knowledge human, reliability and fallibilism

Author: Mariano Artigas
Date of publication: Naples, 1992

One of the main problems we have to face when reflecting on the image of man in our scientific age is that of the value of the human knowledge , which in turn appears in strong dependence on the evaluation of the knowledge provided by experimental science. In this context, questions about the reliability of science occupy a central place. Jürgen Habermas has written that if we were to reconstruct the philosophical discussion of the modern era as a judicial process, the only question to be decided would be this: how can we obtain reliable knowledge ? [Habermas 1968, p. 11].

It is well known that fallibilism is one of the main ideas of Karl Popper's Philosophy and that it implies the denial of any subject of reliability. The extent to which fallibilism is widespread in contemporary epistemology can be seen from the following words, spoken by a character in a dialogue recently composed by John Worral: "I suppose everyone is a fallibilist nowadays". However, fallibilism is not an easy concept, as can be appreciated if we examine the problems involved in the whole quotation : "I suppose everyone today is fallibilist about scientific theories: I mean, not fallibilist in principle (since this position seems to be dictated by logic alone), but fallibilist in the internship, since the history of science sample clearly shows that even the most successful high-level theories can eventually be rejected (although they may survive as borderline cases)" [Worral 1989, p. 268]. As Worral himself points out, such claims require non-trivial qualifications about what it means to reject a theory, and involve not only logical arguments, but also epistemological and historical problems fraught with difficulties.

I will try to analyse what kinds of problems fallibilism seeks to solve, what class difficulties it encounters due to its rejection of reliability, and how these difficulties can be overcome.

1. Fallibilism and the rise of the knowledge

When applied to knowledge, the term "fallibilism" expresses that our knowledge is not perfect. The Collins Cobuild Dictionary (London 1987) explains the meaning of the term "fallible" as follows: "If someone is said to be fallible, this means that their judgement or knowledge is not perfect and can be wrong. If something is said to be fallible, it means that it is not perfect and can be wrong". The adjective "fallible" can be applied to the subjects they know or to the methods they use or to the statements they propose. All three aspects are found in Popperian texts, which gives rise to not inconsiderable difficulties.

However, it would be wrong to concentrate the discussion on the meaning of the terms, since Popper insists that we should avoid this subject of questions, because the central point is always to notice what is the problem we want to solve (and I am quite close to agreement with this remark). The question then is: what is the problem that Popper poses when he speaks of fallibilism?

The answer is not difficult. Undoubtedly, the central problem of Popper's epistemology is the rise of knowledge. Moreover, the perspective that Popper adopts is also clear; he basically states that, in order to evaluate any cognitive claim, what really matters is our attitude towards it: if someone tries to establish it by positive arguments, he adopts a justificationist attitude that will encounter serious difficulties from the point of view not only of logic, but also of history and of the general Philosophy . Thus, fallibilism is closely related to falsificationism, critical rationalism and conjecturalism, which are the positions Popper takes on civil service examination, respectively, verificationism, justificationism and dogmatism.

This explains why fallibilism is used by Popper as a label which expresses which subject of Philosophy is adopted. A central text in this respect is chapter 10 of Conjectures and Refutations, in which Popper classifies philosophers into two main groups. The first is the group of "the verificationists or justificationist philosophers of knowledge (or of belief)", and the second is the group of "the falsificationists or fallibilists or critical philosophers of knowledge (or of conjectures)". Popper describes his own position in the following terms:

"the falsificationists or fallibilists say, roughly speaking, that that which cannot (for the time being) be overthrown by criticism, is not (for the time being) worthy of serious consideration; while that which can be so overthrown and yet withstands all our critical efforts to bring it about, may well be false, but is not unworthy of serious consideration and perhaps even of being believed - though only tentatively so.... The falsificationists (the group of fallibilists to which I belong) believe - as do most irrationalists - that they have discovered logical arguments which show that the programme of the first group cannot be carried to completion: that we can never give positive reasons for a theory to be true..." [Popper 1963, p. 228].

So far, fallibilism is not an entire theory of knowledge. It is a methodological proposal for correctly assessing the claims of knowledge. That proposal is closely related to falsificationism, and thus to the logical impossibility of establishing the truth of a theory by positive corroboration of its consequences. As a methodological precaution, fallibilism constitutes a salutary warning. Problems begin when fallibilism is extended to an entire epistemology, for this involves two extrapolations. The first is to extrapolate a methodological attitude as if it were a complete description of the scientific method. The second is to extrapolate this partial picture of science as if it were a complete theory of human knowledge . I will now briefly consider these two issues.

