I will not be disputing the premises in Harris’s “central argument” in his “public challenge” announcement, nor do I have have to prove that his conclusion is false (though it probably is). The key indefensible bit in his “argument” is the “Therefore” bit and I only need show his conclusion that “questions of morality and values must have right and wrong answers that fall within the purview of science (in principle, if not in practice)” is not necessarily true – ie that “In principle (and probably also in practice), questions of morality and values may conceivably fail to have right and wrong answers that fall within the purview of science”.
But first, with regard to the specific sub-theses that Sam would like to see demolished, let me make the following comments:
(1) A state of “worst possible misery” is not actually something that has been shown to exist (in fact it is exactly as problematic as finding a state of maximal total well-being discussed below), and in any case a state of arbitrarily severe misery is plausibly better in some consistent moral view than one of semi-conscious happiness in which all of the individual’s personal dignity is lost.
(2)&(4) A science of morality can be a perfectly valid branch of psychology in which the objective is to predict what humans in any particular context will judge to be morally correct, but it will not itself determine that moral correctness and may well predict disagreements. Science can also help people make moral (and medical) decisions by predicting outcomes, but the actual decisions often involve trade-offs where the choices made by different people in essentially the same context may well be completely different. What science does not do is tell us which choice to make, with regard to either “health” or morality.
(3) Finally, and most importantly, the “landscape” model is wrong (because it implies the existence of a single real-valued objective function for which there is neither any reason in principle nor any convincing evidence), and useless (because even if there was one the possibility of widely separated local maxima makes it impossible to be sure of having a unique answer to every question).
With regard to the first point, even for a single individual there are various chemical concentrations, neuronal excitation levels, and connectivity parameters, or whatever, that might be associated with different kinds of reported well-being, possibly in conflict with one another and with no known way of “best” weighting them to provide a single scalar ordered measure of “well being” nor any reason for such a “best” weighting to actually exist. (And it is important for it to be a scalar since something with many components cannot necessarily be ordered!). But even if there were some single measure of the well-being of an individual, the issue of how to weight the relative importances of different people and potential people over all of time and space would also defeat the project of maximizing “total” well-being. Amazingly Harris doesn’t seem to understand the difference between questions like these which really don’t necessarily have an answer with ones such as “How many birds are in flight over the earth right now?” which do have an answer but one that is essentially impossible for us to determine.
On the second point, in the unlikely event that there really was a real-valued (ie ordered) measure of total well-being for us to take as the “Objective Function” to be maximized, then the mental model of a “landscape” (over a multi-dimensional state space) could be realistic, but even then the question of what to do in any given context would not necessarily ever be answerable. This is not just a matter of technical detail or computational difficulty but inherent in the general structure of the problem. There may well be states of equal total well-being achieved in completely different ways, or we may be near a lesser “peak” with no route to the higher one other than by descending through a land of great pain and peril.
But, getting back to my main objective, dealing with those four canards was not a complete waste of words because if you understood what I was saying, then from where we stand now it is quite conceivable that no matter how much we learn we will never be able to decide moral questions by appeal to science (or to anything else for that matter!)
Let us imagine two people coming from different tribes one of which has succeeded over many generations as cultivaters by encouraging its members to value bling as the primary virtue while the other has succeeded as herders by fostering the values of blaah. Both evaluations have served their respective cultures well but they conflict on the issue of bong. Let us even imagine that someone has defined a parameter that is believed to truly represent the aggregate success and flourishing of the species (not an implausible scenario if one does not require that belief to be well founded but probably a total crock otherwise), but let us also imagine (as may well happen) that that parameter will not be predictably different under either regime. Of course we can imagine such a situation, so it is clear that we can envisage circumstances in which a moral question can have a “right” answer to one person which is “wrong” to another with science unable to adjudicate. I do not need to find such an example in the annals of anthropology in order to establish that it is conceivable, so I am certain that I have demolished Sam’s claim to have shown that science must ever, even just in principle, be able to provide determinative answers to moral questions.
I also believe that I have shown the error in each of the specific sub-arguments that Harris cites in his elaboration of the challenge.
Of course the prize may go to someone else. But go it must!
(And I wouldn’t mind at least being included in the long list of successful refuters.)