principle we can call “distributive reciprocity.”[6] By reciprocity we mean that the community forms a network of givers and receivers, from which we benefit as receivers and to which we contribute by giving of ourselves for the sake of others. It is by virtue of such a network that the community as a whole can move beyond the limitations of its individual members. This reciprocity is distributive rather than commutative, because it is not based on repaying this or that person for something we have received from them, but rather based on what benefits others and our ability to help them, especially in times of need. Often those to whom we give will be other than those from whom we have received, and rarely will we be required to give exactly what we received, or even something commensurable with it. The point of distributive reciprocity is not to “balance my books” with those around me (as it would be in commutative reciprocity), but to participate in a community of unconditional giving and humble receiving, so that we may together overcome our individual lack of self-sufficiency.

Commutative reciprocity certainly has a place in such a community (it will be important for any trading context, for instance) but it cannot be the sort of reciprocity that governs the community at a fundamental level, for two reasons. First, in a community governed by commutative reciprocity each member must give or receive for the sake of their own private good, which we’ve seen is not self-sufficient enough to constitute happiness. Second, in a community governed by commutative reciprocity, receiving from someone is conditioned on the ability to give to them, which does not produce the sort of stability necessary for happiness. We start out impaired as children, reliant on our parents and teachers to train us up to be independent adults. This benefit is received without us first paying our parents something, and indeed many of us will never be able to pay our parents back for the goods they have secured for us. Beyond this, there is no guarantee about when we will succumb to our vulnerabilities, and so no guarantee that we would be able to give to those people enough to justify us receiving their help. Thus, we are left just as vulnerable as we were as an individual—if not more, because of the possibility of people taking advantage of us when we’re in need.

So, then, happiness is the common (communicable) good of a community achieved in a shared life governed by a principle of distributive reciprocity. This is why Aristotle concludes from our lack of self-sufficiency that the individual is naturally like a part to a whole. This whole goes by various names: Aristotle calls it the city or state, we could call it the political community, or we might call it the complete community in contrast to the many incomplete communities to which we belong. The completeness of a community refers to the range of the human good that the community encompasses. In the complete community the members participate as humans and the common good of the community is therefore human happiness itself, which is the complete human good. But in incomplete communities the members participate in more limited capacities—as a father, as a worker, as a team player, as a student, etc.—and so the common goods of these communities do not extend to the full scope of the human good for its members, but only to scopes that correspond to these limited capacities. The complete community comes into existence when the various overlapping incomplete communities are organized in order to achieve happiness, since otherwise there would be some human goods outside of the purview of the complete community, which would undermine its completeness.

Returning to the question that we posed at the beginning: why should you care about the well-being of the person you encounter in the desert wasteland rather than attack them or forcibly take their possessions? All other things being equal, it’s because as humans you are both ordered towards happiness and happiness is a common good of the shared life of a complete community. Thus, you should work together as the start of such a community. On the other hand, if they start attacking you or show themselves to be untrustworthy, then the wise thing to do would be to avoid them or defend yourself rather than pursuing such a community with them. Or maybe they are trustworthy but not generally virtuous, in which case you can work with them on a utility basis while at the same time trying to reduce their negative influence on you—after all, even in the complete community there are relationships governed only by utility. Of course, there is more to consider, but we can start to see how the Aristotelian framework helps us to evaluate the factors at play, explaining both why seeking the well-being of others is worthwhile but at the same time helping us reason about when it is inappropriate.


[1] I have been thinking about this sort of question on and off over the last few years, but until recently have struggled to make any progress in answering it. It started when I wrote a post on the good of others, and then later came across James Chastek’s post on two reasons for forming political associations. After reading Chastek, I realized I had missed something important in my original post, but I wasn’t entirely clear on what it was. Over the years I have explored various parts of the underlying Aristotelian theory, but was unable to come to any satisfactory conclusions. That is, until at the suggestion of others I began reading Alasdair MacIntyre’s work, particularly his book Dependent Rational Animals, which I would highly recommend to anyone interested in these sorts of questions.

[2] For more on common goods, see Charles De Koninck’s “On the Primacy of the Common Good” and Jeffery Nicholas’s “The Common Good, Rights, and Catholic Social Thought.”

[3] We mention only the cardinal virtues here since these cover all the other virtues. We have left out justice (the fourth cardinal virtue) because it presupposes that we are ordered toward living in community with others, which is what we’re trying to establish. See David Oderberg, “On the Cardinality of the Cardinal Virtues” and Aquinas’s ST II-I, Q 61, A 2 and II-II, QQ 47–170.

[4] Alasdair MacIntyre, Dependent Rational Animals, p. 94–6. The gist of his argument is that any person’s awareness of their own identity is direct and ungrounded by deeper criteria which could be used to form a concept of self, while other people’s awareness of that person’s identity are necessarily criteria-grounded. A person’s own concept of self arises, then, by their recognition of the coincidence of these two ascriptions, allowing them to adopt a generally reliable conception of self through knowledge of them by others.

[5] See MacIntyre, Dependent Rational Animals, pp. 63–128 for a lengthy discussion of all of this.

[6] This is the name I have given to the version of reciprocity discussed but left unnamed by MacIntyre in Dependent Rational Animals, pp. 99–101. He later introduces the name “just generosity” to describe the virtue that is necessary to achieve this sort of reciprocity within a community (pp. 126–8).

God’s act of choosing

Classical theism holds that God is absolutely simple, which is to say that there is no absolute distinction within him, sometimes summarized by the phrase “all that is in God just is God.” For Thomists, this entails that God must be purely actual, which is to say that there is no mixture of potentiality and actuality in him as there are in his creatures—everything in God is one act. But if everything is one act, then how could he have acted otherwise? I have discussed this in the past, where I criticized the reasoning behind this question, so here I want to delve a bit more into the positive account of how I think about God’s freedom.

The basic problem is that the following statements seem incompatible with one another:

  1. God engages in the same act in every possible world.
  2. This act of God occurs in the same circumstance in every possible world.
  3. God determines which possible world is actualized.
  4. A different possible world could have been actualized.