With respect to the first, the scientific method can hardly be described in preferably negative terms. Eugene Freeman and Henryk Skolimowski have lamented "that both Peirce's and Popper's methodology have been designated by such an inept term as fallibilism", because that term suggests "the human tendency to err" and usually means "propensity towards error" or "propensity to be erroneous or inaccurate"; in this sense, "the term is singularly inept, almost a caricature, as a name for scientific method", because "it misses the main point about what is done in science when mistakes are made - that is, not that mistakes are made, but that they are (a) recognised, (b) eliminated, and (c) advanced beyond, so that one gets asymptotically closer and closer to the truth". Freeman and Skolimowski suggest that "a much more accurate designation for identifying the methodology of both Peirce and Popper is to be found in Popper's inspired phrase, conjectures and refutations, which captures much better the essence of the scientific method" [Freeman-Skolimowski 1974, pp. 514-515].

Recognising and eliminating errors implies positive capabilities and achievements that should be reflected in any theory of scientific method. Of course, Popper's methodology alludes to this, but in that case we can conclude that fallibilism is not an adequate label . However, I think there is more to it, because we do in fact obtain true knowledge and we know that we obtain it. The clarification of this problem depends on our ideas about certainty.

With regard to the second problem, although Popper obviously distinguishes between the ordinary, the scientific and the philosophical knowledge , he nevertheless states that

"the scientific knowledge can be studied more easily than the common-sense knowledge . Indeed, in a way, it is the knowledge of common sense writ large. Its problems are extensions of the problems of the knowledge common sense. For example, it replaces Hume's problem of reasonable belief with the problem of reasons for accepting or rejecting scientific theories" [Popper 1935, p. 22].

I do not entirely agree with agreement with this. I do not deny that experimental science and ordinary knowledge share some important features, nor that the method of conjecture and refutation is widely applied in ordinary life. However, I think that the ordinary knowledge includes some capabilities that are used as assumptions of the sciences and constitute genuine necessary conditions for the existence and progress of the scientific research . Such are, among others, the capacity for self-reflection and the sense of evidence (a term rarely used in fallibilist contexts), which are assumed in the argumentative capacity. There is a feedback of scientific progress on these assumptions: scientific progress retro-justifies and extends them, and sometimes corrects them. Therefore, we can use the methods and achievements of experimental science to learn more about how the ordinary knowledge works, but it would be a mistake to forget the special features that the ordinary knowledge possesses, which relate to the instructions of our whole knowledge.

That fallibilism tends to present itself as an entire theory of knowledge can be seen by analysing the 1961 Addendum to Popper's The Open Society, especially sections 4 to 9. I will underline several points that are of particular importance. Popper writes:

"By fallibilism I mean here the idea, or the acceptance of the fact, that we can be wrong, and that the search for certainty (and even the search for high probability) is a wrong search. But this does not imply that the search for truth is wrong. On the contrary, the idea of error implies that of truth as the patron saint that may not be attained. It implies that, while we may seek the truth, and may even find it (as it seems to me we do in many cases), we can never be quite sure that we have found it. There is always room for error, although in the case of some logical and mathematical proofs that possibility may be regarded as small. But fallibilism need not at all lead to sceptical or relativistic conclusions. This becomes clear if we consider that all known historical examples of human fallibility - including all known examples of miscarriages of justice - are examples of the progress of our knowledge. Each finding of a mistake constitutes a real advance in our knowledge... Therefore, we can learn from our mistakes. This fundamental perspective is, in fact, the basis of the whole epistemology and methodology.... " [Popper 1961, pp. 375-376].

It is easy to see that Popper's fallibilism includes a general evaluation of the human knowledge , some specific methodological caveats, and many historical interpretations. I think that these three elements are not balanced, and that this is partly due to the polemical context in which fallibilism was born and developed, and to some assumptions that depend on that context.

