Inspired a bit by Ralph's post critiquing Kant's view of unconditioned or fundamental goods, I've been investigating why Kant arrives at his (in my opinion) dissatisfying view that we have only indirect duties to non-human animals. Here's my tentative analysis and conclusion.
Kant begins part I of the Groundwork with his distinction between unconditioned and conditioned goods. He argues that only the rational will is (or is the source of) an unconditioned good. All other goods, including various virtues, such as courage or intelligence; talents and "gifts of nature"; health, wealth, and other "gifts of fortune"; and happiness itself, are conditioned goods. I take Kant to mean by this that their goodness (or perhaps their specifically moral goodness) is only realized when they are accompanied, or exercised by, the good rational will. Indeed, according to Kant, these goods are positively evil “without the basic principles” of a good rational
will to guide them or correct their influence. (G 4:393-4)
In these opening sections of the Groundwork, neither animals nor their welfare are explicitly mentioned. They are mentioned obliquely when Kant later introduces the Formula of Humanity. (G 4:428-29) There Kant reiterates that all “objects of the
inclinations have only a conditional worth” and hence any states of affairs
achieved through action have only conditional worth. Kant then states that “beings the existence of which rests …on nature” or “beings
without reason” have only relative worth, the worth belonging to things, not
persons. Only rational nature,
Kant concludes, has “absolute" or unconditioned worth, so
only our humanity is an “objective end” or an end in itself.
My diagnosis is that animals and their welfare turn out to be conditioned goods within Kant's axiology because this is logically forced upon him by (1) his belief that the rational will alone is an unconditioned good, and (2) his belief that every good is good either unconditionally or conditionally in relation to what is unconditionally good. Since I have Kantian sympathies, I'm disposed to accept (1). So (2) strikes me as the source of the problem. Kant seems to suppose that the unconditioned/conditioned goodness distinction exhausts all possible goods. But animal welfare in particular seems like a counterexample to this. Suppose that Kant is right about (1). Does it follow then that animal welfare is a conditioned good? Only if Kant's distinction is exhaustive. But here's a reason to think it isn't: What characterizes conditioned goods is that their goodness is realized only when accompanied by, or exercised by, the (good) rational will. If the flourishing or suffering of animals is a normatively significant fact —and I've given no direct argument for that claim here — then its significance is not dependent on its relation to the goodness or badness of any rational will. Animal flourishing is good regardless of whether it is the product of any rational willing. Likewise, animal suffering is bad regardless of whether it is the product of any rational willing. I, qua possessor of a rational will, cannot render the flourishing of an animal bad through exercising my will in a particular manner. Nor can I render the suffering of an animal good through exercising my will in a particular manner. (I'm supposing here that non-human animals are not rational wills in Kant's sense.) In contrast then to the other conditioned goods Kant mentions, the value of animal welfare is invariant with respect to rational volition.
If so, then Kant was wrong to suppose that his unconditioned/conditioned goods distinction exhaustively catalogs all goods. Animal welfare is neither an unconditioned good nor a conditioned good with respect to the unconditioned good of the rational will. If a good with these attributes could be made out on a Kantian view, then perhaps animal welfare could be a source of direct duties despite its not being an unconditioned good. Or so I conjecture.
For those interested in the topic, Patrick Kain has an excellent article on the topic, forthcoming (I think still forthcoming) in a volume on the Metaphysics of Morals edited by Lara Denis. It’s well worth reading.
Interesting. Some rough comments:
(1) Given what you say about the goodness of animal flourishing, how will you resist the conclusion that it is unconditionally good?
(2) An alternative:
Accept that animal flourishing is a conditional good but reject the claim that is something is only a conditional good, then it lacks intrinsic value.
At least at the start of the Groundwork, a conditional good appears to be one which is worthy of choice, approval, or esteem in only some contexts. More generally he seems to be thinking that most things are conditionally good in the sense that their evaluative valence is context-variable.
Let’s say it is true that sometimes animal flourishing is not worthy of choice, approval, or esteem & and even that in some context it would be bad (maybe it is a sadistic animal species?). We can still argue that animal flourishing has intrinsic value in the contexts in which it is good.
