Moral Objects, Moral Agents and Consciousness
February 10, 2008 by Tim Nailer
That’s my what honours thesis is on. This post is an attempt to rid myself of thesis-writing block. First off, let me just explain what I take ‘ethics’ to mean. To simplify, ethics is about how our actions affect other individuals. This is not universally accepted but it is a useful working definition for certain types of ethicist (contractarians, rights theorists, utilitarians) as well as non-philosophers. Now I’ll introduce some jargon: a moral agent is an individual that performs the action while a moral object is the individual that is affected by it. So if Alice were to kick Oliver, Alice would be the moral agent and Oliver would be the moral object. But what if Alice were a horse and she didn’t know the difference between right and wrong? Would she still be a moral agent? And what if Oliver were just a doll and felt no pain from being kicked? Would he still be a moral object? What makes an individual a moral object or a moral agent?
These are the kinds of questions that I’ll be dealing with in my thesis this year. The title of the post mentions ‘consciousness’ so there’s a clue there: I think moral agency and moral objecthood both require consciousness but not the same type of consciousness.
Philosophers make a distinction between phenomenal consciousness and non-phenomenal consciousness. To (over)simplify again, the former refers to things we feel while the latter refers to things we think. Examples of phenomenal consciousness include the experience of a particular shade of green, the smell of roast chicken, a myriad of different types of pain, as well as emotions like joy, guilt and anger. If it feels like something, anything at all, to be in a particular mental state, then it’s a type of phenomenal consciousness. Non-phenomenal consciousness doesn’t feel like anything. Examples of non-phenomenal counsciousness include learning, accessing thoughts and memories, percieving and responding to stimuli, and computing data. Basically, if a computer can do it then it’s probably a type of non-phenomenal consciousness. But what about a computer that’s programmed to detect a particular shade of green? The detection of the colour and all the formal operations it does are non-phenomenal but its experience, what it feels when it looks at the colour, is phenomenal.
Perhaps the most interesting thing is that the two don’t have to go together. It’s possible for the computer to detect the colour and act exactly like a human would but not to experience anything. Beings that are identical to humans but lack phenomenal consciousness are known as zombies and my thesis will probably be invaded by them at some point (not to mention hordes of zombie animals such as the zombats that give their name to this blog).
Now if Alice kicks Zomboliver, it’s not going to hurt him at all. He may show all the external signs of pain but he won’t feel anything. From Zomboliver’s point of view it would be no different to kicking a computer, doll or other inanimate object. It seems pretty clear that phenomenal consciousness is required to be a moral object.
But is it required to be a moral agent? Could Zombalice reasonably say that she’s not to blame for kicking Oliver because, like Alice the horse, she didn’t know the difference between right and wrong? I’m inclined to think not but things are a bit trickier here.
For instance, how would Zombalice know that kicking Oliver would hurt him? Normal Alice knows that it hurts when she’s kicked and can generalise this to other people but Zombalice doesn’t have this reference point. Of course normal Alice doesn’t know that Oliver feels pain; there’s no way to distingiush between Oliver and Zomboliver, but she does realise that her feelings correlate with certain stimuli and responses, and can infer similar feelings in others from corresponding stimuli and responses. Since Alice generalises from her own experience to that of both Oliver and Zomboliver, it is reasonable to belive that Zombalice would do the same. Where Alice infers phenomenal consciousness, Zombalice infers none. For Zombalice, Oliver may as well be Zomboliver and we’ve already seen that there’s nothing inherently wrong with kicking Zomboliver. Zombalice is off the hook, if for slightly different reasons than Alice the horse.
A zombie utilitarian would be strange indeed, trying to maximise something she had no concept of. A zombie Kantian might be more sensible since nothing in Kantian ethics requires moral agents to have any feelings, but I’d like to think of Zombalice as a utilitarian for now. How can we make Zombalice appreciate the difference between good and bad actions? Perhaps the most simple way would be to program it into her; to give her a list of stimuli and responses, a list of the feelings associated with those stimuli and responses, and a set of rules telling Zombalice to provide certain stimuli and elicit certain responses. Zombalice needn’t know what pain feels like to know that it is wrong to cause it.
Another possibility is for Zombalice to have phenomenal consciousness temporarily, like Data’s emotion chip in Star Trek. This would allow Zombalice to generalise from her own experiences in the same way as Alice even if she were no longer capable of having those expereinces.
