This post is in response to a piece in New Atlantis by Austin L. Hughes:
The central question is something like `can science answer all questions?’, a much discussed topic. Note that it is not the question `is science the best way of answering all questions in its relevant domain?,’ a question to which we can be sure the answer is positive. It is more about the scope of that relevant domain, and in particular whether there is anything outside of it.
Scientists sometimes confuse the two. A good example of the sort of area where this may be false is ethics. If a scientist holds that only scientific claims are truth evaluable, you can ask him whether it is true or false that it is wrong to torture cats for fun. The choices are to claim that the statement is not truth-evaluable (not in fact a position to be quickly dismissed but unlikely to be chosen) or to allow that it is. If it is, then one can enquire of the scientist what data could be collected to establish the truth of the claim. There seems to be none. Thus we may make out the claim that science, while superb within its relevant domain, does not answer all questions of interest. That result may be confirmed by observing a group of scientists in the pub, and measuring the proportion of the time that they discuss questions which have an empirical answer: not a high proportion. In fact — why discuss such questions at all? Why not go and perform the appropriate measurement?
Hughes notes that science has taken over a lot of territory from philosophy. This is true, but should concern no-one. Philosophy which conflicts with science is valueless. Where it is useful is in the wider domain of non-empirical questions, which is why the domain question above is of such importance. Science will also tend to develop more than philosophy does — parts of philosophy of mind may be resolved by psychology for example. Note that this is a different process to the one where science potentially gradually answers all questions of interest, or provides a complete account of everything. How will science answer the question as to why there is something rather than nothing? (A question Hughes later raises.) What would a scientific answer to that question even look like?
Hughes is slightly unfair on scientists when he writes that they “confus[e] correlation and causation, or confus[e] implication with a bi-conditional. The former does occur quite often in journalistic reporting of scientific papers. However, while some scientists may be unaware that all they can ever measure is correlation, and that causation is not an observable quantity, I think most working scientists are alive to the possibility of confounds. This is why control groups are so important. The latter error is probably also not very frequently seen. The error precisely stated is proceeding from the claim “if A then B” and observing B, to conclude A. This will not be valid when there may be other causes of B. Scientists however are quite good at looking at several possible explanations for their data.
Hughes claims that philosophy can assist science by deciding what are the proper boundaries of science. This is true, but I think he weakens his case here by introducing Popper’s falsification requirement: only hypotheses which can be falsified are proper science. One problem with this is that it renders some items non-scientific while they certainly appear to be part of that domain. For example, there are string theories in cosmology which appear to be completely non-testable since they would require particle accelerators of immense energy. Thus, many of the string theories would be non-science which seems like an odd result — but not an impossible one. The worst problem though is that Popper may not be being philosophical. Many scientists would agree with the criterion but also count is as part of science — or at least meta-science, in that it is something that emerges from their general practice even if it cannot itself be subjected to scientific test or subjected to itself. That causes a problem of supportability because if the test of what is scientific is not itself a part of science, then what has been achieved, when one is seeking to delineate domains?
Hughes is unfair to Quine in criticising the latter’s use of `nerve endings’ as an explanation for certain biology. This is just a placeholder for the biology that transpires to be true in the relevant domain. Philosophers also speak of `c-fibres firing’ for whatever mechanism transpires to be the physiological underpinning of pain. Philosophers cannot be expected to do the biology better than the biologists, and are interested in different questions, such as: `if we had a complete physiology of pain, would it be sufficient for someone to feel pain that they replicated that physiology?’ Answering these questions does not require a detailed knowledge of the physiology which may not in fact be available. It does not even matter for philosophical purposes if the physiological information never becomes available. We can still have fruitful debates about the meaning of sensation without having a detailed description of how we sense, physiologically. (Or more likely, without having a detailed description of how sense data generate qualia.)
It is definitely correct, as Hughes says, that assuming that scientists are right on non-scientific questions is a mistake. But plenty of action is still properly left to science — dealing with climate change, furthering stem cell research will provide useful results, promoting vaccination, eliminating fears of radio masts — and the problem here is that the public accords too little weight to the views of scientists, for reasons that escape me, but which may be connected to the difficulties in defining the correct domain for science. They may also be related to the current distrust of experts in all fields.