I wonder whether Kant missed this possibility and was lead astray by the fact that ‘conditional’ shows up in ‘conditional good’ and ‘conditional value’ – this might make you think that the former entails the later (so to speak). I do not know the relevant German; but, as a good particularist would tell us, we cannot infer that something has only instrumental value from the fact that it has context-variable valence.
Oops — The second ‘is’ should be an ‘if’ in the first sentence of (2).
Brad,
Answering your two points at once: I resist the conclusion that animal welfare is unconditionally good on the grounds that it falls through the cracks in Kant’s taxonomy, being (as I said in the original post) neither an unconditioned good nor a good whose goodness is realized on the condition that the rational will accompanies or exercises it.
It’s clear that Kant acknowledges there are conditional goods that are nevertheless intrinsically good. Happiness is one: It’s not good unless accompanied by the good will, but its goodness isn’t derived from its relation to the good will, nor is its goodness borrowed from the good will. In a longer paper I’m writing, I try, in the spirit of your (2), to argue that animal welfare could therefore be intrinsically good despite not being unconditionally good. What I’d insist on though is what I said in the original post: that animal welfare, unlike human happiness, is not a conditioned good either, since its value is invariant with respect to rational volition.
Thanks, Michael. I agree with you about happiness.
I am still having trouble seeing why animal flourishing is not an unconditioned good, on your view.
Can you say a bit more about how you understand ‘unconditioned good’ or what being an uncondited good entails (on your view) & say how it rules out animal flourishing?
Are you assuming, e.g., that being an unconditioned good entails being the thing that explains why conditioned goods are good or not?
Michael,
this is interesting but I think there is more going on that this. I think there is a clear idea of constructivism about value in Groundwork.
This is why I think accepting your line would be a bit of a problem for Kant.
Suppose we accept that:
“If the flourishing or suffering of animals is a normatively significant fact —and I’ve given no direct argument for that claim here — then its significance is not dependent on its relation to the goodness or badness of any rational will.”
This means that we have a new category of value which is not dependent on its relation to the rational will. Now, there is no account in Kant about what it is to be good in this way.
There is a story of the value of the good will itself in terms of autonomy. There is also a story of what it is for objects to be good qua being the objects of a will that it is a rational will (see 4:413 for instance). But, there is no story of what it is to be good independently of having a relation to a good will.
Now, Kant might have accepted such independent values, but I think this would have Kant’s position uncomfortable. First, it would have posed the question – why not accept realism about value across the board. Second, many people think that there are anti-realist arguments in Kant.
So, I believe that Kant stuck to his value theory for a reason. He did want to give a story of why animal-welfare is a necessary object of a good will – and this is the indirect one. Being cruel to animals makes cruelty towards humans easier, and this the good will cannot consistently will. Note by the way, that this story doesn’t make the value of animan-welfare conditional in the same way as some character-traits are on his view – it is merely claim of the source of normativity. It’s not in the animals themselves but rather in the objecthood of a rational willing.
Of course, many of us think that this story is too indirect. You might be right that he can give up the story and adopt a direct one. I’m just worried that this makes him too realist for his own liking in value-theory.
Michael,
Even though I am tempted to agree with you that Kant does or at least could accept that happiness has intrinsic and final value (loosely speaking), I still think he may be making the error I mentioned in my first comment & that that might explain his bad view on animals.
Consider your summary: “…Kant reiterates that all “objects of the inclinations have only a conditional worth” and hence any states of affairs achieved through action have only conditional worth. Kant then states that “beings the existence of which rests …on nature” or “beings without reason” have only relative worth, the worth belonging to things, not persons. Only rational nature, Kant concludes, has “absolute” or unconditioned worth, so only our humanity is an “objective end” or an end in itself”
This seems to be the suppressed premise (which is false):
(SP) If something has only conditional worth, then it has only instrumental value.
Otherwise, it is hard to see how to get from the points about conditional worth to the conclusion about animal flourishing not being an end-in-itself.
Hi Michael:
Is that right? Isn’t it Kant’s view that the goodness of happiness is conditional on the happy person’s being worthy of happiness, and also that someone is worthy of happiness only if they have a good will? But then there is a clear sense in which we can say that the goodness of a person’s happiness derives from its relation to their good will, no?
Thanks for these comments, everyone.
