Friday, February 5, 2010

What Science Knows

Jim Franklin’s What Science Knows: And How It Knows It is a wonderfully lucid explication of science and the philosophy of science.

Philosophers are not always the clearest of writers. Sometimes it is a result of the complexity of the ideas they are grappling with. Sometimes it is because their thinking is not all that clear, and their prose reflects that. Sometimes, I suspect, it is because they associate being very clear with being simple: lucid prose does not, apparently, imply profundity of thought.

Jim Franklin’s thinking is clear and he has no such false ego, so we get powerful thinking about science presented very clearly, clearly enough for any reasonably intelligent layperson to follow. It may help that he is a mathematician who also does philosophy.

This book is a defence of science, both against its enemies and misunderstandings by its friends. Science is about reasoning from evidence, so the first chapter sets out an Objective Bayesian (or logical probabilism) view of evidence which:
… holds that the relations of uncertain evidence to conclusion is one of pure logic (p.7)
It is certainly the view I am most comfortable with, that makes the most sense and to which our actual reasoning seems to tend to conform. Franklin sets it out very clearly, interspersed with useful selections from key thinkers. While it is all clear, sensible and informative, I particularly commend his discussion of observation and experiment (Pp21ff).

The next two chapters are on the enemies of science and their criticisms, starting with the Sophists and ending up with what Franklin’s teacher (and mine) David Stove famously called Four Modern Irrationalists (Popper, Kuhn, Lakatos and Feyerabend) who would generally not regard themselves as enemies of science. But, if as Franklin does, you hold that deductivism is false (i.e. you are not a sceptic about induction), and science operates because it is false, then enemies of science they are. Franklin’s writing is full of dry wit that makes it a delight to read. This discussion of the popularity of Kuhn’s analysis gives a taste of its pleasures:
Kuhn’s success is also an instance of the enduring appeal of theomachy, a mode of explanation that worked so brilliantly for Freud and Marx, and, long before, for Homer. What was previously thought to be a continuous and uninteresting succession of random events is discovered to be a conflict of a finite number of hidden gods (or classes, complexes, paradigms, and so on, as the case may be), which manipulate the flux of events to their own advantage, but whose machinations may be exposed by the elect to whom the interpretative key has been given.
Further reasons for Kuhn’s success are not hard to find. He gave permission to anyone who wished to comment on science to ignore completely the large number of sciences that undeniably are progressive accumulations of established results—sciences such as ophthalmology, oceanography, operations research, and ornithology, to keep to just one letter of the alphabet. That certainly saved a lot of effort. Kuhn’s theory had a special appeal to social scientists as well. Political scientists, sociologists, and anthropologists recognized Kuhn’s picture of disciplines relegating the accumulation of evidence to the background while bringing fights about theory to the fore; they were delighted to hear that what they had until then considered an embarrassment was the way things were done in the most respectable sciences. Kuhn even offered something to massage the egos of natural scientists themselves. It might seem at first glance that his claim that most scientists were drones was an insult, but there was good reason why it was met with the same equanimity one notices in fundamentalist religious circle at the news that only 144,000 were saved. The damned may be a majority, but of course they are other people; every scientist had the opportunity to cast himself as a revolutionary hero of a new paradigm, shamefully ill-used by the establishment (Pp34-5).

The next chapter is about postmodernist enemies of science. The Sokal Hoax naturally features prominently, as it ought to (it is both funny and revealing), but just as part of a deadly dissection of postmodernism.

Having cleared the undergrowth of intellectual weeds, and laid out the key intellectual tools, Franklin then takes us through a chapter-by-chapter journey through science.
Starting with the ‘furniture’, the concepts science uses. Franklin cheerfully starts with ‘the cat sat on the mat’ and moves on and up so that, a little over a dozen pages later, we are treated to one of the most lucid explanations of quantum mechanics one is likely to read (Pp66-7). Franklin notes that:
While science is largely in agreement with common sense and our grammatical categories when it comes to the basic furniture of the world, it has made, in the course of time, certain adjustments to the “manifest image” of the world. In certain, though limited, respects, “folk” concepts have turned out to be inadequate (p.68)
He takes us through these adjustments, being careful not to overstate the matter. Franklin comes across as having a certain basic respect for the mass of his fellow humans that sharpens his wit when considering those who give themselves cognitive airs and graces. (No wonder he finds the history of religion such a useful source of historical analogies.) He nominates energy as being where science has the most divergence from folk concepts (Pp70-1). Then there are the genuinely new concepts science has needed (such as acceleration), most of which have come from mathematics (p.72).

Considering the physical sciences, Franklin divides them into fundamental physics and all the others. As the science of the basics of material things, he concentrates on physics: what is concerned with; how we gain knowledge in physics (including what thought experiments are); the general outline of current knowledge concluding with how dark matter is reasonable speculation, but not knowledge.

A sense of the depth and the limits of our knowledge is very vividly conveyed in his next chapter looking at biology and the cognitive sciences, including the travails of Artificial Intelligence. Then it is on to mathematics (which Franklin is a professor of):
Mathematics is the gold standard of knowledge. Mathematics is proved true, and is stays proved true. Proofs convey certainty and (if the proofs are not too long for our minds to cope with) they induce understanding of why mathematical truths must be true (p.107).
This is a robust view that, of course, has come under attack. But Franklin is having none of it. Mathematics is as much about the world as biology:
The subject matter of mathematics is structure, or pattern (p.109)
There follows a very enlightening discussion of what mathematics is about and how proofs work.

