Deontology Is Compatible with Act-Consequentialism (guest post)
“It’s standard to divide the moral landscape into deontology, consequentialism, and virtue ethics, thereby assuming that these three are mutually exclusive and jointly exhaustive. I, like some others, find this deeply problematic…”
The following is a guest post by Douglas Portmore, Professor of Philosophy at Arizona State University. It is part of the series of weekly guest posts by different authors at Daily Nous this summer.
Deontology Is Compatible with Act-Consequentialism
by Douglas Portmore
In philosophy textbooks, it’s standard to divide the moral landscape into deontology, consequentialism, and virtue ethics, thereby assuming that these three are mutually exclusive and jointly exhaustive. I, like some others, find this deeply problematic.
One problem that I have with this “Textbook View” is that it assumes that “deontological theories are best understood in contrast to consequentialist ones” (Alexander & Moore 2020). I, however, believe that deontology is compatible with act-consequentialism—something that Jake Zuehl convinced me of in May 2017. Let me explain what convinced me.
Broadly construed, act-consequentialism is the view that the ultimate right-making feature of an act is that its outcome is not evaluatively outranked by that of any available alternative (Portmore manuscript). And note both that an act’s outcome is to be construed broadly to include everything that would be the case if it were performed and that to evaluatively rank (hereafter, simply ‘rank’) a set of outcomes is to rank them in terms of some evaluative notion, such as ‘good’, ‘good for’, ‘fitting to desire’, or ‘ought to be preferred’.
When we combine act-consequentialism with different ways of ranking outcomes, we get different versions of act-consequentialism. For instance, classical utilitarianism (Bentham 1780) combines act-consequentialism with the view that outcomes are to be ranked in terms of their overall goodness, which classical utilitarians take to be solely a function of the intensities and durations of the pleasures and pains that they contain. This theory implies that agents must always maximize the overall good. By contrast, ethical egoism is an act-consequentialist theory that ranks outcomes in terms of how good they are for the agent. It holds that one outcome outranks another if and only if it is better for the agent. Thus, ethical egoism implies that an agent ought to produce the outcome that’s best for themself even if it’s worse overall.
There are at least three reasons for thinking that this is how ‘act-consequentialism’ should be defined. First, it’s appropriately broad in that it includes theories that are generally regarded to be act-consequentialist, such as ethical egoism (Wikipedia), classical utilitarianism, self/other utilitarianism (Sider 1993), rights consequentialism (Sen 1982), and satisficing consequentialism. Second, it’s appropriately narrow in that it excludes theories that are generally regarded not to be act-consequentialist, such as Kantianism, virtue ethics, motive utilitarianism (Adams 1976), rule-consequentialism, and theological voluntarism (Murphy 2019). (Admittedly, deontology is often thought not to be act-consequentialist, and, yet, this definition fails to exclude all forms of it—or so I’ll be arguing. But I hope to show that this thought is mistaken as well as to diagnose what has led some to make this mistake.) Third, it accounts for the fact that “the term ‘[act-]consequentialism’ seems to be used as a family resemblance term to refer to any descendant of classic[al] utilitarianism that remains close enough to its ancestor in the important respects” (Sinnott-Armstrong 2019)—the important respects being those that account for classical utilitarianism’s tendency to “haunt even those of us will not believe in it” (Foot 1985, 196). For, as Philippa Foot noted, classical utilitarianism’s act-“consequentialist element is one of the main reasons why [classical] utilitarianism seems so compelling” as well as the main reason why “the move to rule utilitarianism seems to be an unsatisfactory answer to the problem of reconciling utilitarianism with common moral opinion” (1985, 196 & 198). And there is indeed something quite compelling about the act-consequentialist idea that we should first tell agents how to rank the various possible worlds that they could actualize through their actions and then direct them to perform whichever act would actualize the one that ranks highest—the one that they ought to prefer to every available alternative (see Portmore 2011, 34–38).
So, that’s how ‘act-consequentialism’ should be defined. How, though, should ‘deontology’ be defined? It is, I believe, best defined as the view that there are agent-centered restrictions that are ultimately grounded in our duty to respect persons—beings who are autonomous in the sense of being capable of employing reason to set and pursue their own ends. (Agent-centered restrictions prohibit an agent from performing certain types of acts even to prevent two or more others from each performing a morally comparable instance of that act-type. Thus, commonsense morality includes such a restriction in that it prohibits you from breaking your promise even to prevent two or more others from each breaking their promises.) Admittedly, some might consider this view to be too specifically Kantian to be definitive of ‘deontology’. But, even so, they should admit that this view is deontological—and, indeed, paradigmatically so. So, if I can show that this deontological view is compatible with act-consequentialism, I will have shown that deontology is compatible with act-consequentialism.
To see that this deontological view is compatible with act-consequentialism, consider Kantian act-consequentialism, or ‘Kantsequentialism’ for short (Portmore forthcoming). It holds both that
(K1) an agent’s refraining from treating persons as mere means is right because
(a) the ultimate right-making feature of an act is that its outcome is not outranked by that of any available alternative and
(b) the outcome of an agent’s refraining from treating persons as mere means is not outranked by that of any available alternative
(K2) the outcome of an agent’s refraining from treating persons as mere means is not outranked by that of any available alternative because persons are ends-in-themselves who are owed respect, and respecting them necessitates refraining from treating them as mere means.
The idea is that, given that agents owe persons respect, they ought to rank the outcome of their refraining from treating them as mere means above the outcome of their treating them as mere means, and this, in turn, explains why, given act-consequentialism, agents ought to refrain from treating persons as mere means.
From K1 and K2, it follows that
(K3) an agent’s refraining from treating persons as mere means is right ultimately because persons are ends-in-themselves who are owed respect.
And, given K1–K3, Kantsequentialism is deontological. For it includes an agent-centered restriction against treating persons as mere means that is ultimately grounded in the duty to respect persons. Thus, it implies that I should not murder one person even to prevent the mafia from murdering five others and that this is ultimately because my would-be victim is a person whom I have a duty to respect.
Of course, some may object that if Kantsequentialism holds that an agent’s refraining from treating persons as mere means is right ultimately because persons are ends-in-themselves who are owed respect, then it can’t be an act-consequentialist theory. They are assuming that act-consequentialists cannot hold this and must instead hold that what ultimately explains the fact that refraining from treating persons as mere means is right is something like the fact that its outcome is not outranked by that of any available alternative. But this assumption is false.
To demonstrate this, I’ll need to appeal to the distinction between the ultimate right-making feature of an action and the fact that ultimately explains its rightness.
To illustrate this distinction, consider the act of flipping the switch in the Trolley Problem, and assume that this act would maximize the good by maximizing the number of lives saved. According to classical utilitarianism, the ultimate right-making feature of flipping the switch is not that it would maximize the number of lives saved, but that it would maximize the good. After all, classical utilitarians hold that it would be wrong to maximize the number of lives saved if doing so would fail to maximize the good. Now, some classical utilitarians (e.g., Moore 1903) believe that ‘right’ just means ‘produces the best outcome’. And, on their view, the fact that ultimately explains the rightness of flipping the switch (given that doing so would produce the best outcome) is the fact that ‘right’ just means ‘produces the best outcome’. Thus, classical utilitarians hold that whereas the ultimate right-making feature of flipping the switch is that it maximizes the good, the fact that ultimately explains the rightness of flipping the switch is the fact that ‘right’ just means ‘produces the best outcome’.
