Our Knowledge of the Growth of Knowledge:  Popper  or  Wittgenstein?
Peter Munz . Routledge and Kegan Paul

Munz's book is welcome because it takes the debate over Popper's ideas past the point where it bogged down on falsification and the problem of induction.  For some years after The Logic of Scientific Discovery appeared in 1959 critics offered two main lines of argument.  First, they pointed out that no falsification by empirical evidence could be decisive (for example, due to the fallibility of observational techniques), therefore the falsification criterion of science was useless.  Second, they claimed that scientific rationality depends on justification of theories and so Popper's theory of conjectural knowledge and his critique of inductive proof are twin daggers pointed at the heart of scientific rationality and a denial of the notion of scientific progress.   

Popper and others replied that he had drawn attention to the limitations of falsification and his methodological strictures should be regarded as guidelines or conventions to maximise the exposure of theories to empirical tests.  As to inductive proof, Popper argued that this was defective on purely logical grounds and so the rationality of science had better depend on something else which he attempted to provide with a theory of conjectural knowledge which shows how knowledge can grow without ever being proved certain or even probable.  Each side stood firm on these points and it seemed that little gain would come from protracted restatements of the rival positions.

Munz explains how Popper's writing since the early 1960s has increasingly treated matters of biology and evolution.  This move provides some additional arguments in favour of the idea that rationality does not consist in the avoidance of errors but in the elimination of error and the correction of mistakes. In the wild this occurs through the elimination of organisms and species by natural selection; in human affairs the primary method of error elimination can be the process of critical discussion.  Munz points out that  'mistakes' of a certain kind are necessary for biological evolution.  These are the mutations that provide variations (new cards) beyond the range provided by the recombination of genes  (shuffling the deck).  Most of these mistakes are not helpful for the organism but some are advantageous.   Desirable  mistakes are thus analogous to the successful products of inventors, scientists and artists.  We need these mistakes because if all sociocultural objects, ideas, rituals, procedures and traditions were reproduced faultlessly, the system would be closed, with no variation, no novelty and no possibility for change and progress.

In his chapter on Imre Lakatos and Paul Feyerabend, Munz argues that the creative problem-solving process has two distinct phases: first, the generation of tentative solutions to problems; second, the process of evaluation, testing and error elimination. The first phase has been called the context of discovery, the second the context of justification.  They are analytically distinct although they usually alternate at high speed in the mind of a person at work on a problem.   Any attempt to elevate one phase over the other results in a partial and one-sided view of the problem-solving venture.   Hence, Feyerabend's celebrated  'anything goes' principle is entirely appropriate at the stage of generating tentative solutions because brainstorming and lateral thinking are required to generate new ideas but if this approach persists at the evaluation phase, then everything stays. 

Munz contends that a healthy culture will encourage experimentation and exploration at the risk of many false starts and dead ends.  At the same time high standards of criticism will be maintained, though as Bartley has pointed out, criticism needs to be optimal, not too strict or hasty.  To maintain this precarious balance a theory of criticism is required that focuses on objects or ideas and not on the personality or motives of the artist. In addition, an advance in cultural evolution is also needed permitting people to communicate with one another despite major and perhaps even fundamental differences in their respective belief systems. Such an advance is facilitated by Bartley's work on metacontexts (see 'Bartley's Critique of Reason', Monthly Review October 1985).  For most of recorded history, Munz suggests the basis of social and cultural bonding has been shared belief systems that are exempt from criticism. 'Where knowledge is used a social bond, people cannot afford the luxury of exposing it to criticism, lest their co-operation be endangered or cease'.

But now, in certain places, a threshold has been crossed: some people; '...have managed to establish societies which are not dependent on the purity of any given cultural strain and which are  bonded by  criteria  other than the adherence to any  particular  belief system and its rituals.'  The essential feature of such a society is that some aspects of its evolution can be regulated by critical discussion in a way that was previously not possible.

However if this does not happen there may instead be further fragmentation, with the proliferation of self-contained and exclusive sub-cultures. These developments undermine the impulse to maintain communication and a sense of community between groups with diverse backgrounds. 

Munz sees the tendency towards closed systems as an emotional and obscurantist reaction to the evolutionary step towards the systematic critical appraisal of beliefs and taboos, which he himself advocates.  Philosophically, it has its counterpart in sociological theories of knowledge whose defence of relativism expresses frustration with the failure of positivism and  'mirror' theories of knowledge (as Munz calls them.)   Mirror theories portray the mind as an inert vessel, or perhaps an induction machine, which accumulates knowledge as a result of passive exposure to the outside world.   The extreme form of the mirror theory occurs in Wittgenstein's Tractatus  Logico-Philosophicus with its 'picture' theory of language in which  all valid,  meaningful  and  justified sentences mirror  some  extra-linguistic fact. Russell similarly aspired to do something or other with foundations in logic.

