Imre Lakatos: A Critical Appraisal

3. History of Science and Its Rational Reconstructions

3.3 Improving Lakatos' Meta-criterion

In the light of my criticisms advanced in the previous section, I suggest the following revision and further explication of Lakatos' criteria for a progressive theory of rationality. I propose that the theory of rationality must be:

  1. 'internally' empirically progressive by either postdicting or novelly deriving[7] corroborated 'basic value judgements' of scientists from known historical circumstances, or by successfully postdicting other types of novel historical facts from known 'basic value judgements' of scientists, in the process of writing the history of science as the application of method, and
  2. 'externally' empirically progressive by successfully postdicting novel historical facts in its explanation of the seeming inconsistency between the expected 'basic value judgements' of scientists with the application of method and the actual 'basic value judgements'.

I am unsure about what relative weight to give to these two conditions when judging rival theories of rationality. This will not normally be a problem, for usually one methodology will be far superior 'internally' and 'externally' compared with its rivals, although there will be difficulties in assessment with the appearance of a new powerful theory. This is a detail that will need to be worked out at some later time.

Book cover: Philosophy of Science: A Very Short Introduction by Samir Okasha

My revised criteria have completely eliminated the need for Lakatos' troublesome condition 1. His condition 2. is now encapsulated under both my conditions i) and ii). It will be served by i) if the new 'basic value judgement' is shown to be irrational by empirically progressive developments in 'internal' historiography and by ii) if it is shown to be irrational by empirically progressive developments in 'external' historiography. Condition 3. is obviously superseded, once again, by my conditions i) and ii). Condition 4. is superseded by that part of my condition i) that refers to the novel derivation or postdiction of corroborated 'basic value judgements'.[8]

In this revised form, I think that the MHRP is of value in judging rival theories of rationality, although I doubt its persuasive force. This is because the arguments in its favour are a meta-level application of the MSRP and so will not be persuasive to those not already partial to this particular methodology. The first argument, that the MHRP recognizes that a rational reconstruction of the history of science can never be complete because of human fallibility, although derived from the MSRP's notion of progress in an ocean of anomalies, has independent weight. However, the second argument, that the MHRP can judge progress by the extent to which a methodology successfully postdicts novel facts, will only have weight for those methodologists that already recognize the epistemic worth of corroborated novel facts. This is a weakness of Lakatos' meta-level defence of the MSRP.

It is at this point that I want to correct a misinterpretation of Lakatos' second thesis. It is not too difficult to be misled by Lakatos' dialectical presentation into thinking that he argued that one test of a methodology is to see how adequately it rationally reconstructs the history of science using itself as a meta-criterion of adequacy, and that on this test of coherence the MSRP passes while inductivism, conventionalism and falsificationism fail.[9] However, a more careful reading will show that no such coherence test was advocated by Lakatos, either in his [1978a: ch. 2] or in his earlier version [1978a: ch. 3].

His point was not to test the adequacy of each rational reconstruction with its own meta-methodology, but to propose a new meta-criterion for judging the adequacy of rational reconstructions of history. To do this, he adopted a dialectical approach by first showing the inadequacies of meta-falsificationism by applying it in practice, and then by proposing a better historiographical method of criticism; namely the MHRP. Lakatos wrote:

I shall try to develop this historiograpnical method of criticism in a dialectical way. I shall start with a special case: I first 'refute' falsificationism by 'applying' falsificationism (on a normative historiographical meta-level) to itself. Then I shall apply falsificationism also to inductivism and conventionalism, and, indeed, argue that all methodologies are bound to end up 'falsified' with the help of this Pyrrhonian machine de guerre. Finally, I shall 'apply' not falsificationism but the methodology of scientific research programmes (again on a normative-historiographical meta-level) to inductivism, conventionalism, falsificationism and to itself, and show that —on this criterion— methodologies can be constructively criticized and compared. This normative-historiographical version of the methodology of scientific research programmes supplies a general theory of how to compare rival logics of discovery in which (in a sense carefully to be specified) history may be seen as a 'test' of its rational reconstructions.

[Lakatos 1978a: 122f]

This is a clear, concise and precise summary of Lakatos' intention. We have here not a single word of 'coherence tests'. For Lakatos, it is the MHRP that is to supply the general theory of how to test comparatively methodologies against history. That Lakatos applies meta-falsificationism and the MHRP to other than their object-level counterparts also belies the notion that he had any coherence test in mind.