Now, to see that it’s a mistake to think that act-consequentialists must hold that what ultimately explains the fact that a certain act is right is the fact that its outcome is not outranked by that of any available alternative, consider William Paley’s utilitarian theological voluntarism (1802). It holds both that
(UTV1) an agent’s maximizing utility is right because
(a) the ultimate right-making feature of an act is that it accords with God’s will and
(b) God wills agents to maximize utility
(UTV2) God wills agents to maximize utility because God is perfectly good, and his perfect goodness necessitates his willing agents to maximize utility—utility being what’s good.
And, from UTV1 and UTV2, it follows that
(UTV3) an agent’s maximizing utility is right ultimately because God is perfectly good.
In other words, utilitarian theological voluntarism holds that what ultimately explains the fact that maximizing utility is right is the fact that God is perfectly good, not the fact that God wills agents to maximize utility. Yet, Paley’s theory is clearly theological voluntarist, for it is committed to UTV1a—the idea that the ultimate right-making feature of an act is that it accords with God’s will.
It seems, then, that moral theories are to be differentiated in terms of what they take the ultimate right-making feature of actions to be, not in terms of what fact they hold ultimately explains their rightness. And if Paley’s theory counts as theological voluntarist given its commitment to UTV1a and despite its holding that what ultimately explains the fact that maximizing utility is right is the fact that God is perfectly good, then Kantsequentialism must count as act-consequentialist given its commitment to K1a (the idea that the ultimate right-making feature of an act is that its outcome is not outranked by that of any available alternative) and despite its holding that what ultimately explains the fact that refraining from treating persons as mere means is right is the fact that persons are ends-in-themselves who are owed respect.
Thus, Kantsequentialism is both deontological and act-consequentialist. And that means that deontology is compatible with act-consequentialism.
Shameless self-promotion but perhaps people might be interested in the considering the compatibility of Kant’s ethics and consequentialism from a historical perspective: https://philpapers.org/rec/WALKAC
Also relevant is Jens Timmermann’s article about how ‘deontology’ is a bad word for categorizing a moral theory: https://philpapers.org/rec/TIMVWWReport
I also felt the word “deontology” is confusing especially when talking about Kantianism. Kant had many formulas so it’s hard to pigeon hole his ethics into an abstract category as deontology.
I’ve always thought Kant’s formulas were somewhat consequentialist albeit with specific results: avoid practical contradictions, avoid treating others as mere means, and making sure actions can be universalizable.
Confucianism is also lumped into deontology since it does prescribe certain actions an agent should do although there is no justification for such actions e.g. rectification of names: a father must be a father, a mother must be mother, a child must be a child, a teacher must be a teacher, etc. These are prescribed roles without any justifications.Report
Thanks for the interesting post, Doug. I think I understand the view. What you’re saying is: act consequentialists believe that right acts are made right by how their outcomes compare, BUT: what makes one outcome outrank another can be an intuitively deontological thing, like the duty to respect persons.
This first struck me as a bit unusual. The duty to respect persons is not itself a consequentialist duty. So aren’t you giving up the consequentialist game?
But now I think you’re right: the view does count as an act consequentialist theory of right actions. It’s just unusually perched upon a nonconsequentialist theory of respect.
I have another worry, though. Most deontologists don’t think the goodness of outcomes directly determines which acts are right. They believe in other sources of reasons: in particular, reasons that have to do with fitting one’s action to the situation.
Take Ross’s view of promises — ground zero for modern deontology. Ross says:
For Ross, the reason why you should keep the promise isn’t “the outcome of promise-keeping would be good,” but instead it’s something like “I promised to do it.” Another example: Judy Thomson would say that your body rights themselves make it wrong for me to pinch your nose.
For you, these views count as deontological but not consequentialist. I would agree. But I’m inclined to say that Ross and Thomson-type views are the only sorts of deontological views. We can define consequentialism and deontology directly in terms of the most fundamental right-makers.
As I see it, consequentialists think all acts are made right by outcomes, and deontologists think some acts are made right by their fittingness. This view of deontology (i) correctly classifies the main views and (ii) preserves the traditional contrast with consequentialism.
You’ve convinced me that Kantsequentialism should count as consequentialist. But I still wouldn’t count it as true deontology. It just mimics deontology, by basing its ranking of outcomes on the fittingness of respect for persons.Report
You write: “consequentialists think all acts are made right by outcomes, and deontologists think some acts are made right by their fittingness.” I agree, of course, with the first clause. Regarding the second clause, wouldn’t all deontologists and all consequentialists agree that an act is right if and only if it is fitting to perform it? It seems to me that they just disagree on what makes an act fitting to perform. In any case, are you suggesting that a theory is deontological if and only if it holds that the ultimately right-making feature of an action is that it is fitting to perform it? I don’t think that this is right. But before I say why, please confirm whether this is your view. And if not, please state what you take to be necessary and sufficient conditions for a moral theory’s being deontological.Report
Short answer: basically yes! Long answer:
Deontologists think some rightmakers aren’t features of outcomes (which the action promotes), but instead features of the situation (to which the action is fitting). That’s how I use the term, anyway!
I’m surprised that you’re so skeptical of the contast between “fittingness” and outcome-based reasons. It’s got a history:
I suppose nowadays you’d agree with Ross, who came to think all reasons are fittingness based (see Hurka’s British Ethical Theorists, p. 74). But like Hurka, I think there’s a useful distinction between two types of reasons here. Only some have to do with fitting actions; others, with good outcomes.Report
I’m fine with the contrast between fittingness reasons and outcome-based “reasons.” For instance, I’m fine with one saying that some fact that counts in favor of P’s being true is a fittingness reason for you to believe that P, whereas the fact that an evil demon will kill you if you don’t believe P is an outcome-based reason for you to believe that P. (But, although I’m fine with the distinction, I actually think that so so-called outcome-based “reasons” are not genuine reasons at all. That is, I’m error theorist about one half of the distinction.) In any case, I think that all moral theorists should accept that an act is right (or ought to be performed) if and only if it is fitting to perform it (or, more accurately, fitting to intend to perform it) — or, if you like, there is most fittingness reasons to perform it. Why can’t an act-consequentialist accept that an act ought to be performed if and only if there is most fittingness reasons to perform it? Isn’t the difference between those who accept and those who deny act-consequentialism just whether they think that how an act’s outcome ranks is what determines how much fittingness reason there is to perform it?Report
In other words, you seem to be assuming that the fittingness reasons for performing an act can’t be a function of the consequences of performing that act just because it can’t be a function of the consequences of intending to perform it.Report
I agree that reasons for attitudes are all about fittingness. But I don’t think the same about reasons for action.
If I donate cash to promote utility, the utility isn’t a fixed feature of the situation to which I’m responding; it’s something that might or might not be brought about, depending on what I do.
The natural way to express my reason is by saying: I donate in order to promote utility. “In order to” is the sign of an outcome-based reason rather than a fittingness reason. (We wouldn’t say “I believe p in order to…) The reason fits snugly into a non-finite infinitive clause: “to promote utility.”
By contrast, we tend to express fittingness reasons with finite clauses in the indicative mood, as in: “I believe in dinosaurs because I saw a bunch of fossils at the museum,” or “I picked her up from the airport because I promised to give her a ride.”
That’s why I think consequentialist reasons for action aren’t based in fittingness: they’re not fixed features of the situation, and they’re naturally expressed with “in order to.” I wasn’t assuming anything about intention. (Where did I mention intentions?)