The classical alternative to the mirror theory of knowledge is the 'lamp' theory - the view, which Wittgenstein himself later came close to adopting, that the outside world is endowed with qualities projected upon it by the linguistic framework of the observer.     Two of the more influential philosophers today who hold what amount to lamp theories of knowledge are Thomas Kuhn and Richard Rorty, but Munz argues that both the mirror and lamp theories of knowledge contain the same structural error, namely the assumption of the need for some critical or epistemological authority that will provide a grounding for indubitably justified beliefs:

After establishing that knowledge cannot be a relationship because it is not like 'mirroring', Rorty argues that it must be justified by something else.  Here, then, is the heart of the matter.  For Rorty, knowledge is not knowledge unless it is justified...  In Rorty's view, the great divide in philosophy is between the upholders of mirror philosophy and the believers in the authority of speech communities.  As against this, I would argue that the great divide is between justificationists of all persuasions and Popperian falsificationists, who believe that we can have knowledge but that no knowledge can ever be justified.

Evolutionary epistemology, Munz contends, provides a way out of the impasse created by true belief theories of knowledge, whether of the mirror or lamp variety.  Instead of falling back on justification, he advocates the employment of criticism leading to selection and tentative critical preference.  But any shift from traditional theories of knowledge to an evolutionary epistemology requires theories of criticism to replace theories of justification. This is where Popper and Bartley have made their mark in exploring the avenues of non-justificationist criticism. 

At least five forms of criticism can be discerned; the test of evidence, the test of logical consistency, the test of comparison with other accepted theories, the 'check on the problem' and the check on the metaphysics.  The test of evidence has been the traditional obsession in the various 'mirror' theories of positivism and empiricism that have dominated the philosophy of science.  However is was the failure of these theories to survive criticism from Popper, Kuhn and Lakatos, which precipitated the modern crisis in the philosophy of science.  It is now generally accepted that things other than evidence need to be taken into account in evaluating theories.  The test of internal consistency is important and logic is also a tool for drawing out the consequences of a theory for experimental testing or for technological application.  Another way to evaluate a theory is to subject it, or its consequences and implications, to comparison with other well tried or generally accepted theories. The provisional nature of this type of appraisal is especially obvious because even 'well accepted theories' are liable to be found wanting at any moment by one or more of the five forms of criticism. 

The fourth type of critique, the check on the problem, is in some ways the most fundamental of all for two reasons; first, it can reveal that a protracted debate has parted company with the problem that it purports to be about, as when a discussion about freedom and civil liberties shifts to focus on the meanings of the word freedom or  democracy.  Second, problems always arise in the context of metaphysical assumptions or presuppositions and so a thorough review of the problem leads to the fifth form of criticism - the check on the metaphysics.  

From the time of Hume the Western empirical tradition has been hostile to metaphysics and so this is the least developed form of criticism.  It requires sensitivity to theories such as determinism and reductionism which make claims about the structure of the world but cannot be subjected to empirical tests. These theories are a constant source of irritation to philosophers in the anti-metaphysical tradition and to hardheaded, matter of fact scientists.  This is revealed by the reaction of most scientists to a suggestion that reductionism is in principle not capable of providing all the answers.  They claim that the success of their science proves the validity of reductionism; they do not realise that it is the reductionism concept of success itself that is under threat.  As R.  G. Collingwood noted in An Essay on Metaphysics (1940) scientists are likely to violently resist any invitation to examine their absolute presuppositions because they provoke 'numinous terror' and so they are hard to handle in an age that claims (falsely) to have no superstitions, hence no methods to dissipate the terror and face up to the source of it.

In Our Knowledge of the Growth of Knowledge, Munz did not draw upon Popper's theory of metaphysical research programmes which appeared in Quantum Theory and the Schism in Physics  (1982). Development of the implications of this theory remains the great task of Popperian exegesis, though even without this addition to the Popperian arsenal, Munz clearly believes that the evolutionary account of knowledge is more fruitful than the programs offered by Wittgenstein and the positivist mirror theories of knowledge.
London, 1985.

Printed in the Age Monthly Review, November 1987

Full index of articles
the Rathouse