Lakatos' dialectical approach is further explained in his earlier version. Referring to his meta-falsificationist 'falsification' of falsificationism, he wrote, 'I only chose this Socratic-Popperian dialectical way of developing my position because I think this is the best way of developing a complex argument . . . ' [Lakatos 1978a: 148] No mention here of Popper's fatal incoherence. In fact, Lakatos explicitly stated that his meta-falsificationist 'falsification' of falsificationism was no reason to reject falsificationism: 'Popper's theory of scientific rationality need not be rejected simply because it is "falsified" by some actual "basic judgements" of leading scientists.'[Lakatos 1978a: 132][10] This is because it is an inadequate meta-criterion and there are good arguments for a better meta-criterion, namely the MHRP, Lakatos explained. My key aim here has been to dispel this notion that Lakatos was presenting a general method of testing methodologies for coherence.

Although Lakatos did not use such a coherence test on theories of rationality,[11] it seems that such a test as advocated by Fox [1981: 100] and mistakenly attributed to Lakatos is warranted. The problem of coherence is an acute one for monistic methodologies such as the MSRP.[12] If a theory of rationality posits that a theory is rationally justifiable iff it satisfies the demands of a particular set of methodological criteria, then what is the epistemic status of this theory of rationality? In order to answer this question, methodological monists have no choice but to apply their own methodological criteria to the methodological criteria themselves. The alternative non-monist methodologies are as seriously burdened with the unavoidable charge of arbitrary selection of methodological criteria.

So, I propose two separate tests for monistic methodologies qua rational reconstructions; a test of adequacy of the rational reconstruction as judged by the MHRP (as given above) and a more general test of coherence. As explained by Lakatos, we cannot judge the adequacy of a falsificationist rational reconstruction, once falsificationism is put into a monistic form, by a meta-falsificationist standard, just as we cannot justifiably test any rational reconstruction by this meta-criterion. (We must use the MHRP for this purpose.) But we can use meta-falsificationism to test falsificationism for coherence. On this test, falsificationism, as with some or most other methodologies in monistic form, fails. The MSRP, however, as Lakatos had argued, is the most progressive historiographical research programme and hence passes the test of the adequacy of rational reconstructions and the coherence test.

My early impression of Lakatos' [1978a: ch. 2] essay on the rational reconstruction of the history of science was that it was of little value. However, on closer scrutiny, I have come to regard it as being of some considerable merit. In this discussion, I have outlined my reasons for this change in my appraisal. In some important respects, though, it is still deficient. The testing of methodologies against the history of science is dependent on a prima facie case for the rationality of science and such dependence serves to weaken the force of any such test. (The same can be said for the above coherence test.)

Secondly, the arguments for Lakatos' meta-criterion and my revised version partly presuppose the cogency of the MSRP as a theory of rationality. In this respect, it is of limited effect in independently supporting the MSRP. It is for these reasons that I develop independent arguments for the MSRP in my Allan [2016b]. I do think that once Lakatos' criteria for rational theory appraisal are shown to fall out of an analysis of the requirements of an objectivist epistemology, his MSRP is on solid ground.


  1. [7] By 'postdicting' or 'novelly deriving' a fact I mean deriving a statement from the theory that was known to be true to the constructors or revisers of the theory during construction or revision, but was not used to construct or revise the theory.
  2. [8] Worrall [1976: 164–68] has produced a valuable clarification and revision of Lakatos' meta-criterion. In the last paragraph of his discussion [1976: 168], he briefly mentioned that on his account a better methodology may rationally reconstruct fewer 'basic value judgements' of scientists. However, he did not develop this idea as I have done here.
  3. [9] See, for example, John Fox's review of Lakatos [1981: 100].
  4. [10] See also Lakatos [1978a: 151]. Here, Lakatos' text is in agreement with Fox's complaint [Fox 1981: 100] that Lakatos has not falsified falsificationism. Fox's misunderstanding on this point and on Lakatos' main thesis may be due to his falsificationist sensitivities.
  5. [11] It must be pointed out that Lakatos here did not criticize Popper's falsificationism for incoherence, but because he 'never offered a theory of rational criticism of consistent conventions'. See Lakatos [1978a: 123].
  6. [12] For a brief discussion of the problem of coherence for epistemologies, see Fox [1981: 96].

Copyright © 2016

You will be interested in

Share This

  • twitter
  • facebook
  • linkedin
  • googleplus
  • gmail
  • delicious
  • reddit
  • digg
  • newsvine
  • posterous
  • friendfeed
  • googlebookmarks
  • yahoobookmarks
  • yahoobuzz
  • orkut
  • stumbleupon
  • diigo
  • mixx
  • technorati
  • netvibes
  • myspace
  • slashdot
  • blogger
  • tumblr
  • email
Short URL:


PDF Download Imre Lakatos: A Critical Appraisal

Download this essay

PDF Download Imre Lakatos and a Theory of Rationality

Download complete essay on Imre Lakatos and a Theory of Rationality