Thanks for the replies, by the way. I’ve been wanting to ask you about fittingness reasons for a while, so it’s nice to hear your thoughts.Report
Could you say what you mean by ‘fittingness’? I think of the fittingness of something as a matter of the accuracy of its representational content. But actions don’t have representational content, only attitudes do. So, I’m unclear what it means to say that an act is fitting. That’s why I interpreted you as talking about intentions. Intentions, unlike actions, have representational content. Also, do you think that reasons for action are distinct from reasons for intending to perform that action? If you do, could you say how the two can come apart? I’m with Scanlon in thinking that there are no reasons for action apart from the reasons there are for intending to perform them. After all, we only have control over actions insofar as we have control over our intentions and the world happens to cooperate by allowing our intentions to be efficacious.Report
Why think that?
Suppose I have an unjustified true belief. My belief is accurate, but it doesn’t fit my evidence, so it isn’t supported by fittingness reasons. (I’m only using that term because you used it; my preferred term is “responsive reasons.”)
What content does an intention have that an action doesn’t?
I’m with Dancy in thinking that reasons for action are prior.Report
An unjustified true belief that P represents P as being true. So, that belief is fitting on my understanding of the term of art. In general, the belief that P is fitting if and only if P is true. In any case, I’m still just wondering what you proposed alternative definition of ‘deontology’ is — the one where Kantsequentialism doesn’t count as deontological. I thought that it was that a theory is deontological if and only if it holds that the ultimately right-making feature of an act is that it is fitting to be performed. And this is why I was wondering what you meant by ‘fitting’ and why you thought that the Kantsequentialist can’t hold that an act is right if and only if it is fitting.Report
Deontologists believe some reasons for action come from features of the situation, not outcomes. You can call these “responsive” reasons, or “fittingness” reasons, or whatever you want. The key contrast is between fitting the situation and promoting an outcome.
This conception of fittingness is almost a century old. I gave sources, glosses, examples, and a grammatical test. (Fittingness reasons aren’t naturally expressed with “in order to.”)
I hope that’s enough!
This will be my last comment–back to work. Thanks again for your post.Report
Not enough for me. For I still don’t have a good grasp the distinction between fitting the situation (too metaphorical for me) and promoting an outcome. But thanks for your comments. I’ll give them more thought.Report
Small point: You say that you’re only using the term ‘fitting’ because I used it. But I was just following you in your very first comment where you said: “Most deontologists don’t think the goodness of outcomes directly determines which acts are right. They believe in other sources of reasons: in particular, reasons that have to do with fitting one’s action to the situation.” So, again, I’m just trying to get a clear statement from you of what your definition of deontology is. Is it that a theory is deontological if and only if it holds that the ultimate right-making feature of an action is that there is most fittingness reason (where a fittingness reason is a reason that has to do with fitting the action to the situation) to perform it? But if that’s it, it seems more like a trivial metaethical position rather than a substantive normative ethical position. After all, don’t all normative theories accept that you ought to do what there is most reason to do?Report
As I see it, the reason this doesn’t allow one murder to stop the Mafia from five is because the consequences are agent-centered: my treating someone as a means counts as worse (for me) than others’ treating them as means. But this doesn’t take care of the case where you murder one person to prevent yourself from later murdering five. (Unrealistic case: an oracle tells you if you don’t murder the one you’ll murder the five; more realistic case: if you don’t murder one, you will end up in circumstances where you will face daily temptation to murder, and knowing your psychology you expect you will sometimes fail to resist )
You can save deontology in that kind of case by making the consequences be agent and time centered, but I don’t know how plausible the resulting theory is.Report
I’m a bit confused by the point of Kantsequentialism (beyond earning bonus points for the awesome name!). I see the value of creating a new theory of morality because it is a good one–that addresses shortcomings of existing theories, by being more justified or yields fewer unpalatable conclusions (allowing one to lie to protect an innocent person from a maniac, or avoiding the menace of a utility monster). I also see the value of creating a new theory because it is clearly distinct from existing ones–it offers a new perspective, even if not a better one. But what exactly is the value of creating a theory that overlaps with two existing theories and blends elements of both? Is it to force others to refine the boundaries of deontology and consequentialism, and thus make them more clearly distinct perspectives, and thus more useful? The piece seems mostly to emphasize that it is possible to ground a theory in one perspective while mimicking the conclusions of the other. It’s a clever counterexample, but for what purpose?Report
The point of this post is only to argue that act-consequentialism is compatible with deontology. If you’re interested in how Kantsequentialism addresses shortcomings of existing theories, please see my “Consequentializing agent‐centered restrictions: A Kantsequentialist approach” at https://philpapers.org/rec/PORCCA-4.Report
Interesting post! Like some others, I’m not quite convinced though. It seems like an essential feature of consequentialism that the moral status of an action depends only on its consequences: on things downstream from the action. Your argument (seems like it) assumes that “having not been treated with due respect” counts as a consequence of an action. But that’s only a consequence in the sense that “being such that A Theory of Justice was written 100 years prior” is a future fact about 2071. Or so it seems to me. (If I thought the sole thing of intrinsic value was carefully made decisions, and claimed that we should maximize that value, my view wouldn’t be act-consequentialist, I claim, since it isn’t evaluating actions based on their consequences, except in a derivative and trivial sense.) I guess I would say that the interesting point your example raises is that an act consequentialist theory of the right is only compatible with certain theories of the good (since some goods are inherently “historical” or whatever we want to call them): one can’t just mix and match as one pleases.Report
Very few philosophers these days think that an essential feature of consequentialism is that the moral status of an action depends only on its consequences: on things downstream from the action. Here is just a partial list of philosophers working on consequentialism who deny that this is an essential feature of consequentialism: Elizabeth Anderson, Campbell Brown, Jamie Dreier, Paul Hurley, Michael Smith, David Sosa, etc. Indeed, I can’t think of a single contemporary philosopher who explicitly insists that this is an essential feature of consequentialism. Can you? And why would we want to restrict it to consequences downstream from the action?Report
Thanks for this helpful clarification (and interesting post). Quick follow-up: Is the Anderson piece you’re referring to Value in Ethics and Economics? Or does she argue for the view you’re ascribing to her somewhere else?Report
Yes. It’s Value in Ethics and Economics that I had in mind.Report
Huh. I’m not actually an ethicist myself, but I think that’s what I was taught consequentialism was long ago, and the first line of the SEP article on consequentialism says “Consequentialism, as its name suggests, is simply the view that normative properties depend only on consequences.” Furthermore, as your argument suggests, if we don’t define ‘consequentialism’ that way, we don’t get a distinctive genera of moral theory. With that being said, what would you (and Anderson et al) call the genera of moral theory that says the moral status of an action depends only on its consequences?Report
Just speaking for myself:
I do think consequentialist theories are those according to which the rightness (and other deontic status) of an act is a function of its consequences. I thought the point at issue in this little sub-thread was whether consequences in this taxonomy must be causal consequences. (I take it that’s what ‘downstream’ is supposed to indicate.) And as Doug notes, my view is that they needn’t be causal for the theory to be consequentialist.