So my thesis will really be two theses: that phenomenal consciousness is sufficient for moral objecthood and that it is not necessary for moral agency. I know of other objections to these claims (the experience machine, zombies lack motivation) but I’ll leave it there for today.
I wonder if you have made a distinction for the sake of making a distinction since I am unsure how it helps you separate non-phenomenal consciousness with phenomenal consciousness. I should also say that my intuitions on this matter follow Husserl, and I don’t see how the non-phenomenal is separating out of the phenomenal since it is a fact that certain experiences, including moral experience, are given in terms of the object and the feeling constituting those experiences.
What do you get out of the distinction?
Hi Ed,
That’s thanks to a lack of clarity on my part. When I said that phenomenal consciousness referred to things we feel, I meant experiences. You mention that some experiences are ‘given in terms of the object and the feeling constituting those experiences’. I agree, many experiences do have both affective and representational aspects. For instance, the smell of roast chicken is both chickeny (representational) and appetising (affective). My view is that both the chickeny smell and the appetising feeling are examples of phenomenal consciousness.
Non-phenomenal consciousness is different from either of these. I’m sort of following Chalmers’ idea of psychological consciousness or Block’s idea of access consciousness - a purely functional process that needn’t be accompanied by any subjective experineces.
For me, the distinction is important because it seems self-evident (admittedly as a hedonistic consequentialist) that phenomenal consciousness has implications for moral objects whereas non-phenomenal consciousness does not.
Zombalice is functionally (indeed, physically) identical to Alice, recall. She’s not going to behave any differently. So it’s not like we need to give her any special stimulus in order to convince her to play nicely. (I can’t tell whether that’s what you had in mind.)
But it might be argued that Zombalice’s utterances lack meaning, or at least do not have the same meaning that Alice’s would. Depending on your theory of mental content (among other things), you might deny that Zombalice has moral concepts at all. Or you might grant that she (it?) has the concepts, but deny that her moral beliefs can be justified in the way that ours can.
As for moral objects (is ‘moral patient’ a more standard term?), the key question seems to be whether Zomboliver has *interests*. So that’s going to depend on your theory of welfare. It’s easy enough if you assume hedonism, but otherwise we might ask why it should matter so much that Zomboliver can’t have experiences. He can (arguably) still do other important things, like fulfill his desires, or pursue projects at which he may succeed or fail, etc. Is that enough to suggest that his life can go better or worse; that he can be harmed?
I’m afraid I’m better at asking these questions than answering them. Good luck with your thesis!
Hi Tim,
You might want to take a look at Phillip Robbins and Tony Jack’s 2006 paper “The Phenomenal Stance” (if you haven’t already). They give an argument for thinking that moral agency requires assuming a phenomenal analogue of Dennett’s Intentional Stance. That is, part of being a moral agent is seeing the object/patient as a locus of experience, which (arguably) would also require the moral agent to possess phenomenal consciousness. If you’re looking for someone to represent the opposition, Robbins & Jack seem to fit the bill quite nicely.
An entirely separate worry I have is about the rather vague definition you’ve given for ethics. Surely my actions have effects on other individuals in all sorts of ways that aren’t relevant to ethics: my walking down the sidewalk casts a shadow on the shoes of others walking on that sidewalk (or having coffee at the sidewalk cafe); and if we are to take Lewis and physics seriously, any action I perform will cast an ever-widening causal “light cone” and effecting an indefinite number of future events. Surely, ethics isn’t concerned with all my shadow-casting actions or with all of the effects that lie in my causal light cone. Or put another way, surely I’m not acting qua moral agent in such cases. Presumably, there’s a difference between my walking down the sidewalk in the everyday context and my walking down the beach past the baby drowning in the shallow waters.
Thanks for the comments guys. Since this is going to be my work for the year I appreciate any pointers and criticism you’ve got.
Richard,
I’ve not studied it in any detail but my view is that when Alice’s own phenomenal states can be the intentional objects of her utterances. Obviously this isn’t true for Zombalice but I’m unsure if there’s a difference between the intentional objects of Alice and Zombalice’s utterances when they talk about Oliver’s phenomenal states.
I know Tom Regan uses the term ‘moral patients’ but it seems that he is specifically referring to moral objects that aren’t agents. Under Regan’s view a dog is a moral patient but a normal adult human is not. I prefer ‘moral objects’ as a sort of shorthand for ‘objects of moral actions’.
