Philosophy and History of Science and Technology

Course Home

Thursday, 09-14-17: Pseudoscience III



We began today by considering revisiting Popper and Kuhn's radically differing accounts of demarcation criteria before turning Lakatos' notion of a scientific theory neither as interpreted formal language (Popper) nor as conceptual scheme or paradigm (Kuhn) but as a research program.

For more on Lakatos' notion of a research program, see my notes above. Suffice it to say that a research program is rarely abandoned altogether as Popper's notion of a scientific theory under the threat of falsification would have it. Rather, a research program is adjusted and amended as need be. The important feature of a research program which distintuishes science from pseudoscience for Lakatos is the program's progressivity in terms of its capacity to predict novel facts.

What counts as a novel fact? Lakatos does not have a general account to help us here. He merely points to a couple of examples of what he has in mind:

  • Halley's prediction under Newton's research program of the time and place of the return of Halley's comet.
  • Eddington's expedition to verify Einstein's prediction of starlight being bent by the Sun's gravitation.

So it seems a progressive research program has the capacity to predict what would otherwise have come as surprises for us, and to do so with considerable precision. That is not, unfortunately, much to go on.

Next we took up Thagard's account of demarcation criteria. While he starts in very much the same place as Lakatos, he adds in a good deal more to make for necessary and sufficient conditions on pseudoscience--conditions he later recants.

Thagard takes up Astrology because it has been unfairly dismissed by Popper, Kuhn, and Lakatos as a pseudoscience. Let us be clear that Thagard is not endorsing Astrology as a science. Rather, he is condemning the demarcation criteria Popper, Kuhn, and Lakatos propose. Thus,

  1. Since some astrological claims are testable, it cannot be the case that falsification alone keeps Astrology out of the science club (as it were).
  2. Similarly, just as we never reject any scientific theory whole cloth because a prediction turns out false, it would not have been fair to hold Astrology to such a strict standard.
  3. Astrology nonetheless presents a number of unsolved problems and some technical resources to solve them.
  4. Nor can we say that Astrology is a pseudoscience simply because it historically made supernatural assumptions; alchemy did as well, yet alchemy eventually developed into one of the more rigorous and sophisticated sciences.

Following on Popper, Kuhn, and Lakatos, we can locate Thagard by pinpointing his subsequent points of convergence and divergence.

  • Thagard agrees with Kuhn and Lakatos' criticism of Popper that science does not in fact progress by falsification and theory rejection. The process is much more complicated and, for want of a better word, organic. Good but not perfect theories can be made better by careful ammendment and adjustment so as to save the phenomena.
  • Against Kuhn, however, Thagard argues that Astrology enjoys a long puzzle-solving tradition.
  • With Lakatos and against Kuhn, Thagard argues that science must be progressive in its explanatory and predictive power.
  • Yet against Lakatos, Thagard argues that a lack of progress in predicting novel facts may occur (and has occurred) in sciences lacking a serious competitor.

We proceeded to examine the demarcation criteria Thagard draws from these considerations and discussed one his more surprising conclusions:

It was unreasonable for critics to reject Astrology as a pseudoscience prior to the development of Psychology

To understand why his demarcation criteria have this implication, please review my notes on Thagard. In particular, the first of Thagard's necessary and sufficient conditions on pseudoscience require that the putative science be less progressive for long periods relative to its competitors. It follows that a theory which is the sole player on the field, if you will, or the only game in town is rejected by Thagard's necessary and sufficient conditions on pseudoscience. It is by default a science. To follow my really terrible metaphors out to their gory conclusion, it wins the game by forfeit: No other teams have shown up to challenge it, so it gets a free pass into the science club.

It is also important to appreciate that, for Thagard, a theory being scientific (or pseudoscientific, as the case may be) is not an absolute property of the theory. It is a relative property of the theory--relative, that is, to the broader context of scientific endeavor. Note that this entails that some whacko ("whacko" here being a philosophical term-of-art) theories like pyramidology or ufology cannot be dismissed as pseudoscientific until scientific practice generates progressive alternatives.

To be sure, Thagard was taken to task for these counterintuitive implications. Although being counterintuitive is not simpliciter a reason to reject a philosophical position, so many counterexamples and counterarguments were shown to follow from these implications that Thagard eventually gave up his original position in favor of a more nuanced position (as outlined in the notes.)

At this point in our discussion we turned to illustrating why the debate over science-vs-pseudoscience is not a debate relegated to the erudite corners of academic philosophy. Rather, the many court battles over whether Intelligent Design ought to be taught as a competitor to evolutionary theory (typically under the slogan, "teach the controversy!") brings out the practical importance of understanding what legitimate science requires.

For example, acting as a friend of the court in the Arkansas Creation Science case, Ruse proposed (and Judge Overton followed his proposal in his opinion) a set of five necessary conditions on a theory T being a science:

If T is a science, then (Judge Overton's list)

  1. T is "guided by natural law."
  2. T is explanatory by reference to natural law.
  3. T is confirmable using empirical methods.
  4. T is tentative.
  5. T is falsifiable.

It follows that T is a pseudoscience (viz., not a science) if one or more of these conditions are not met.

As we shall see next time, however, there are good reasons to oppose most, if not all, of these conditions.