Sunday, February 9, 2014

The Moral Landscape – A Response

Being a huge fan of Sam Harris' book, The Moral Landscape, I was intrigued when he announced a contest for people to try and disprove the central argument of his book. You can see the contest details here:

Given that this is a topic I care greatly about, I spent some time thinking through his arguments along with various ideas on the science of morality that I've been pondering for a while now, and found that I actually had some objections to his thesis after all. So I decided to write an entry in the competition, which I've reproduced below.

I'd love to hear opinions, critiques, or any other feedback from anyone who has read the Moral Landscape. Of course I'm happy to hear from anyone, but this essay probably won't have much meaning if you're not familiar with the book.

The Moral Landscape – A Response

The main flaw that I can see with the central argument of The Moral Landscape is not the concept of using science to uncover truths about how to maximize well-being of conscious creatures, but the extension of this to science being able to give definitive answers, even in principle, about what we ought to do. The reason why I see this as a problem is because it doesn’t seem possible to provide a single, obviously correct definition of what it is that we’re trying to optimize.

Using the Moral Landscape metaphor, in order to work out which peaks are higher than others, or how any two positions on the landscape relate to each other, we need to define what this ‘altitude’ axis actually represents. Since we’re talking about the collective well-being of all conscious creatures, we need some way to aggregate the individual suffering and flourishing of each conscious creature into a single value that can be compared with any other possible state on the moral landscape.

This may sound like a problem in practice rather than in principle, but I would argue that it is very much a problem in principle. I'm not talking about how we would go about, in practice, calculating the well-being of each individual creature in order to arrive at some collective well-being value. The problem is how to define a formula for collective well-being that is clearly correct, and is not just one of many possible ways to define collective well-being. If there is no single, obviously correct way to define the collective well-being of all conscious creatures, then we have multiple ways to compare different states of collective well-being that will result in different answers, and therefore no valid, scientific way that we could say one is better than the other.

Assuming that we can devise some way to measure the well-being of an individual creature, what then? Do we take the average of all creatures and assign ‘heights’ on the moral landscape based on this average, thus making the goal be the highest average well-being of all creatures? Or do we add up all the individual well-being values and simply go for the highest total score? How can we decide, even in principle, which is the more valid way to measure collective well-being?

Further, how do we account for deviation from the mean? Say we have two different states with the same average well-being for all conscious creatures, but in one case, all creatures have the exact same value, i.e. the average value, while in the other one, some are much higher, while some are much lower, but they average out? Should these be equivalent peaks on the moral landscape? Is each as good as the other? How can you scientifically determine that answer, even in principle?

The other major problem here is how to account for creatures with different ‘levels’ of consciousness. How much should we value the well-being of a dog compared to a human? How does our aggregate value of well-being weight the individual values of every type of creature that has some form of consciousness? If we only cared about the well-being of humans we could get away with weighting each creature equally, but it seems reasonable that creatures with the capacity to suffer more and/or the capacity to have more profound positive experiences, should somehow be weighted greater in the global well-being equation than those that do not.

How many puppies should suffer before it’s preferable for a human to suffer? Should human beings intervene in the predator/prey relationships in nature, and the subsequent suffering that these relationships cause to prey? Nature has evolved some truly horrible behaviours that various conscious animals have to endure, and all of this suffering and flourishing needs to be taken into account, along with humans and the animals we choose to look after as pets.

We need to remember that well-being is often a zero sum game. Many choices that affect well-being positively for some creatures, will affect it negatively for others. This is what makes moral science a very different thing to other areas of science such as health. Imagine if, whenever a person took actions that increased their life expectancy by one year, some other person would have theirs reduced by one year? How would the science of health look then? It would suddenly look a lot more like moral science, where we would need to justify exactly what it is we were trying to maximize, and a single definition of health would suddenly become much more important.

Moral science is the study of collective well-being. We can discover all kinds of truths about how to maximize specific aspects of well-being. But there does not appear to be any single, unassailable definition of collective well-being of all conscious creatures, and without this, we cannot rely on moral science to tell us what we ought to do, because moral science can’t tell us what definition of collective well-being we must use.