How can we most accurately distinguish what’s true from what isn’t?

People pondering this topic have been stymied since Sextus Empiricus wrote Outlines of Pyrrhonism, in 200 A.D., by the fact that we can’t give incontrovertible defenses for our contentions about what is true or false. We can’t because:

(1) The problem of induction: (In the words of Max Born) “. . . no observation or experiment, however extended, can give more than a finite number of repetitions”; therefore, “the statement of a law — B depends on A — always transcends experience.” In this sense, there is no possibility of a method of positive inference which will logically lead from limited observations of similar phenomena under similar circumstances to the formation of a law. Statements of law are of stronger content than the evidentiary statements used in defending them. “Yet,” Born says, “this kind of statement is made everywhere and all the time, and sometimes from scanty material.”

— and —

(2) The problem of ultimate justification: Any reason for an idea is always founded upon a prior assumption, merely moving the justification one step previous in an infinite regression. We never get to an ultimate justification no matter how deep we dig; there is no foundation at the bottom.

Philosophers have been greatly concerned by the situation, because they’ve believed that, if there are no solutions to the problem of induction and the problem of ultimate justification, then there is no firm basis for rationality.

Stuck, many philosophers tried harder to get around these problems than to solve them. Some, such as P.F. Strawson, argued that the problem of induction was misconceived, that there was no need to justify inductive inference. They asserted that the relationship between evidence and conclusion isn’t illogical, but merely nonlogical. They believed there were two fundamentally different kinds of inference: deductive, defining logic; and inductive, defining natural sciences. They viewed inductive inference as inductively valid, just as deductive inference is deductively valid. The problem was “dissolved” by learning not to apply the standards of deductive logic to inductive inference.

Ludwig Wittgenstein started from these premises. He assumed that the problem of induction was insoluble. Then, attacking the concept of scientific unity, he not only accepted the argument that inductive inference is inductively valid, he greatly extended the application of that strategy of thought. Wittgenstein asserted that, since the standards of logic couldn’t be permitted to judge science, neither the standards of logic nor the standards of science could be permitted to judge other disciplines. Wittgenstein’s followers extended this into the idea that no discipline could be judged by application of the rules of any other discipline. Wittgenstein declared that such cross-disciplinary judgments were errors, which he called “category mistakes.” He called people who made category mistakes “bad pupils.” He then switched from the concept of disciplines of knowledge to “language games” and “forms of life”, each self-validated by their own rules and beholden to none else; and he left as the only function of philosophy to merely subdivide categories into unreconstructible fragments and describe each new one’s “first principles” or “grammar of logic.” Each area of knowledge was viewed as a closed system, with its own rules and logic, irreducible to and incompatible with any other area of knowledge.

Thomas S. Kuhn’s sociology-of-knowledge built upon these Wittgensteinian notions. Whereas Wittgenstein viewed fields of knowledge as incomparable spatially — not in contact with each other to apply the standards of one to another regardless of apparent closeness, due to an absence of a universal, logical underlying framework to gird them to each other; Kuhn added that fields of knowledge were also incomparable temporally — without the same framework, necessary for applying the standards of any period’s state of knowledge in any discipline to its predecessor or successor. Kuhn asserted that “paradigms” — his phrase for systems of thought on a given subject — are incommensurable. They can’t be compared with each other, and so, one cannot be used to judge another. Paradigms are, in his perception, relative: none inherently better or worse than another. The shift from one paradigm to another — say, from Aristotelian mechanics to Newtonian mechanics, or from Newtonian mechanics to Einsteinian relativity and quantum mechanics — are supposed to be nonrational, and not at all indicative of progress towards truth. With the possibility banished that rationality plays some role in knowledge growth, Kuhn chose to explain knowledge solely with sociology. “In Kuhn’s account of scientific knowledge, most of science is bound by precedent, tradition, and commitment to reigning paradigms that are guarded, licensed, and franchised by scientific elites — elites concerned to train, indoctrinate, supervise, socialize and politicize initiates into the scientific enterprise. The main activity for which initiates are to be trained is the solving of relatively minor problems — Kuhn calls them “puzzles” — set by, and in conformity with, the reigning “paradigm”. Intellectual revolution is a rarity, more disruptive than enlightening, something against which guardians of paradigms should protect themselves — even weeding out dissidents — for the sake of the larger enterprise. Kuhnian sociology-of-knowledge is strongly normative: what is popular is right,” notes W.W. Bartley, III.