As for Zomboliver’s interests, I’ll be focusing on hedonism and preference-satisfaction in the thesis. Hedonism seems to require phenomenal consciousness but I don’t think preference satisfaction does. Personally, this makes me lean toward hedonism.
Adam,
Thanks for the suggestion. Most of the ethicists I’ve talked to don’t like the idea of zombie moral agents but it’s nice to have something concrete to refer to.
Yeah, I’m not happy with my definition for ethics either (it’s probably thanks to me not knowing what audience I was writing for). I’d amend it to say something like this:
Ethics (for consequentialists) is about how the actions of moral agents affect properties or states of affairs that are deemed to have intrinsic value according to some theory of value.
You may be interested to read Julia Tanney’s “On the Conceptual, Psychological, and Moral Status of Zombies, Swamp-Beings, and Other ‘Behaviourally Indistinguishable’ Creatures” from Philosophy and Phenomenological Research, 2004
Available on her website here: http://www.kent.ac.uk/secl/philosophy/staff/tanney.html
Thanks PfH. Now that it’s in front of me I remember skimming it last year for an essay but I’d forgotten about it because I didn’t need it then. It’ll definitely be useful this year though so I’m glad you brought it back to my attention.
Actually, when combined with your argument cited above about deontological systems, your argument on this page seems like a great argument in favor of Kantianism.
In the (possible) presence of zombies, there is no rational way to distinguish who or what is and isn’t a moral object. If you had convincing knowledge of zombies as possible in principle, there is no argument or observation that can rationally convince you that I am, or am not, a worthy moral object, so, for example, if I were drowning, you would not only have no way of absolutely knowing whether you should run to save me, or whether you should stay where you are so you don’t block anybody’s sun on the way over; there wouldn’t even be any evidence that could weigh on the matter in a ‘limited-rationality’ way.
From a hedonist position, of course, you could decide that your feeling of personal discomfort at watching me drown would outweigh blocking a few people’s sun in the aggregate utility function, but there are certainly examples that can deal with that. Should you jump in front of a bus to push two people out of the way? What about two zombies? How would you ever know the difference?
I know anyone who’s given serious consideration to a straight utilitarian position is comfortable with a certain degree of uncertainty in ethical judgement, but this seems absolutely beyond the pale. Acting as a complete egoist, or for that matter sociopath, would be a perfectly good-faith interpretation of such a utilitarianism!
I really just don’t think hedonism is at all plausible if zombies are metaphysically possible. (Myself I’d jettison the zombies first–they just cause too many problems. But, given your blogname …)
Thanks Xplat,
There’s no doubt about it, zombies cause huge epistemic problems for hedonism and your argument certainly is persuasive. I can’t really do justice to it here but when I’ve given it some more thought I’ll write a post on it.
Tim,
Do Zombies have goals? These states seem more like thoughts than sensations to me. If so, then a zombie can have its goals blocked, and that might be a moral issue. The zombie has rationally formed its goals based on its beliefs and desires (does it have desires?) and then someone willfully blocked it from achieving its goals. If someone did that to me, I’d see it as a moral issue.
Also, if we follow a strict sort of Kantian line, isn’t all that matters for moral objecthood that you are rational? An end in a kingdom of means?
Sounds like an interesting project–good luck!
PS I have serious doubts that zombies are possible in any relevant sense. Needless to say, I think Chalmers et al are wrong on this.
Hi Josh,
I’d say zombies have goals insofar as they have desires. But a zombie desire would lack any phenomenal component; they might desire food but they wouldn’t feel hungry, nor would they feel satisfied once the desire was fulfilled. I don’t have a problem with seeing desires in this way but it does seem that blocking a zombie’s goal is quite different from blocking a human’s goal. Now I’m pretty much a straight hedonist so I’d say it’s a moral issue in the case of the human but not in the case of the zombie. Preference utilitarians are concerned with the frustration of preferences so they might say that it’s an issue for both humans and zombies (though it would depend on whether they thought non-phenomenal desires were possible).
I probably should be a bit careful talking about Kantian ethics, because I don’t really know much about modern Kantians. Still, it does seem objecthood is tied up with rationality for Kantians (in as far as they find talk of moral objects useful).
Zombies, at least as Chalmers describes them - atom by atom replicas of humans but entirely lacking phenomenal consciousness - don’t seem possible to me either. But even though I don’t belieive in *physically* indistinguishable zombies, *behaviourally* indistinguishable ones seem like a possiblity.