2. The assumptions of fallibilism

Indeed, Popper's epistemology can only be fully understood if it is confronted with inductivism, verificationism and scepticism. Popper notes the difficulties that these positions have to face and sample their relations to some general aspects of the theory of knowledge. Popper's arguments are mainly directed towards the clarification of the problem of demarcation between science and pseudo-science, which is largely equivalent, in the Popperian context, to the problem of rationality:

"But, unlike the irrationalists, we falsificationists believe that we have also discovered a way to realise the old ideal of distinguishing rational science from the various forms of superstition.... We hold that this ideal can be realised very simply by recognising that the rationality of science lies not in its habit of appealing to empirical evidence to support its dogmas - astrologers also do so - but only in the critical perspective.... For us, therefore, science has nothing to do with the search for certainty or probability or reliability. We are not interested in establishing scientific theories as certain, or certain, or probable. Aware of our fallibility, we are only interested in criticising them and putting them to test, hoping to find where they are wrong; in learning from our mistakes; and, if we are lucky, in making progress towards better theories" [Popper 1963, pp. 228-229; italics mine].

I wonder whether it is necessary to deny any subject of certainty or reliability in order to clarify these problems. I rather think that, in this way, old misunderstandings are replaced by more refined perspectives, which, however, do not correspond to the real achievements of the sciences.

Perhaps fallibilism depends too much on the basic rationalist perspective. It reacts to it, but nevertheless accepts the equation between legitimate certainty and a perfect and absolute certainty that would be a simple consequence of linear logical arguments. Like an iceberg that only occasionally appears before our eyes, the idea of a perfect and absolute knowledge is the target against which the attacks of fallibilism are directed. Already in his early work Popper concluded that:

"The old scientific ideal of knowledge - an absolutely certain and demonstrable knowledge - has been proved to be an idol. The requirement of scientific objectivity makes it inevitable that any scientific statement must always have a tentative character" [Popper 1935, p. 280; italics mine].

This idea is always latent in Popper's works, and is sometimes made explicit. For example, the word "fallibility" in the table of contents of The Self and Its Brain contains a single reference letter, which refers to a chapter graduate "Conjectural explanations versus ultimate explanations". There Popper opposes his conjectural method to the essentialist perspective that seeks ultimate explanations. According to Popper, an ultimate explanation "neither needs nor can refer to further explanations. On the contrary, any conjectural explanation can give rise to a new problem". Ultimate explanations would also be infallible [Popper 1977, p. 172]. (Incidentally, it seems to me that the reference letter there to Aristotle's scientific ideal in the Second Analytics, repeated several times also by Hans Albert in the same sense [Albert 1984, pp. 19-30 and 37; 1987, p. 69], would require more careful consideration).

Explaining the main lines of evolutionary epistemology, inspired by Popper's ideas, Gerhard Vollmer puts the same idea as follows:

"An absolute justification of the human knowledge is not possible. Any such attempt to rise above the quagmire of uncertainty leads to a triple impasse, namely either to a circle that is logically flawed, or to an infinite return that is practically impossible, or to an arbitrary suspension of the postulate of justification that leads to dogmatism. This triple alternative of dead ends was aptly named by Hans Albert Münnchausen's trilemma" [Vollmer 1987, p. 174].

Münnchausen's trilemma [Albert 1968, I.2; 1987, p. 69] has been examined in a clearly Popperian context by William Warren Bartley III, under a slightly different form and name ( Fries' trilemma ). Bartley has claimed that this is not a genuine trilemma and that, apart from those mentioned, there are many other possibilities [Bartley 1962, pp. 211-216]. In my opinion, what Albert's and Vollmer's formulations of the trilemma show is only that the problem about the foundation of knowledge is posed as if every claim of knowledge must be based on logically linear proofs. It seems to me that the issue is much more complex, and that it also calls for a rethinking of the role played by the asymmetry between verification and falsification, which is one of the central points in the defence of fallibilism.

Vollmer argues that "the goal of epistemology is not to provide absolute justifications of knowledge and truth claims, and adds:

"If we had a knowledge of that subject, true, reliable, universal, goal, epistemologists might feel obliged to explain how it is possible. But so far, no one has achieved a single sample of perfect knowledge. Therefore, there is nothing to explain; the problem simply does not exist" [Vollmer 1987, p.175; italics mine].

Following this argument, Vollmer uses the classical example of Newton's theory, and goes on to insist on opposing the perfect and absolute knowledge to the fallible knowledge :

"But is Newton's theory absolutely true? No!... we even know that it is in fact false....