Kuhn considered objections that the degree of correspondence with reality of the contents of ideas must surely have some influence on their acceptance or rejection by the scientific community to be sufficiently undone by the assumed irrational basis of science and by the fracturing of knowledge into incommensurable language games.

Most contemporary philosophers have followed this line of reasoning. And so we’ve been left in the modern world with the popular views that (1) there really is no firm basis for rationality, and (2) interdisciplinary criticism is invalid. With the foundations removed from rationality, the common next step is to assume that acceptance or rejection of an idea is necessarily nonrational. The assumption is that, since no idea can ultimately be justified, the choices come down to either accepting any given idea on a nonrational basis, or denying the idea on a nonrational basis, Either way, according to Wittgensteinian reasoning, one’s ideas would ultimately be nonrational, founded on faith. This is essentially what Ludwig Wittgenstein meant when he famously wrote, in On Certainty, “At the end of reason comes persuasion.”

All of these ideas of Kuhn’s and Wittgenstein’s, as well as the current post-rationalist attitudes of epistemic relativism, hinge upon the insolubility of the problem of induction and the problem of ultimate justification, and hinge upon this insolubility precluding a tenable version of rationality — but it’s not so.

It does appear that the problems of induction and ultimate justification are indeed insoluble; however, it turns out that doesn’t actually pose a problem for rationality. As Karl Popper figured out in 1934, rationality doesn’t depend upon the reason behind the formation of an idea being justified. This is because criticism, not justification, is the basis for rationality. Whether an idea came from careful scientific research or from a Ouija board may be valuable for its insight into how ideas are conceived, but not for vetting whether ideas are true. What is important to determining the truth of an idea is not the justification behind it, but its ability to survive criticism.

It’s like the difference between Lamarckian evolutionary theory and Darwinian evolutionary theory. In Lamarck’s theory of evolution through the inheritance of acquired characteristics, an organism’s physiology had to be fully justified prior to the organism’s birth. Whereas, in Darwin’s theory of evolution through natural selection, an organism’s physiology is only relevant due to its ability (or inability) to survive to reproduction within whatever environment it finds itself. Similarly, in justificationist epistemologies, an idea’s veracity has to be fully justified prior to the idea’s generation. Whereas, in critical epistemologies, an idea’s veracity only needs to survive our best efforts to refute it.

The following explanation of how knowledge can be held, with logical consistency, is part of an epistemological system called Pancritical rationalism:

1) Induction, as a method of positive inference which verifies the correctness of statements, doesn’t exist. Theories are never securely founded.

2) While statements of law cannot be verified, they can, in principle, be falsified. If I hypothesize that all flamingos are pink, no amount of pink flamingos that I see will prove that all are pink, but I only need to see one black flamingo to falsify the hypothesis (provided we accept the legitimacy of the data). If one exposes one’s ideas to severe criticism and rigorous testing designed to reveal non-instances of a proposed law, and the ideas survive, there is no reason to suppose that they are wrong. But, if one exposes one’s ideas and they are duly refuted, then there is a rational reason to dismiss them. Falsifiability provides a foundation for rationality.

Justificationists, afraid to be taken off of firm foundations, might object that there must be some ultimate standards for the criticism to which hypotheses are exposed. But that’s wrong. The standards of criticism are always open to criticism, themselves. Isomorphically, there is no ultimate set of standards for survival to reproduction which an organism’s physiology must meet: when the weather is hot and dry, an organism must be able to survive heat and drought; but when the weather changes to cold and wet, those standards no longer apply, and the organism must now be able to withstand coldness and moisture, instead.

3) We may be thinking that we are basing our rationality on induction, but we’re actually just taking guesses, and our rationality is based on the results of deductive tests against our guesses. These guesses may be inductively conceived, may truly originate from perceived cause and effect, and a hypothesizer may feel confident of the rightness of a guess — but this is unrelated to the logical methodology of how knowledge is held.

4) Without accepting an idea is definitively correct or definitively incorrect, reserving prejudice, one can still credit whether an idea best fits with available knowledge compared to any known alternative hypothesis.


With a workable method of holding ideas rationally that renders the problem of induction and the problem of ultimate justification irrelevant, we can see that the entire footing of the aforementioned work of Wittgenstein and Kuhn is a misstep. Logically consistent rationality is possible. The acceptance or rejection of ideas can indeed be based on correspondence with reality.

