![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
As usually understood, ‘conceptual engineering’ is a form of conceptual inquiry aimed at diagnosing problems with extant concepts and finding better concepts to replace them. This can seem like an appropriate response to a skeptical concern that our concepts are cognitively deficient: unsuitable for use in serious inquiry. We argue, however, that conceptual engineering, so understood, cannot reasonably be motivated in this way. The basic problem is that on the first hand, since conceptual engineering is itself a form of inquiry, it cannot succeed by using the problematic concept itself in inquiry (since it is unsuitable for use in inquiry); but, on the other hand, methods for carrying out inquiry directed at concepts without using those concepts are constrained in such a way as to make conceptual engineering very unlikely to succeed. The upshot is that conceptual engineering has no reasonable chance of addressing the skeptical concern about cognitive deficiency. This is an important and previously unarticulated result, about what conceptual engineering can and cannot reasonably be expected to do.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 For some key texts on conceptual engineering, see Haslanger (Citation2012); Cappelen (Citation2018); Burgess et al. (Citation2019). We describe conceptual engineering more exactly in Section 3 below.
2 Thus the ‘mere’ in Jennifer Nado’s taxonomy is misleading: ‘Conceptual engineering is a recently (re-)popularized methodological approach which aims at the improvement, rather than the mere analysis, of our current conceptual repertoire’ (Citation2023, 1). Conceptual analysts aim at improvement, too: improving our judgments about the world itself, through the conceptual clarity afforded by conceptual analysis.
3 See Cappelen (Citation2019, §9.2), Sawyer (Citation2020), Belleri (Citationforthcoming), and Queloz (Citation2022), who sees Strawson’s Challenge as one form of a broader ‘authority challenge’.
4 See, e.g. Machery (Citation2021, 7) for discussion.
5 The point of this is not to convince the reader that conceptual analysis is cognitively valuable, but rather to give a concrete comparator that, at the same time, vies as an alternative interpretation of how some conceptual inquiry currently described as “conceptual engineering” is or could be cognitively valuable. Conceptual analysis thus works both to give clearer sense to the question of how conceptual inquiry could be cognitively valuable and as part of some possible error-theoretic explanation of how a form of conceptual inquiry which might appear to have some claim to be being conceptual engineering could still be cognitively valuable: because it is actually more plausibly a case of conceptual analysis.
6 Cappelen is talking about Frege of the Begriffsschrift, while we cite Frege’s Foundations of Arithmetic. Though we don’t think Frege has changed his mind between Begriffsschrift and the Foundations, our concern is not really with Frege scholarship. We use Frege and Carnap only to provide a familiar historical ground from which to articulate a useful distinction between two conceptions of conceptual inquiry.
7 In a related but distinct context, Frege acknowledges the lack of ‘new truths’ in his work, but counters that methodological developments like his ‘also further science’ ([Citation1879] 1967, 6). At least part of what he has in mind is the improvement that increases the quality of our knowledge of old truths.
8 For a recent overview of views of conceptual analysis, see Kölbel (Citation2021). Kölbel tries to construct a dilemma for how to understand the methodology of conceptual analysis but far as we can tell, the account of conceptual analysis we give avoids his dilemma. For a different puzzle for which the account given here is also a solution, see Rattan (Citation2016).
9 We rely here on the intuitiveness of this point, but there is also a principled argument why there is no need for the salient additional premise, ‘<C> applies to x iff x is C’: an argument which builds off the regress worries that have led theorists to foreground the importance of trading on identity (Campbell Citation2002) or semantic coordination or representation as the same (Fine Citation2007). To require the additional premise in the inference from ‘<C> applies to x’ to ‘x is C’ also launches a regress, avoiding which depends on a coordination relation between the singular concept referring to the concept, and the concept itself. This relation is a conceptual sameness rather than a de jure coreference between <C> and C, extending the standard semantic coordination framework to include relata of different logical types. For related discussion, see, e.g. Rattan (Citation2019).