Hughes goes on to challenge Hawkings’s admittedly very poor grasp of philosophical questions in three areas: metaphysics, epistemology and ethics; these are basically the questions `what exists?’; `how do we know anything?’ and `what should we do?’. As he points out, Hawkings’s idea that by positing the multiverse as an explanation for the existence of this and perhaps other universes merely moves the `why is there anything’ question one level back. Hawking is rather embarrassing on these sorts of topics because he attempts to take a line of which the starting point is “now physics has progressed thus far we no longer need philosophy” and then proceeds to do philosophy very poorly. One one occasion, I have seen a Nobel prize winning scientist deliver a lecture on ‘the method of doubt’ without mentioning Descartes, which would be seen as extremely bold in philosophical circles — by which I mean you would fail the essay as a first year undergraduate.
Hughes is also right to dismiss the idea that evolution explains why we can understand the universe because it does not explain why the universe is governed by stable scientific laws. We could have evolved in any way at all in a universe with unstable scientific laws — assuming we could survive in such a beast — and be unable to understand anything because conditions would change all the time. There could never be any replication and reproducibility is crucial to the scientific method.
Hughes is also right to note that the claims that evolution should select us to believe only and always the truth are questionable. He might go further though and note the work of McKay And Dennett in `The Evolution of Misbelief’ on the fitness enhancing nature of certain false beliefs. For example, a healthy level of delusion about one’s own capacities is associated with optimal mental health. Hughes might also consider the extensive work of Gilovich (`How we know what isn’t so’) on the vast amount of things people believe — in religions, in astrological dreams, in the predictive power of tarot cards, in the healing power of crystals, that interviewing candidates produces better hires — which are simply empirically false. Hughes does usefully note Trivers’ work in this context though.
Hughes is wrong when he says “it is very hard to see such an advantage to DNA sequence analysis or quantum theory”. While he is right, strictly speaking, to say that evolution could not possibly have selected us to be good at quantum theory, it is entirely plausible that it could have selected us for ever more and improved information gathering capacities and intelligence, and that those have after many years led us to quantum theory. We might even think that extreme intelligence is like the peacock’s tail. Not necessarily essential in itself, but a powerful method of attracting mates because it signals “I can afford to carry this huge brain around.” It might just be simpler though to hold that more intelligence is always better for survival in an uncertain world.
Hughes makes his strongest point when he attacks naturalism, the view that what is natural is right. Why would it be? If what our ancestors did was selfish and driven by natural selection, why would our current views be any less driven by natural selection or any less wrong? Note that none of this means that natural selection is not correct. It is indeed the correct explanation of how we came to be here. It is right if right means correct, but it is not right if right means good. That is not the appropriate question to ask about something like natural selection. It is.
Hughes is wise to challenge utilitarianism, because it tends to be widespread outside philosophy and especially appealing to scientists, because they think they can measure things to answer the question as to what is the greatest good for the greater number. Anyone in philosophy can provide dozens of fatal counterexamples. One of the simplest ones is the christians and the lions: if the value to each of 100,000 spectators of seeing a christian torn apart by a lion is greater than the disvalue/100,000 to the christian, utilitarianism (in simple versions) supports it. Avoiding that conclusion requires some fancy footwork, particularly when one notes that the number 100,000 can be made arbitrarily large in a thought experiment. Also one can ask utilitarians why they think an infinite number of people living on the planet is a better result than a smaller number.
There is an immediate counter to Hughes’s claim that “both states and countries with high rates of reported “happiness” also have high rates of suicide suggest that people’s answers to surveys may not always provide a reliable indicator of societal well-being, or even of happiness.” To be sure, he may be right in his premise that high happiness countries are high suicide countries. But a more reasonable interpretation of this is that people do not know if they are happy, and to bear in mind that the number of people who commit suicide is very small, and perhaps inequality in societies is a factor? Perhaps a society in which some people are very happy because they are rich, and most people are quite happy because they are quite well off, is very difficult for very poor people to live in?
I do not understand Hughes’s complaint that “Harris seems to think that free will is an illusion but also that our decisions are really driven by thoughts that arise unbidden in our brains.” He seems to think that these claims are in tension when in fact they are entirely consistent. Perhaps Hughes thinks that if a decision arises in my brain, then ipso facto I have free will, but that is in need of an argument and contrary to views in which the subconscious is prominent.
Hughes’s conclusions is well made and supported by his arguments. He writes: “Continued insistence on the universal competence of science will serve only to undermine the credibility of science as a whole.” This is surely true. In fact, would not the credibility of science be enhanced by acknowledging the proper restriction of its domain? Within the empirical arena, it is supreme and unchallenged. But that’s not all there is.