It might not be clear why (2) is necessary, but if it weren’t then we could easily explain how the same act results in different possible worlds with reference to the difference in circumstances in which the act occurs. Such a move is not open to the Thomist, however, because if the circumstance of God’s choice differed between possible worlds then it would be contingent, and we hold that God is the cause of all contingent truths. Denying (3) amounts to denying that God is the creator and denying (4) amounts to a modal collapse, both of which are problematic. In my opinion, the answer isn’t to be found in denying any of these statements, but in clarifying what is involved in this single act of God.

The one act of God involves many things that for us obtain in different acts: by this one act, God knows, and chooses, and causes, and loves, and so on. These things are diversified in us because they result from the actualization of diverse potentials, but in God there is no potentiality and therefore no such diversification. Since the one act of God involves various such aspects, one aspect might help us to understand something about this act that another does not. In particular, if we consider the act in terms of free choice, then I think we can understand how the above four statements are compatible with one another.

Something that is often discussed in the context of free choice is the ability to choose otherwise. As I’ve mentioned in a previous post, I don’t think we should elevate this to a universal requirement of free choice, but it does apply in many cases. Regardless, we should certainly say that it applies in the case of God’s choice of which possible world to actualize. In cases where it does apply, we say that an agent in circumstance C freely chooses A only if they could have chosen something other than A in the same circumstance. The “circumstance,” here, is taken to include the entire state of affairs causally prior to the choice by the agent, including the agent’s own psychological state. Understandably, then, one objection that is sometimes leveled against such views is that it seems impossible to explain why the agent chose as they did rather than the alternative.

It seems to me that we can give an explanation that aligns well with our intuitions about our own choices. If we suppose that a reason for preferring one option over another results in a corresponding power for choosing that option rather than the alternative, then the fact that we can have reasons for various options means we will have the corresponding powers for choosing between those options. Imagine that Alice is choosing between two incompatible options, A and B. Then the circumstance C prior to her choice will include reason RA for choosing A over B and reason RB for choosing B over A. In this case, RA will result in a power in Alice to choose A rather than B, and RB will result in a power in Alice to choose B rather than A. Assuming that Alice in fact chooses A rather than B then RA will explain her choice, and conversely if she chooses B rather than A then RB will do the explaining. In either case, then, C has the sufficient resources to explain the choice that Alice makes, even though she could have made a different choice.

However, there needs to be more in the picture if we are to fully account for Alice’s choice. After all, if Alice had all these reasons but then something external forced her to choose one way or the other, then we wouldn’t say that she freely chose anything. Because of this, we need to include an additional condition, to the effect that a free choice is something that arises from the agent themselves rather than being externally imposed upon them. From an Aristotelian perspective, this means that a free choice must result from an act within Alice, in which the agent moves themselves from being impressed by reasons for her different options to pursuing one of those options rather than the others. More generally, this “act of choosing” or “act of arbitration” is one in which the agent comprehends all the reasons for the various options, arbitrates between them, and exercises the relevant power for choice that is grounded by the relevant reasons.

Now, this act of choosing will be the same regardless of which option ends up being chosen. We can retroactively qualify in terms of the particular choice that was made—as the act of choosing A or the act of choosing B—but this will just be a world-relative way of thinking about an act which in itself is unqualified and the same in each world—the act of choosing between A and B. After all, the act of choosing is part of the explanation which is supposed to be compatible with either choice. And if there were distinct acts of choosing A and of choosing B, then neither would be compatible with the contrary choice.

So, in the case of human choice, we can distinguish between three stages. In the first stage we understand something as worthy of pursuit for some reason, which includes it among the options of our choice and constitutes a power for us to pursue it. In the second stage, we apply the act of choosing, whereby we comprehend the options, arbitrate between them, and exercise one of these powers. This transitions us to the third stage, wherein we actually pursue the option we chose in the second. We may speculate that in some choices (such as the choice to believe something) there is really no third stage. But at least in most cases, the third stage is necessary in humans because we will need to move or change ourselves in some way in order to pursue anything.

Human choice is broken up into stages like this because everything we do is achieved through the successive actualizations of various potentials. By contrast, in God there are no potentials, and therefore no need for such a succession. God does not need to discover anything, or deliberate over the options by considering one feature and then the next. Instead, in his one act he knows the options immediately and altogether. As such, both first and second stages of the above process occur together: God comprehends the options and arbitrates between them all at once. Furthermore, God does not need to change himself in order to pursue the option he chooses. He doesn’t need to get his body in position or start thinking about something else before he can do something, but immediately causes his choice. So either we should say that there just is no third stage with God, or that it occurs along with the first two.

The upshot of this is that God’s one act involves an act of choosing, and an act of choosing is something which allows for alternate outcomes. So, while it’s true that God does the same thing in every possible world, the thing he does involves choosing which world to actualize, and is therefore compatible with any number of worlds resulting from it.

A simpler argument about essentially ordered series

In the previous post I explained what an essentially ordered series is and gave an argument to the effect that every such series must have a first member. While I still think the argument works, discussions about the post with others have led me to realize that we can give a considerably simpler argument for the same conclusion.

Start with a case of one thing causally influencing another, say Alice picking up her mug from the table. The thing doing the causing (Alice) is the agent and the thing being influenced (the mug) is the patient. Both the agent and the patient contribute something that enables the causation to occur, namely the actuality in the agent and the potential in the patient,[1] and the causation itself consists in these contributions coming together in such a way that the potential can be actualized by the actuality.

A causal series is a series of successive causations. When this occurs, the members of the series fall into one of three categories: the first member (if there is one) is an agent without being a patient; the last member is a patient without being an agent; and each intermediate member is in some sense both an agent and a patient. For any intermediate member, its agency and its patiency will coincide either accidentally or essentially. They coincide accidentally when the actuality whereby the member is an agent is to some extent distinct from the potential whereby it is a patient. For instance, when Bob lifts Charlie so that he can grab an item off the top shelf, Charlie has the potential to be lifted and the actuality of grabbing, where the latter is distinct from the former.[2] An intermediate member’s agency and patiency coincide essentially when the actuality whereby it is an agent just is the actualization of the potential whereby it is a patient. For instance, when Alice moves a stone in a shape on the ground using a stick, the actuality in the stick whereby it moves the stone just is the actualization of its potential to be moved in that way.

