(This post is the second in a two-part series on the relationship between thought experiments and empirical experiments. The first post can be found here.)

I am going to knock out all of the design theory up front so bear with me for one paragraph and then we can get back to the fun stuff. According to Roger Kirk, who wrote one of the most widely used textbooks for research design in the behavioral sciences, “Experiments are characterized by the: (1) manipulation of one or more independent variables; (2) use of controls such as randomly assigning participants or experimental units to one or more independent variables; and (3) careful observation or measurement of one or more dependent variables. The first and second characteristics—manipulation of an independent variable and the use of controls such as randomization—distinguish experiments from other research strategies” (Kirk 1968). A good experiment isolates the variables being studied to remove outside influence and allow researchers to make strong, causal claims about relationships between variables.

Now that the basic theory is out of the way, lets see how it applies to empirical experiments. Good empirical experiments start with at least one scientific hypothesis (and usually one or more statistical hypotheses) that is falsifiable and offers one or more predictions that can be experimentally tested. Because good empirical experiments isolate from and control for outside influences, they are consistently reproducible. Good experiments allow scientists to assess the accuracy of the predictions they made based on their hypotheses and accordingly determine whether the hypotheses themselves are supported or rejected by the data.

Thought experiments follow a similar model. Like empirical experiments, thought experiments isolate particular variables in order to study the relationship between them. This generally takes the form of creating a hypothetical situation in which all of the parameters of the experiment are set by the philosopher in order to limit consideration to the specific decisions or phenomena being studied. The purpose of this is generally to understand how something would be in an ideal case, free from the confines of circumstance, so that decisions themselves can be isolated from their circumstances and general statement of principle can be made. These statements of principle, like predictions from hypotheses, are then tested with further experiments to find circumstances in which they do not hold or to show that they do hold in cases where they were previously believed not to. To see an example of this, check out my series of posts on the Trolley Problem/Surgeon Problem (Part 1, Part 2, Part 3).

These two types of experiment compliment each other well because the things thought experiments are good for, making falsifiable predictions and assessing the implications of facts and principles on real-world circumstances, are the same things empirical experiments do poorly. Because empirical researchers face a tradeoff between internal and external validity, the most internally conclusive experiments are the ones that require the most conceptualization to put in context and evaluate. Similarly, empirical research provides a way for philosophers to determine if the assumptions on which they construct their theories are consistent with what we “know” about how the world works and human behavior. Neither type is useful without the other, and everyone would do well to become familiar with both.

It is not a coincidence that many of the greatest thinkers in the history of humanity were both scientists and philosophers. Many of humanity’s greatest achievement have come from people who were willing and able to harness the analytical and predictive power of both disciplines and use them in tandem. It is vitally important that scientists have a working knowledge of philosophy and analytical thinking, and that philosophers appreciate empirical research and the sciences. In a society that tends to sort people by academic discipline and force them to specialize early, there is perhaps no rarer but more vital skill than that of thinking across disciplines and recognizing that knowledge is most powerful when it is most complete.

——

This brings us to the conclusion of my blog on thought experiments. Thank you to everyone who has read, commented, and given me feedback; I sincerely hope that you have enjoyed reading my posts as much as I have enjoyed writing them. When I started this blog, I had to chose something to put under the title at the top of the page. I settled on the quasi-official motto of western philosophy: “the unexamined life is not worth living.” I believe very strongly that thought experiments are an excellent tool for this kind of critical examination, and I am glad to have been able to share this tool with you over the course of the last several weeks.

If you are a natural-science person (or a psychology person), you will most likely have to take a class in experimental design while you are here at Penn State. Experiments are the cornerstone of the scientific method, so it makes sense to devote significant attention to understanding how they work, what makes some better than others, and what their limitations are. For the last several weeks, I have talked about thought experiments and the role they play in philosophy and science (If you haven’t already and you are interested in thought experiments in science, check out my posts on Relativity and The Importance of Thought Experiments to Modern Physics). For these last two posts, I want to shift gears and talk about the relationship between thought experiments and empirical experiments.

Over the course of writing this passion blog, one of my major objectives has been to dismiss the notion that thought experiments are something people use to speculate about the world from their armchairs without actually observing anything, and that empirical experiments are “better” for learning about “real” things. Empirical experiments and thought experiments are both important tools people can use to understand our world, and they each have their own distinct purposes. Especially in fields like Physics and Neurology, thought experiments can help scientists both to determine what kind of empirical experiments to perform and how to make sense of the results of those experiments. There are certainly places where empirical experiments are “better” than thought experiments, but there are also places, even in the natural sciences, where thought experiments are “better” (one of which is in determining what makes a given method of examination “better”, but this is a subject that requires a more thorough treatment than I am prepared to offer here).

In the 21^st century, one of the places where empirical and thought experiments have both come into conflict and complimented each other tremendously is in neurology. If you are interested in neurology, philosophy of science, or intellectually stimulating conversation in general, I highly recommend that you take 13 minutes and 22 seconds to listen to a Philosophy Bites podcast interview with Barry Smith on the interaction between these two tool sets for exploring the mind (the interview can be found at http://philosophybites.com/2008/09/barry-smith-on.html) and, if you enjoy that, as many of their neuroscience focused interviews as you can handle (which can be found here http://philosophybites.com/neuroscience/). Essentially, Dr. Smith lays out the role that each method has in advancing both neuroscience and the various related philosophical disciplines. To him, philosophers are too skeptical of evidence from empirical experiments that contradicts their expectations, which he believes is hampering the discipline’s progress. On the other hand, he recognizes that philosophers were the first people to raise the kinds of questions neuroscience explores, that their thought experiments provided an important starting place for the empirical research, and that, in his own words, when it comes to “pathologies and neurological breakdown, help is needed by the neurologists and biologists from the philosophers to help to explain how to characterize these experiences, to understand what they’re like, and to contrast them with normal experience.”

In my next post, I am going to discuss what makes a good empirical experiment, what makes a good thought experiment, and compare and contrast the two.