Of course, as Preston points out, there’s no good reason to think that ‘consequentialist’ as used by ethicists has had a very precise analysis, and even less reason to fight over the labels. So more substantively: it seems to me that the difference between theories that make rightness a function of causal consequences and theories that make it a function of (both causal and) non-causal consequences is not very interesting and hasn’t been important in the history of ethical theory. As Shyam Nair puts it, there’s no ‘fault line’ there.Report
That’s right, Jamie. That’s why I clarified as the original poster did “on things downstream from the action.”Report
There’s no good reason to fight over labels, but there is good reason to get clear on what the possible views are. In this post, I’m not interested in fighting over labels; I’m interested in showing that a theory can hold that the evaluative ranking of outcomes is prior to the right of acts and yet also hold that the rightness of respecting persons is prior to the evaluative ranking of outcomes,Report
There’s no problem with saying that consequentialism is concerned solely with consequences. The problem is that you take consequences to be “things downstream from the action.” But nowadays most philosophers recognize that consequentialists are not just concerned with the causal (i.e., downstream) consequences of acts.Report
Sorry, when I said “downstream” I didn’t mean to imply that only causal consequences counted: there’s negligence, omissions, maybe reasons aren’t causes, maybe libertarianism is true, etc. With that being said, maybe we’re equivocating on what a consequence is. I was thinking that everyone would agree that consequences are temporally posterior to the things they are consequences of: a consequence of x is something that happens or obtains as a result of x, and so if we don’t know what happens after x occurs, we don’t know what x’s consequences are. I’m guessing that’s not true based on the above replies, but any clarity on the question would be helpful. Along similar lines, it would be helpful to know the paradigm examples of non-causal consequences in the literature, if it isn’t a bother. Thanks!Report
Temporally posterior: that’s okay with me, although of course logical consequences needn’t be temporally posterior.
As a result of: to my ear that’s causal. The only results of an act are things the act causes. But it’s okay with me if we use ‘as a result of’ to include other kinds of dependence than causal dependence.
You’re right, though, that a couple of examples should help clear things up!
I prefer to avoid trying to differentiating causal and non-causal consequences because I wish to avoid getting into the metaphysics of causation. This is why I talk about the outcome or possible world that would be actual if an act were performed. So, what’s wrong with taking act-consequentialism to say that an act is right if and only if the agent ought to prefer the possible world that it would actualize to that which would be actualized by any available alternative act? But, in any case, here’s an example: a non-causal consequence of your running is that you’re running. There’s a long history of consequentialists and non-consequentialists alike allowing that the act-consequentialists can be concerned with the value of the act itself. And that the act has been performed is not a causal consequence of performing it. It’s a logical consequence.Report
Thanks for all the helpful comments Doug and Jamie! In the end, I think we’re thinking of “consequences” differently: I would only say that logical consequences are consequences in a trivial sense, since everything entails itself. Hence, if we “count” logical consequences, the claim that the moral status of actions depends only on their consequences is compatible with moral status depending only on intrinsic facts about actions themselves, which seems paradigmatically deontological (and hence non-consequentialist) to me. (Is that an even more straightforward argument for the compatibility of deontology and act-consequentialism?) I’m sure your conception of consequences is more aligned with the literature, but I do think the difference in our conceptions is relevant to the plausibility of Doug’s argument. I also wonder if “my” conception is motivating the “textbook view” according to which deontological and act-consequentialist theories are incompatible. Anyhow, thanks again for your helpful comments, and patience with a non-specialist.Report
Thanks for the post, Doug!
Let me first say that I think your reasoning here is relatively flawless, but, much like with much of the ‘consequentializing’ literature, my takeaway is quite different.
Basically, I think what this kind of thing shows us is that its a mistake, born out of our (analytic philosophers) deeply ingrained desire to give precise conceptual analyses for our terms of art, to try to give necessary and sufficient conditions for what views count as ‘consequentialist’ and what views count as ‘deontology’. Why not think of these as kinds of family resemblance terms or ‘ways’ of approaching ethical theory?
Note – I am certainly not saying precision or conceptual analysis are bad, I think they can be incredibly useful. But don’t conclusions like the one you defend here make you, at least for a moment, wonder whether we may have taken the wrong fork in the dialectical road a few miles back?
To some extent you do address this by flagging why your definitions of the terms meet some of the relevant desiderata, so that’s fair enough. So let me frame my hesitation as a sort of objection: Why should I care whether Kantsequentialism counts as consequentialist or not? Its going to have some things to be said in its favor, and some against.
But if I’m the kind of person moved by the “spirit” of consequentialism when I read people like Mill and Sidgwick, should I really see this as capturing the thing that I think matters about my pre-theoretic “consequentialist” intuitions? And if not, then why isn’t the interesting question just whether this theory is a good theory or not, however we engage in the moral theoretic book keeping?Report
I think that it’s a mistake to try to come up with fully illuminating necessary and sufficient conditions for all ordinary words of English, but why is it a mistake to try to do so for our terms of art? The meaning of any ordinary word in English is simply a function of how it is used by competent speakers of English. And their usage might not be so tidy as to allow us to provide fully illuminating necessary and sufficient conditions. (I add ‘fully illuminating’ because it seems that we could at least say, for instance, that ‘X is game if and only if it resembles certain paradigmatic instances of games in the relevant respects’.) But when it comes to terms of art, I think that it’s important that each of us to specify how we’re going to use them in terms of fully illuminating necessary and sufficient conditions. Otherwise, we risk talking past each other. Now, I think that many others use the terms ‘act-consequentialism’ and ‘deontology’ precisely as I’ve defined them here. But I think that what’s often overlooked is that, if defined in these ways, they are not incompatible with each other. And, so, I agree that we have taken the wrong fork in the dialectal road a few miles back. We did so by tying act-consequentialism with the view that the good is prior to the right and tying deontology with the view that right is prior to the good. The problem with this is that it could be that what attitudes it is right for us to have (e.g., what kinds of things it is right for us to respect) could be prior to what outcomes are good while what outcomes are good is prior to what acts are right. And thus the rightness of attitudes is prior to the goodness of outcomes which is prior to the rightness of actions. And, in that case, it’s not that the right is prior to the good or that the good is prior to the right; it’s both. Lastly, I think that we may just disagree on what lies at the heart of views such as Mill’s and Sidgwick’s. As I see it, the spirit of act-consequentialism is, as I noted above, “that we should first tell agents how to rank the various possible worlds that they could actualize through their actions and then direct them to perform whichever act would actualize the one that ranks highest—the one that they ought to prefer to every available alternative.” And Kantsequentialism certainly captures this idea. So, you should care whether Kantsequentialism is act-consequentialism, because if it is, then it can capture this compelling idea while also capturing the idea that persons have a special sort of value as end-in-themselves and are consequently owed our respect.Report
Interesting stuff, but it seems like the most important move happens implicitly. Initially you write “Broadly construed, act-consequentialism is the view that the ultimate right-making feature of an act is that its outcome is not evaluatively outranked by that of any available alternative.” And then later “The idea is that, given that agents owe persons respect, *they ought to rank* the outcome of their refraining from treating them as mere means above the outcome of their treating them as mere means, and this, in turn, explains why, given act-consequentialism, agents ought to refrain from treating persons as mere means.” The idea that rankings are agent-relative and that it’s the agent’s own ranking—or what it should be, at least—that’s the basis of the ultimate right-making feature of the action seems to be critical here, but it also seems to be introduced without fanfare, if not covertly. Wording the first statement differently would have made it more explicit, e.g., “Broadly construed, act-consequentialism is the view that the ultimate right-making feature of an act is that its outcome should not be evaluatively outranked by that of any available alternative *by the agent*.” But then I think that more people might bumped on the question of whether that really is how they understand act consequentialism.Report
Dale: I see where you are coming from, but I think that the idea that act-consequentialism can accept an evaluator-relative ranking of outcomes is pretty old hat these days and not all that interesting. More interesting, it seems to me, is whether deontology is compatible with act-consequentialism even if we allow that there can be agent-relative act-consequentialist theories. After all, Alexander and Moore (in their SEP entry) insist on contrasting deontology with consequentialism and yet they explicitly allow that act-consequentialist theories can be agent-relative.Report
Great post. Despite being sympathetic to your view, Doug, I have a couple of questions about the details of the argument.