This essential coincidence of agency and patiency captures a unique middle ground between the “pure” agent and patient that we started with, and for this reason we introduce a third term to describe this sort of member: instrument. Similar to an agent, an instrument has an actuality in virtue of which it actualizes the potential of something else. Unlike an agent, this actuality is something the instrument receives from another as part of the functioning of the causal series rather than something that it contributes itself. Similar to a patient, then, the only thing the instrument contributes to the series is the potential. But unlike a patient, the potential contributed by the instrument is the potential for actualizing a potential in the next member of the series.

By contrast, the accidental coincidence of agency and patiency does not require us to introduce another sort of member, since in this case the intermediate member is separately a patient of the member preceding it and an agent of the member succeeding it.

We are now in a position to introduce the distinction between two types of causal series. An accidentally ordered series is a causal series in which the agency and patiency of every intermediate member coincide accidentally, whereas an essentially ordered series is one in which they coincide essentially. In other words, none of the intermediate members in an accidentally ordered series are instruments, while in an essentially ordered series all of them are instruments.[3]

Given this distinction, we can see quite easily that every essentially ordered series must have a first member. After all, if any such series didn’t have a first member, then the only members in the series would be the instruments and the patient. But instruments and patients only contribute potentials to the series, and therefore this series would not have any actuality in it. But without any actuality, no potentials can be actualized to bring about any effect. And therefore, this series wouldn’t in fact cause anything at all.


[1] More technically, a contribution of a member M to a causal series S is some intrinsic feature of M which makes it possible for S to obtain and which it does not receive as part of S obtaining.

[2] There can still be some overlap or dependency and still be considered distinct, so long as the actualization of the potential is not identical to the actuality.

[3] We are restricting ourselves to causal series where there is at least one intermediate member so that we can give succinct versions of these definitions.

Every essentially ordered series has a first member

A causal series or chain is an ordered collection of members, where earlier members cause or explain later members. Every member other than the first is preceded by one other member, and every member other than the last is succeeded by one other member. We can distinguish between accidentally ordered series and essentially order series. These don’t exhaust all the casual series there could be, but they do pick out important groups.

The difference between these two types of causal series can be illustrated with paradigm examples. A paradigm case of accidentally ordered series is that of parents begetting children: x begets y, y begets z, and so on. On the other hand, a paradigm case of essentially ordered series is that of someone moving a stone in a shape on the ground using a stick.

In both of these cases the later members depend on earlier members, but they do so in different ways. In the stick-and-stone case the stick moves the stone only insofar as it is moved by the person, whereas in the begetting case it is not true that y begets z only insofar as x begets y. Put another way, in the begetting case there is a measure of independence between members that is absent in the stick-and-stone case. Or to put it yet another way, in the stick-and-stone case we have an example of caused causing — the person causes the stick’s causing the stone to move — whereas this is not so in the begetting case.

Contrary to the impression we might get from these examples, the difference is not simply about whether the causes are simultaneous or not. Consider the example of Bob holding Charlie up while Charlie reaches for and picks up something off a high shelf. In this case, Charlie’s reaching is simultaneous with Bob’s lifting, but the latter does not cause the former and so this is an example of an accidentally ordered series.

In my opinion, the best description of the difference between the two types is the third difference we mentioned above: essentially ordered series involve caused causations while accidentally ordered series do not. To put it a bit more generally, we could say that in an accidentally ordered series each member causes the next without causing the members after it, whereas in an essentially ordered series each member causes the next and all subsequent members at once, by causing the later ones through the earlier ones.

Among other things, this description is best because it explains the other two descriptions we gave. The independence that obtains in accidentally ordered series has to do with the fact that members are not caused by members not directly preceding them. And what we mean by “only insofar as” is captured by the fact that intermediate members cause later members because they are caused to do so, like the movement of the stone is caused through the movement of the stick.

Given this difference between these two types of series, we can argue that every essentially ordered series must have a first member as follows:

  1. The entirety of any essentially ordered series forms an act.
  2. This act must arise from one of the members within the series.
  3. The member from which this act arises must be the first member of the series.
  4. Therefore, every essentially ordered series has a first member.

Regarding premise (1), every causal relation belongs to some act but it is possible for multiple causal relations to belong to a single act. When x causes y’s causing z, there are three causal relations (x causing y, y causing z, x causing y’s causing z), but only one act: x causes y and z at once, by causing z through y. We might think about y causing z in isolation from anything x does, but properly speaking it is part of the act arising from x. On the other hand, it is of course possible for multiple causes to work together while belonging to separate acts. Every member in every accidentally ordered series acts separately from the other, since we have said that it causes the next member of the series without causing any of the subsequent members.

With this in mind, premise (1) says that for any essentially ordered series, there is some act to which the entire series belongs. If this weren’t the case, then there would need to be two or more acts such that any part of the series is contained in at least one of them, but the entire series is not contained in any of them. To see why this cannot be the case, assume to the contrary that x and y are two members of the series, such that x is part of some act X but not part of some act Y and y is part of Y but not X. We can assume, without loss of generality, that y is subsequent to x in the series. Now, because we’re in an essentially ordered series, x causes the next member as well as all subsequent members at once, by causing the later ones through the earlier ones. Among these, therefore, x will cause y and y’s causing whatever it does. Therefore, y will be part of the act to which x belongs, namely X, which is a contradiction. Since every member must be part of some act (otherwise it would not be doing anything), it follows that all members must belong to the same act. Therefore, the entire series forms an act.

Regarding premise (2), we note that whatever a causal series does is reducible to what its members do, and therefore we have three options for whence the act of the series arises: (a) from nothing, (b) from one member, or (c) from a multitude of members. It can’t arise from nothing, since nothing comes from nothing. It also can’t arise from a multitude of members, since any multitude will contain at least one member from which it cannot arise, even in part. After all, any multitude will have at least two members with one being subsequent to the other, and therefore one caused to do what it does by the other. In this case, the act will in no way arise from this later member, not even partially. Thus, the act does not arise from that particular multitude, but some proper part of it. Since the same could be said of any multitude, it follows that the act arises from no multitude. This leaves us with option (b), that the act arises from one member within the essentially ordered series.

Regarding premise (3), if the act arises from a single member then that member is not caused to act by members prior to it. Since every member other than the first is caused to act by members prior to it, this member from which the act arises cannot have any members prior to it. Thus, it is the first member in the series.

