Scientific Method in Practice

Rating:
Summary: Excellent discussion and overview
Review: This book provides an excellent introduction to the theory and practice of scientific method. Basic concepts and procedures are discussed ranging from the role of experiment, probability, and statistics to the nature and structure of hypothesis and theory. Appropos of that topic, I thought I'd discuss somemore concepts related to that here. Specifically, I recently came across a discussion of the "structure of science" and how that differs from pseudoscience and non-science activities. But just exactly what does "structure" mean in regard to science?

First of all, the word 'structure' is much too general and doesn't really mean much by itself. However, if we take it to mean the actual structure and nature of theory building in a particular discipline and also the scientific method actually employed, we are getting somewhere, as even the physical sciences differ more than one might think, even in regard to something as seemingly basic as scientific methodology, and we can contrast and compare them with regard to these two requirements very easily.

But in order to determine and compare their differences, we will need a couple of basic definitions. First, what is science and what is scientific method? A good starting position is that science is a combination of at least two things--the hypothetico-deductive method combined with some sort of empirical validation. Notice I say "empirical" and not "experimental," since this will come up later. A stronger definition would require some form of experimental validation, such as what occurs in experimental physics and experimental psychology. A good basic definition of the experimental method is that it consists of systematic variation under controlled conditions in order to determine causal relationships. So a scientific discipline could be said to be one that meets these two criteria. So far so good.

The problem with this definition, although it is almost universally accepted, is that even among the physical sciences, there are interesting exceptions. For example, take the science of planetary astronomy. No-one can start and stop the planets in their orbits to make experiments, and yet no-one doubts astronomy is a science, because it can make predictions about eclipses down to the second that are valid, and yet planetary astronomy seems to lack one important aspect of science. Hence, astronomy is an empirical rather than an experimental discipline. But it's still considered a science since it produces models that make very accurate and verifiable predictions about reality, and there is no better test of a science that its ability to make accurate predictions on the basis of observations. (Also, anybody who thinks planetary astronomy isn't a real science should try picking up a textbook on something known as Lagrangian Mechanics).

This implies that a better definition of science is the ability to make and validate predictions. This is not a bad idea, and is basically the definition of a theory. There is a lot of confusion about what a theory is in science, but if you keep that one criterion in mind, you can't go far wrong. Occasionally you will see it said that theories differ from hypotheses in being more complex and broader in scope, but although this might be true sometimes, this isn't a really a true distinction between hypothesis and theory. The only real difference between a theory and a hypothesis is that a theory has been tested more times and has more experimental validation, and so greater confidence is therefore placed in it.

But getting back to the our discussion of comparative differences in theory construction and method, we run into further problems when we come to the historical sciences, which include disciplines such as historical geology, many areas of psychology, and many areas of biology. But first, we need a definition of historical versus non-historical science. A good definition is that sciences like physics and chemistry are concerned with phenomena controlled by presumably universal natural systems that are non-historical, that is, independent of the time at which they operate. However, biological organisms and even the earth itself are historical entities whose characteristics may change through time, and whose workings may depend on historical laws that are not unchangeable and invariant, as in physics. This is the difference between historical and non-historical science.

So now let's consider the situation in historical geology. Instead of the deductive prediction of future events from known, present causes, it turns the scientifc method on its head to become inductive inference of ancient causes from their historical results. Hence, historical geology doesn't appear to be a science in the usual sense, and yet nobody doubts that it is indeed a real science, because, well, basically because it works, and no-one is too worried about its rather strange logic and methology.

I could go on further about this, but this is a pretty long review already and will do for a basic discussion of the topic of how several of the sciences differ in regard to their theoretical structures and scientific methods. There is another distinction one could make, say, between the mathematical sciences and statistical sciences because there we find differences in a science's ability to connect causes and effects reliably, with the difference being one of deterministic versus probabilistic causation, but as this is already a longer review than I intended, as I said, I will stop here. I hope you found my little discussion of the nature of theory and structure in science useful.