Thursday, May 10, 2012

Torture - Part I: Sam Harris' Moral Theory

A member of the studio audience has asked me to address the subject of torture.

I was wondering what your thoughts about Sam Harris' arguments for torture were. He seems to proffer the view that, by accepting collateral damage from war as inescapable yet demonizing the collateral damage from torture, we are holding to a very irrational double standard. I find his defense of torture emotionally disturbing, but I can't seem to produce a logical argument that counteracts his thesis.

I am going to do this in two parts.

For the first part, I am going to address the moral foundations on which Harris built his defense of torture. Harris uses an act-consequentialist moral theory. An act-consequentialist theory says that the right act - the act one ought to perform - is the act that produces the best overall consequences. Torturing a suspect to prevent a bomb from going off is legitimate when it produces the best consequences - which, when the suffering of the person tortured is stood up against the killing and maiming that will be prevented, is almost always the case. In this post, I will raise objections to that foundation. Most of my objections will apply to all forms of act-consequentialist morality.

One thing that an act-consequentialist needs is an account of what counts as good consequences. Harris has proposed that good consequences can be cashed out in terms of the well-being of conscious creatures. The act that produces the best consequences is the act that produces the most well-being for conscious creatures.

What Is Goodness?

The first problem we have here is that Harris’ concept of good is empty. The term “well-being” is, itself, a value-laden term. It does not answer the question of what goodness is because now we are burdened with the task of unpacking "well-being" to try to get at the value contained within.

Intrinsic Value

The second problem with Harris' theory applies to all theories that postulate some sort of intrinsic value. Harris' theory is a member of a large family of theories in which the theorist claims that value is intrinsic to some state - that certain organizations of matter are simply inherently good. Other members of this family identify happiness, pleasure, eudaimonia, the absence of pain, life, existence, preference satisfaction, desire fulfillment, self-realization, autonomy, health, wealth, power, or various combinations of these as the organizations of matter having this property of intrinsic worth.

I reject all of these proposals on the grounds that intrinsic value does not exist.

If intrinsic value did exist, how would it work? What would it require for a state of affairs such as happiness or "the well-being of conscious creatures" to have this intrinsic property of goodness? Is there some sub-atomic particle – some sort of “goodon” – that matter emits when matter organizes itself into the form of a happy creature or a conscious creature with well-being? How are we detecting this "goodon radiation?" Do we have some sort of goodon detector built into the brain? Is it reliable? How is it that being a goodon emitter implies that all intentional creatures are obligated to maximize goodon emissions? How do goodon emissions generate duties and obligations?

Any argument that value resides entirely in some state and our duty us to maximize realizations of that state needs to tell us how that state acquired this property. If it is not through goodon emissions, then what is it?

Ultimately, I hold that there is no answer to this question. Furthermore, when a theory that requires some sort of intrinsic value is held up against one that does not, we should go with the simpler theory. Intrinsic values do not exist. A person who attempts to build a theory grounded on intrinsic value is in no better shape than a person who attempts to build a theory grounded on God. Both myths only produce moral fictions.

Note that my list of things that do not have intrinsic value includes "desire fulfillment". One objection I often hear to my own account of value is that I am assigning intrinsic value to desire fulfillment - and intrinsic value does not exist. That is not the case. Desirism says is that a desire that P provides an agent with a motivating reason to act so as to realize a state of affairs S when P is true in S. Desire fulfillment has no intrinsic value - desire fulfillment is not the goal of intentional action. A state of affairs S where P is true in S is the goal of intentional action – for an agent with a desire that P.

This "state of affairs S" might not contain any desire fulfillment at all. An agent might desire that a particular planet – containing only plant life (no animals, no intentional agents of any kind) continue to exist long after the agent is dead. This would motivate the agent to realize a state of affairs in which that planet continues to exist. However, a state of affairs in which that planet continues to exist is not a state of affairs in which desire fulfillment has been maximized. In fact, it is a state of affairs where no desire fulfillment exists at all. Desire fulfillment has no value. A state of affairs S where P is true in S that has value to agents with a desire that P.

Action-Based Theories

A third problem with Harris' account of value applies to all theories that make action the primary object if moral evaluation. They ignore important facts about how intentional agents work.

Intentional agents seek to fulfill the most and strongest of their own desires, and choose that action that would have fulfilled the most and strongest of their own desires in a world in which their beliefs were true and complete. (Note that this does not imply "selfishness" as some people argue. The desires that an agent seeks to fulfill can and often does include desires for the well-being of other people and aversions to the harm or suffering of other people.)

This means that the only way that an agent can always and only act so as to realize a state of affairs that maximizes some "intrinsic good", be it pleasure or happiness or the well-being of conscious creatures, is if the agent has only one desire. An agent must have a desire to maximize happiness, or pleasure, or whatever it is that the act-consequentialist claims to be the holder of ultimate value - and no other. If the agent has any other desire – an aversion to pain, a love for his own children, a desire for sex, a preference for chocolate over vanilla, then this desire will, under some circumstances, motivate an agent to perform something other than the act-consequentialist best act.

