[An earlier version of this appeared on the Extropy list on April 17, 1999. The indented passages are from the Extropian FAQ that inspired these comments. There's also a paper connected to the FAQ. I've not had time to read it, but anyone seriously interested in this subject might want to read that paper and the works cited therein. To my recollection, no one responded to this on that list.]
What are pancritical rationalism (PCR) and comprehensively critical rationalism (CCR)? Two phrases for the same thing. I'll use PCR for purposes of this answer. PCR is a prescription for rationality that emerges from the work done in evolutionary epistemology based on the work of Karl Popper, William W. Bartley, and Donald Campbell.
What does it mean to say that it emerged from evolutionary epistemology? Well, PCR is a solution to a problem, and to explain the solution, first we have to understand the problem. The problem was formulated as the basis of ancient Greek Skepticism, and has plagued philosophers ever since, particularly in the last century. The Skeptics correctly pointed out that no idea could be justified except in terms of some other idea, and that ultimately this process of justification had to come to an end in some unjustified idea or set of ideas. (i.e. everyone has to take something on faith.) Some concluded from this that one could believe anything one wants.
The question here should be: Is non-propositional, non-ideational justification possible? The problem for PCR (simplifying here for lack of space) is the underlying acceptance of representationalism – the view that the mind is not directly aware of reality but only of its own contents, the latter are called "ideas." If this is not so, then it is possible to justify ideas in terms of non-ideas, e.g., in terms of sense perceptions.1 In fact, this is what pancritical rationalists and evolutionary epistemologists do through the back door.
In other words, they rail against foundationalism (the view that knowledge can and must be grounded), then they go on to accept sense perception (to some extent; see below) as a foundation as well as others (such as logical consistency).
Evolutionary epistemology, however, denies that our knowledge is grounded in justification. In fact, our knowledge is grounded in a combination of conjecture and criticism. A piece of knowledge begins as a lucky guess, so to speak, and then becomes a piece of knowledge through surviving the process of intellectual criticism. Our knowledge is constituted of successful survivors of criticism, not deductively produced from a set of "self-justifying" or "nonpropositionally justified" axioms.
While I admit "conjecture and criticism" extend knowledge, they do not justify it – as they are both in need of justification. Why choose these? Why not choose faith? Or why not stick with whatever philosophy has been handed down from your parents or teachers? If the pancritical rationalist claims that these are not as successful (and I agree with her here), then we must ask how she knows this? In measuring success along a certain scale, she is already accepting a foundation outside of "conjecture and criticism" – one outside of pancritical rationalism!
The alternatives here create too narrow a scope for knowledge. The alternatives are not "conjecture and criticism" on the one hand and deduction on the other. (Even Popper called his method "hypothetico-deductive" – meaning hypothesizing a theory then deducing its consequences and testing these!) What is forgotten here is induction, though evolutionary epistemologists generally believe induction is impossible. (Here is not the place to defend induction.)
Also, it should be mentioned that criticism is always criticism according to some standard – even if that standard is not openly stated in the criticism. For example, when geologists early in the twentieth century criticized Alfred Wegener's theory of continental drift, they did so partly by finding what they believed were inconsistencies between it and physics. (To be sure, they also made many invalid criticisms, but this particular one was valid. Wegener's explanation for what powered continental drift was wrong.) The standard here was physics as understood at that time by those geologists. It was not anything goes, else they might have used phrases from Shakespeare's sonnets or the latest stock reports to criticize his theory.
The primary criticisms that we use to separate truth from falsehood are these: Is the idea logically coherent, or is it self-contradictory? Does it contradict reproducible pieces of sensory data? (That is, is it empirically feasible?)
One of these "primary criticisms," not "contradict[ing] reproducible pieces of sensory data" is foundational! What's more, it is "nonpropositionally justified"! As I wrote above, the pancritical rationalist sneaks her groundings in through the back door – or, here, the front door when nobody is looking.
But of course, we sometimes have the problem of distinguishing between two ideas both of which are logically self-consistent and both of which accord to the empirical facts. There are further criticisms to help us here. Which of the ideas better solves the problem that the theory was intended to address? Which of the ideas is simpler? (This is the old test called Ockham's razor.) Which of the ideas is more elegant? These other "technologies of criticism" (such as empirical evidence, scientific theories, coherence, explanatory power, Ockham's razor, esthetics as in mathematical beauty, and the question of whether or not a theory actually solves the problem at hand, and how well it does so) emerge from being critically compared (by the means at our disposal) to their allelic principles. Now that we better understand how important these tests are, much work needs to be done in categorizing and understanding them.
