Cry of The Faithful Reductionist
I recently read an essay by John Horgan, author of “The End of Science”, in the September 18, 1999 edition of New Scientist about his latest book, “The Undiscovered Mind”. He proposes that if the problem of understanding the human mind might lie outside the limits of science, then it may never be solved by the traditional reductionist methods.
When confronted with such statements, my first reaction is to probe the source for an explanation of what he or she means by “solve” and “understanding”. Horgan passes this hurdle effortlessly by citing many sources from neuroscience, and sets the criteria at being able to predict and explain human behavior and diversity (in particular, mental illnesses). However, his outlook is very bleak. He entertains the possibility that there may be no explanation of these phenomena that we can find in the scientific realm, and advises us to pursue a more literary approach to the problem of the mind. I will explain below why I think that a training in the natural sciences is incompatible with the acceptance of such a view.
Let us take a pedagogical example in trying to understand the way science works. Imagine a very large smooth table with balls rolling around. If we observe the way the balls move and interact with each other, we can discern patterns of their motions. We are conducting an experiment. Soon, just as Newton did, we can formulate a model which enables us to predict the motions of the balls in advance. As we all know, this model will incorporate concepts like conservation of energy and momentum, and a few other things, depending on how complicated our setup is.
Like every model in science, this one too has its limitations. By a stretch of the imagination, let us suppose that we can test the limitations simply. We find out that our model is really very good, unless the balls are moving too fast, or they are too small, or they are too sticky. In all these limits, we can see our model fail, and thus we can establish its range of validity.
Now suppose that we take our mathematical model as applying to, say, one hundred balls and show it to our mathematician friend. He will stare at our gigantic system of equations in horror, and, as a first reaction, will ask us for some evidence that the equations in fact have any solution at all. We need not worry. We know they do, simply because if we let the balls go, they do something. It is impossible that they do not. This feature of our model is very special, and it differentiates it from a random jumble of equations which might not have a solution. There cannot be some mysterious property of nature which tricks us and in fact leads to no solution of elementary mechanics in the case of one hundred balls.
The power of reductionism, the method of explaining phenomena in terms of ever-simpler and interacting objects, lies in its strict relationship with experiment. Our model of mechanics operates with the concepts of speed and position. Because these are measurable quantities, it is impossible that the balls could not be described in terms of these quantities. It is impossible for a ball not to have a position or speed. Again, we have to be careful that we stay inside the stated (indeed measured) limits of our model. It is important to understand that quantum mechanics and relativity do not replace Newtonian mechanics in the sense that the latter would not be true anymore. Rather, quantum mechanics extends our knowledge beyond the limits of the original theory.
I believe this argument can be extended to any system that is studied by natural science. Let us take the brain for example. If we believe that the constituents of the the brain are neurons which interact chemically and electrically, then it is impossible that the brain cannot be described using these objects. Now some people believe that neurons and their classical interactions are not enough to describe the brain. In particular, Roger Penrose speculated that quantum mechanics may be involved in the workings of the mind. This is a very legitimate question. But it has to be investigated on the level of the components of the brain. Experiments have to be carried out which establish the existence or non-existence of non-classical interactions. Before this is done, no one has the right to claim that the brain cannot be explained using the currently available components. Certainly the lack of an explanation today, which is what John Horgan is complaining about, is not a valid argument.