Epistemic relativism is bunkrapt, failing to congrue with reality. If the only sense in which knowledge progressed was chronologically — if scientific change was merely a social phenomenon meandering aimlessly through the geography of possible thoughts — then our models of reality would be infinitely far from actual, objective reality, and would not ever be getting any closer (except, perhaps, temporarily, by chance). If this were the case, then behavioral adaptation as a method of survival would be an impossibility. As would any kind of technology.

Cross-disciplinary criticism remains a valid form of judgment. Wittgensteinians object to this, often pointing out instances when the rules of one scientific discipline supposedly don’t work for another. For example they may tell you that physicists would absolutely insist upon reproducibility for the Michelson-Morley experiment, but anthropologists have accepted non-reproducible data from first encounters with isolated tribes in the Amazon and New Guinea. But this is a misapprehension, and the implication that this suggests different science disciplines are operating by fundamentally different rules is erroneous. Both physics and anthropology demand the results of experiments must be reproducible when the experiments themselves can be repeated, but neither branch of science thinks that data is only acceptable when it comes from repeatable events. When Comet Shoemaker-Levy 9 collided with Jupiter, no physicist said, “Throw away all the data we collected; the collision’s not replicable.” When repeatability is available, it’s a convenient and powerful opportunity to test and criticize; when it’s not, there are still other possible grounds upon which criticism can be based. For example, criticism can be based on noncorroborability with related research. Some disciplines may study things that are easier to repeat than other disciplines do, but they are, nonetheless, thematically similar in their process of advancement — the generation of conjectures followed by rigorous attempts to refute them. Attempts to show the alienation of disciplines, and the inapplicability of cross-disciplinary criticism, are bogus.


But what about what scientists actually say they do? For example, what about when Richard Dawkins said, “Have you ever wondered how we know the things that we know? How do we know, for instance, that the stars, which look like tiny pinpricks in the sky, are really huge balls of fire like the Sun and very far away? And how do we know that the Earth is a smaller ball whirling round one of those stars, the Sun? The answer to these questions is ‘evidence’.”

Is this not inductivism and justificationism — with evidence “proving” hypotheses through inductive inference — in direct contradiction with everything I outlined, above? And is this not faith-based and irrational?

Yes. Many scientists and science advocates, such as Richard Dawkins, Carl Sagan, and Bill Nye, tend to discuss how we know what we know in inductivist and justificationist terms.

Insofar as these methods extend finite repetitions into universal laws held as facts: yes, they are faith-based. Having said so, it’s pertinent to point out the distinction between this kind of faith and blind faith. The inductive methodology these scientists suggest begins with pattern recognition of observed phenomena, then extends it into inferred generalizations. Contrarily, blind faith does not begin with pattern recognition of observed phenomena. Whereas scientists are expressing faith based on empirical data, blind faith is based on nothing objective whatsoever, and often in contradiction to all available data.

Nevertheless, extending these patterns into certitude about universal laws is still faith. The notion that limited observations can “verify” or “prove” hypotheses is not logically defensible. So — yes, what these scientists and advocates are suggesting is irrational. In the quote above, Dawkins is mistaken; evidence helps us form our ideas, but it doesn’t get us to knowledge.

It’s unfortunate that scientists and advocates popularize a flawed model of how we know what we know, when better ones are available. They’re writing bad checks, so to speak, when they assert that science is trustworthy because evidence grounds their ideas in rationality. Such mistaken statements engender the preventable problem of irrationalist critics arguing that science is based on faith — if you champion an inductivist model that is necessarily based on faith, then it’s appropriate for critics to point that out. Moreover, advancing this unworkable model leads people to misunderstand how science, and the enterprise of increasing humanity’s knowledge, work.

However, the scientists and advocates who espouse this faulty model, wherein evidence proves hypotheses through inductive inference, suffer more from a lack of self-awareness than from an actual problem with how they conduct scientific investigation or how they hold ideas as true or false. After all, they still infer predictions from their hypotheses, whereby it is possible for their ideas to fail if the predictions fail. They still put them to the test. They still dismiss the ideas that fail the tests. They still submit their work to their peers for critical review and attempts at refutation. In sum, they still try to grow knowledge through the process of conjectures and refutations. Which is to say: they are operatively Pancritical rationalists, regardless whether they proclaim the opposite.