On objectivity, we are now in a much better position. Our knowledge - as uncertain, imperfect, conjectural, preliminary and fallible as it may be - finally has a chance, at least, to be goal, to be true about the real world as it is.

knowledge perfect about nothing, or knowledge imperfect about the real world: which would we prefer? Of course, there is no choice (Newton's theory is, in fact, false); but if there were, wouldn't we choose the second alternative?" [Vollmer 1987, p. 176; italics mine in the case of "absolutely" and "perfect"].

Fallibilism excludes any subject of certainty and concludes that all empirical knowledge is hypothetical. But it should be noted that there are different types and Degrees of knowledge, of truth, of certainty and of theoretical constructs: for example, we can think of experimental laws, of general principles, of models of spatial Structures , of phenomenological theories. It would be inappropriate to call all these constructs hypothetical in the same way. I am sure that fallibilists know these elementary distinctions well, but I wonder if they sometimes forget them.

Fallibilism stresses that the absolute and perfect knowledge is beyond our reach. I agree with agreement. But from this we cannot conclude that any subject of certainty is unattainable.

3. The consequences of fallibilism

The complexity of fallibilism becomes clear when we consider the inevitable problem about its scope: is it possible to extend the fallibilist thesis to include fallibilism itself? It was Bartley who proposed an affirmative answer and called it, first comprehensive critical rationalism, and later pan-criticism. From agreement with Bartley:

"Although the problem about the limits of rationality can, in my view, eventually be solved only within the context of a Popperian-style fallibilism, Popper's own first explicit attempt to solve the problem is inadequate, is as fideistic as those of Ayer and Putnam, and seems to operate within a justificationist context alien to the dominant themes of his own thought."

Bartley explains that, in 1960, he discussed these issues with Popper, who introduced some changes in chapter 24 and "a polemical Addendum on relativism" in The Open Society and its Enemies, and also some changes in other works. Bartley goes on to say that:

"Despite these alterations, Popper's earlier fideistic outlook has only been patched up, removing some old notions but retaining the old terminology - e.g., critical rationalism - and the old slogans. The result is a confused status " [Bartley 1962, pp. 104-105].

In fact, Bartley refers to a text in chapter 24 of The Open Society in which Popper writes:

"While an uncritical, comprehensive rationalism is logically untenable, and while a comprehensive irrationalism is logically tenable, this does not mean that we should accept the latter. There are other sustainable attitudes, especially that of a critical rationalism that recognises the fact that the fundamental rationalist attitude results from an (at least tentative) act of faith - from faith in reason. Our choice is therefore open. We can choose some form of irrationalism, even some radical or comprehensive form. But we are also free to choose a critical form of rationalism, one that frankly admits its origin in an irrational decision (and which, to that extent, admits a certain priority of irrationalism). The choice before us is not simply an intellectual matter, or a matter of taste. It is a moral decision..." [Popper 1945, ed. 1977, vol. 2, pp. 231-232].

The discussion of this problem is difficult, as is evident from several contributions in Radnitzky-Bartley [1987]. Vollmer puts it in the following terms:

"Evolutionary epistemology is inseparably connected with hypothetical realism. This is a modest form of critical realism. Its main thesis are: All knowledge is hypothetical, that is, conjectural, fallible, preliminary.... According to agreement with this position, all knowledge is hypothetical, i.e. uncertain. This affirmation is itself part of a theory, specifically of a theory of knowledge. It must therefore be applied to itself. But this is said to lead to contradictions" [Vollmer 1987, p. 188].

Vollmer's reply first stresses that hypothetical realism claims that all synthetic statements are hypothetical, and continues:

"Now, this statement is itself either analytic or synthetic.... It is not evident, however, whether it is analytic or not..... Let us suppose, then, that our statement 'all synthetic statements are hypothetical' is itself synthetic. It should then be true for itself and therefore self-applicable. Then it claims to be itself as synthetic also hypothetical, i.e. either false, or true and unprovable. In particular, it could be false: there could exist, as already said, demonstrable systemic statements. If they did exist, then the main thesis of hypothetical realism would be false. But this, in spite of all self-reference letter, does not lead to any contradiction. From agreement with hypothetical realism, it belongs to the essence of synthetic statements to be possibly false. However, our statement might not be provable. For, supposing it were provable, it would become untrue and, therefore, it would again be unprovable. On the other hand, the assumption that it is unprovable does not lead to contradiction. Indeed, that is precisely what it asserts about itself" [Vollmer 1987, p. 189].