And we are all awash with these other desires and interests. If we have a desire to perform the act-consequentialist best act, it is one weak voice in a massive chorus of concerns and interests.

Taking seriously the charge that we must also perform the act-utilitarian best act would mean deciding what to buy at the grocery store according to which purchases produce the best consequences, planning the family vacation around what would produce the best consequences, having sex when and only when and only because) having sex at that time and in that manner produces better consequences than anything else one can be doing at that time.

Insisting that everybody have only one desire - the desire to produce the best consequences - isn't even a possibility.

In this post, I have provided offered criticism of the underlying theory that Harris used in his analysis of torture. The post is getting a bit long, so I will save the next part for tomorrow. I will look at the issue of torture specifically, at the problems that arise from Harris' moral theory, and present an alternative that avoids those problems.


Unknown said...

Let me start by saying I usually love your stuff. I'm an atheist looking to study moral decision making (with an emphasis on stereotypes against atheists & other groups), so I greatly appreciate your blog! With that being said, I found this post very poorly thought-out.

I would have to call myself an idealist "act-consequentialist," such that we cannot know, right now, what will provide the best consequences, nor do we even know how to measure said consequences. On that much I agree with you; that is not a new claim, though. Even well-being researchers acknowledge that whole-heartedly (I work in a world-renowned neuroscience lab studying a lot of meditation, so I hear plenty of well-being talk). As Sam Harris has pointed out, we will get there. There's no question: Neuroscience is advancing at such a rate that we will surely have various truly valid & objective measures of "well-being" for individuals in the very near future, despite the slipperiness, or near-emptiness, of that term today.

However, my qualm with the act-consequentialist stance is that the emotional bias toward those close to you (i.e. being willing to sacrifice 40+ people on the other side of the world in order to save the life of your daughter), although not purely "rational," is perfectly ethical. That is not to say we are completely OK to sanction genocide in the name of those we love, only that it's OK for one person's life to be worth more than another's (potentially a LOT more). We cannot possibly intend to further every single human being's well-being. We don't have enough room nor resources on this planet for every human to be perfectly Happy. To wish Happiness, with a capital H, on every human is to wish doom on those to come, and Happiness should include the knowledge that your children and their children will be just as happy or happier than you. Obviously, the striving for well-being includes overcoming that limitation, but as things stand, it IS a limitation.

Also, your breakdown of Action-Based Theories seems to claim that the only drive in humans should be for a more moral world. Clearly there are decisions that have nothing to do with morality, and never will. Chocolate vs. vanilla? If there were evidence for it (I have absolutely no idea), one could claim that vanilla production is the direct cause of more deaths than chocolate production. In that case, sure--vanilla would be a more immoral flavor than chocolate. Otherwise, it just plain is not a decision with any moral bearing. Also, knowing that the groceries you buy, so long as you're buying a generally considered "healthy" lot, and nothing produced with the blood of its workers, might only contribute to ~.5% of the moral variance in your moral capacity would make that a trivial decision, despite having some impact. You're simply taking the argument ad absurdum, making you sound something like a theologian trying to dismantle this.

Sorry if that wasn't perfectly coherent or sound in the explanation of my reasoning. Perhaps I'll extend my views into an independent post sometime soon...

Thanks for all your great posts, despite my slightly negative reception of this one!

Alonzo Fyfe said...

Luke Hinsenkamp

I think it is a bit early for you to say that you disagree with me.

I will agree that we will come up with a more objective account of well-being. However, when we do, we will discover that it (like pleasure, happiness, or the others) cannot legitimately be called the sole end of human action. There will be no intrinsic value in well-being.

That we will someday have a reduction for well-being does not answer the objection. When that some-day gets here, the objection will go away. Until then, the use of "well-being" as a way to unpack "goodness" remains empty today.

Also, as you will see tomorrow, I do not advocate that the only drive is for a more moral world. In fact, I would object to any theory that made such a claim - for much of the same reasons that I object to intrinsic value theories. "A more moral world" would be one of those things that have no intrinsic value.

Desirism allows for a broad category of interests that have no moral bearing at all. In fact, one of the arguments I provide in favor of desirism is that, unlike act-consequentialism, desirism accounts for the moral categories for evaluating action - obligatory, prohibited, and non-obligatory permission.

Act-consequentialism, on the other hand, has the implication that there is only one permissible act (the act that produces the best consequences) and all other acts are impermissible. That is another reason why I reject act-consequentialism. It cannot account for all three categories.

Jesse Reeve said...

The term “well-being” is, itself, a value-laden term. It does not answer the question of what goodness is because now we are burdened with the task of unpacking "well-being" to try to get at the value contained within.

That we will someday have a reduction for well-being does not answer the objection. When that some-day gets here, the objection will go away. Until then, the use of "well-being" as a way to unpack "goodness" remains empty today.

