This review originally was posted to the Mathematical Association of America's Book Review Site.
Many
years ago my wife and I hosted a married couple we didn’t know for dinner. He
was a law school student but she was a local Ph.D. student studying the philosophy
of science. Having recently obtained my doctorate in engineering I commented
that there was no such thing as a philosophy to science. Rather, I naively
said, science flowed from observations of facts, logical and mathematical
arguments, or simply from one’s own thoughts of how the world behaved. Was I
wrong.
This
woman explained that there was, indeed, a philosophy to science. She told me
how scientists held particular views on what to study and how to study it and
that the scientific method was subjected to biases and underlining prejudices
of the scientist himself. I was
flabbergasted, surprised to hear such thoughts because I believed science was,
shall we say, independent of the scientist.
Along
comes this splendid book, Philosophy of
Science for Scientists by Lars-Goran Johansson: a lovely text book for
undergraduates. The book is a highly readable introduction to how one can view
the practice of science. I wish I had read such a book while in school and I
wish my professors had spent some time, even just a lecture or two, explaining
how students can see science from a view point a bit removed from our studies.
It would have expanded everyone’s view of science and opened us to a better way
to see what we were doing.
The
book begins with a look how science started in ancient Greece where Thales
believed water was “responsible for the change in all things.” A reasonable
position, writes Johansson, because all living things require water. This approach
is consistent with a modern view of science that is based on observation. We
observe the world, form hypotheses, and conduct tests and experiments to verify
our hypotheses. The ancients did not have access to the complicated
experimental apparatuses we have (no super colliders, few controlled experiments)
but they were excellent observers of their world.
The
text shows how knowledge can grow with propositions to justify a true
belief. For example, ancient astronomers
observed the planets and formed an argument of planets revolving around our Earth.
We have better knowledge today but at that time, their geocentric model fit
their observations quite well. While we now subscribe to a heliocentric solar
system, ancient tables depicting the orbit of the moon about earth are very accurate
up to this day.
There
is a lovely treatment of hypothesis testing with a slightly contrived example
of a medical test for an AIDS medicine. The authors state the null hypothesis
to test, provide example data, and discuss how one can either accept or reject
the null hypothesis. It’s a simple example, but illustrative. (In engineering I
never learned of hypothesis testing in school but learned of it in operations
research. This would have been good to hear first in a classroom.)
Here’s
a great example of observation bias of the observer. Robert Rosenthal, a
psychologist, asked students to experiment with mice to see how well they performed
in solving a maze to find a food bowl. The students measured the time it took
each mouse to get to the bowl. They were told some of the mice were “gifted”
and would solve the maze quicker than the other mice. The students found the
gifted the mice did, indeed, find the food quicker than the other mice. The
students also found the gifted mice refused to move only 11% of the time but
the other, non-gifted, mice refused to move 20% of the time. Of course, there
were no differences between the mice, only the bias of the students from the
original instructions.
The
book touches on what I think is a crucial topic: paradigm shifts. A short discussion from Thomas Kuhn’s The Structure of Scientific Revolutions,
now over 50-years old, shows how Kuhn viewed changes in scientific thought. For
example, there is Newton’s analysis of lunar motion as continual free fall.
Interestingly, earlier in the book we met Ptolemy’s geocentric view of the
solar system, and the inevitable use of epicycles to explain, say, the
retrograde motion of Mars. I believe it would have been beneficial for the author
to have gone from epicycles to the concept of encrustation within Complexity
theory to show a current view of the shift in paradigms. (This is one of the
few shortcomings I found.)
The
next part of the book explores Causes, Explanations, Laws and Models. Cause is
wonderfully illustrated by an experiment on the wing length of fruit flies with
a genetic defect. Fruit flies with this defect will have shortened wings if the
temperature is around 20-degrees Celsius when the flies are maturing. If, however,
the temperature is 32-degrees Celsius the wings grow to normal length. Does the
genetic defect cause the wing length shortening or does the temperature do so?
The answer depends on how we compare populations of fruit flies: If we compare
at the same temperature then the defect is caused by the gene. If we compare
populations at different temperatures then we would say temperature is the
cause of the defect. It’s an interesting discussion point and worthwhile for
students to think about.
The
author explores Explanations with an example of storks and birth rates; the two
declined remarkably between 1966 and 1980. Figure 1 shows the plot from the
text and notice the whimsy of the drawing. The correlation between birth rates
and the presence of storks was so consistent it was calculated that the
probability of coincidence was 0.1. The
author suggests the correlation was due to industrialization Again, this is an
interesting discussion point for students.
The
author goes on with causes and effects, and Bayesian probabilities. The author
presents a good argument for why explanations need to show the reason for a
phenomena and not just some relationship. For example, ancient observers could
predict a new moon accurately but they did not know why their predictions were
true. Along comes Newton with his theory of gravity to explain orbital motion as
a mathematical law, not just a table of observations. New moons are predicted with
reason, not just tabular notions.
The
author discusses other topics and the reader will find all of them of
interest. In
short, this is an excellent introduction to understanding science in a general sense.
Students and practitioners will find it worthwhile to read and discuss. I wish
I had read such a text long ago but I am glad to have benefited from it even
now.
Figure 1: Correlation of births and stork sitings. |