The conclusion in (4) is sufficient for many of the uses of essentially ordered series, such as some arguments for God’s existence or Aristotle’s argument that there must be some chief good (see the last section here). But we could also use it to say something about the finitude of essentially ordered series. Since every member has one predecessor and one successor, and since there must always be a first member, it follows that the only way for the series to be infinite is if it has no last member. Conversely, then, any essentially ordered series with a last member must be finite.

Feser on Koons on relative actuality

In a recent blog post, Edward Feser outlines a problem he sees with Rob Koons’ usage of relative actuality/potentiality to explain how an Aristotelian B-theorist might explain the reality of change. If you want to see Koons’ proposal then take a look at section 6 of his paper or take a look at Feser’s brief summary of it on his blog post; I will assume a general familiarity with it in this post. The objection is as follows:

Suppose someone suggested that time was nothing more than a spatial dimension.  I reject this view, and Rob wants to avoid it too.  One problem with it is that it also seems incompatible with the existence of real potentialities in the world.  If past, present, and future events are equally real in exactly the same way that the spatially separated hot and cold ends of a fireplace poker are (to borrow a famous example from McTaggart), then it seems that they are equally actual, just as the hot and cold ends of the poker are equally actual.  There is no real potentiality in a spatialized conception of time, and thus no real change — and thus, really, no time either.

But suppose our imagined spatializer of time defended his view by saying that there is potentiality of at least a relative sort on his conception of time.  Suppose he appealed to the poker analogy, and suggested that we could say that relative to the left side of the poker, the poker was actually cold but potentially hot, whereas relative to the right side, it was actually hot and potentially cold.  Suppose he suggested that there was therefore a kind of “change” in the poker from left to right.  And suppose he suggested that a similar kind of relative actuality and potentiality could be attributed to things and events at his spatialized points of time, and that a similar kind of change could therefore be attributed to them too.

In my view this would be clearly fallacious, involving little more than a pun on the word “change.” … I imagine Rob might agree, since, as I say, he too wants to avoid spatializing time, or at least denies that the B-theory need be interpreted as spatializing it.  But I fail to see how Rob’s notion of relative actuality and potentiality captures real potentiality, and thus real change, any more than my imagined spatializer of time does.

Feser’s point here is that the relativization of actuality/potentiality doesn’t help the spatializer to account for change, and he struggles to see how it could be comparatively any better for Koons. But I think he misdiagnoses the problem with the spatializer’s suggestion, and in fixing the diagnosis the difference between the spatializer and the Aristotelian B-theorist becomes clear.

As the last sentence of the quote makes clear, Feser thinks that the problem with the spatializer’s account lies in the fact that there are no real potentials anywhere, and so there can be no real change. But in the poker example, there surely are real potentials in the picture: the cold end is actually cold and potentially hot, whereas the hot end is actually hot and potentially cold. Anyone who accepts the Aristotelian account of change will affirm this, since we all know that pokers can change from hot to cold and back again. And they will likewise tell you can’t deny the presence of potentials simply because of the presence of actualities. Thus, if we suppose that time is spatially spread out like the poker, then the mere fact that all events are actual doesn’t prevent them from including various real potentials. Moreover, these real potentials and actualities seem to be adequately demarcated by using the notions of relative actuality and potentiality, so the problem isn’t there either.

The true problem with the spatializer’s answer is not that there are no real potentials, but that the potentials and the actualities are not connected in such a way so as to constitute change. Change, after all, is the actualization of a potential, it is not the spatial coincidence of a potential with an actuality. We can think of it like this: one reason someone might spatialize time is in order to explain why time is extended rather than collapsing into a single moment — it is extended, they say, for the same reason that space is extended, whatever that reason happens to be. But, of course, the extension of space doesn’t depend on the actualization of potentials, and therefore we know that simply pointing at spatially separated potentials and actualities doesn’t give us real instances of change. The problem, then, is that the spatializer spatializes time, and in doing so removes the possibility of an actuality ever being the actualization of a potential.

By contrast, the Aristotelian B-theorist need not spatialize time in order to explain its extension, because they have a ready alternative: the extension of time is explained by the actualizations of potentials. That is, the fundamental reason why things exist through multiple moments is that at earlier moments they have potentials which are actualized in order to bring about later moments. Of course, we’ve explained before that this is only the beginning of the account, since things can also move through time because of their real relations to other things. But even after we fill out all the details, the extension of time will always ultimately be explained with reference to the actualization of potentials — which, you will notice, is another way of stating the Aristotelian position that time is the measure of change.

So, then, the difference between the Aristotelian B-theorist and the spatializing B-theorist is that the former accounts for the extension of time in terms of the actualizations of potentials, whereas the latter has no room for the actualization of potentials because of how they account for the extension of time.

Cooperation with evil

When one person cooperates with the act of another they can “share” in that act in different ways. Knowing these differences is important, because the degree and manner of sharing can influence whether the cooperator is morally culpable for their cooperation. Here I wish to outline the different modes of cooperation by deriving them from another principle of moral decision-making, the principle of double effect.[1]

Desire, intention, and foresight

We are to pursue good and avoid evil,[2] but the complexities of our world mean that our actions will often have both good and bad effects. The principle of double effect outlines the conditions that need to be met in order for an action with both good and bad effects to be morally permissible. The principle itself presupposes a more basic distinction between intention and foresight, which we will outline before proceeding to the principle itself.

It’s easy to think of cases where we foresee that our actions will have certains consequences which we find undesirable. Imagine Alice punches Bob under duress — perhaps someone is threatening to kill Bob if Alice doesn’t punch him. Alice punches Bob out of the desire to save him, and although she foresees that this will cause him pain, she does not desire to cause him pain. We might even acknowledge this difference in the less extreme case where Alice merely desires to punch Bob in order to test her strength, and foresees that this might cause him pain even though this is not what she desires. The point here is that we can recognize that our actions will produce all sorts of effects that we do not desire.