The following chapter examines the enemies of mathematics, primarily Lakatos, and how their criticisms either misconstrue mathematics or confuse application with mathematics itself—that our universe may not be entirely Euclidean does not make Euclidean geometry false as mathematics, just (possibly) not perfectly accurately applicable to the world. What is true of a formal structure as a formal structure is a different question from whether that formal structure applies in the world. As Franklin notes:
The writings of the irrationalists about mathematics have had absolutely no impact on mathematicians (p143).
With Lakatos himself being re-interpreted by mathematicians so his ideas have become harmless. Mathematics education has suffered some inroads, however.

The last 70 years or so has seen the emergence of a series of new formal sciences variously labelled ‘computer science’, ‘control theory’, ‘signal processing’ … They have the further charm that:
They supply a number of concepts, like “feedback”, that permit in-principle explanatory talk about complex phenomena without demanding too much attention to detail (Pp146-7).
One can see the attraction. Discussion of verification in programming leads to a nice discussion of the difference between necessity and certainty (p.156).

Chapter 10 takes us through probabilities and risks, a necessary but easily confusing area of reasoning. Franklin both sets out the logic involved and the connection to acting in the world.

Chapter 11 wrestles with the question of whether social sciences are sciences. The examples of Adolphe Quetelet’s pioneering 1835 work on the predictability of the level of suicide and modern analysis of traffic dynamics (both very nicely explained) are used to establish that the scientific method applies to the social world. Which is a different matter from claiming that the current social sciences “have made it”:
So the answer to the question, “Are the social sciences sciences?” is “Yes and no”. Their logic of confirmation of theories is the same Bayesian reasoning as is found in science, and theories in science can be established beyond reasonable doubt. Statistical methods can predict short-term trends from data just as well as in the harder sciences. But finding the appropriate concepts and the causal connections between them is much more difficult, while imaginative insight into the minds of individual actors gives an extra mode of understanding unavailable outside the sciences of the human (p.196).
One only has to contemplate the parlous state of macroeconomics (nicely discussed here), in contrast to the much stronger situation of microeconomics, to see what he is getting at.

Science is (as irrationalists and postmodernists make much of) done by actual people in specific institutions and processes. Franklin’s chapter on “actually existing science” shows a nice appreciation of the realities of “doing science”, particularly his discussion of the necessary, but inevitably flawed process of refereeing papers (Pp203ff). But processes of inferring from evidence to conclusions are not invalidated by being done by fallible humans: fallibility does not entail irretrievable failure. That people sometimes fail or do things badly does not mean that they never succeed.

In his penultimate chapter, Franklin tackles the complexity barriers to knowledge. Not all complexity is a problem—modular (able to be dealt with in bits), averaged and turbulent complexities can be dealt with. But if there are a lot of unknowns in a complex system, so that both the evidence and the logical reasoning involved are complex, that is another matter. Franklin uses the cases of evolution and global warming as examples, concluding that:
In both evolution and climate change, the majority view of the scientific experts is well ahead. In neither case is there any known coherent alternative. But the complexities of the evidence are such that a higher standard of politeness in skeptics who raise serious problems would be well-advised (p.235).
Events in the realm of climate science since the book was published have not made this conclusion less true.

The final chapter considers limits to science. As he points out, this is not a scientific but a philosophical question:
… science itself does not make any findings about its limits. (p.237).
Franklin then considers why consciousness may, and ethics will, remain (at least in part) beyond the ambit of science. He has a nice response to D. M. Armstrong’s materialism about the mind:
An awareness of one’s own sensation, as sensation, is basic. And not only basic to philosophy, but to science, which depends entirely for its ultimate inputs on observations, that is, human sensations. It would not be easy, logically, for science to start with the sensations, infer the existence of a physical world, and then cut off the branch it was sitting on by deleting the sensations (p.240)
More generally, it would not be sensible to be so impressed by the achievements of science that one rendered what scientists do as people engaged in the activity of science otiose, as eliminative materialism appears to do.

Franklin is happy with the notion of a science of consciousness, it is just that:
None of that serious work on consciousness and physics has any tendency to support materialist view of the mind (p.242).
His position on ethics is somewhat similar, though more so. Science has discovered interesting facts about ethical behaviour in people. But these facts do not generate ethical conclusions, since one cannot infer an ‘ought’ from an ‘is’.

As to what the universe must be like to include beings of irreducible moral worth:
That is an extremely difficult question. Science has taken us to the brink of another kind of knowledge, by opening up to us a vast range of natural facts but also exposing us to the limits of its own ways of knowing. We cannot believe that what science knows is all there is (p.251).
Not merely because science has large areas of unknowns even in its realm of concern, but because the realm of what can be known may well be larger than what can be known by science.

Franklin is thus an enthusiast for science, but not for scientism. What Science Knows is a great source for understanding the soaring achievements of science: that they are achievements, and how they have been achieved. In particular, it is an excellent starting point for the basic concepts of science in general and various sciences in particular.

No comments:

Post a Comment