I must admit that I do not consider this subject of reasoning too important, and I suppose Vollmer would agree with agreement. What I do consider important is that the statement "All knowledge is hypothetical" is, in my opinion, simply false, unless we accept some rationalist patterns that I do not share. We do not have a perfect and absolute knowledge . However, we can obtain a knowledge that is contextual and partial but, at the same time, authentically true.

With respect to the anthropological and social consequences of fallibilism, I share many aspects of Popper's ideal of the open society, but I wonder whether fallibilism provides an adequate foundation for them. It also seems to me that evolutionary epistemology can stimulate the study of some interesting problems, but I think it does not provide a real solution to the main problems about human knowledge ; for example, saying that our knowledge corresponds to reality because it is the result of selective processes [Vollmer 1975, pp. 102-106], can hardly be considered as a real explanation of our capacities. As a consequence of the limitations of fallibilism on the epistemological level, we will obtain too narrow a picture of man and society if, in order to study them, we only use the resources assumed and implied by fallibilism.

I think fallibilism would provide a much truer picture if it were purified of its polemical implications. After all, what fallibilism highlights are the limits of our knowledge, and it would be logical to recognise that it is not a universal thesis , but a partial and limited perspective on some particular problems.

4. Reliability and fallibilism

Fallibilism denies that we can reach a reliable knowledge . I will try to show that we can accept that experimental science provides reliable knowledge, and that we can even explain this fact to some extent.

The term "reliable" is applied to something that offers guarantees that it will achieve something goal. It can be applied to experimental science in two ways. The first refers to the means used to achieve the goals of science, i.e. scientific methods and constructs; in this sense we can speak of internal reliability. The second refers to science as a human business that coexists with other enterprises such as the ordinary knowledge and Philosophy, and leads to a comparative assessment of experimental science with respect to other human activities; in this sense we can speak of external or comparative reliability. While the two senses are different, they are nevertheless closely related to each other.

Without the need to adopt too compromising perspectives, we can assume that the general goal of experimental science is to achieve a knowledge of nature that can be subjected to empirical control. The requirement of experimental control is a necessary condition for the acceptability of theoretical constructs, and can be carried out agreement in a variety of ways. However, it imposes severe limitations on scientific constructs: they must be constructed in such a way that, at the very least, they can be logically related to other constructs that are ultimately related to the results of repeatable experiments.

Therefore, the construction of scientific concepts must include empirical aspects related to experiments, so in this sense, the context of finding plays an important role in the assessment of scientific constructions. Even if one accepts that the psychological aspects of finding may be irrelevant for that assessment, the process of construction includes the establishment of a whole set of definitions to be used in the subsequent work . Experimental science involves creativity and interpretation at all stages. The scientist can never be replaced by impersonal devices, except for routine work.

The constructivist dimensions of science are omni-comprehensive. Scientific constructs are evaluated by applying them to situations that relate to specific problems, and this involves creative planning and interpretation. Experimental science is a hermeneutic business in all its aspects. Objective patterns exist, but they are never free of interpretation. It is always possible to reinterpret them.

However, we can sometimes obtain rigorous evidence about our constructions, so that we can attribute to them a truth that will always be contextual and therefore partial, because of the constructive and interpretative elements mentioned above, but it will also be authentic, since it somehow reflects aspects of the real order of nature.

The possibility of attributing truth to our constructions must face the logical difficulty posed by the characteristics of the hypothetico-deductive method which, due to its structure, does not allow us to infer the truth of the premises from the truth of their logical consequences. However, even admitting that by means of this method we will never obtain perfect demonstrations from the logical point of view, in many cases we can use criteria which, although not totally conclusive, are nevertheless sufficiently strong. These are, above all, the explanatory and predictive power, the accuracy of explanations and predictions, the mutual support between different theories and the convergence of independent evidence. Of course, the application of these criteria leads to different Degrees of reliability of agreement with the types of constructs and situations; it is not the same, for example, to examine the validity of experimental laws, general principles, spatial Structures or phenomenological theories, to cite some relevant cases.

