Tuesday, January 07, 2014

science, limited (Hilary Putnam)

Ch 8 The impact of science on modern conceptions of rationality. In Reason, Truth and History (1981) by Hilary Putnam

Putnam is suggesting that it is better to think rationally about both science and our values, without regarding one as more important than the other, rather than elevating science thinking above values thinking.

Putnam is trying to create a space in between two extreme, opposite views that tend to dominate our cultural discourse. The extreme views are:

Scientism: that the scientific approach is the only productive approach to take towards all of our issues, including social issues. Science, unlike all other viewpoints, provides us with the opportunity to understand the world objectively.

Relativism: Absolute scientific truth is an illusion. Truth, if you want to call it that, resides in an individual's interpretation of the world, which depends of their current knowledge and cultural context. There are many ways to interpret the world, we should respect all of them and not be so arrogant as to think anyone knows the best way.

This dichotomy (scientism / relativism) can be expressed with different labels, eg., objective / subjective; facts / values; materialist / idealist; reductionist / holist; monist / pluralist. In some debate these become swear words, that each side throws at the other. Putnam wants to collapse these dichotomies, which have become rigid ideological labels, into distinctions. We can talk rationally about both science and values without presuming that science represents some higher form of truth.

Our concept of progress tends to be scientific progress. The Industrial Revolution and the Computer Revolution represents progress. You can convincingly argue that Newton knew more science than Aristotle. But it is harder to argue that Shakespeare was a better poet than Homer. Science has progressed dramatically whereas literature and the arts have not.

History, as interpreted by the positivist August Comte (1798-1857), shows us that science is a heroic success story. We have moved from primitive myths, to high religion, to metaphysical theories (Plato, Kant), to positive science. This represents intellectual success as well as the obvious material and technological success.

From the scientism viewpoint, value judgements are viewed as suspect because they can't be verified by the methods of science. We can't obtain universal or majority agreement on ethical questions about abortion or homosexuality. Therefore science must be superior because the correctness of scientific theories can be demonstrated publicly.

Of course it is not really true that there is agreement on scientific theories. But most of us agree that scientific theories have testable consequences. Scientific language refers to publicly verifiable observations and not subjective private introspection. If we perform such and such actions then we will obtain such and such observable results. Much of the maths and science might be too difficult for the public to understand. But often the experts seem to be in agreement. And the public usually defers to the experts.

Not always. But current controversies such as the anthropogenic global warming debate or the reading wars debate are fought out with both sides claiming that science is on their side. Both sides think that science and not something else will deliver victory to their cause. When in doubt do more science.

Can science deliver us an objective view of the real world? Quantum physics interpreted as pop science is seen to be cool because it delivers us imaginative, mind boggling views of how the world really is that defy common sense. I have seen real scientists on TV arguing for parallel universes or that we are just holographic projections from the nearest black hole. Does anyone really believe this? I prefer common sense. If someone down the pub thinks they are just a holographic projection then they can buy their own drinks.

Instrumentalism is the idea that scientific theories are instruments to predict observations rather than attempts to describe the real but hidden structures of the world. Instrumentalism of itself is not a tenable or fully rounded concept of rationality. Sure, it is very valuable to know efficient means to attain certain ends. But it is also valuable, probably more important, to know what ends to choose. Scientific rationality applies to public means-ends connections. If this defines science then it also confines science. Why would anyone be satisfied with such a narrow conception of rationality?

Complex, real life, judgements require a high level of rationality but cannot be proved scientifically. It's strange that the fact that some things are impossible to prove (our value judgements) should become an argument for irrationality of belief about those things.

Nevertheless, instrumental success is appealing to the contemporary mind. Industrial society, both capitalist and socialist versions, have promoted their legitimacy on the basis of rising productivity and increasing standard of living. There is no doubt in my mind that this has some real merit. If a billion people or so live on a dollar a day then obviously we need to increase their standard of living. That is why I can't abandon my socialistic sentiment, since capitalism might increase standard of living but at the same time it also increases the gap between rich and poor. But, it is also true, that once their basic material needs are met, humans also look for more, a deeper concept of human flourishing or eudaimonia (Aristotle's term).

Empiricism is the belief that experience is the only source of real knowledge about the world. (I plan to discuss empiricism more fully later). One extreme version of empiricism is phenomenalism, that all we can talk about scientifically are sensations. JS Mill, for example, described physical objects as the "permanent possibility of sensation" (1865).

It follows from phenomenalism that all worthy facts are ultimately instrumental. If phenomenalism is correct then instrumentalism is elevated from the mundane and practical to high science. If you perform such and such actions then you will have such and such experiences. The only worthy knowledge is means-ends connections. This form does not fit "good", "bad" or other ethical judgements. "Good" people who take "good' actions have many divergent experiences and outcomes. Such ethical statements have no cognitive meaning, they are purely "emotive". This way of thinking (phenomenalism) was promoted by the Logical Positivists and Logical Empiricists [Ernst Mach (1838-1916), Otto Neurath (1882-1945), Rudolf Carnap (1891-1970 ), Hans Reichenbach (1891-1953)] and flourished from the mid 1920s to the mid 1930s.

