In Pursuit of Truth

Blog post by Jeffrey Thayne on October 14, 2008

Jeffrey Thayne

In our age, science has attained a respected status that has largely gone unchallenged. I suspect there are many good reasons for this. Clearly, we have many conveniences today that we commonly attribute to the scientific endeavor. This is certainly a good thing. Brent Slife and Richard Williams believe that “it is important to understand how science has come to command such respect.”1 They explain,

The obvious answer is that science is persuasive. We are persuaded through scientific experiments that their results are trustworthy and accurate. The next question, of course, is why it is that scientific experiments are persuasive. When we look at the issue carefully, it seems clear that scientific experiments are persuasive because they follow the form and structure of a logical argument. The persuasive power of science, in this sense, is simply the persuasive power of logic.1

The relationship between logic and scientific discovery, Slife and Williams claim, “can be traced to Aristotle. It seemed to him … that following logical procedures is our best guarantee against unwarranted conclusions and errors in thinking.”1 A scientific experiment, they explain,

is essentially a logical argument. An experiment is set up much like a logical argument of the form “If _____ , then _____ .” Essentially, the researcher says, “IfI measure these certain variables in this way, and if I control for those other variables in this way, and if I make observations under these specified conditions, then I will observe that particular result.”1

The question we may ask is, “On the basis of observation, can this logical argument make any indubitable or genuine truth claims?” Let’s consider. Imagine we have a theory, and on the basis of this theory, we predict that we will observe a particular phenomenon under certain conditions. We do, indeed, observe the phenomenon as we had predicted. Can we then conclude that our theory is true? Let’s lay this out in the form of a logical argument:

If theory x is true, we will observe y.
We observe y.
Therefore, theory x is true.

This experiment is set up to verify a hypothesis on the basis of observation. Schooled students of logic will see an elementary fallacy in this argument, commonly known as affirming the consequent. The conclusion does not follow from the premises. The argument is invalid. Slife and Williams use a counterexample to show easily how this argument is fallacious:

If Socrates is a man, then he is mortal.
Socrates is mortal.
Therefore he is a man

“This is obviously not a good argument,” Slife and Williams explain, “because showing that Socrates is mortal does not necessarily show he is a man. He could be a dog or a bush or any other mortal thing.”1

Certainly, this does not mean our hypothesis is not true. We have certainly marshalled evidence in support of our theory. We just can’t claim to have proven our hypothesis. Can any scientific experiment make this claim? Slife and Williams continue:

That experimentation cannot prove anything true has been known for a long time. … The very way empirical studies are set up can always and only demonstrate the consequent. Thus, it is impossible—by the rules of logic implicit in the experiment itself—to prove any hypothesis true.1

It is not my intention here to “disprove” science, or to dismiss science in any way. I believe that well-performed scientific studies ought to be persuasive to us. My only goal here is to change the way we think about science. We can’t think of science as an indubitable, royal road to truth. In my next post, I will explain how a philosopher of science named Karl Popper proposed an alternative to the philosophy of verificationism, and I will also discuss some of the challenges of Popper’s point of view.

1. Brent Slife and Richard Williams, “Science and Human Behavior”, in What’s Behind the Research? Discovering Hidden Assumptions in the Behavioral Sciences (Thousand Oaks, CA: SAGE Publications, 1995), pp. 167-204.

comments so far
  1. I was talking with Dr. Richard Williams about this issue of uncertainty in science. His skepticism regarding the ability of science to provide mankind with necessary, certain, and universal knowledge awoke me from my dogmatic slumber. So I challenged his skepticism and told him that I could think up an experiment that would allow me to ascertain with certainty the status of an empirical hypothesis.

    I said: if I hypothesize that all swans are white, and then I find at least one black swan, then I have absolutely disconfirmed the hypothesis that all swans are white. He replied that I may be wrong because perhaps I have not performed a *crucial* experiment. I asked him what was meant by a crucial experiment. He told me that it was an experiment where all possible variables are controlled.

    While I still hold some reservations about this response, he is absolutely right about the importance of crucial experiments.

    People like to put evolution on par with relativity because, they claim, both were discovered through rigorous inductive processes, like an inference to the best explanation. This is true, but what they fail to acknowledge is that relativity has been subjected to crucial experiments in the hypothetic-deductive tradition and survived, but macroevolution has not.

    I look forward to reading your portrayal of Popper’s falsification doctrine.

  2. Thanks for your comment, Dave!

    I think we should also distinguish between theories and predictions. It is certainly true that experiments can confirm or disconfirm predictions. For example, “I predict that I will observe x.” I observe x. My prediction was true. However, what we can’t do is prove or confirm any theories about why we observe x. It is theory y in “If theory y is true, I will observe x” that is unprovable.

    I don’t know if this helps at all. I will be talking more about the crucial test in my next post on the subject.

  3. Dave, your swan example seems to me to disprove the theory that “All swans are white”; I’m not sure why Dr. Williams didn’t agree about disproving. But I think the more important point is that a theory that begins with “All …” can’t be proven. I might be missing something, though.

  4. Also, Nathan, as I tried to distinguish before, I’m not sure that “all swans are white” could even be classified as a theory; rather, it is more of a prediction of the sort, “we will never observe a non-white swan.” Whatever evolutionary or genetic theory you may have that leads to this prediction is not validated or disconfirmed by the observations in question.

  5. I recently read a quote in George Kelly’s book, A Theory of Personality, in which he expresses a similar opinion to the one in my post:

    The function of a scientific theory is to provide a basis for making precise predictions. These predictions are formulated in terms of hypotheses and are then subjected to test. The outcome of the test may be essentially that which was predicted. …

    The substantiation of hypotheses is really not quite as simple as this. The catch is in the design of the experiment. If the experiment is so designed that other obvious hypotheses would have expressed the same prediction, the question arises as to which hypothesis was verified. As a matter of fact, in scientific research, one never finds the ultimate proof of a given hypothesis. About the time he thinks he has such proof within his grasp, another scientists comes along with another hypothesis that provides just as plausible an explanation of the experimental results.

    … The relevant point for the purposes of this discussion is that even the precise hypotheses which one derives from a good scientific theory are never substantiated with absolute finality, no matter how many experiments are performed.

Leave a Reply