In this way the four main aspects of the reliability of science can be explained: intersubjectivity, empirical contrastability, predictability and progressivity. We have no guarantee of success in our attempts to explain nature, but if we use the scientific method of construction and control and, moreover, are fortunate, then it is easily understood that we obtain a subject from knowledge that possesses these characteristics. I have elsewhere set out in detail my views on this topic [Artigas 1988 and 1989], which are closely related to the objectualist realism developed by Evandro Agazzi [Agazzi 1969, 1978, 1985 and 1988].

We can also use these ideas to explain the relationship between the scientific perspective and the perspectives adopted in other human endeavours. It is understandable that if we study those aspects of human existence that cannot be subjected to experimental control, we will not obtain the special reliability that is characteristic of experimental science. But this has nothing to do with scientism. On the contrary, it can be shown that experimental science is not entirely self-sufficient, since it relies on philosophical assumptions that constitute the very conditions of possibility of science.

In particular, experimental science assumes an ontological realism that affirms the existence of nature as a hierarchy of interrelated levels, whose entities possess a dynamism that is independent of our intervention and that unfolds from agreement with patterns. Science also assumes a gnoseological realism that affirms our capacity to know natural patterns, as well as to evaluate the truth of our constructs; this implies the capacity for self-reflection and a sense of evidence. These assumptions are not studied thematically by the sciences, but are used by them at least implicitly. Moreover, scientific progress exerts a feedback-action on these assumptions: it retro-justifies them, expands them and makes them more precise. There is a feedback of science on these assumptions, which correspond to a realist Philosophy that goes hand in hand with experimental science [Artigas 1992 a].

Therefore, there is no civil service examination, but mutual complementarity, between experimental science and a realist Philosophy . Scientism is untenable because experimental science, which would be the paradigm of any reliable knowledge , relies on assumptions that have a validity of their own. Scientific naturalism is untenable because it attributes to a particular, admittedly very effective, perspective an absolute character that it does not actually possess. There is no civil service examination between the scientific knowledge and the pretensions to know metaphysical dimensions that do not only refer to the problem of transcendence and the spiritual dimensions of the person, but also to purely natural entities [Artigas 1991]; in fact, as has already been pointed out, science assumes the existence of a natural dynamism whose explanation points towards dimensions that are not trapped by the scientific method and can be properly understood as ontological and metaphysical dimensions [Artigas 1992 b].

I think this perspective is compatible with some central ideas of fallibilism, as long as one does not understand fallibilism too rigidly and does not turn it into an entire epistemology. I will conclude with an example that illustrates my point. From agreement with Bartley,

"Ordinary fallibilism asserts that certainty cannot be achieved, and that anything we can say, even anything that can be incorporated into science, may be wrong. Ordinary fallibilism also includes the sense that much of what we have already discovered only touches the surface, and that many new things remain to be learned about ourselves and the universe in which we live, things that far transcend our current knowledge . I am a fallibilist in this ordinary sense" [Bartley 1989, p. 207].

On this quasi-definition of fallibilism, I will point out four qualifications. (a) Its validity is restricted to a total and absolute certainty that would correspond to a perfect knowledge in the absolute sense. (b) While any empirical statement has limited validity, we can talk about truth in concrete, i.e., we can assert the contextual and partial, but authentic, truth of many empirical statements. (c) The validity of the claims of knowledge depends on our goals, so we can often reach a Degree of certainty that is sufficient for our purposes. (d) Although experimental science shares many logical aspects with ordinary knowledge and other cognitive claims, assertions about truth and certainty must always be interpreted agreement with the different goals and corresponding patterns of each human business .

It seems to me that these qualifications can be accepted even by the most orthodox fallibilists. I think fallibilism would provide much more truthful explanations if it were separated from some of its polemical implications. Gerard Radnitzky has formulated some valuable clarifications in this respect [Radnitzky 1987, pp. 84-86], which are in line with my reflections. After all, we are co-authors of a study on the reliability of science [Agazzi-Artigas-Radnitzky 1986], and I have not found any difficulties in our common work .