10 As Cappelen puts it, conceptual engineering ‘is the project of assessing and then ameliorating our concepts’ (Citation2019, 132). David Chalmers notes that ‘[m]any or most of the standard examples in the recent conceptual engineering literature are cases … [in which] you somehow fix or replace the old concept, so the old concept is no longer around’ (Citation2020, 6, 13). The survey article by Isaac, Koch, and Nefdt (Citation2022) takes it that ‘conceptual engineering is a branch of philosophy concerned with the process of assessing and improving our concepts’.
11 See, e.g. Cappelen (Citation2018), Flocke (Citation2020) and Nado (Citation2023). As Nado notes, ‘the exact relation classification procedures have to [concepts] is … left almost entirely open’, in her work (2).
12 As Stalnaker (Citation1984, x) puts it, ‘The goal of inquiry is to find out how the world is’; and as Friedman (Citationforthcoming, §2) notes, successful inquiry is epistemically responsible: ‘traditional epistemic norms speak to … what [we] should do over the course of inquiry’.
13 When Amia Srinivasan suggests that some conceptual engineers function more as prophets than theorists – arriving at better concepts by ‘simply see[ing] the world as no one else (yet) sees it’ (Citation2019, 150), – she may have in mind a conceptual engineering project pursued without inquiry.
14 Actually, things are a little bit more complicated, though not in a way that affects our argument. It is well known that false assertions can be provisionally made in inquiry for the purposes of deriving something absurd, thereby showing the falsity of the asserted claim: this is the reductio ad absurdum form of reasoning. Analogously, we should be able to provisionally use any concept, no matter what is wrong with it, to derive some absurd result, thereby gaining evidence that something is seriously wrong with that concept. So, we must qualify the description of cognitive deficiency: strictly speaking, a cognitively deficient concept is one which renders any inquiry in which it is used irresponsible unless it appears only for this limited purpose, in a reductio-style context. We will typically leave this qualification off, since it does not affect the main line of argument. (Footnote 20 below explains, at the point when it might seem to become relevant, why this is.)
15 For example: Cappelen typically associates essentially these motivations for conceptual engineering with Carnap and with Chalmers and Clark, respectively (see Citation2018, §§ 2.1.1–2.1.2 and Citation2019, 137–139).
16 See Chalmers (Citation2020), and for a similar point, Deutsch (Citation2020). One might wonder: which features might make a concept cognitively deficient? It is worth reviewing some possibilities, though to endorse a particular answer would require committing to a particular theory of concepts, which we should not do. Among the ‘fairly rare’ cases in which a concept really is ‘better gone’, Chalmers mentions ‘an inconsistent concept’, which, according to Kevin Scharp, is one such that, if we use it, we can ‘reason[] our way to contradiction’ by valid inferential steps (Citation2020, 399). This seems like a reasonable candidate for cognitive deficiency, since it is plausibly epistemically irresponsible to believe any conclusions reached on the basis of inquiry involving such concepts, given that those results would be contradicted by others reached through equally valid inferences. The same holds for the focus in Machery (Citation2021) on ‘unreliable’ concepts, ‘in the sense that they often lead to false conclusions from true premises’. Use of such a concept is epistemically irresponsible.
17 E.g. Nietzsche’s claim that we ‘must no longer accept concepts as a gift’, because any such concepts ‘are … the inheritance from our most remote, most foolish as well as most intelligent ancestors’ (Citation2019, 1). Whether Nietzsche is really expressing such skepticism is a separate question. As a statement about Nietzsche’s views about how genealogical considerations function and what they target, we think that this reading has Nietzsche acquiescing in metaphysical and semantical presuppositions that would have been foreign to his most mature thinking on these topics. For discussion see Clark (Citation1990, Chapter 4) and Williams (Citation2002, 12–19).
18 For the first approach, see, for example, Descartes (1641); for the latter, see, for example, Williamson (Citation2004).
19 A classic description, in linguistic form, comes from Grice (Citation1958) 1989, 172): ‘If I philosophize about the notion of cause … I expect to find myself considering … in what sort of situations we should, in our ordinary talk, be willing to speak (or again be unwilling to speak) of something as causing something else to happen’. See Thomasson (Citation2007) on the central place of this method in both analytic philosophy and phenomenology.
