The subtitle of Sam Harris’ The Moral Landscape: How Science Can Determine Human Values can be read in two ways. One would point to a book I might be interested in reading, the other to one I could dismiss in advance as nonsense. My problem is that the review by Russell Blackford recommends the book in spite of criticizing it on the grounds that it fits the second interpretation, and Blackford’s criticism is so cogent that his recommendation despite that problem might convince me to be less dismissive. On the other hand Harris’ responses to the review and to others who have expressed the same concern have been so facile as to reinforce my doubt that he has anything useful to tell me.
The key is how we interpret the word “determine”. It is the job of science to predict observations and one interpretation of “determination” is as prediction, as in “determine the precession of the orbit of Mercury”. In that sense, to “determine human values” is to predict with some reliability what an average human individual or group will decide to value under specified circumstances, and any theory which helps with the understanding (and manipulation) of that would certainly be interesting (and perhaps dangerous). The other interpretation of “determine human values” is to establish the truth of a value proposition. But science neither establishes “truth” nor has any construct which has the imperative or exhortative content which is implied by a statement of moral value.
A related confusion is with how the word “should” is sometimes used in science – but without any kind of exhortative implication.
When a chemist says that if a certain mixture is heated then a specified result should form, she does not mean that if it is heated and the specified result does not form then the chemicals will have done a bad or wrong thing, but rather that she will have been wrong in her theory and/or her understanding of the situation.
Similarly when a social scientist (assuming such things actually exist) says that people in a certain social context should prefer a certain type of behaviour as morally correct, he does not mean that if they don’t prefer and act as he predicts they will have committed a moral error, but rather that his theory of their morality will in that case have been deficient. He can (and should) strive to optimize his efficiency in predicting what they will choose, but he cannot (and should not) claim that his prediction of their judgement actually entails that judgement. (In certain contexts, of course, it may well be true that their judgement will in fact be influenced – or even determined – by his prediction, but that will only be as a result of whatever level of respect his opinions are accorded in the ambient culture. This is but one way in which science might be used in a moral argument to help people determine their moral values and guide their behaviour, but it should be clear that this is not the same as saying that the science itself determines them.)
Harris, however, in his responses referred to above, really does seem to think that science can determine human values in the second sense. And he specifically alludes to some concept of “total well-being of sentient entities” as a definable numerical (or at least totally ordered) value that should be maximized. (Of course I doubt that such a goal would be universally accepted by humans if faced with a competing species who claimed the capacity for a higher level of consciousness and well-being and felt that our continued presence was an impediment to the maximization of their happy numbers, but even if it is so accepted, there is no scientific statement that corresponds to that being the right thing to value.)
Of course the idea that there is a well-defined scalar objective function which is maximized by moral decisions is really so simplistic as to be laughable, but even if there were one it should prove of little comfort to Harris.
Ironically, his image of a mountainous “Moral Landscape” provides the ammunition for his defeat even if one does accept the idea of a well-defined objective function, for if the situation was simple enough for there to be one it would still not necessarily have a unique maximum. So there may well be moral positions from which the process of increasing “morality” leads to completely different maximally “moral” positions. And if such things were actually computable, perhaps a world of paternalistic tribes with violently enforced adherence to social taboos but with freedom from doubt and a sense of moral righteousness would indeed lead to equal or greater total happiness and stability as one of sexual and intellectual freedom. (As might also the world of drugged-out zombies which he regards as a frivolous example from where he stands now.) In either case, one has to give up what one holds to be of great value to achieve what the other claims to be a greater value – and there is no way of comparing values without experiencing and observing both universally applied and at their peak.
My frustration with Harris is that, from my position part way up one of the two mountains mentioned above, I need a way to deal constructively with those on the other one and hopefully to persuade them to climb down (or across some kind of bridge) to join me. There may be ways in which science can help with that, but by misstating its role in an easily refuted way Harris discredits and undermines that endeavour.
(Note to self: add more re last para of Blackford’s review)