Again, the pancritical rationalist assumes more than she let's us on to. If everything is "conjecture and criticism," then where do we get all these methods of criticism? What makes one method better than another? Or why not any method or no method? The problem is the same as above: if you know a certain method works, then there is a reason for this. One task of epistemology should not be to blindly accept it, but to find out why it works. The pancritical rationalist is unable to do this without stepping outside of pancritical rationalism as a methodology – not merely in terms of content. (A philosophy that contained all knowledge would not be a philosophy at all; so I do not ask any philosophy to do this. Even so, it should be able to handle this problem of justification.)
To give an example, if the pancritical rationalist claims Ockham's Razor (i.e., entities should not be multiplied beyond necessity) is better than reading tea leaves when deciding between two theories because it gives better results – i.e., theories decided on the former turn out to match the facts better than those decided on the latter – then she is accepting correspondence with sense perception which is a form of nonpropostional foundationalism. If she does not accept sense perception from the start, little will convince her that the theory is better or worse in terms of sense perception. That much should be obvious.
My point is not to show that we must embrace skepticism to escape foundations, but only to illustrate that pancritical rationalism cannot answer the skeptic. I believe other philosophies can answer the skeptic, but only ones that provide a system that goes further than pancritical rationalism allows.
PCR tells us that we need both rich, fruitful conjecture and disciplined, effective criticism to increase our base of knowledge. What distinguishes PCR from the older idea of critical rationalism is that critical rationalism could not defend itself against the accusation of an irrational commitment to rationality. Most proponents of critical rationalism who addressed the topic conceded that there could be no rational justification of rationalism itself. By doing away with all recourse to justificationism, PCR can be seen as rational as long as it exposes itself to its own criticisms. Although critical rationalism was not itself justifiable, PCR is itself criticizable, and therefore meets its own critierion of rationality. PCR starts by conjecturing the very simplest ideas (the logical principle of noncontradiction, for instance), testing them with whatever other means of criticism are at hand, and then using these tested ideas to test more complex conjectures.
The problem here is that to accept criticism or criticizability as a criterion is still foundationalism; it is still a form of justification. That being so, it does not escape the need for justification or validation – it merely pushes the problem from the foreground.
I bring up validation here because justification is one means of validating an idea. Some ideas are self-evident – such as the Axiom of Identity. We don't accept the Axiom of Identity because it works or is justified by appeal to some other idea, but because it is presupposed in all knowledge (all ideas and perceptions). Any attempt to deny it also presumes it.
Also, as argued above, criticism itself relies on a lever with which to move the criticized.
An example of PCR would be examination of whether critical argument would be possible, whether discrimination between true and false would be possible, without the logical law of noncontradiction. In fact, it would be impossible. It's possible without the law of the excluded middle and other less-central parts of logic, but not without noncontradiction. So noncontradiction must be accepted (conjectured), if one wishes to engage in critical discussion, if one wishes to criticize or persuade. Noncontradiction is not a given, but any alternative to it we can come up with is clearly inferior to it for one or more reasons. So we start with noncontradiction, conjectured, and accepted on the basis of surviving strong intellectual criticism better than any of its competitors, and now we can in turn use it (as the criterion of logical consistency) as a technology of criticism in evaluating and selecting among more complex competing conjectures. In this light, PCR looks like an epistemic bootstrapping scheme.
The problem with this line of reasoning is that the Principle of Non-contradiction and the Law of Excluded Middle are corollaries. And both are corollaries to the Law of Identity. They are given by all forms of awareness and implied in all knowledge. Even PCR assumes them. After all, PCR is what it is, and not something else, and not something else and itself at the same time and in the same way.
In the end, we wind up with an impoverished methodology – one that to confute the skeptics surrenders too much and asks too much of itself, after pretending to erase its foundations.
Notes:
1. Alternatively, if representationalism is so, then all bets are off. The mind has no discernible connection to the external (extra-mind) world and nothing can more be said about it.