Simon: It’s important to my view that the intrinsic/extrinsic distinction not be identified with the unconditioned/conditioned distinction. The former is about the source of something’s goodness, the latter about its logical relationships to other goods. Yes, you’re right that happiness’ value is, according to Kant, conditional on the good will. And ‘derived’ was sloppy talk on my part. So here’s the idea: Happiness’s goodness is conditional on the presence of the good will, but its goodness isn’t as a means to, or as a constituent or component of, the good will. An analogy, admittedly problematic: Imagine a modular house, consisting of two components, the frame and the roof. Take the roof in isolation. Its goodness qua roof is conditional on the presence of the frame upon which the roof sits. But the roof’s goodness doesn’t stem from any feature of the frame, even though its goodness requires the presence of the frame to be realized. (This is of course complicated by these in turn being part of some third thing of value, the house. Perhaps the house = Kant’s highest good?) So the roof’s value is intrinsic but conditioned. Does that clarify matters at all?
Brad: I agree that (SP) is false and that if Kant held it, this would explain why he thought that animals could only have instrumental value, namely, since he believed they aren’t unconditionally good, they must be conditionally good, and that, in combination with (SP), results in animals being only instrumentally good. But my strategy is different: In fact, I want to argue that animals or their welfare belong to neither category. Part of the reason is the ‘invariance’ argument I gave in the original post: Animal welfare’s value doesn’t seem to depend on its relationship to rational volition. Since this seems to show that animal welfare is not a conditioned good, you then understandably ask why it doesn’t then follow that animal welfare is an unconditioned good. The problem here is the ambiguity of ‘unconditioned’: If it simply means ‘not conditioned,’ then, yes, I have to concede that animal welfare would be unconditioned in that sense. But I think Kant’s intended usage is that ‘unconditioned good’ means not only not conditioned on the presence of any other good, but also a good that serves as the condition of other goods being good. In other words, ‘unconditioned’ is an implicitly relational notion for Kant and is therefore not identical with the particulartists’ notion context-invariability. This accords well with his usage of the term and with other things he says about the good will (that is the supreme good, etc.). So animal welfare is not a conditioned good, but nor is an unconditioned good in the sense I interpret Kant to use that term. As I said, it falls through the cracks of Kant’s distinction, but turns out to be an intrinsic and final good.
Jussi: As time has past, I’m increasingly unhappy with constructivism as a theory or as an interpretation of Kant. Though most of the objections leveled at constructivism are theoretical, that it has difficulty providing us plausible explanations of the value of, among other things, animal welfare is a reason to reject it and opt for a realist interpretation of Kant. I find Korsgaard’s talk of our ‘legislating’ the value of animals extremely unsatisfying, for example.
Also, I wanted to mention that Adrienne Martin’s post (“Too many distinctions in value,” June 09) is helpful in sorting out these axiological distinctions.
Hi Michael, yes, I think your comment does help (although I must admit, I have some difficulty imagining a roof to have intrinsic value)! Your main point, if I understand it right, is that even if it is agreed that the goodness of happiness is conditional on its being enjoyed by a good will, this leaves the question of whether its goodness is intrinsic or extrinsic open. Point taken.
But now I’m inclined to agree with Jussi and say that Kant is a constructivist – and thus thought that happiness enjoyed by a good will is good precisely because that state of affairs (i.e. happiness of a person with good will) would be chosen/legislated by a good will. But this is to explain the goodness of happiness as extrinsic. Is there something in Kant’s text that shows this interpretation clearly mistaken? I would love to hear more from you (maybe another post, if I’m going dragging you off on a tangent) on why you think Kant is not best interpreted as a constructivist.
Michael,
(Sorry for the long-winded post). I like your idea of not taking the unconditioned/conditioned distinction as exhaustive, at least when we try to apply it to things we might think of as ‘a good thing’. But I might take it in a little different direction. It seems to me that there is something downright misleading in Kant’s presentation in the Groundwork. He seems to say “Look, there’s only one really fantastically valuable thing, a good will. Everything else rates below that in some way — perhaps because everything else gets its value by being related in the right way to it. So let’s base our ethics on that really valuable thing.” That just invites the question, “Oh yeah? What’s so great about a good will?” “Can’t I steal or lie to protect a work of great beauty from permanent destruction?” And, of course, “What about animals?” These do seem to be real showstoppers as questions, at least when the view is put that way.