Now, in everyday speech we treat desire and intention as virtually synonymous, but when it comes to the principle of double effect we need to distinguish them. Desire is an orientation of our will wherein we recognize something as worth pursuing — this is more or less what we mean in our everyday speech. Intention, in the sense we’re using it here, includes desire but goes beyond it to include everything that makes an action what it is.[3] For instance, if Alice’s hand moves into Bob’s face, this could be because Alice is punching Bob, or because she is reaching for something near him and mistakenly hits him, or something else. The motion underlying each of these actions is the same, but the intention which informs this motion determines which action is being performed.

To put this difference another way, desire is a psychological feature within the agent performing the action whereas intention is a feature of the action itself. An action includes the psychological features of the agent performing it, but is not exhausted by them. If I’m very angry I will desire some outlet for this anger, which can be achieved by punching a wall, shouting loudly, throwing a pillow across the room, and any number of other actions. In each of these cases, the desire is the same, but the intention varies and includes all sorts of things that are beyond my desire, such as risking breaking things in the house, or damaging my hand.

Despite being distinct, desire and intention are still connected to one another. One way of articulating this connection is as follows: when we perform an action, we intend to do so under a description of the action that makes it seem desirable to us in some way — the intention is for the action as such, but the desire attaches to the action insofar as it is captured by this description.[4] When Alice intends to punch Bob under the description of testing her strength she might not desire to perform an action which causes pain as one of its effects, but she cannot avoid intending to do so because punching him is exactly this sort of action.[5]

Another way intention can be wider than desire is when we engage instrumental reasoning, where we do one thing (the means) for the sake of another (the end). Whenever we perform X as a means to Y, we intend both X and Y even though we might not desire either on their own terms, but for wholly extrinsic reasons. The case of Alice punching Bob under duress is an example of this: Alice punches Bob as a means to saving him, but there is nothing about punching Bob that she finds desirable apart from the fact that it will save him. Nevertheless, because she engages in the act of punching Bob, she intends to punch him alongside intending to save him.

Even though intention is wider than desire, it doesn’t necessarily extend to every foreseen effect of an action. This is the intention-foresight distinction. When Alice punches Bob to save him, she might foresee his pain without intending it. This is possible because an action is distinct from its effects: in this case, Alice intends an action which is such that it will cause both Bob’s pain and his salvation, but she intends one of these effects and not the other.

Principle of double effect

Having distinguished desire, intention, and foresight, we can state the conditions that make up the principle of double effect. These are the conditions under which an action with both good and bad effects is morally permissible:[6]

  1. The act intended by the agent is morally permissible.
  2. The bad effect is foreseen, but not intended.
  3. The bad effect is not a means to the good effect.
  4. The good effect is proportionate to the bad effect.

As we can see from this, rather than reducing the question to rule-following or outcome maximization, the principle of double effect ties them together into a unified set of considerations. Notice also that the principle presupposes that we already have a wider theory of how to determine when acts and effects are good or bad, or how to weigh these against each other (in the proportionality condition). It’s goal is not to give us such a theory, but to help us apply it in complicated situations.

We won’t spend all our time defending these conditions, but we can say something about each of them. Condition (1) captures the idea that intention determines the action, and so an evil intention will make the action evil. Condition (2) says that our intention must not extend to the bad effect, since in this case we would be intending something evil. Based on what we’ve already said, our intention could extend to the bad effect if the action is performed as a means to it. Condition (3) is concerned with the reverse of this, namely that the bad effect shouldn’t be pursued for the sake of the good effect. In this case, we would again be intending the bad effect, which would be a problem. A more general consequence of this is that the ends do not justify the means. Finally, condition (4) tells us that even after satisfying the other conditions, there is still the question of whether the good effect provides a sufficiently strong reason for doing something that will result in the bad effect. If the good effect doesn’t in some sense compel us to pursue it to a degree that warrants the bad effect, then we are to that extent choosing the bad effect itself, and thereby intending it. Making the judgment required of (4) can be very complicated, and involve considerations of things beyond guaranteed consequences, such as risk and danger, weighing up nice-to-haves vs need-to-haves, and so on.

Let’s apply the principle of double effect to the two scenarios involving Alice punching Bob. In the first case, where Alice punches Bob to save him, the act is justified: punching is not intrinsically immoral, she does not intend Bob’s pain, she does not use his pain as a means to his salvation, and the value of his life outweighs the cost of his temporary discomfort. In the second case, however, the act is not justified, at least because testing her strength is not sufficient reason to cause Bob pain.

Modes of cooperation

When we apply the principle of double effect to the issue of cooperation, we find these conditions result in successively more nuanced modes of cooperation. In any cooperation, there is a principal agent performing the principal act, and the cooperator is someone who performs a cooperative act which is in some way contributing to the principal act. In our case, we’re interested in the permissibility of cooperation with an evil principal act. Throughout this discussion, we will consider Charlie’s cooperation with Alice punching Bob where there are no overriding reasons, such as saving him.

The modes of cooperation in tree form. As we’ll note later, proximity and dispensability really lie on a spectrum, but we include them here to illustrate where they fit in the bigger picture.

The first distinction is between formal cooperation and material cooperation. In formal cooperation the cooperative act is the same as the principal act, and the cooperator is effectively just another principal agent. Put another way, with formal cooperation the cooperator shares the intention of the principal agent, and therefore the two of them perform the same act for the same end.[7] For example, Charlie formally cooperates with Alice when together they attack Bob. Formal cooperation with an evil act fails on the first condition of the principle of double effect, since the action intended is the principal act, which is morally wrong.

In material cooperation the principal act is not identical to the cooperative act, but in some sense an outcome of it. Stated in the terms of the principle of double effect, the principal act is an effect of the cooperative act. Material cooperation can be divided into immediate and mediate. In immediate material cooperation, though they are not identical, the cooperative act is still in some sense part of the principal act. This happens when the cooperator doesn’t share the intention of the principal agent, but nevertheless engages in part of the principal act for some other reason, such as out of fear. Since intentions determine actions, this divergence in intention results in two distinct actions rather than only one. From the outside, this kind of cooperation might look similar to formal cooperation, since in both cases the cooperator is actively engaged in the principal act. For instance, we might say that Charlie cooperates with Alice in this way when, out of fear, he holds Bob down for Alice to punch him. Immediate material cooperation with an evil act fails on the second condition of the principle of double effect. This is because whenever an act is performed as part of another, then the former is performed as a means to the end of the latter.[8] Since the principal act is evil, the cooperative act is being performed as a means to an evil end.[9] As we mentioned earlier, it follows that the bad effect of the cooperative act (ie. the principal act) is intended by the cooperator.

