The most general thesis of fallibilism may be stated as (F) No beliefs are absolutely secure (Haack 1982:145). However, (F) may be supplemented by very different epistemological statements, among others: (a) anti-inductivism, (b) falsificationism, (c) the Duhem-Quine thesis (our beliefs are neither verifiable nor falsifiable), (d) conventionalism (our beliefs are based on certain conventions), (e) the view that all empirical statements are interpretative, (f) pragmatism, (g) relativism, (i) skepticism, (j) coherentism, and (k) sociologism. Each of (a)-(k) as well as other possible supplements to (F) explains that none of our beliefs are ever secure. According to (a)-(c), there are no conclusive procedures which might prove the absolute correctness of our beliefs. Conventionalism explains (F) by pointing out that, if we change some conventions, we also have to change the beliefs based on those conventions. Then, if all empirical statements are interpretative, they are correct relative to the assumed principles of interpretation; this conclusion is stated in (e). Furthermore, our beliefs are not secure because there are no absolute utilities as standards of epistemic evaluation (f) or because everything is relative (g), or because there is no knowledge at all (i), or because the coherence of systems of belief can not be established absolutely (j), or, yet again, because social contracts regarding the correctness of our beliefs may be changed (k).¹