I think maybe approaching the issue of value in a different way might work better, (and I think it has lots of precedent in Kant’s writings). Start with the question, not what is the supreme value, but, instead, what sort of guidelines (rules, restrictions, etc) are or should be ultimately authoritative for us in dealing with each other as rational beings. We are all beings who act on, respond to and acknowledge reasons, who plan both individually and together. We also have our own individual and collective goals. We don’t face that question with animals, since they don’t in the same way, or at least not in a sufficiently sophisticated way, plan and reason with us about our joint behaviors. This seems to me to be the crux of the idea that animals have a ‘conditioned worth’; they’re just not a part of this particular, admittedly narrow, though really serious, project.
So we look at the fact that we have this plurality of rational agents, each with goals, thing we care about, and we will (if the Kantian arguments are right) conclude something like “We should treat each others goals (provided those goals don’t undermine our mutual rules) with a certain sort of respect or acknowledgement – not willfully interfering with them, for instance, helping each to achieve them on occasion, and so on – and we say that those goals are ‘goods’ or ‘valuables’, though of course they are so because they are the aims adopted by rational agents (and for different reasons, some for their intrinsic interest, some not). Here, ‘rational agency’ of a certain sort – the kind that is capable of manifesting ‘good will’ – appears to be in the driver’s seat in a certain sense, guiding our deliberations about each other’s goals, etc. Again, although animals also have goals, and we also refer to them as ‘goods’, there’s something different going on here, since we don’t have this more general backdrop of trying to come up with the appropriate set of guidelines that we’ll all be adhering to having to do with how to treat each other’s goals and interests.
This is not yet to say anything about the ‘intrinsic’ value of animals and their needs and interests. Those may well be something we should also pay serious attention to. They’re just not a part of the sort of project we’re engaged in with each other. Some (well, perhaps many) will say that thinking there is a special “separate” question about guidelines dealing with rational agents already wrongly elevates we rational agents above other beings. I don’t know what there is to say to that, except that it seems to me that it really is a distinct and separate, and, yes, in some sense more important or at least prior question, to that of guidelines for deal with everything else.
Robert, I see your remarks as continuous with what I would say here too. In a longer paper on this topic, I try to show that it is possible for us to have direct duties to animals without significantly modifying the scheme you provide, wherein rational agency is in the driver’s seat, as you say. In essence, I see Kant’s ethics — the official, orthodox stuff he wrote down — as accounting for what Scanlon called what we owe to each other. The trouble is that animals don’t have a place in that scheme. I’m trying to defend direct duties to them without placing them within that scheme. Rather, animal welfare is appended to that scheme. (The paper is here, if you’re interested: http://michael.cholbi.com/wpcontent/uploads/2010/01/DirectDutyToAnimals.doc)
Simon, that’s probably a bigger topic than I’d want to tackle here. But here’s how I see the issue concerning constructivism: I understand Kant as a rationalist, someone who thinks moral judgments state reasons for action. If reasons for action are mind-independent or fact-like, a la Dancy, then Kant’s a realist. If reasons for action are mind-dependent attitudes, then Kant’s a constructivist. My reservations about constructivism are not novel, namely, that its procedural account of moral deliberation underdetermines truth and that it ultimately seems confused about the order of explanation with respect to judgments concerning reasons for action. But I was also unsure about something else you said: You said that the fact that the happiness enjoyed by a good will would be chosen by a good will shows that the goodness of happiness is extrinsic. I don’t get that inference. The good will would also choose to act from a motive of duty, but that’s not extrinsically good. It’s intrinsically good.
Michael,
Interesting post. I think, with other commenters, though, that your solution can’t work for Kant. Of course this doesn’t mean that you can’t have some sort of Kantian theory. To explain:
Here’s how I understand the story about happiness for Kant: we each will our own happiness, and this is a natural, say, psychological, fact about us. But, is my happiness a good thing? On reflection, I see that I don’t always count happiness as good (as valuable) – the classic example is when a murderer is made happy by his murder, I don’t count his happiness as good. So, there must be some condition under which happiness is good or bad, and of course Kant says this condition is the good will. What makes the will good, or not, is the kind of reasoning we employ when we reason practically about what to choose as our ends.