You say that “that moral theories are to be differentiated in terms of what they take the ultimate right-making feature of actions to be, not in terms of what fact they hold ultimately explains their rightness.” Why isn’t it a mistake, then, to define deontology as “the view that there are agent-centered restrictions that are ultimately grounded in our duty to respect persons”?
Separately, when I first read that definition, I took it to mean that our duty to respect persons provides the complete ultimate grounds for restrictions according to deontology. And you seem to suggest, for each view *, that *1a is not part of the ultimate grounds for the fact that certain acts are right according to *. For instance, if I understand you correctly, you think that K1a is not part of the ultimate grounds for the fact that refraining from treating persons as mere means is right; instead, the ultimate grounds are provided by the explanans in K2. But wouldn’t K1a and the explanans in K2 together be the ultimate grounds of the fact that refraining from treating persons as mere means is right? Alternatively, the argument could still work if deontology is defined as the view that there are agent-centered restrictions that are ultimately grounded in part in our duty to respect persons. But, in that case, I imagine deontologists might want there to be limits on what else can count as the ultimate grounds of restrictions, and that they might be especially concerned if outcomes count as part of the ultimate grounds. That is, deontologists would want to build into the definition of deontology that deontic facts are not ultimately grounded even in part by facts about outcomes.Report
Good question regarding why it isn’t a mistake to define ‘deontology’ as I have. Here’s my thinking. By ‘moral theory’, I mean a criterion of rightness (the rightness of actions, that is). And I don’t think of deontology as a criterion of rightness. In this respect, deontology is unlike, say, Kantianism: the view that the ultimate right-making feature of an act is that it doesn’t treat humanity as a mere means. To my mind, then, deontology is not a criterion of rightness but a class of criteria of rightness that hold in common a commitment to agent-centered constraints that are ultimately grounded in our duty to respect persons. An analogue of this would be a class of criteria of rightness that hold in common a commitment to impartially justifying our actions to each other. Call these theories “what-we-owe-to-each-other theories.” These include Scanlonian contractualism and Hooker’s rule-consequentialism, where Hooker says that the rule-consequentialist is ultimately committed, not to the maximization of the good, but to impartially justifying their actions to each other. Thus, I think that there’s something broader than a criterion of rightness, and deontology, unlike act-consequentialism, falls into this broader category. Another example of this broader category might be consequentialism, where this is understood as the class of criteria of rightness that are committed to maximizing the good. This would include act-utilitarianism, motive utilitarianism, and cooperative utilitarianism, but not Hooker-style rule-utilitarianism. Thus, I would say that Kantianism is a criterion of rightness that is deontological and not act-consequentialist. But I would say that Kantsequentialism is a criterion of rightness that is both deontological and act-consequentialist.
I have to go now, but I’ll reply to your second paragraph later today or tomorrow morning.Report
On your last paragraph: If deontologists want to build into the definition of ‘deontology’ that deontic facts are not ultimately grounded even in part by facts about outcomes, that’s fine by me. Ultimately, what I’m interested in is not what we label theories but whether our labels obstruct us from seeing how we can incorporate various plausible ideas into a single theory. Kantsequentialism combines the idea that there are agent-centered constraints that are ultimately grounded (at least, in part) in our duty to respect persons with the idea that we should first tell agents how to rank the various possible worlds that they could actualize through their actions and then direct them to perform whichever act would actualize the one that ranks highest—the one that they ought to prefer to every available alternative.Report
One possibility is that different normative principles apply to different concepts which, collectively, comprise moral thought.
Imagine a fat-man-trolley case in which there are a million people on the tracks. Almost everyone will concede that in such a case you ought to push the man. Obviously consequentialists will. At the same time, does anybody think that pushing the man is just? That that is what justice demands? That sounds forced to me, and I suspect to many consequentialists, too. It is more natural to say that it is unjust to push the man in this case, even though it is morally permissible (even obligatory). Consequences may compete with justice (which is deontological). In this case, the consequences win out.
There is research on people’s intuitions about ethical distribution which supports such a pluralistic conceptualization of morality (see, e.g., Konow’s “Fair and Square”). When it comes to how we ought (all-things-considered) to distribute scarce goods, people want (1) justice to be done, which is to say give people what they deserve (a deontological principle); (2) promote efficiency (i.e. consequences–more output, surplus, utility); and (3) attend to people’s minimal needs.
In this way, consequentialism and deontology are compatible like two people in a relationship: each has his or her own legitimate interests; in some cases those are consonant but other times they conflict; and sometimes one person’s wins out, sometimes the other person’s does, and that is OK.Report
I, like many deontologists, don’t think that it’s permissible to push the fat man off the bridge to stop the trolley and save the other five. To do so, would be to use the fat man as a mere means. Of course, this doesn’t get to the main point of your comment. Regarding that, let me just say that I’m concerned only with the concept of moral permissibility in this post.Report
Fascinating post! I’m tempted to tollens your ponens and thus become less ecumenical than the Textbook View rather than more ecumenical.
I’m interested in the second and third reason you offer to accept your definition of consequentialism. By the end of the post, the most important formulation seems to be: A view is act-consequentialist if and only if
(A) “the ultimate right-making feature of an act is that its outcome is not outranked by that of any available alternative”
where “outranked” has an appropriate evaluative interpretation. If that’s right, won’t some versions of the excluded views –
“Kantianism, virtue ethics, motive utilitarianism, rule-consequentialism, and theological voluntarism.”
– be compatible with act-consequentialism at least in the same sense that deontology is compatible with act-consequentialism?
For example, suppose our theory of “outranked” is cashed out in terms of what we ought to prefer and what we ought to prefer in each case is to actualize the possible world in which we perform whatever act would conform with a system of rules that if everyone conformed with, would maximize expected utility. Act-consequentialism would thus be compatible with at least a form of rule-consequentialism. Act-consequentialism wouldn’t be compatible with a rule-consequentialist view that rejected (A). But by a similar token, any form of deontology that rejects (A) won’t be compatible with act-consequentialism. So I wonder whether deontology’s relationship to act-consequentialism is different from the other theories you mention? If it’s not, then while (A) excludes some theories, it may not exclude as many as we would have expected act-consequentialism to exclude. That’s my worry about the second reason.
My worry about the third reason builds on the second. It seems a bit like (A) ushers act-consequentialism from the scene. Once it’s seen that (A) is compatible with all moral theories in one important sense, act-consequentialism will no longer “haunt” those who don’t believe it. The connection to classical utilitarianism will loosen dramatically. Kantsequentialism, after all, doesn’t look much like a descendent of classical utilitarianism in a way that I think would keep Foot up at night. Of course, that may be all the better for Kantsequentialism.
So I worry – because I found the argument otherwise very compelling – that I shouldn’t want act-consequentialism to be understood in a way that’s “appropriately broad”. Instead, I should stop inviting rights-consequentialist to my act-consequentialist dinner parties.
