The Scandal of “Science”
Epistle 6. Posted on 2019-03-08.
The philosopher C. D. Broad (1887 – 1971) famously said that “induction is the glory of science and the scandal of philosophy.” This article briefly explores induction as the scandal of “science,” and its implications for counseling.
The Problem of Induction
Most introductions to the problem of induction refer to the work of David Hume (1711 – 1776), although I like to also consider Sextus Empiricus (c.160 – c.210) and Chrysippus of Soli (c.279 BCE – c.206 BCE).
There are two (or more) categories of inference: deduction and induction. I purposely disregard abduction as redundant, but that discussion belongs elsewhere. In what follows, analytic logic is used rather than mathematical logic.
The following is a deductive argument:
If it breathes, it lives.
But John breathes.
Therefore John lives.
The conclusion is necessarily true, it must be true, given that the premises are true, and given the form of the argument is correct. This is not to say that it is a guaranteed fact that John lives, but merely that John must live if he breathes, and if it is conceptually impossible that something can be both breathing and not living.
But consider the following argument:
If it moves, it lives.
But John moves.
Therefore John lives.
This is invalid as a deductive argument, because the premises and argument do not make it necessarily true that John lives. The problem is that it is possible that something can both move and not live, such as a rock passing through a window.
Most people would gladly use probability to work with an induction. Since it cannot be deduced that John lives because he moves, it would be popular today to induce that John probably lives, because he moves and is human. Although there are more moving and non-living things in the cosmos, there are probably more moving and living humans than moving and non-living humans.
There are several definitions of induction, such as proceeding from particular to universal claims, but it is easiest to simply regard an induction as any inference that was not deduced.
Induction is a problem because the inference cannot be regarded as valid, unless it is justified with another induction. Consequently, all induction is circular reasoning.
It is popular to prefer induction because deduction is more limiting, and because it is more attractive to move on to new information. A deduction, after all, does not contain new information. The conclusion that “John lives” is contained entirely within both the syntax and semantics of its argument. What deduction does allow is an inevident conclusion to be deduced from premises that are evident; in this way only is deduction capable of dazzling the seeker of new information.
Although deduction is more limiting, it is also the only valid form of reasoning available. Someone may estimate a complex statistical model that predicts a probability interval of investment prices in the near future for a financial security, because they seek the best estimate, given the information available. Nonetheless, the estimation of that probability interval involves circular reasoning. By referring to a deduction as reasonable, any induction is unreasonable by definition. To deduce the future price of that financial instrument, I would conclude the entry price is non-negative, because I can deduce that a negative entry price is conceptually impossible. Of course, this isn’t very helpful to the modeler who seeks the optimal entry price given the information. There are many possible avenues with deduction toward a better answer that remains reasonable, but it also requires a change in perspective.
Consider any journal article that uses a sample population to estimate an optimal dosage. There are many additional inductions, such as which sources of data are affordable, the decision when to stop collecting data, the representativeness of the sample population, the appropriateness of the statistical model selected, the fact that any regression model is misspecified when a relevant predictor is omitted (and how often are all relevant predictors both known and available). Each model has numerous assumptions. Some variables are assumed to be F-distributed but never tested — let alone reported — because such testing usually requires MCMC sampling.
But let us ignore the vast number of assumptions involving circular reasoning, let us even pretend for a moment that it is, or even could be, responsible reasoning. If the goal of this medical journal article is to consider the optimal dosage for future use, perhaps instead deduction could be used to report that under such and such circumstances that it is conceptually impossible for a given dose to kill any patient, or for another dose to cause any minimal change in female patients over age 67, or you get the idea. If it is unreasonable, given the information available, to make a certain statement, there may be other useful and related statements that can be made, and can be made responsibly, which is reasonably.
The Demarcation Problem
The demarcation problem refers to the distinction between science and pseudoscience. The areas of philosophy of science and epistemology are vast, and I am certainly leaving out a lot when I declare the distinction is, relevant to this article, mainly the difference between deduction and induction, between science and pseudoscience.
Karl Popper (1902 – 1994) criticized psychoanalysis as pseudoscience because its theories were not falsifiable. The academic field of psychology, as well as other fields, agreed with the criticism and adopted falsifiability as a criterion.
Falsifiability was based on modus tollens, which has the following form:
If p, q.
But not q.
Therefore not p.
Given theory (conjecture) p, data q should be falsifiable. If data q is falsified, theory p has also been falsified. By substitution:
If theory, implied data.
But not implied data.
Therefore not theory.
To return to an earlier example:
If it breathes, it lives.
But not John lives.
Therefore not John breathes.
Although the above may be regarded as practical rather than scientific, it is deduced that John does not breathe because John does not live, and because all living things are breathing things.
This brings us back to induction. Karl Popper was a deductivist. He regarded induction as having no place in science. Our recent history of science involves embracing both falsifiability and induction.
Through the decades, numerous scholars have regarded induction as resulting in pseudoscience. Nonetheless, it is the norm. It was embraced, and it is the norm, because its use has been successful. Everything from smartphones to quantum computing is due to induction. But so are countless deaths, mistakes, tragedies, wars, and so on. A medical study that reported an optimal dose given a statistical model and a sample population surely has resulted in a particular doctor recommending that dose to a particular patient for whom it resulted in death.
Implications for Counseling
The implications for counseling are no less troubling than for the rest of “science.”
There certainly exist people for whom psychiatric medicine is appropriate. But before accepted, it is also important to consider the vast network of circular reasoning involved in the development of the theories and the studies conducted. It is worth considering almost every “evidence-based” (i.e. data-dependent) claim you have heard as literally being unreasonable. This is not to say that it is true or false, but rather to consider what can be concluded reasonably.
Medicine aside, other social scientific theories – this includes the field of psychology – are also data-dependent, and therefore include long laundry lists of inductions. Even if a counselor abandons the circularity of pseudoscientific theories that may or may not apply, if deduction is not embraced solely, unreasonable conclusions may certainly still result from relying on expertise based on experience. Does someone else’s advice based only on their experience apply to you? The correct answer is: not necessarily.
Within philosophical counseling, Stoic Therapy specializes in using analytic logic. Deduction is embraced and induction is avoided. The pursuit of wisdom is, as I define it, the pursuit of reason. As with anything else, I can make no guarantees about being right or wrong, but I guarantee that I avidly pursue reason and am always wary of its contrary.
Please be wary of anything unreasonable (inductive); be wary of the scandal of “science.”
Vale (pronounced WAH-lay is Latin for “Farewell”),