In mediate material cooperation the cooperative act is not part of the principal act, but merely supplies one or more conditions necessary for the principal act to occur. Mediate material cooperation can be divided into instrumental and non-instrumental forms. These aren’t commonly mentioned in discussions on cooperation with evil, but we will see that they fall out of the principle of double effect like everything else, and fit naturally into our taxonomy. In instrumental mediate material cooperation the principal act is enabled by the cooperator as a means to a benefit (or avoidance of loss) for the cooperator. For example, Charlie cooperates with Alice in this way if at Alice’s request he lures Bob into a room so that she can punch him, with the promise that she will give Charlie any money she finds on Bob’s unconscious body. This mode of cooperation fails on the third condition of the principle of double effect, since in this case the bad effect (the principal act) is a means to the good effect (the benefit, or avoidance of loss).

One possible reason this mode of cooperation isn’t often discussed is that it’s requirements are not often met in cases of cooperation. For a meditate material cooperation to be instrumental, it’s not enough for the cooperator to benefit from an evil act. Rather, the cooperator needs to benefit specifically from the principal act which is enabled by the cooperative act. In order words, the evil needs to be the bad effect of the cooperative act which is used as a means to the good effect. After all, the principle of double effect considers the effects of our actions, not the antecedents to our benefits.

Proximity and dispensability

If the cooperation is material, mediate, and non-instrumental, then it satisfies the first three conditions of the principle of double effect. This leaves us with the proportionality condition to consider. In general, the factors that need to be considered when making a proportionately judgement will vary greatly from one situation to the next. Nevertheless, the fact that we’re considering proportionality in the context of cooperation means that we can lay out two general factors that will always need to be considered alongside anything else we might consider, namely proximity and dispensability.

The proximity of a cooperation is how causally near it is to the principal act, which is to say how few actions there are between it and the principal act. Imagine Charlie invites Bob to come over to his house the next day, and when Bob eventually arrives Charlie distracts him, which gives Alice the opportunity to punch Bob. In this case, the distraction is more proximate with the punching, whereas the invitation is more remote. An important caveat is that causal proximity doesn’t always correspond to temporal proximity: if Charlie had set up the distraction to go off automatically days prior to inviting Bob, it would still be causally more proximate than the invitation.

The dispensability of a cooperation is the degree to which the principal agent is in need of the cooperator’s help specifically, rather than being able to get it equally from others. For example, if Bob is skeptical of invites from everyone except for Charlie, then Charlie would be more indispensable to Alice’s plans than if Bob would happily accept invites from a number of different people.

Proximity and dispensability are applicable whenever cooperation is material, mediate, and non-instrumental. Moreover, they are morally relevant because they determine the degree to which the cooperator has influence over whether and how the principal act might occur: the more proximate and indispensable the cooperator is, the more likely they are able to prevent the principal act from occurring at all, or at least reduce the chance or degree of its success. We describe this all with terms like “the degree to which”, “more likely”, and “reduce”, because neither proximity nor dispensability are not discrete notions, but fall along a spectrum. Together, they make up a 2-dimensional spectrum which we can visualize as follows:

The 2-dimensional spectrum created by proximity and dispensability.

When it comes to considerations about proportionality, proximity and dispensability will always be among the things that must be considered, but they will rarely (if ever) be alone. That said, all other things being equal, the closer we get to proximate and indispensable cooperation the more significant we need the good effect to be in order to warrant the cooperation; conversely, the closer we get to remote and dispensable cooperation the less significant we need the good effect to be. Somewhere in the middle are the cases of remote and indispensable cooperation, and proximate and dispensable cooperation.

It’s tempting to suppose that any proximate and indispensable cooperation will fail the proportionality condition, but this is mistaken. Consider the cooperation between a dentist’s assistant and the dentist while they are doing a root canal on a patient. Certainly this is proximate, and if we suppose that this is a private practice with only one assistant then it will also be indispensable, but it does not fail the proportionality condition since the pain caused by the procedure is warranted by the benefits it will secure for the patient. It seems, then, that the most we can say in cases where the cooperation is proximite and indispensable, is that these factors make it harder (or at least don’t make it easier) for the good effect to be proportionate to the bad effect.

Conclusion

These, then, are the various ways that we might cooperate with evil. By connecting them to the conditions of the principle of double effect we have seen that only non-instrumental mediate material cooperation with evil has a hope at being morally permissible. And even then it will ultimately depend on whether the good effect is proportionate to the evil of the principal act with which we cooperate.


[1] The inspiration for this derivation comes from David Oderberg, “Further Clarity on Cooperation and Morality”Journal of Medical Ethics 43 (2017): 192-200.

[2] Aquinas considers this a corollary of the principle Aristotle begins his ethics with, namely that the good is that which is pursued in every activity (ST II-I Q94 A2 corp, cf. NE I.1).

[3] In terms of classical philosophy, the intention is the form of the action. I have discussed this idea in my blogpost “Form vs structure, and what it means for virtual existence”.

[4] This is the so-called “guise of the good thesis”. The description under which we consider the action is the guise, and this description makes the action desirable because it casts it in a good light, thereby highlighting it as something worthy of pursuit. I have discussed the relationship between desire and goodness in my blogpost “How Aristotle starts the Nicomachean Ethics”.

[5] For a lengthy discussion of intention in the context of the principle of double effect, see Matthew B. O’Brien & Robert C. Koons, “Objects of Intention: A Hylomorphic Critique of the New Natural Law Theory”, American Catholic Philosophical Quarterly 86 (4): 655-703 (2012).

[6] For a more detailed discussion, see David Oderberg, “The Doctrine of Double Effect”, in T. O’Connor and C. Sandis (eds) A Companion to the Philosophy of Action (Oxford: Wiley-Blackwell, 2010): 324-30. The order I have used here is from David Oderberg, “Further Clarity on Cooperation and Morality”.

[7] This account of formal cooperation corresponds to our earlier point that intentions are the forms of actions.

[8] See my discussion of superordinate and subordinate activities in my blogpost “How Aristotle starts the Nicomachean Ethics”.

