![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
Abstract
This article argues that inquirers should adopt an active orientation to the limits of their knowledge, an approach referred to as “Transformational Fallibilism”. Drawing on the Popperian tradition, this approach treats the fallibility of knowledge as more than a philosophical nicety, rather seeing the questioning of claims, including those that have been successful, as a key way to improve the understandings of inquirers. This is illustrated with reference to the example of Newtonian and Einsteinian understandings of gravity and time in the natural sciences, and debates about the role of fathers in child-rearing in the social sciences. I compare this approach to that found within critical realism, arguing that while defenders of the latter acknowledge fallibility to some extent, certain of their arguments also place problematic limits on it. Examining Sayer’s work, I argue that his view that practically successful understandings reveal something about the structure of the world involves an overconfidence in these understandings. I suggest that Elder-Vass’s use of the idea of approximate truth involves a similar difficulty. By contrast, I argue for the importance of treating all elements of the understanding, including those bound up with practical successes, as potentially in need of reconstruction at any point.
Acknowledgements
I would like to thank Christoforos Bouzanis and Nathan Coombs for their comments on versions of this paper.
Notes
1. This is my (fallible) recollection of his remarks.
2. The arguments in this article might have been presented in a different way, such that what I call “transformational fallibilism” here was argued to be the genuine content of any commitment to fallibilism. The argument would then be that others who claim to be fallibilists simply do not work through the implications of their commitment properly. This would have been a possible way of presenting the argument but given that it is common for adherents to fallibilism to hold a cautious, limited view of its consequences, it seemed clearer to me to identify the approach I am defending as one sub-category of fallibilism.
3. In making this argument I am not subscribing to a coherence theory of “truth”, rather looking at the relevance of coherence to processes of inquiry.
4. A similar concern is addressed by Popperians in their considerations about what kinds of developments are legitimate “scientific” moves to make when a theory is faced with an anomaly and is thus falsified (e.g. Popper Citation1959; Lakatos Citation1978). One possible Popperian treatment of this is that the replacement theory should have greater “content” than the theory which was falsified. I cannot go into the complexities of the issue here but suffice to say there has been much debate about this idea including what are, to my mind, telling criticisms of such an approach (see for example Holmwood and Stewart Citation1991; Bamford Citation1993). I should also note that I am not attempting to set up an idea of scientificity but rather looking to elucidate one way in which a set of understandings can plausibly be said to have undergone an improvement.
5. It’s important to be clear that in arguing that any element of the understanding may require revision I do not mean that an inquirer should be prepared to change the totality of their understandings in one go, as it is hard to see how a rationale for such a change could be given. One of the limitations of Kuhn’s analysis was his tendency to see competing approaches as self-enclosed and mutually exclusive totalities such that converting from one to the other involved something like a complete conversion from one set of theoretical and factual understandings to another (see Demir Citation2006).
6. It is worth noting that some philosophers have worked to “pick off” cases identified Laudan by questioning whether they were really “successful” (see for example Leplin Citation1997; Psillos Citation1999). I cannot properly address their arguments here, but I would comment that the criteria for success that these later philosophers use are often both highly restrictive (excluding very much apparently “scientific” knowledge down the ages from meeting these criteria, let alone any “non-scientific” knowledge) and also at a questionable distance from the criteria that seem to be have used by the knowledge-producers at the time.
7. Thanks to Sharani Osborn for suggesting this example.
8. Popenoe cites a wide range of studies, and it is possible to question whether all of them support his claims. Nevertheless, in the areas I have identified, it seems to me that at least some of the studies he cites are consistent with his arguments.
9. It might be wondered how the views defended here relate to Popper’s doubts about the idea of justification and his questioning of the giving of “positive evidence” for knowledge (e.g. Popper Citation1959). I would say that the approach I am taking is broadly Popperian in that I agree that knowledge claims are not able to be justified in the sense Popper often focuses on. That is to say, Popper rejects the idea that claims can be shown to be true or highly probable because of positive evidence in their favour, and I follow him on this. A more complex issue is whether I might differ with Popper on the extent to which positive support of any kind can be given to a knowledge claim. Quite what Popper’s views on this matter are have been discussed and disagreed upon in debates that address the related question of whether Popper is tacitly committed to induction in some form (for one important contribution see Lakatos Citation1978). All I can do here is give the reader a flavour of my own view: that “positive support” for a claim is not an intrinsically problematic notion, and indeed is part of reasonable argumentation, but that no matter how much “positive support” a claim has, this does not make it unquestionable.
