However, Popper's falsificationism, according to Roy Bhaskar, cannot solve the problem of an "implicit _ceteris paribus_ clause" rationally:
***** Both Lakatos and Kuhn agreed, however, that the Popperian system could not, at least in its original form, stand up to the material provided by history. (Popper, Lakatos grumbled, 'does not raise, let alone answer the question: "under what conditions would you give up your demarcation criterion?"') Every theory was always immersed in an 'ocean of anomalies'; so that, strictly speaking, every theory was always falsified. In this context actual scientists had to be much more dogmatic, or tenacious than the Popperian model allowed. Moreover, as Duhem had pointed out, every theory was formulated subject to an implicit _ceteris paribus_ clause, so that the hypothesis of an intervening or disturbing influence could always be invoked to explain away apparent counter-instances. Conditions, therefore, could not be specified in advance as to when a theory should be given up, just because of the possibility of the implicit _ceteris paribus_ clause breaking down.... Then again in real history falsifications never issued from a simple dyadic confrontation between a _single_ theory and a set of facts; but between _two_ (or more) theories and their facts; that is, in real history falsifications were _replacements_. And the replacement, when it came, normally consisted in a refinement and modification of the existing theory, rather than its complete rejection. The original Popperian model had left a mystery: after the refutation -- what? Or to put it another way it could not account for the genesis of any new conjecture or research line. In real history scientific theories do not spring from the void -- but from the development and reworking of cognitive material that pre-exists them, necessitating the creative employment of ideas from adjacent fields, Bachelard's 'scientific loans'. (Roy Bhaskar, _Reclaiming Reality: A Critical Introduction to Contemporary Philosophy_ (NY: Verso, 1989), p. 31) *****
On the other hand, neither Kuhn nor Feyerabend manages to sustain the intelligibility of over _what_ incommensurable descriptions clash, thus falling into super-idealism. Nor has Lakatos shown, Bhaskar argues, "how unless nature were uniform, it would be rational to work on progressive rather than degenerating programmes" (p. 11).
Bhaskar writes that theories of scientific change have failed to make "a distinction between causal laws and empirical invariances (patterns of events)" implied by "experimental activity" (p. 39). The problem of the "implicit _ceteris paribus_ clause," in Bhaskar's view, results from a "_fusion_ of the world and experience (a conflation of the ontological and epistemic orders)," to which both Popper and his critics are committed (p. 38). In other words, a distinction must be drawn between "the _real_ structures, mechanisms and systems of relations at work in nature (and society), providing the ontological basis of causal laws, and the manifest (or _actual_) patterns of events they generate, whether humans are causal agents in their productions or not" (p. 40).
Yoshie