[9] We can distinguish the “primary” end of the cooperative act from the “superordinate” end of the cooperative act. The former is the end for which the cooperator intends to play their part in the principal act in the first place, such as fear. The latter is the end for which the cooperator intends to act insofar as this act is part of the principal act. These two ends may be distinct from each other or subsumed into a single end that encompasses both of them. Either of these options is consistent with what we’ve been saying, since in either case the cooperative act is intended for the sake of at least one end that differs from that of the principal agent’s act.

The sense in which an agent is directly involved in producing the effect

In my second post on omni-instrumentality I said the following with respect to God’s essential cooperation with the natural powers of his creatures:

… the influences of both God and creature on the final product are direct (or immediate), complete, and total, but according to different modes. That is, they both extend to the final product (direct/immediate), to all of the final product (complete), and to every detail of the final product (total), but not in the same way (different modes). It’s because they contribute according to different modes that both God and creature can influence the effect without rendering the other superfluous.

The difference, I claimed, lies in the fact that the creature causes the effect whereas God causes the causing of the effect. God is the divine agent whereas the creature is the divine instrument. Given the breadth of topics covered in that post, it’s understandable that I didn’t spend too much time dwelling on this point. Nevertheless, it is a puzzling claim to make and reader Mashsha’i has drawn attention to this in comments on the post. After some back and forth with Mashsha’i in the comments, I’ve decided it’s worth to put my thoughts together in a short follow-up post.

Starting with an imperfect but intuitive example, imagine Alice and Bob are hanging up a big painting. This can work in at least three ways, which correspond to the three modes of cooperation. First, it could be that the two of them are holding the painting together at the right height while trying to hook it against the wall, which would be coordinate cooperation. Second, it could be that Alice gave Bob the painting as a gift which he is hanging it up himself, which would be accidental cooperation. And third, it could be that Bob is busy holding up the painting while Alice instructs him how to move so as to get the painting in the correct position, which would be essential cooperation — by guiding him Alice causes Bob to cause the painting to hang properly.

With these three cases in hand, we may ask: in which of them is Alice directly involved in the hanging of the painting? At least in the first two scenarios the answer is clear: Alice is directly involved in the case of coordinate cooperation, and only indirectly involved in the case of accidental cooperation. But what about the essential cooperation? It seems to me that there is a sense in which Alice is directly involved in the hanging — since the hanging of the painting occurs by virtue of her and Bob working together — but it is equally clear that this cannot be the same sense which applies in the coordinate cooperation.1

Reflecting on the painting example, I think we can distinguish between two senses of what it means to be “directly involved”. Something is directly involved in the production of an effect if (1) it directly contributes to the effect, or (2) is contributes to an act which directly brings about the effect. These two senses are not incompatible but neither are they coextensive: whenever we have (1) we also have (2), but not vice versa. In the coordinate cooperation, both Alice and Bob directly influence the motion of the painting, and so both (1) and (2) are true of both Alice and Bob. In the accidental cooperation, both (1) and (2) are true of Bob, and neither are true of Alice. And in the essential cooperation, both Alice and Bob contribute to the act of hanging the painting but only Bob directly influences the motion of the painting, meaning that (1) and (2) are true of Bob but only (2) is true of Alice.

Using the categories of my post on cooperation, what’s common between these two senses of direct involvement is that there is combination between Alice and Bob in the production of the final product. Within this commonality, the difference between the two lies in the fact that coordinate cooperation is combination without dependence while essential cooperation is combination with dependence. The introduction of dependence in essential cooperation means that of the two things cooperating in the act, the second cause will directly contribute to the effect while the first cause will merely contribute to the act that directly brings about the effect.

In the case of God’s essential cooperation with nature, we can say two things. First, the sort of direct involvement he has in the final product of the cooperation will be the second one outlined above, but with an important qualification: apart from the particular way in which Bob depends on Alice in our essential cooperation example, he is largely independent of her; on the other hand, a creature is entirely dependent upon God for all of its being. Thus, the “distance” between Alice and Bob is far greater than that between God and the creature.

Second, while I was right (in my post on concurrence) to say that God is directly involved in the production of final products, I was wrong to equate “direct” with “immediate”. In the cooperation, God is directly involved in the production of the effect precisely by virtue of mediating his causal power through the creature in the act which directly brings about the effect.2 Thus, while God is directly involved in the production of the effect, his involvement is mediated rather than immediate. In fact, this might well another way of characterizing the difference between the two senses of direct involvement: the first sense involves direct and immediate involvement, while the latter direct and mediated involvement.


  1. In my post on cooperation I defined “direct” as “having no intermediate product” (“This occurs between two things when a product arises directly from the influence of both of them, which is to say that there are no intermediate products between the causes and their mutual product”), but we still need to explain the difference between coordinate and essential cooperation in this respect.
  2. The qualification “in the cooperation” is important, for there are many cases in which God does immediately cause the effect. For instance, parents mediate the becoming of a child, but God immediately causes its continued being.

Existential inertia in terms of inertial actuality

This is part two of my follow-up to my original post on potentiality and inertia, where I respond to comments made by defenders of existential inertia. In the previous post we delved deeper into the nature of potentials, and saw that not everything we call a potential is equally deserving of the name. This helped us clarify the indifference of potentials, and how they relate to their actualizations. In my estimation, the indifference of potentials is the core reason for thinking that potentials depend on some other actuality for their continued actualization. In the course of this post, we will have an opportunity to further unpack these two ideas and the connection between them.

The objection of interest for this post is raised both by Joe Schmid and Graham Oppy in different contexts. Oppy, in responding to a paper by Feser, says the following:

Yesterday, throughout the entire day, there was a red chair in my room. Pick some time t around noon yesterday. At t, the chair existed, and the chair was red. Moreover, at t, the chair had the potential to exist, and to be red, at t + ε, where ε is some relatively short time interval (say, a millionth of second). Do we need to postulate the existence of some distinct thing that exists through (t, t + ε) that actualizes at t + ε the potential that the chair had at t to both exist and be red at t + ε? I do not think so. Given that, at t, the chair has the potential to exist and to be red at t + ε, all that is required for the realization of this potential is that nothing intervenes to bring it about, either that the chair does not exist, or that the chair is not red, at t + ε. Potentials to remain unchanged do not require distinct actualizers; all they require is the absence of any preventers of the actualization of those potentials. In particular, things that have the potential to go on existing go on existing unless there are preventers — internal or external — that cause those things to cease to exist.1

In private correspondence, Schmid has raised a similar objection: even if we grant the continued indifference of potentials, all we need is some kind of explanation, not necessarily an explanation in terms of a concurrent sustaining efficient cause. The persistence in actuality could be explained by a whole host of explanatory factors that don’t invoke sustaining causes — for example, the prior state and existence of the object combined with no sufficiently destructive causal factors, or in terms of an existential inertia tendency, or something else.