Since I would accept (A) along with some additional theses as the true definition of “act-consequentialism” or insist on being an “act-consequentialist 3.5e,” perhaps I’m merely engaging in a verbal dispute. I’m not sure. I found the third reason offered compelling. I’d like my conception of act-consequentialism to have some robust, intimate links to a tradition of normative ethical thought which has been at the center of a variety of debates for the better part of the last century. And (A) alone seems too sparse for that. But I also feel like a bit of stodgy conservative for table-pounding over tradition.Report
I’ll address each worry separately. Regarding your first worry: I don’t think that any version of the excluded views will come out to be compatible with act-consequentialism. Take rule-consequentialism. Rule-consequentialists are, in virtue of being rule-consequentialists, committed to the view that the ultimate right-making feature of an act is that it accords with the ideal code of rules (i.e., the code of rules that if accepted by the vast majority of everyone everywhere would produce at least as much goodness…). Given this commitment they must deny what the act-consequentialist is committed to: that the ultimate right-making feature of an act is that its outcome is not outranked by that of any available alternative. So, these two give different criteria of rightness. Now, you’re correct that we can act-consequentialize rule-consequentialism by combining the idea that the ultimate right-making feature of an act is that its outcome is not outranked by that of any available alternative with the idea that one outcome is outranked by another if and only if the act that produces the one accords with the ideal code whereas the act the produces other doesn’t. But this doesn’t show that act-consequentialism is compatible with rule-consequentialism. Rather, it shows only that there is a version of act-consequentialism that will necessarily get exactly the same deontic verdicts that rule-consequentialism gets. So, the sort of compatibility that I’m interested in is not that of yielding the same verdicts, but rather whether their essential commitments are compatible. As I see it, act-consequentialism essential commitment is that the ultimate right-making feature of an act is that its outcome is not outranked by that of any available alternative, whereas rule-consequentialism essential commitment is that the ultimate right-making feature of an act is that it accords with the ideal code of rules.Report
Thanks for replying. I’m still having trouble seeing why deontology doesn’t have a commitment to an ultimate right-maker, but rule-consequentialism does. Before reading your post, I thought deontology at least rejected the view that the ranking of outcomes was the only ultimate right-maker. After reading your post, the idea that I should demand the deontologist understands their view as involving commitments about the ultimate right-maker strikes me as wrongheaded. But then I think about the rule-consequentialist. Why is it no less wrong-headed to demand something similar of them? I would have thought before your post they should have a commitment to views about the ultimate right-maker. But after your post, I think maybe that too is wrongheaded. Or at least no more principled than requiring the deontologist to have such commitments. Is there a principle which allows me to sort essential commitments ahead of time? I don’t see what the sorting mechanism on your view is. Report
Whether deontology has a commitment to an ultimate right-maker depends on how it’s defined. I’ve defined ‘deontology’ as the view that there are agent-centered restrictions that are ultimately grounded in our duty to respect persons. So defined, it has no commitment to what the ultimate right-making feature of an action is. Now, you are clearly skeptical that my definition is the correct definition of ‘deontology’. Fair enough. I do think that my definition comports well with how most philosophers use the term, but clearly not everyone agrees — see for instance Daniel Munoz’s comments. In any case, despite entitling my post “Deontology Is Compatible with Act-Consequentialism” for the sake of its click-baitness, I’m not that interested in how we define ‘deontology’ or ‘act-consequentialism’ and am much more interested in whether a theory can accept both (1) the view that there are agent-centered restrictions that are ultimately grounded in our duty to respect persons and (2) the view that the ultimate right-making feature of an act is that its outcome is not evaluatively outranked by that of any available alternative.Report
Ah, fair. I see how the point is fundamentally about logical space rather than the exact groupings under different headings.Report
Regarding your worry about the third reason: I don’t see how (A) — that is, the view that “the ultimate right-making feature of an act is that its outcome is not outranked by that of any available alternative” — usher act-consequentialism from the scene. Indeed, (A) just is act-consequentialism. Moreover, it seems that only those who accept (A) can accept the compelling idea that we should first tell agents how to rank the various possible worlds that they could actualize through their actions and then direct them to perform whichever act would actualize the one that ranks highest—the one that they ought to prefer to every available alternative. And, as I see it, it’s this compelling idea that explains why utilitarianism tends to haunt even those who do not accept it.Report
Whoops – I did equivocate on act-consequentialism there. Your conception of act-consequentialism certainly hasn’t been ushered from the scene. But I do worry that the conception of act-consequentialism that haunts people has been ushered.
I’m not sure how to adjudicate our disagreement, but here’s one thought: I’m not sure it would be a struggle to get all of my deontological interlocutors to agree to (A). And if they did, they would shrug and say all of my work is left ahead of me in resolving the disagreements between act-consequentialism and deontology. But I don’t see how your view makes sense of that without positing that those of us disagreeing about whether an act-consequentialist or a deontological theory is a better theory of a given moral domain are all in error.
As a fan of consequentialism, I’m the wrong one to speak to this, but I find it difficult to believe that the idea that
“…we should first tell agents how to rank the various possible worlds that they could actualize through their actions and then direct them to perform whichever act would actualize the one that ranks highest—the one that they ought to prefer to every available alternative.”
is what haunts people, especially once we tell them that there are very few (perhaps no) further structural or substantive constraints on what determines the ranking. If our conception of ranking, outcome, what we ought to prefer, etc. are capacious enough, not much is ruled out. Indeed, it makes me worry about how much content of (A) really has. Do you think there are more constraints on the meanings of the terms “ranking,” “outcome,” and “ought to prefer” than you’ve laid out here. For example, does the ranking need to be a quasi-order? Thanks.Report
You’re absolutely right that “if our conception of ranking, outcome, what we ought to prefer, etc. are capacious enough, not much is ruled out.” But those who disagree with (A) don’t disagree with (A) because they think that it’s not capacious enough to generate the deontic verdicts that they think are correct. Rather, they think that it gives the incorrect explanation for why those are the correct deontic verdicts. According to (A), an act is made right by the fact that its outcome is not outranked by that of any available alternative. But those who disagree with (A) — e.g., Paul Hurley, Frances Kamm, Elizabeth Anderson, Chris Howard, etc. — think that what makes an act right has nothing to do with how its outcome ranks relative to those of the available alternative. So, your worry that (A) doesn’t have any content vis-a-vis the deontic verdicts it generates is right. For (A) is compatible with any set of deontic verdicts. But you seem to be overlooking its explanatory content. It tells us what makes/explain why right acts are right. The situation is analogous with the situation with many scientific theories. For instance, the various interpretations of quantum mechanics can all explain the data but they give different explanations for the data. And this is just the point that data underdetermines what the correct theory/explanation is — see Duhem and Quine.Report
I seem to keep giving this false impression – neither of my worries are about the extension of deontic properties. My worry is about what the explanatory content of (A) is.
Here’s an attempt to reformulate: My worry is that if the load-bearing concepts of (A) are capacious enough, then (A) has insufficient explanatory content. To illustrate, suppose someone claimed that
(B) The ultimate right-maker of an act is X, Y, and Z.
I ask, “What are ‘X,’ ‘Y,’ and ‘Z’?” And they replied, “Oh, they’re just the names of capacious variables. They could be anything: Justice, the year 1984, the mereological fusion of my nose and the Eiffel tower, etc.” I think (B) doesn’t really have any explanatory content.
I have a non-zero credence in the view that (A) in effect just is (B), but let’s set that uncharitable hunch aside.
If the various load-bearing concepts of (A) are too capacious, then (A) is closer to (B) than the act-consequentialists (in the sense that haunt others) should want. My worry, in other words, is that (A) falls prey to an objection you posed to Daniel above: “…it seems more like a trivial metaethical position rather than a substantive normative ethical position.”