Hence, phenomenalists could still claim to be curious about the "big questions" such as black holes, the Big Bang and evolution unlike the "vulgar" form of instrumentalism that only appealed to practical outcomes. For the phenomenalists it was just a fact of life that worthy knowledge was ultimately instrumental in form: If you perform such and such actions then you will have such and such experiences just happened to be the only sort of knowledge worthy of true science.

The logical empiricists were not worshippers of practical, narrow minded instrumentalism. Their healthy motivation was to eliminate obscurantism and metaphysics from intellectual discourse. But by drawing a sharp line between the factual (observational science) and the evaluational (talk about values) they ended up giving a distorted picture of the factual.

Putnam discusses how phenomenalism unravelled from within. One of its chief theoretician, Rudolf Carnap, tried but failed to show that statements of science are translatable one by one into statements about what experiences we will have if we perform certain actions.

Hence, since such science can't be attained (translation from science statements to certain experiences) then it is not justified to exclude value judgements (as emotive) on those same grounds.

It is a correct generalisation about the practice of scientists that their observation statements are couched in a certain type of language, a public physical thing language. But it is an error to turn this into an epistemological absolute. For example, if you are not allowed to talk about sensations because they are private then all introspection is ruled out.

Here is a danger which Putnam is articulating. Some people are in awe of the instrumental success of science which for them is free of the interminable debates we find in religion, ethics and metaphysics. In an age of various forms of quackery and snake oil salesmen (Tarot cards, Palmistry, homeopathy, astrology etc.) the best antidote is to stress the success of the scientific approach (achieved through rigorous controlled and / or double blind experiments). But there is another type of problem. In a culture hypnotised by the success of science a philosophy emerges (Scientism) which can't conceive of useful knowledge and reason outside of what we call science. What else could real knowledge be, except for science?

Modern versions of attempts at scientific objectivity have replaced the now rejected efforts of the Logical Empiricists. One of them is the Bayesian school. The problem here is that prior probabilities are subjective, or, the time taken for them to converge may be very long. You can't draw a sharp line between the actual beliefs of scientists and the scientific method.

Putnam also discusses Nelson Goodman's (1906-1998) critique of inductive projection. I'll leave that to another time since Nelson Goodman's work seems important enough to warrant more study in its own right (see references).

How then do we account for the success of science?

Scientists did develop a new set of methodological maxims between the 15th and 17th Centuries. Robert Boyle (1627-1691) played a major role here (see Leviathan and the Air Pump by Steven Shapin and Simon Schaffer). Before Boyle experiments were conceived as illustrations for doctrines believed on deductive and / or a priori grounds, not as evidence for and against theories. Boyle advocated several things which changed all of that:
  • he distinguished between actual experiments and thought experiments
  • all experiments should be completely described, including failed experiments
  • he wrote manuals for experimental procedures
This shift from a focus on a priori beliefs to testing theories by controlled experiments was a significant methodological shift. And it did lead to very successful science.

So Putnam agrees there is a successful scientific method but argues that it is not a rigorous set of formal rules. It also requires informal rationality or independent intelligence to be successful. It presupposes rationality and does not define rationality.

There are other limits to what science can achieve. It is not always possible to perform controlled experiments and sometimes only passive observation is possible, eg. for ethical reasons. Evaluating evidence for alternative theories is often an informal matter.

Karl Popper argued that science should proceed by putting forward highly falsifiable, risky theories, testing them, until only one survived. One problem with this is that it is not possible to test all strongly falsifiable theories, there are too many of them.

Furthermore, the falsifiability criterion would rule out one of science's favourite theories: Darwin's theory of evolution by natural selection. Darwin's theory does not imply definite predictions. We accept it because:
  • it provides a plausible explanation for an enormous amount of data
  • it has been fruitful in suggesting new theories
  • it provides links to other fields such as genetics and molecular biology
  • there is no satisfactory alternative theory
The method being used to evaluate Darwin's theory is inference to the best explanation even though the best explanation is not strongly falsifiable.

So, what remains of the scientific method? We can think about it either in the Boyle sense of strict experimental procedure or as a vague thing like:
"Make experiments and observations as carefully as you can and then make inferences to the best explanation and eliminate theories which can be falsified by crucial experiments"
Putnam argues that such a vague definition of the scientific method could be used for ethical induction just as readily as scientific induction.

This post is mainly a summary of one chapter of a book by Hilary Putnam, although I have thrown in a few examples and brief comments of my own on the way through. I'll write up how Putnam's thinking has impacted and changed my thinking in a later post.

Putnam's chapter also awakened or in some cases reawakened my interest in reading other books some of which I have already bought but haven't had time yet to read. Here they are.


Goodman, Nelson (1978) Ways of Worldmaking

Goodman, Nelson (1983) Fact, Fiction and Forecast (4th Edition)

Pielke, Roger A, jnr. (2007) The Honest Broker: Making Sense of Science in Policy and Politics

Sarewitz, Daniel (1996). Frontiers of Illusion: Science, Technology and the Politics of Progress. Temple University Press

Sarewitz, Daniel (2004). How science makes environmental controversies worse.

Shapin, Steven and Schaffer, Simon (1985) Leviathan and the Air Pump: Hobbes, Boyle, and the Experimental Life

No comments: