Where exactly is the ‘real’ in critical realism? Plus, a Dewey-James alternative

Click to increase image sizeClick to decrease image size

Acknowledgements

I wish to thank the Journal of Critical Realism for the invitation to respond to Elder-Vass’ thoughtful reflections and for a chance for a fruitful exchange. Comments from and conversations with Chris Winship were also essential.

Notes

1 See Patterson [1982] 2019 for a rich comparative history.

2 To be clear, there is a connection between Habermas and Pragmatism, especially Pierce (see Habermas 1992). The concept of the trans-historical ‘community of inquirers’ from Pierce was important to Habermas especially, but the Pragmatism from James-Dewey explicitly historicizes inquiry.

3 Westbrook (in Dickstein 1998, 128) recounts, ‘ … over the last several years historians such as Kloppenberg and I have found ourselves participating regularly with Rorty in conferences and symposia in which our role is to say to him, often repeatedly, ‘Gee, that argument that you say that you and Dewey make is very provocative, but Dewey never made it and I do not believe he ever would make it since it is at odds with arguments he did make.’ Rorty then shrugs his shoulders and acknowledges genially that the Dewey he is talking about is one of his ‘imaginary playmates’ a ‘hypothetical Dewey’ who says the sort of things Dewey would have said had he made the ‘linguistic turn’ … ’

4 An example Rorty (see especially Rorty 1998) uses in several pieces is abortion. The ‘Kantian’ says that human life must always be treated as an end, not a means. The fetus, inasmuch as it has the potential to be a human, must be respected and therefore we should outlaw abortion. A ‘rights liberal’ argues that bodily autonomy is the most fundamental right and thus a woman’s right to choose cannot be infringed upon. The ‘utilitarian’ argues that society in the aggregate benefits from families who can plan. So, what philosophical method allows us to adjudicate between the duties, rights, or utilities? Rorty says none. So, we come up with some solution where the fetus is human after the second trimester which sort of makes the Kantian happy, sort of makes the Liberal happy, and the utilitarian is probably satisfied. Is there some fundamental ontology about what it means to be human that we are appealing to? None exist; and thus at best we can come up with a compromise that democratically ‘works’ for everyone.

5 Furthermore, Rorty characterizes his position as a 'Peircelike claim' which parallels Habermas’ that 'our criteria for truth is one which refers to “undistorted communication”, but I do not think there is much to be said about what counts as “undistorted” except “the sort you get when you have democratic political institutions and the conditions for making these institutions function” (Rorty 1989. p. 86).' Thus, unlike Habermas as noted here, Rorty does not have a systematic theory of human cognition and communication and instead appeals to practices in existing bourgeoise liberal societies. This 'realist Habermas' position again strikes me as quite similar to Elder-Vass’.

6 See also MacIntyre (1981) on this point, that is, the gap between social practices and the language of moral values.

7 Rorty in The Consequences of Pragmatism (1982) is his perhaps more direct formulation of this idea, ‘To accept the contingency of starting-points is to accept our inheritance from, and our conversation with, our fellow-humans as our only source of guidance. In the end, the pragmatists tell us, what matters is our loyalty to other human beings clinging together against the dark, not our hope of (Certainty)’ (Rorty 1982, 166).

8 That combination, of fallibilism and anti-skepticism, is according to Putnam, ‘the unique insight of American pragmatism’.

9 Rorty’s self-characterization as a Deweyan is one source of this misunderstanding. Furthermore, Dewey was a prolific writer throughout his long life and thus one can find any number of quasi-contextual quotes to the effect of ‘the community makes truth’. However, Kloppenberg (in Dickstein 1998) argues that Rorty is certainly not a Pragmatist on precisely these grounds: linguistic agreement is Rorty’s basis for truth, while for Dewey, experience is the foundational category.

10 See especially Marcellesi (2013) on this point.

11 I often find the work described as positivist by some CR scholars as actually quite attentive to emergent entities (multi-level modeling for instance), attentive to complex causal mechanisms (see debates over frontdoor and backdoor criteria in causal inference designs or the specifics of ‘the causes of effects’ relative to the ‘effects of causes’), and computational modeling in the vein of either exponential random graph models or agent based models is explicitly motivated by the idea that multiple layers of social reality interact with one another in a non-linear fashion.

12 This distinction is analogous to difference between the use of gradient descent algorithms to compute optimizations of many machine learning algorithms and the closed-form analytic solutions in traditional ordinary least squares. When used in a typical social science setting, a linear regression has a single solution to minimize the error term; but finding a similar solution in the case of many machine learning applications – when the data can be several orders of magnitude greater in terms of both scale and dimensions – is computationally prohibitive. Consequently, varieties of gradient descent algorithms are employed which iteratively minimize the error term by finding local minimums. Likewise, the pursuit of a single correct recipe of social science is the pursuit of a linear solution, but as one encounters many more dimensions of data, the Pragmatist urges partial experimentation – short descents down the curve.

13 A Pragmatist might suggest that this is the case because we are reacting to the same empirical problems.