10. In case it be thought that this is an anomalous formulation by Sayer, we can also consider his statement that “the usefulness [of a theory] is not accidental but due to the nature of the objects of knowledge” (Sayer Citation[1984] 1992, 70). See also his criticism of instrumentalism which states that “… it is the structure of the world, rather than our theories about it that make practices possible or impossible”. (Sayer Citation[1984] 1992, 71).
11. It might be considered illegitimate to draw parallels from Sayer’s example, which deals with practical everyday knowledge, in order to extrapolate conclusions about his treatment of scientific claims. However, I would argue that this is not the case. I am not aware of Sayer marking any fundamental gap between practical everyday knowledge and scientific claims. Furthermore, the views that I have cited are raised precisely in a critical discussion of alternative philosophies of science – conventionalism and instrumentalism. My sense is that Sayer chose a somewhat “everyday” example for the sake of simplicity rather than to distinguish between everyday practical action and scientific analysis in this regard.
12. I am not arguing that the implication of Sayer’s arguments about conventions and reality is that science is redundant. Even if my interpretation of this strand of Sayer’s argumentation is correct, Sayer could still see science as relevant for investigating other kinds of phenomena where successful conventions have not yet been established, or for generating new levels of precision in relation to known phenomena. My point is rather more specific: it is that the strand of Sayer’s approach which is less fallibilist in character implies that once success is established this tells us something about the world and thus encourages us to move science on to other areas or towards refining existing successful understandings rather than seeing these as up for a more fundamental reconceptualization and revision.
13. See Holmwood (Citation2001) for a related critique of Sayer’s ideas which argues that the latter’s defence of a distinction between system and lifeworld (Sayer Citation2000b) reifies the categories of analysis in a problematic way.
14. It’s worth acknowledging that there are variations on this within analytical epistemology, with some writers arguing for further conditions, and others wishing to dispense with the relevance of justification. For an interesting discussion of some of these positions see Williams (Citation2001).
15. An analogy would be as follows: if we are going to test out recipes by working out which of them results in a moist cake, we would make the cakes and then test them for their moistness. In such a case, we have to be able to know which cakes are moist and which are not.
16. I note here that Psillos’s analysis of approximation to truth is intended to avoid the difficulties found in that put forward by Popper (see Chapter 11 of Psillos Citation1999). It is perhaps also worth noting that it is questionable whether Popper’s own defence of the notion of verisimilitude is really consistent with his fallibilism. But this is too large an issue to address here.
17. It might seem puzzling that the approximate truth would be of interest if the truth was already available. Psillos is interested in approximate truth because it provides him a means to say that although a general theory may not be strictly true, it can be very close to the truth (the established facts of the matter). For example, Kepler’s first law is not strictly true because it does not capture planetary orbits absolutely accurately; nevertheless, for Psillos, it is approximately true because it is close to the truth about those orbits (Psillos Citation1999).
18. The enthusiast for the concept of truth might at this point say “Ha! Doesn’t this example show that you accept an idea of approximate truth in relation to cases such as the truth of the length of the table?!” My response would be that what I believe myself to be identifying about the table using my ruler is based within a (simplistic) framework of ideas about length. My sense of what is approximate and exact is within this framework. My understandings are unlikely to be compatible with the deepest considerations about length and measurement currently available in physics and philosophy, and nor do I see the current state of knowledge in those disciplines as the truth, the last word, on the question of length. Thus, I do not think I am capturing the truth of the length about the table by measuring it.
19. Holmwood and Stewart (Citation1991, 18, 19) make a similar argument in relation to Bhaskar’s analysis of theory change.