![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
On the traditional relationalist conception of singular thought, a thought has singular content when it is based on an ‘information relation’ to its object. Recent work rejects relationalism and suggests that singular thoughts are distinguished from descriptive thoughts by their inferential role: only thoughts with singular content can be employed in ‘direct’ inferences, or in inferences that ‘trade on identity’. First, this view is insufficiently clear, because it conflates two distinct ideas—one about a kind of inference, the other about a kind of process that grounds inferences—under the title, ‘trading on identity’. Second, this leaves us without a notion that can be used as an alternative to relationalism about singular thought. The first notion is no more applicable to singular than to descriptive thought. The second might help us to better understand singular thought. However, it does so, not by replacing the view that singular thoughts are information-based, but by helping us to understand the nature of information-based thought.
Disclosure Statement
No potential conflict of interest was reported by the author.
Notes
1 There’s disagreement about which particular relations count (e.g. perception, testimony, etc.). See Evans [Citation1982], Bach [Citation1987], and Sawyer [Citation2012].
2 In a formal language, it’s encoded by the way that the system works: the rules say that sameness of syntactic type entails sameness of semantic type. For Campbell, it’s encoded by sameness of sense.
3 Note that, if we add a premise to Argument 2 stating that a = b, the argument will trade on the identity of the referents of tokens of ‘a’ and ‘b’. This illustrates that trading on identity is a presupposition of formally valid arguments involving repetition of singular terms across premises.
4 I’m not sure that this is what Campbell and others would say (they don’t discuss the case) but it helps to make sense of positing both notions.
5 Campbell [Citation1987: 279–80] talks of thinkers trading on identity.
6 Dickie uses ‘direct coordination’ to refer to the activity of bundling information to produce direct coordination relations, and ‘unity relations’ for direct coordination relations.
7 For Dickie (see also [Citation2020]), beliefs formed and bundled in this way give rise to cognitive focus, which brings particular objects before the mind in a relation of robust ‘aboutness’ (what singular thought involves).
8 Dickie calls non-singular thoughts ‘merely descriptive’ because she claims that some singular thoughts involve ‘description-based cognition’. This is her version of the claim that non-relational (in Bach’s sense) thoughts can be singular.
9 This shouldn’t be confused with the claim that singular thoughts can be ‘satisfactional’, a claim rejected by Dickie.
10 The claim that only singular thoughts can be used to trade on identity is not always explicitly discussed by MFC proponents, but emerges in remarks like Jeshion’s [Citation2010: 129] claim that descriptive, unlike singular, thoughts occur ‘discretely in cognition’.
11 Campbell’s discussion is packaged as about the role of non-descriptive sense, or ‘the notion of the sense of a singular term’ [Citation1987: 275, 278]. See also Dickie [Citation2015: 80] and Recanati [Citation2016: 5–7].
12 This satisfies our ground rule: Argument 3 doesn’t count, because its derivation lacks independently specifiable features.
13 Dickie agrees [Citation2015.: 244].
14 Note the restriction to ‘thinkers like us’. I won’t claim here that a ‘swampman’-style thinker (or one whose history and interaction with her environment were substantially different from ours) couldn’t have directly coordinated thoughts.
15 A ‘tracking process’, in my sense, need not be perceptual, or even causal/informational. A fully inferential process that works by deciding, because of a general belief that all Ø’s are F, that some property F applies to the object that is the unique Ø is a tracking process.
16 ‘Matching’ is used broadly, to include relations mediated by principles pairing different descriptive information.
17 It first appears in Pylyshyn and Storm [Citation1988].
18 A descriptive matching algorithm could also produce correspondences between different instances of descriptive information. For example, a process employing an algorithm applying the principle of looking at a later time for the object at the nearest location, and thereby matching descriptions specifying different locations, would be descriptive. See Pylyshyn [Citation2006: 224–5].
19 You might object that property discrimination must have some role in object-tracking: how else could we account for the kind of figure-from-ground discriminations that are presupposed? Pylyshyn’s claim is that property discrimination plays a merely causal, rather than computational, role [Citation2006: 218, Citation2007: 17, 33, 73]. It might be that the system’s states are responsive to certain properties, but that the tracking process doesn’t rely on encoding and employing property information in computations producing correspondences. Whether or not this is true, it’s coherent and doesn’t imply that the mind somehow tracks haecceities.
20 The crux is that human attention-scanning speeds would prevent the successful application of any descriptive procedure. Given the manner and speed of motion of objects in MOT experiments, a descriptive procedure would mix up objects, distractor objects ending up in locations producing the assumption that they were targets. See also Pylyshyn [Citation2006: 223–4, Citation2007: 36–7]
21 Recall that Dickie [Citation2015: 238] calls this an ‘information-marshaling strategy’.
22 This is inspired by Evans’s [Citation1982] idea of a controlling conception. See also Dickie’s [Citation2015] idea of a proprietary means of justification.
23 There is space for a still-weaker classification whereby singular concepts are those that involve, or permit, some degree of singular update.
24 Thanks to Imogen Dickie, Aidan Gray, Robin Jeshion, Josef Stern, Mark Sainsbury, Ken Taylor, Malte Willer, and audiences at the University of Texas at Austin, the University of Wisconsin at Milwaukee, and the University of Chicago, for discussion of the issues and questions. I also thank two anonymous referees for AJP for very helpful comments, and Stephen Hetherington for especially valuable editorial advice that helped me to shorten and improve the paper.