20 We acknowledged (in footnote 14 above) that strictly speaking, cognitively deficient concepts could be used in special reductio-style contexts, for the limited purpose of gaining evidence that something is wrong with the concept. This qualification does not interfere with our point here, since as standardly practiced, the above methods employ these concepts in normal contexts. Indeed, use in reductio-style contexts alone could not give the kind of diagnostic information that the engineer needs: not just that the concept is bad, but what is bad about it and what is nevertheless good.
21 Like an anthropologist and historian, such an inquirer would use metarepresentational concepts in her thinking about the target concept. Since the engineer does actually grasp the target concept itself, these could be transparent metarepresentations. But since she cannot engage in the inferential transitions with first-order uses that such representations licence, her inquiry is still from the external perspective. Her transparent metarepresentations eschew the cognitive value transparency affords and so are, with respect to cognitive value, like opaque and not transparent metarepresentations.
22 For example: suppose again that concepts can be cognitively deficient because inconsistent, in Scharp’s sense that we can ‘reason[] our way to contradiction’ using them. (See footnote 12 above.) If subjects appear to reason their way to contradiction using a concept, this might seem like evidence for a diagnosis of the concept’s inconsistency. But since this interpretation would rule out that the subjects are epistemically responsible when using this concept in beliefs to support other beliefs in inquiry, it would be more charitable to offer an alternative interpretation: say, that though the concept itself does not license these inferences, some subjects falsely believe that it does. This interpretation preserves the subjects’ epistemic responsibility in their use of the concept while also making sense of their error.
23 See, for example, Chapter 8 of (Williamson Citation2007), for the point about knowledge.
24 Daniel Garber explicitly links the proper methodology for the history of philosophy, which incorporates ‘the principle of charity’ (352), to the methodology of anthropology (Citation2013, 361). Garber’s disputes with Bennett, like other major disagreements about charity in this discipline, do not concern whether to be charitable, but about whether proper charitable interpretation involves maximizing truth, reasonableness, knowledge, or what. For discussion of Garber and Bennett’s differences, see Lin (Citation2013).
25 Rorty et al. make this point about anthropology, aiming to draw morals for the history of philosophy: ‘The anthropologist is not doing his job if he merely offers to teach us how to bicker with his favorite tribe, how to be initiated into their rituals, etc. What we want to be told is whether that tribe has anything interesting to tell us’ (Citation1984, 6). Anthropologist Tim Ingold sees ‘participant observation as the key to the practice of anthropology’, which allows the anthropologist to avoid ‘the great mistake [of] confus[ing] observation with objectification’. For Ingold, ‘to practice anthropology … means to study with people, not to make studies of them … We do so in order that we may grow in wisdom and maturity, in our powers of observation, reason, and critical thinking, in the hope and expectation that we can bring these powers to bear on whatever problems we may tackle in the future’ (Citation2017, 23).
26 As Sawyer (Citation2020) suggests, the distinction between concepts and conceptions can help here: engineering projects don’t replace and ameliorate concepts, but conceptions. Ball’s (Citation2019) temporal externalism is also congenial to the view described here. Note that recasting things in this way involves undertaking some commitments about concepts – namely, that revising even deeply held conceptions or inferential patterns is consistent with grasping the same concept. This runs against certain conceptual role or inferentialist views that make certain beliefs or inferences constitutive of grasp of a certain concept.
27 Sally Haslanger, for example, initially suggests that the concept associated with the word ‘woman’ is deficient from the point of view of the pursuit of justice (see, e.g. Haslanger Citation2000). Later, however, observing that her ‘own discussion of these issues … has been confused and confusing’ (Citation2010, 429), she considers reinterpreting it along the lines we suggest, noting that ‘our meanings are not transparent to us: often ideology interferes with an understanding of the true workings of our conceptual framework and our language’ (Citation2006, 383).
28 Thanks to participants at the Conceptual Engineering Workshop at the University of Toronto in the fall of 2017, the mid-Atlantic Philosophy of Language Conference in the summer of 2018, and the University of Toronto conceptual engineering reading group that met online in the winter of 2021, for questions, comments and discussion. Thanks as well to two anonymous referees from Inquiry for comments that helped clarify and improve the paper.