So, this is his constructivism: values are “constructed” by a process of practical reasoning – they are the outputs of practical reason, not its inputs. Thus, when you say that the welfare of animals is something whose goodness does not depend in any way on the good will, I worry that this portrays values as a very different kind of thing: something that can exist in the world without dependence on a process of practical reasoning. Depending on how the story is told, this could look like a kind of reification of values, analogous perhaps to empirical properties. This is precisely the sort of view that Kant wants to reject – I think it is a paradigmatic case of a moral theory of heteronomy that employs an empirical principle.
Pete – I’m under no illusion Kant would endorse my view concerning our duties to animals. After all, he explicitly rejects it. And you’re correct that my view puts animal welfare in a more empirical or realist category of value that Kant may have been uncomfortable with. In your terms, I’m treating animal welfare as an input to practical reasoning, rather than an output. My aim to is show that we need not reject the general Kantian picture of interhuman ethics to accept a direct duty concerning animal welfare.
But as to the contructivist claim: I’m always puzzled by the sort of move you make in your comment. How is “the sort of reasoning we employ when we reason practically” about our ends the explanation of the good will’s goodness? It seems rather that in the case you describe, we discover something about normative relations, namely, that the goodness of happiness is conditional on the presence of the good will. But I don’t see that this relation is the product of my practical reasoning. It’s a discovery within my practical reasoning. I’m happy to grant that practical reasoning can ascertain values. What I don’t see is how values are the outputs rather than inputs. It seems to me that the constructivist is confusing order of explanation with order of discovery, assuming that since values are discovered via practical reasoning, reasoning has values as its outputs. What am I not understanding fully?
Michael,
I’ll give a shot at an explanation, but with the caveat that this is an interpretive dispute that has gone on for years and I’d be a little surprised if I were the one to dissolve it.
The radical claim that values are the products of practical reasoning, rather than things that pre-exist practical reason and inform it (values are discovered) follows from the point about happiness, if this point generalizes.
The reason that the murderer’s happiness doesn’t count as objectively good is that he is engaged in a morally wrong act (note that the murderer might think his happiness is good, but we are concerned in morality, I’ll take for granted here, not simply with what we subjectively think is good, but with what is objectively good). This implies that considerations of right inform and determine our value ascriptions. But this isn’t an epistemological point, it is a normative one. Whether some end counts as a good or valuable one depends on our reasons for pursuing it – on whether our wills are good. Now, Kant wants to say this isn’t true only about happiness, but about everything. Even the virtues aren’t always good – it depends on whether the ends to which they are put conform to our conception of moral rightness – again, on whether we are courageous in conformity to a good will (an Aristotelian might take this to be definitional of courage, in which case I mean something like “bravery in the face of danger, regardless of the source of the danger”).
So, happiness is only objectively good when it is the object of a good will. I agree that this relation is something we can discover within practical reason. But what makes a will good? Our wills aren’t always or necessarily good – they can be bad. The answer can’t be, “Our wills are good when (that is, because) we will the good ends.” It can’t be this because we’ve just seen that the source of the goodness of our ends is the good will. Still, it is at least a condition of my will’s being good that I only will objectively good ends – my will certainly isn’t good if I intentionally will bad ends. (It turns out that this condition is a formal property of the will itself, not some other good thing. So the will, when it is good, is still unconditionally good – no other good thing is the condition of its goodness.) What property or properties do ends that are objectively good have? Well, when my ends are objectively good, it is at least the case that every other practically rational agent must be able to will my end for the reason I do in the context in which I take it to be good. Perhaps every objectively good end has some other property or properties, too. But what properties? The happiness argument can be applied to every other candidate property (so I think the argument goes), such that the only good-making property is this formal property, which leads immediately to the first formulation of the CI.
Then, to determine whether some candidate end of mine is a good one, I need to engage in a procedure of practical reasoning. I need to reason through whether my end is the kind of end that any practically rational agent could pursue for the reason I do in the context in which I take it to be valuable. So, this is the sense in which it is the sort of reasoning we engage in that not only explains the will’s goodness but actually makes it good. The good will, in turn, is the condition on happiness (and anything else) being good. Therefore, the things in the world that are valuable are valuable in virtue of a process of reasoning by practically rational beings.