It may come as a surprise, but for the most part I am in agreement with Oppy and Schmid: it is often the case that something with an actualized potential P does not require some external sustaining cause to sustain the actualization of P. As I mentioned in my original post, Newtonian inertia is a great example of this: physical objects remain maintain their rectilinear motion until some other thing causes them to stop. And the example of the chair’s “inertial color” given by Oppy is another good example: the chair maintains its color until something causes it to be another color. The trouble is that these examples don’t undermine a more nuanced rejection of existential inertia.

In my original post, we saw that the indifference of potentials is consistent with inertial behavior because of what we termed “inertial actualities”. There we arrived at these in a fairly roundabout way, but we can get to them more directly as follows. Say we have some S with potential P and its actualization A at time t, and we want an account of how A persists until some later time t+ε. Given that potentials are indifferent to their actualizations, it can’t be that P keeps A around. Of course, the continued existence of P is necessary for the continued existence of A, but its indifference to A makes it insufficient for the task. There must be something else in the picture, that isn’t indifferent to the continued actualization of P. At this point, a proponent of existential inertia might note that this could be an intrinsic aspect of S to maintain A in the absence of contrary causes. And it certainly could be, but we can know more about it. Since it’s not indifferent to A, it cannot be another potential. And since being is divided into potentiality and actuality, it must therefore be an actuality. Moreover, this actuality must be closely connected to P in a way that other actualities in S are not. This is what I call an “inertial actuality” since it’s an actuality that explains the inertial behavior of the actualization of P.

Going further, I have suggested that we understand an inertial actuality as an incomplete determination to some behavior, in such a way that it allows for variation in the details of how that behavior plays out.2 This incomplete determination gives us potentials for the actuality to express itself in concrete ways. The example of interest in my original post was Newtonian inertia: by virtue of something existing as a physical thing it is determined to continue its rectilinear motion (behavior) but does not fix the vector of that motion (incomplete). Let’s say we’re considering the potential some S has to be moving along some vector V. This potential arises from the incomplete determination of S to continued rectilinear motion, but is (as all real potentials are) indifferent as to whether S is actually moving along the relevant vector. However, when the potential is actualized in S, by virtue of the underlying inertial actuality S will continue along V until stopped by something else. This isn’t because P has somehow stopped being indifferent to its actualization, but because P exists in virtue of an underlying inertial actuality that was never indifferent to S’s motion. Similarly, the material constitution of a typical chair entails that it will have some color, but is consistent with this color being one among a wide range of options. Again, this incomplete determination gives rise to a variety of potentials, each of which is indifferent to whether it is actualized. But when one is actualized, the material constitution of the chair (not the potential) ensures that it will remain actualized in the absence of contrary causes.

Thus, the inertial behavior we observe in the world points us to the lower-level inertial actuality that gives some potentials existence and maintains their actualizations in the absence of contrary causes. The relationship between the two isn’t ad hoc, either, for it is the incomplete determination to the relevant behavior that gives rise to the potentials in the first place. Furthermore, such incomplete determinations have long been recognized when thinking about change, apart from questions of existential inertia.3

All of this helps us see that Oppy’s “potential to remain unchanged” is just another example of a pseudo-potential, this time made up of the potential and the inertial actuality from whence it arises. And it also helps us see the truth of the point made by both Schmid and Oppy, that the indifference of potentials doesn’t necessitate some external sustaining cause outside of the thing with the potential, for the inertial actuality is neither external nor an efficient cause.

Where does this leave us on the question of existential inertia? I think there are two things we can say here.

First, we are still left in need of external causes for actualized potentials that can’t be sustained by some lower-level inertial actuality. Aristotelians have traditionally recognized the distinction between prime matter and substantial form, and Thomists have recognized the real distinction between essence and existence.4 Now, both prime matter and essence are potentials that have no lower-level actualities, and therefore there can be no inertial actuality to maintain their actualizations. In these cases, then, we would need some external cause to sustain the actualizations of these potentials.

Second, we can shift the question of existential inertia from substances to their potentials and actualities, and thereby render moot the question of external efficient causes. The question of existential inertia typically arises in the context certain arguments for God’s existence, such as Aquinas’s argument in On Being and Essence5 and his second way.6 These formulations frame the argument in terms of efficient causes of existence, but strictly speaking all we need is the actualization of potentials by other actualities. That is, we can start with some actuality and build an essentially ordered series of actualizers leading back to a pure actuality. The question of whether this requires a separate substance from the one we started with can be settled as a corollary of what it means to be pure actuality. Despite sounding similar to Aquinas’s first way — which is also construed in terms of actuality and potentiality — this is in fact a different argument. The first way proceeds from the reduction of potentiality to actuality, whereas this argument proceeds from the continued actualization of potentials.

The upshot of all of this, I think, is a more nuanced understanding of existential inertia and its rejection. When we discuss the continued existence of some aspect of reality, we need to get clear on whether it can be explained by some inertial actuality within a substance or not. And if we wish to establish a separate sustaining cause of something, then we need to be clear on which potentials we’re considering within that thing, for not all require such an explanation.


  1. Graham Oppy, “On stage one of Feser’s ‘Aristotelian proof’”.↩︎
  2. There we spoke about this in terms of form and matter, but for this post we will try to restrict ourselves simply to language of actuality and potential.↩︎
  3. For example, they arise in the distinction between substantial and accidental forms in the course of making sense of the difference between substantial and accidental change.↩︎
  4. For my previous discussions on these, see my posts on God, matter, and necessary existence and the real distinction.↩︎
  5. See the bottom of page 240 here.↩︎
  6. See the corpus of ST I Q2 A3.↩︎