Kuhnian sociology-of-knowledge proponents sometimes try to frame the peer review process as an example of sociology’s role in irrationally influencing the growth of knowledge. Peer review, they’ll tell you, is an area where prestige “sells” scientific knowledge, where a big-name scientist at a high-status institution makes research findings practically irresistible. But this is laughable, when you note that peer reviews tend to be double-blind, and note that peers must concentrate on reproducibility, corroborability, falsification, broader and more parsimonious explanations of the data, errors in reasoning, and the like. It’s not the social order of things which pressures peer reviewers, but rather the objective reality which surviving theories must successfully model.

Thomas Kuhn published The Structure of Scientific Revolutions in 1962, giving us the last fifty-three years on which to test how scientific revolutions actually did proceed against how Kuhn’s model would predict that they will proceed. Aside from and more importantly than all of the errors upon which Kuhn’s ideas rest, his ideas have been refuted by the actual course which science took — which is composed overwhelmingly of non-instances of the predictions of the Kuhnian model. Here’s an example (quoted from Frank J. Tipler): “Steven Weinberg and Sheldon Glashow, two of the main creators of the Standard Model [of particle physics] — revolutionaries, according to Kuhn — have both sharply criticized Kuhn’s theory, pointing out that the Standard Model revolution did not occur the way Kuhn would lead one to expect. A former leading proponent of a non-Standard Model theory, John Polkinghorne — an antirevolutionary, according to Kuhn — has said exactly the same thing. All three physicists agree that the physics community accepted the standard model on purely objective, rational grounds. They all present evidence that the Standard Model is in fact a theory closer in accord with objective reality than all other theories which have been developed to date.”

Does this mean there’s no possibility that factors outside of objective criticism can influence the progress of science and the growth of knowledge? Of course not. It’s specifically the sort of irrationally-based sociology-of-knowledge which Kuhnians suggest that fails — where content-free persuasive elements govern the fate of theories, irrelevant of their degree of approximation with reality. For example, one such influence that really does impact the procession of human knowledge is economics. The pursuit of knowledge has costs in terms of time, energy, and money; and there are both limited resources and unlimited possibilities for resource distribution, requiring that choices be made. Economics-of-knowledge, which doesn’t contradict rationalism, was realized by Karl Popper and expanded by W.W. Bartley, R. Coase, F. Hayek, and others. Predictions of how science and the growth of knowledge will occur, according to models of economics-of-knowledge, have been tested and have not been falsified.

One might argue that Thomas Kuhn and and his adherents, on the side of sociology-of-knowledge, versus Karl Popper, William Warren Bartley, III, and others, on the side of rationalism, are addressing two different issues. Kuhnians are trying to describe the way science does proceed, while Popperians are trying to describe the way science ought to progress. (What Thomas Kuhn pejoratively called “idealized” accounts.) This is true, however, there is scant corroboration, and numerous refutations, of Kuhn’s description of the way that science purportedly “really” happens.

In a limited sense, Thomas Kuhn and his followers may be right. As a description of the way that scientists, as individuals, try to act, the views of the sociologists-of-knowledge may accurately describe a significant proportion. Scientists can be as self-interested and self-serving as any other category of humanity. Yet the fact that some scientists really do scheme, misrepresent each other, plagiarize, hold grudges, strive for prestige, stay loyal, foment controversy, condemn others to conspiracies of silence, and so forth — while profligate refutations of the course predicted by the Kuhnian model nevertheless occurred — only testifies to the impotence of sociological influences on the course of science over the long-term.

Thomas Kuhn wrote, “There is no standard higher than the assent of the relevant community.” However, this statement is a failure to see below the surface, and can be dismissed on rational grounds. The course that science has taken shows that the assent of the relevant communities is strongly oriented toward how accurately an idea models reality. And this is so because they recognize that closeness of approximation of truth, estimated through objective and rational methodologies, is a higher scientific standard than mere assent.


Some people think your choices are limited to their simple, binary alternatives: justification and induction as the basis for rationality, or no possibility of rationality at all; reproduction of experiments as a basis for criticism of hypotheses, or no basis for criticism; total acceptance of any given idea on a nonrational basis, or total rejection of the idea on a nonrational basis.

There are wider choices available, including Pancritical rationalism, whereby no appeals to irrationality are necessary.

[Please note that no parts of this are intended to be any kind of swipe at Richard Dawkins, Carl Sagan, or Bill Nye — all of whom I respect, admire, and mostly agree with.]