Most, if not all, explanatory claims that I and others disagree about come back in when constructing the ranking. And that remains a genuine debate over explanation (not only the extension of deontic properties). I assume you allow for these disagreements about less ultimate explanation. I think this is what the debate has always been about. So, “Act-consequentialism” has been ushered from the scene in the sense that it’s been relegated to a more fundamental explanatory level – a seemingly more metaethical level – than the level that’s usually at stake in first-order debates in moral philosophy.
I do, however, feel as if I’ve missed something. In your forthcoming book, do you say more about the formal/structural properties of “rankings”? This is why I asked about quasi-orders in a previous comment. Some constraint like that may still be fairly minimal in one sense, but would rule out enough that I would no longer be worried that capaciousness leads to explanatory emptiness. And your point about logical space is a great one. So I may just be getting hung up on the further claims about what act-consequentialism is. I do think it’s important to know what act-consequentialism is, but perhaps that’s not really the main. project.Report
Sorry that I misunderstood you. Note that, in (A), ‘outrank’ is short for ‘evaluatively outrank’, as noted in the original post. And note that “to evaluatively rank (hereafter, simply ‘rank’) a set of outcomes is to rank them in terms of some evaluative notion, such as ‘good’, ‘good for’, ‘fitting to desire’, or ‘ought to be preferred.” And I believe that all evaluative notions can be analyzed in terms of what it is fitting for the relevant subject to desire. For instance, I think, very roughly, that the good is what it is fitting for an impartial benevolent spectator to desire. And, very roughly, what’s good for a subject is what it is fitting to desire insofar as one rationally cares for that subject. Consequently, I believe that X outranks Y on an agent’s evaluative ranking if and only if it is fitting for that agent to prefer X to Y. And this, I take it, is quite substantive (or “load-bearing”). Many deny that what makes it the case that an agent ought to perform an act is that it’s fitting for that agent to prefer its outcome to that of every available alternative. Now, I don’t take a stand on the formal characteristics of this fitting-to-prefer ranking, but I don’t think that I need to. It’s enough for my purposes that people disagree on whether the rightness of an action is a function of the agent’s fitting-to-prefer ranking.Report
Ah, I see – that’s really helpful. Thanks for all of the follow-ups!Report
It seems to me that any ethics will sort right actions from wrong actions. So any ethics could be rendered consequentialist (in a sense) by specifying that the ranking of outcomes is done so as to rank the consequences of right actions higher than the consequences of wrong actions.
Yet this seems like a cheap trick. Paradigm cases of consequentialism can rank outcomes without reference to the actions which might produce them. For example, classical utilitarianism can tally up total happiness. Egoism can tally up satisfaction for the agent.
Kantsequentialism does not seem to have any independent way to characterize the ranking. Instead, “the outcome of an agent’s refraining from treating persons as mere means” just seems to be doing the cheap trick.
Perhaps the distinction between ultimate right-making and ultimate explanation is supposed to resolve this, but I don’t see how. The worry is about how the view can be formulated rather than about ultimate grounding.Report
I think that moral theories must do more than just sort right actions from wrong actions. To see this, consider that both Paley’s utilitarian theological voluntarism and Mill’s utilitarian act-consequentialism sort acts that maximize utility as right as well as sort acts that don’t maximize utility as wrong. But, from this, we shouldn’t conclude that Paley is just performing some cheap trick wherein he claims that God is perfectly good and so commands each of us to maximize utility simply so that his theory can mimic the deontic verdicts of Mill’s theory. Rather, it seems that Paley disagrees with Mill in an important way. Unlike Mill, Paley thinks that what makes it right for us to maximize utility is that God commands us to do so. Likewise, Kantsequentialism claims that what makes it wrong for an agent to treat a person as a mere means to preventing two others from each treating a person as a mere means is that the agent should prefer the outcome in which the two others each treat a persons as a mere means to the outcome in which they themself treat a person as a mere means. This is to disagree with the non-consequentialist who holds that what makes it wrong to treat a person as a mere means to prevent two others from each treating a person as a mere means has nothing to do with what the agent should prefer. So, I fail to see why this is just some cheap trick. Do you think that Paley’s move is a cheap trick? And if not, do you think that Paley’s move is in some important respect different from my move?Report
I think that there may be a problem in the inference from (UTV1) and (UTV2) to (UTV3). Presumably the inference is only good if the word ‘because’ is univocal in (UTV1) and (UTV2), but I don’t think it is. In (UTV1) ‘because’ indicates a making relation. (UTV1) says that what makes maximizing utility morally right is that the ultimate right-making feature of an act is that it accords with God’s will and that God wills agents to maximize utility. In (UTV2), by contrast, ‘because’ indicates an explanatory relation. (UTV2) says that what explains the fact that God wills agents to maximize utility is that God is perfectly good (and that he is perfectly good necessitates that he wills that agents maximize utility).
So, (UTV1) and (UTV2) use ‘because’ in different senses. Thus, it is illegitimate to infer (UTV3) from them. [What does ‘because’ mean in (UTV3)? Does it indicate a making relation? Does it indicate an explanatory relation? Or does it indicate a completely different relation? Regardless, since ‘because’ means something different in (UTV1) than it does in (UTV2), (UTV3) does not follow.]Report
Throughout ‘X because Y’ should be interpreted as equivalent to ‘Y explains X’. Thus, the argument is as follows.
I am not sure this helps. It does help if explaining the rightness of some kind of activity is the same kind of thing as explaining why an agent wills as she does. But I am not sure that these are the same. When we explain the rightness of some act (or type of act), we are stating the features that make the act (or act-type) right (call this kind of explaining, “explaining-1”). But when we explain why an agent wills as she does, we are either stating the character traits of the agent or stating the reasons/motives that the agent is responding to (or both) (call this type “explaining-2”). So, when we are explaining-2 an agent’s willing something, we are not explaining-1 her willing it. That is, when we cite character traits and/or reasons/motives we are not citing features that make it the case that the agent wills as she does.
(1) What explains the rightness of act is that God wills that agents perform the act.
(2) What explains God’s willing an act is that the act maximizes utility.
Thus, (3) What explains the rightness of an act is that the act maximizes utility.
A theological voluntarist is going to reject (3), right? On theological voluntarism, we cannot explain the rightness of an act without citing the fact that God wills the act. So, (3) must be false on theological voluntarism. But the voluntarist accepts (1) and can accept (2). So, what’s gone wrong? I think it’s that the argument equivocates on ‘explains.’ When we explain the rightness of an act we are doing something quite different than when we explain the willings of an agent. (2) explains God’s willing an act by citing God’s motives/reasons. (1) explains the rightness of an act by citing the feature that makes actions right. We are doing different kinds of things, both called “explaining,” in (1) and (2).Report
The theological voluntarist needn’t accept (3), but I don’t see why the theological voluntarist couldn’t accept (3). Indeed, it seems to me that Paley would gladly accept (3). If you asked him “Why is right to flip the switch in the Trolley Case?” then he could rightly answer: “Because flipping the switch would maximize utility.” Of course, conversational maxims might suggest that he should add: “And it’s right to maximize utility because God commands us to do so.” But, in any case, it seems to me that he would accept (3). Why wouldn’t he?
So, I would deny: “on theological voluntarism, we cannot explain the rightness of an act without citing the fact that God wills the act.” For instance, a theological voluntarist might explain the rightness of honoring thy parents by pointing to the fact that it’s written on a certain stone tablet. Perhaps, though, your point is that, on theological voluntarism, we cannot FULLY explain the rightness of an act without citing the fact that God wills the act. Fair enough. But I don’t see where this presents a problem for my argument. I wasn’t intending to give full explanations. I was merely giving explanations based on certain background assumptions.Report
(I) What fully explains the rightness of an act is that God wills that agents perform the act.