Alonzo, how do you address the same objection to your use of the word "desire?" Can you reduce human "desire" to a description of biological, let alone physical, phenomena? If not-- and if you consider this inability to "unpack" or "reduce" a term to be a serious problem-- then what justifies your assumption that humans have "desires" in the first place?

Alonzo Fyfe said...

Well, I cannot give a full account in the space of a blog comment. Nor will I deny that there are known problems with the belief-desire theory of intentional action. However, I can generally report that the terms 'belief' and 'desire' report functional brain states. They describe how the brain is structured so as to relate input to intentional action given other brain states.

The suggestion that beliefs and desires exist is justified by their role in explaining and predicting a large set of important real-world events - the behavior of intentional agents such as other people. While it is possible that a better theory will come along that we can use to explain and predict the actions of others, it has not happened yet. As with everything else, its existence is justified by its usefulness in explaining and predicting observed phenomena.

Though neuroscience s a very young discipline, it does seem to be supporting the thesis that intentional action is motivated by brain processes that at least roughly correspond to the properties we are familiar with under the concepts of belief and desire. Of particular importance is the progress that has been made in understanding hoe praise, condemnation, reward, and punishment work to strengthen some desires and weaken others.

Jesse Reeve said...

All of these justifications for using the concept of "desire" apply equally well to using the concept of "well-being."

Alice states, "I have a desire to eat ice cream." No reasonable person would respond by demanding she unpack "desire," and otherwise rejecting her claim. Even if "desire" is an imperfect model of human motivation, it is strong enough for Alice's purposes.

Bob states, "I refuse to inject heroin into myself, because it would harm my well-being." No reasonable person would respond by demanding that Bob unpack "well-being," and otherwise force him to inject heroin.

On a larger scale, claims that people "desire" ice cream and that heroin use harms "well-being" lead to useful, accurate predictions about people's willingness to spend money on ice cream and the effect of heroin use on someone's life. Both concepts are, as you put it, useful in explaining and predicting observed phenomena.

I would be surprised if you hadn't seen it already, but Luke Muehlhauser has written a relevant article, A Crash Course In The Neuroscience of Human Motivation:

Alonzo Fyfe said...

Jesse Reeve

The important difference is that "well-being" is a value-laden term.

"Desire" is not.

Alonzo Fyfe said...

I should add, one of the reasons why I am confident that well-being can, eventually, be unpacked is because desirism provides us with a way to do so.

All true value claims relate states of affairs to desires. Therefore, all value-laden terms can be understood in terms of how they answer four questions:

What is the set of state of affairs this term is being used to evaluate?

What are the relevant desires?

Do those states of affairs fulfill or thwart the relevant desires?

Do they fulfill or thwart those desires directly or indirectly?

Usefulness - can be used to evaluate any state of affairs. Relevant desires are determined by the context in which the term is used. A useful item fulfills the relevant desires. Insofar as it is useful, it fulfills those desires indirectly.

Beautiful - applies to states of affairs that are seen or heard. The relevant desires are those regarding visual or audio perception (what a person likes to see or hear). Beautiful items fulfill desires. Insofar as the item is beautiful, it fulfills those desires directly.

Health, Injury, and Illness. Both of these terms are used to evaluate states of physical or mental functioning. The relevant desires are those of the agent whose health is being evaluated. Health is a state of functioning that tends to fulfill desires, whereas injury and illness tend to thwart desires. Direct and indirect relationships both are relevant to the use of these terms. The significant difference between illness and injury is that, with an illness, the cause of a change in functioning is a micro-cause (imperceptible without special instruments), whereas an injury has macro cause such as falling off a ladder where a person can actually know the cause.

Well-being refers to an agent's over-all state (not just phyiscal and mental functioning but security, economic and political power, autonomy). The relevant desires are those of the agent. Well-being is a state where the desires in question are or can be fulfilled. Both direct and indirect fulfillment are relevant.

Alonzo Fyfe said...

Now, using what I wrote in the previous comment, I had said that once we unpack "well-being" we will discover that it represents only a subset of our interests, and that no argument is available for concluding that it is either the sole, or that it ought to be the sole end of all human actions.

In the unpacking done above, this is precisely the result. Our interests regarding our own state - though certainly important - represents only a subset of our interests.

It would take an intrinsic value claim to conclude that these are the only legitimate interests and that all others are illegitimate - and intrinsic values do not exist.

Geoff said...

Why do you care about "desires that tend to fulfill other desires" if "desire fulfillment has no value"?

A goal to realise a "state of affairs" without an accompanying desire is not possible. Your planet and experience machine examples are not convincing.

Desires have intrinsic value. Well-being only makes sense when reduced to desires. Emotion/feeling is the only source of end values.

Alonzo Fyfe said...


Because it will help me to broadly explain some things about desirism, I will answer your comment in a post sometime this wekk.

Please stay tuned.

GeoPorcupine said...

Excellent post, Alonzo. You capture a lot of the problems I have with The Moral Landscape especially trying to shove all the conflicts under the rug of the term "well-being of conscious creatures."