Bibliography

• Agazzi, E. [1969] Temi e problemi di filosofia della fisica, Manfredi, Milano. 2nd ed., Abete, Rome, 1974.
• Agazzi, E. [1978] "Eine Deutung der wissenschaftlichen Objectivität", Allgemeine Zeitschrift für Philosophie, 3, pp. 20-47.
• Agazzi, E. [1985] "Commensurability, Incommensurability and Cumulativity in Scientific Knowledge", Erkenntnis, 22, pp. 51-77.
• Agazzi, E. [1988] "L'objectivité scientifique", in: E. Agazzi, ed., L'objectivité dans les différentes sciences, Éditions Universitaires Fribourg/Suisse, Fribourg, pp. 13-25.
• Agazzi, E. - Artigas, M. - Radnitzky, G. [1986] "The reliability of science", research y Ciencia, nº 122, pp. 66-74.
• Albert, H. [1968] Traktat über kritische Vernunft, J. C. B. Mohr (Paul Siebeck), Tübingen. 3rd ed. 1975.
• Albert, H. [1984] "Transcendental Realism and Rational Heuristics: Critical Rationalism and the Problem of Method", in: G. Andersson, ed, Rationality in Science and Politics, Reidel, Dordrecht, pp. 29-46.
• Albert, H. [1987], "Science and the Search for Truth", in: J. Agassi - I. C. Jarvie, eds., Rationality: The Critical View, Nijhoff, Dordrecht, pp. 69-82.
• Artigas, M. [1988] "Objectivité et fiabilité dans la science", in: E. Agazzi, ed., L ' objectivité dans les différentes sciences, Éditions Universitaires Fribourg/Suisse, Fribourg 1988, pp. 41-54.
• Artigas, M. [1989] Philosophy de la ciencia experimental, Eunsa, Pamplona.
• Artigas, M. [1991] "Science and Transcendence", in: E. Agazzi, ed., Science and Sagesse, Éditions Universitaires Fribourg/Suisse, Fribourg, pp. 87-101.
• Artigas, M. [1992 a] "Three Levels of Interaction between Science and Philosophy", in: C. Dilworth, ed.,Intelligibility in Science, Rodopi, Amsterdam, pp. 123-145.
• Artigas, M. [1992 b] La inteligibilidad de la naturaleza, Eunsa, Pamplona.
• Bartley III, W. W. [1962] The Retreat to Commitment, Knopf, New York. 2nd ed., Open Court, La Salle (Illinois)-London 1984.
• Bartley III, W. W. [1989] "Unfathomed Knowledge in a Bottle", in: F. d'Agostino - I. C. Jarvie, eds, Freedom and Rationality, Kluwer, Dordrecht, pp. 207-223.
• Freeman, E. - Skolimowski, H. [1974] "The Search for Objectivity in Peirce and Popper", in: P. A. Schilpp, ed, The Philosophy of Karl Popper, Open Court, La Salle (Illinois), pp. 464-519.
• Habermas, J. [1968] Erkenntnis und Interesse, Suhrkamp, Frankfurt. 4th ed., 1977.
• Popper, K. R. [1935] Logik der Forschung, Springer, Vienna. The Logic of Scientific Discovery, 14th printing, Unwin Hyman, Boston 1990.
• Popper, K. R. [1945] The Open Society and Its Enemies, Routledge, London, 1977.
• Popper, K. R. [1961], "Facts, Standards, and Truth: A Further Criticism of Relativism", in: The Open Society and Its Enemies, Routledge, London 1977, Addenda, I, pp. 369-396.
• Popper, K. R. [1963], Conjectures and Refutations: The Growth of Scientific Knowledge, Routledge, London, 1974.
• Popper, K. R. [1977] (with J. C. Eccles)The Self and Its Brain, Springer, Berlin - Heidelberg - New York.
• Radnitzky, G. [1987] Entre Wittgenstein et Popper. Détours vers la découverte: le vraie, le faux, l'hypothèse, Vrin, Paris.
• Radnitzky, G. - Bartley III, W. W., eds. [1987] Evolutionary Epistemology, Rationality, and the Sociology of Knowledge, Open Court, La Salle (Illinois).
• Vollmer, G. [1975] Evolutionäre Erkenntnistheorie, 3rd ed., Hirzel, Stuttgart, 1983.
• Vollmer, G. [1987] "On Supposed Circularities in an Empirically Oriented Epistemology", in: G. Radnitzky - W. W. Bartley III, eds, Evolutionary Epistemology, Rationality, and the Sociology of Knowledge, Open Court, La Salle (Illinois), pp. 163-200.
• Worrall, J. [1989] "Why both Popper and Watkins Fail to Solve the Problem of Induction", in: F. d'Agostino - I. C. Jarvie, eds, Freedom and Rationality, Kluwer, Dordrecht, pp. 257-296.