Fellow-Feeling

Dr. Reynolds has a good point.
[I]ronically, it's not rational to be too rational.

Imagine that you're thinking of getting married. Would you want a spouse who sticks with you for purely rational reasons, or one who forms an irrational attachment — let's call it "love" — that doesn't depend on rational factors?

Most people would say the latter. A purely rational attachment is nice, but if things change — say, if you become sick, or unattractive, or broke — a rationally attached person might rationally choose to leave. A person who loves you, on the other hand, might stick around anyway, because being parted from you, even if some of your charms have vanished, would cause emotional pain, while helping you feels good.

Likewise, you'd like to hire an honest employee, one who will feel guilty about stealing from you. A rational employee won't steal if there's a danger of being caught, but an honest one won't steal even when he can get away with it, because if he does he will feel guilty, while if he resists temptation he will feel virtuous.

A person who is perfectly rational about costs and benefits, with no irrational constraints like loyalty or honesty (or patriotism), is a person who will lie, cheat and steal whenever he or she can get away with it. A sociopath, basically.
This is a position that the Enlightenment tried to do away with, but it's worth noting that even Kant came around to it. While maintaining that ethics was an exercise of practical reason, in the Doctrine of Virtue he ends up describing a number of "moral feelings" that are the ground of caring about doing the right thing at all (Ak. 6:399-403). This was published late in his life, many years after his more famous Groundwork that appears to downplay the role of feelings in morality to an extreme degree.

On careful and long consideration, a certain set of feelings are presupposed in caring about doing the right thing. A purely rational being just might not care about going above and beyond the demands of law for the benefit of other people.

11 comments:

E Hines said...

A rational employee won't steal if there's a danger of being caught....

and

A purely rational being just might not care about being virtuous at all.

I disagree with both, which are saying much the same thing. A purely rational being will insist on being virtuous. Less than virtuous is imperfect, and so irrational. That we often disagree on what is virtuous, what is honorable, what is rational is our failing, not a failing of rationality, virtuousness, honorableness.

Eric Hines

Grim said...

I altered that last sentence on reflection, because Kant means something different from "virtuous" than Aristotle does (which means that the word, for Kant, means something different from what I almost always mean by it). The new version of the sentence reads:

A purely rational being just might not care about going above and beyond the demands of law for the benefit of other people.

That's what Kant really means by 'virtue.' He doesn't mean 'strength' or 'excellence,' as Aristotle does and I do: those are objective, and thus of course a rational being will see the strength for a strength.

What Kant means is that you'll do things for others that they have no claim on you to demand at law.

E Hines said...

What Kant means is that you'll do things for others that they have no claim on you to demand at law.

That's exactly what our Judeo-Christian heritage requires of us as individuals to do: to provide succor to the least of us who cannot provide their own. Independently of the law. And wholly rational.

Eric Hines

Grim said...

Not in the sense that a computer is rational.

The emotional states Kant identifies are "moral feeling" (by which he means the ability to feel affected by moral duties), conscience (the capacity to recognize moral duties), love of human beings (which is the real ground of Jesus' commandment), and respect ("a feeling of a special kind," for which he gives the Latin reverentia).

It's not clear that, without those emotions, you'd recognize or feel bound in the ways Kant thinks you should. Even if the calculations are based on "pure practical reason," you need some emotional content to feel bound.

Grim said...

Though, as I've argued here several times, I think Kant is wrong to impute as much to rationality as he does. Computers can never be ethical until they develop a way to formalize analogies, the best system for which is still quite primitive. This is because ethical and political thought is analogical, not logical: and analogies always break.

Therefore, we need to go beyond logic even to begin talking about ethics. No two cases are exactly alike, so we can't apply logical forms, which depend on the existence of logical objects.

Grim said...

"Computers can never be ethical until they develop..."

Unclear antecedent. I mean, computers can never be ethical until human developers develop a way of formalizing analogy in a solid way.

E Hines said...

Computers are never rational. They can't even do random, much less irrational.

... you need some emotional content to feel bound.

Hence the rationality of including, and feeling, emotion.

Eric Hines

Grim said...

Computers can't do random or irrational, and they can't do illogical or analogical. They can apply a rule, or even a very complex set of rules. Computation is any system in which outputs can be predicted by inputs from the rules.

"Rational" is used in a number of senses, but it tends to carry the connotation of calculation. Computers can do that, but that's not the kind of process that is useful for ethics or politics.

douglas said...

"It's not clear that, without those emotions, you'd recognize or feel bound in the ways Kant thinks you should."

It's pretty clear- we call it sociopathy.

Grim said...

You mean it's pretty clear in the other direction. :)

douglas said...

ah, yes- :)