(II) What fully explains God’s willing an act is that the act maximizes utility.
Thus, (III) What fully explains the rightness of an act is that the act maximizes utility.
The theological voluntarist can accept (I), I think. On voluntarism, it is necessary and sufficient for an act to be right that it be willed by God. If “fully explains” in (II) indicates God’s reasons/motives, then the voluntarist can also accept (II). That is, if our task is to cite a complete account of God’s motives for willing an act, the voluntarist can accept that the task is complete when we assert that the act maximizes utility. But the voluntarist must reject (III).
If we agree that the move from (I) and (II) to (III) is illegitimate, we must ask what has gone wrong. I would assert that what’s gone wrong is that (I) and (II) are using a different sense of ‘fully explains.’ But, if fully explaining the rightness of an act is doing something different than fully explaining an agent’s willing an act, then explaining the rightness of an act is also something different than explaining an agent’s willing an act.
The voluntarist can accept
(3) What explains the rightness of an act is that the act maximizes utility
only if ‘explains’ means something like it does in
(2) What explains God’s willing an act is that the act maximizes utility.
Since in (2) we are explaining by citing motives, it follows that in (3) we must be doing something similar. So, (3) must mean something like: What motivationally explains the rightness of an act is that the act maximizes utility.
Now, on voluntarism, the idea of motivationally explaining (i.e., explaining by citing motives) the rightness of an act might make sense. But motivationally explaining rightness is different than explaining rightness in the sense of citing the feature(s) that makes acts right. Premises (1) and (I) are best interpreted as explaining in this latter sense.Report
You write: “If we agree that the move from (I) and (II) to (III) is illegitimate….” I don’t agree.
It seems to me that this is a valid argument form:
(A) What fully explains X is Y.
(B) What fully explains Y is Z.
Therefore, (C) what fully explains X is Z.Report
Earlier you said, “Perhaps, though, your point is that, on theological voluntarism, we cannot FULLY explain the rightness of an act without citing the fact that God wills the act. Fair enough”
I took that to mean that you think that the voluntarist would reject
(III) What fully explains the rightness of an act is that the act maximizes utility.
And you are saying that the move from (A) and (B) to (C) is legitimate. So, the move from (I) and (II) to (III) is, on your view, legitimate.
But then, if the voluntarist accepts (I) and (II) and the move from (A) and (B) to (C) is legitimate, then voluntarist myst accept (III).Report
But (i) and (II) are clearly false. Take (II). What fully explains God’s willing an act is not solely that the act maximizes utility. A full explanation would have to include the facts that God is perfectly good, that a perfectly good being would will agents to do what maximizes the good, and that utility is good.Report
It depends on what it is to fully explain God’s willing an act. If to explain why an agent wills an act is to cite their motives, then we can fully explain why an agent wills an act be citing a motive without citing anything about the agent’s character.
If you are thinking that we must include facts about the agent’s character in a full explanation of why the agent wills an act, then you are thinking of a full explanation as a causal explanation. But not all full explanations are causal because not all explanations are causal.Report
I agree that Kantsequentialism is both deontological and act-consequentialist. That said, I think Kantsequentialism is false because there are cases where I could maximize goodness only by violating agent-centered restrictions. The Jim and Pedro case seems to me to be such a case. If you disagree about the Jim and Pedro case, then I suspect readers can come up with other cases that will be more convincing. Perhaps, then, ethics textbooks might now say the following: deontology and act-consequentialism mutually exclude each other, except in the case of one (probably false) theory, namely, Kantsequentialism. (I know, I’m being tendentious in calling Kantsequentialism a probably false theory here.)
Doug, do you think Kantsequentialism is true? Or instead are you just arguing that it is a possible theory, one worth at least some discussion and consideration? If the latter, how much discussion and consideration are we talking about here?Report
As formulated here, I do not think that Kantsequentialism is true. And one of the reasons for that is the one you cite. As formulated here, Kantsequentialism is much too simple a theory. The version of Kantsequentialism that I accept is non-absolutist and is compatible with the commonsense consequentialism that I argued for in my 2011 book.Report
I see. Thanks for clarifying. I’ll check out the book.Report
Thank you for the provocative piece, Douglas. Having read through it once, I am concerned primarily with two issues: first, the distinction you draw between the ultimate right-making feature of an action and the fact that explains its rightness; and second, what you take to be the sufficient conditions of an ethical theory being compatible with Kant’s. True, at face value, there seems to be a distinction between features and facts. But does Kant’s theory not provide us with an answer to what the fact(s) in the latter category are? Consider:
“The fact that ultimately explains the rightness of an act is that it was performed from the motive of duty (or respect for persons, as you put it–the two are mutually entailing, as I understand Kant)”
I believe that this, and not K1a, is a necessary condition of a theory’s being Kantian and/or compatible with Kantianism. It is not enough that respect for persons be the ultimate right-making feature; the Kantian also has strict views on what the ultimate right-making facts are. I take it you think this is not necessary for a theory to be compatible with Kantian ethics. Why not?
(I hope my comment made sense and is readable. Again, thank you!)Report
Note that I’m not concerned with whether Kantsequentialism is compatible with Kant’s moral theory. I’m concerned with whether Kantsequentialism is compatible with a theory that is committed to agent-centered restrictions that are ultimately grounded in our duty to respect persons. Also, the relevant distinction is not between features and facts. The relevant distinction is between things that a proponent of a given criterion of rightness is committed to and things that they are not committed to. Paley, qua theological voluntarist, is committed to the ultimately right-making feature of an act being that it accords with God’s will. Paley, qua theological voluntarist, is not committed to all and only acts that maximize utility are right or that God wills us to maximize utility. There are, after all, many theological voluntarists who, unlike Paley, deny that all and only acts that maximize utility are right. Likewise, I, qua act-consequentialist am committed to the ultimate right-making feature of an act being that its outcome is not outranked by that of any available alternative. Now, I happen to think in addition to this that our duty to respect persons affects how agents should rank outcomes. But this isn’t something that an act-consequentialist must necessarily accept. Lastly, I don’t believe that Kant thinks that what ultimately explains the rightness of an act is that it was performed from the motive of duty. That’s what explains the moral worth of an act, not its rightness.Report
Trying to fit Kantsequentialism into Campbell Brown’s “Consequentialize This” framework: Does K-ism as here described escape his “No Moral Dilemmas” feature of consequentalism? Or more generally the problem of treating-as-ends-not-means having infinite goodness?Report
It’s hard to fit what Brown says with what I’m saying here because the two of us are talking about very different beasts. As I defined ‘(act-)consequentialism’ here, it consists in a claim about what makes an act right. By contrast, Brown takes ‘consequentialism’ to be “a purely extensional thesis [that]…makes no claim of causation, determination, or explanation.” Also, Brown defines ‘consequentialism’ such that it’s committed to agent neutrality by definition. My definition, by contrast, allows that act-consequentialism can be agent-relative. In any case, though, I believe that Kantsequentialism, as defined here, could accommodate moral dilemmas. But it does have the problem of being absolutist. But, as I noted to another commentator above, I don’t accept Kantsequentialism as formulated here. As formulated here, it is too simplistic a moral theory.Report
Isn’t your claim about Kantsequentialiam just the old hat constrained maximizer stuff? Why or why not?Report