Now, the above is highly condensed and certainly not uncontroversial, but I hope it helps to explain how I understand constructivism and the claim that values are the output of practical reason.
This:
“My aim to is show that we need not reject the general Kantian picture of interhuman ethics to accept a direct duty concerning animal welfare.”
makes me uneasy. The original problem seemed to be how can Kant account for duties concerning animal welfare. Now, the answer is that, *he cannot*, but there is a realist view of duties to animals that is not at least incompatible with Kant.
I have my doubts about this, but mainly my worry is that this seems to place animals in a better position than other human beings. Imagine that there is a human and an animal in front of me. We then ask, why shouldn’t I torture them just for the sake of it?
In the animal case, we are allowed to say that the animal would suffer. This is a normatively significant fact as such – which gives me reason not to torture the animal.
In the human case, we are not allowed to say this. That the human would be suffering is not a normatively significant fact in itself in the same way. It’s significance must be grounded on and derived from my inability to universalise a maxim of torture.
This seems inconsistent – after all, we are talking about the very same kind of suffering. Either Kant has to say that human suffering is normatively significant and we don’t need his CI account of interpersonal duties (like the one not to torture others), or he has to give up to give up your claim about the intrinsic significance of animal suffering.
So, I’m not sure we have coherent view yet.
Hi Michael,
I didn’t make that inference: I said that the view that “happiness enjoyed by a good will is good precisely because that state of affairs … would be chosen/legislated by a good will” implies that the goodness of happiness is extrinsic.
Also, forgive me for being pedantic if you were just speaking imprecisely, but if when you say “constructivism” you really mean that “reasons for action are mind-dependent attitudes”, as your remarks suggested, then I agree with you that Kant is not well interpreted as what you would call a “constructivist”. But perhaps our dispute is merely semantic.
Jussi – You ask why animals don’t enjoy a better more “realist” standing than human agents on the view I’m proposing. My answer is that Kant reserves dignity for that which is unconditionally good (in the second sense I mentioned in my comment directed at Brad). Animals, on the view I’m after, don’t have the pricelessness associated with dignity; rational wills do. (I can elaborate on this if you’d like.)
You also seem concerned that, on my view, the accounts of why torturing animals is wrong and why torturing human agents is wrong would differ. I see this as a strength of my view rather than a source of concern. One of the problems that has dogged (ha! I punned) philosophical ethicists’ approach to animals is a tacit assumption that we can only have direct duties to animals if the source or nature of those duties is the same as the source and nature of our obligations to human agents. This is what gets Kant in trouble: Animals don’t have the properties that ground the moral status of human agents. Similarly, this seems to be an important selling point for many utilitarian defenders of duties to animals — that we need no special account of value to explain these duties, on their view. Humans and animals enjoy moral status because of their capacity to flourish or to suffer. But I think we should reject that assumption, and accept that our duties to animals may have a different source and logic than our duties to human agents. In the particular case you mention, the wrongness of torture will have different explanations: raw suffering in the case of animals, torture’s relationship to the rational will in the case of humans.
Simon – Thanks. You cleared that up nicely. But I guess I’d still press the claim that happiness being an object of choice for the good will doesn’t show that happiness’ value is extrinsic. It might show that something being an object of choice for the good will and an object’s being good are coextensive. Whether that object is extrinsically or intrinsically good will depend on the relevant direction of explanation. Or at least I think so…
As for the “constructivist” point: There’s a lot of different accounts of “constructivism” out there. It’s definitely a term in flux. I’ve always thought that the most helpful characterization is that constructivists believe that moral discourse is cognitive and truth-apt, but they are not realists. Instead, moral statements are made true by mind-dependent facts, namely, our evaluative attitudes. Constructivism goes ‘Kantian’ when the evaluative facts are facts rooted in the formal structure of practical deliberation and endorsement. My own take is that as a reading of Kant, this is mistaken. Moral facts are facts about rational agency, not facts concerning what we have reason to do or endorse. But, yes, we might just be carving up semantics here.
Pete – Thanks; that’s very thorough. I’m afraid you have to color me either confused or unconvinced. But I’ll need some time to figure out where exactly I jump ship.