Imogen Dickie's Fixing Reference articulates a novel and interesting thesis about reference-determination for certain of our mental states, focussing in particular on belief.Footnote ¹ Consider a subject who believes that Obama is president. This belief refers to, or is about, Obama. This is not a brute fact. So in virtue of what is the belief about Obama, rather than anything else, or even about nothing at all? Dickie's answer has two components.

The first component is a thesis Dickie calls Reference and Justification, introduced in chapter 2. To understand the thesis, we need two ideas. The first is Dickie's notion of proprietary justification for a belief. This is a form of justification that the belief's subject treats as overriding or trumping other forms of justification for contrary beliefs. Dickie offers three examples. (1) For perceptual demonstrative beliefs <F(that)>,Footnote ² the proprietary justification is uptake from a perceptual connection with an object. (2) For name-based beliefs <F(NN)>, the proprietary justification is careful uptake from testimony. (3) For description-based beliefs <F(DD)>, where ‘DD’ is a name introduced for the G, the proprietary justification is inference from <F(the G)>. The second idea we need to understand Reference and Justification is that of a body of belief, i.e. a collection of a given subject's beliefs. Reference and Justification concerns bodies of beliefs; it determines, for a given body, a referent for all the beliefs in it. Although Dickie is not explicit about how the relevant bodies are individuated, the idea appears to be that (a) each body contains beliefs of exactly one of the three kinds just mentioned, and (b) each body comprises all of the subject's beliefs involving the same perceptual demonstrative, or name, or descriptive name.Footnote ³ We can now state Reference and Justification. Roughly, it says: for any body B of beliefs and ordinary object o, the beliefs in B are about o iff, for any belief b in B, if the subject S has proprietary justification for b that renders b rational, then (i) S is not merely lucky that b is true, if o has the property predicated of an object in b, and (ii) S is unlucky that b is false, if o lacks the property predicated of an object in b.Footnote ⁴

The second component of Dickie's view is the subject of chapter 3: the thesis that we have a need to represent things outside of ourselves. By a ‘need’, Dickie means a kind of motivational, or goal-setting, mental state much like an intention. Needs differ from intentions only in that they lack conceptual content, in the sense that a subject can need to Φ despite lacking the concept <Φ>. (Example: a newborn infant may need to feed despite lacking the concept <feed> and hence being unable to articulate <feed>-involving thoughts.) Dickie hypothesises that we each have a need to represent objects in the external world. In chapter 3, Dickie also develops an account of justification on which a subject's behaviour, including belief-forming behaviour, can count as justified by virtue of being a skilful means towards satisfaction of a motivational state. Combining this account of justification with the need to represent external objects, Dickie obtains a novel account of the justification for certain of our beliefs: they're justified because forming them is a skilful means toward satisfying our need to represent external objects. I discuss this idea further below. Given Reference and Justification, this account of justification for the relevant beliefs is also an account of reference-determination for them.

Once this framework is in place, Dickie considers three applications of it: perceptual demonstrative belief (chapter 4); description-based belief (chapter 5); proper name-based belief (chapter 6). Chapter 7 continues the discussion of description-based thought, locating Dickie's approach relative to some rivals and arguing that it yields genuinely object-dependent description-based thought. Chapter 8 concludes by considering the role of subjective consciousness within Dickie's framework.

To my mind, the book articulates an interesting and promising approach to reference-determination. Dickie's discussion is nuanced, sophisticated, and full of interesting ideas. My main criticism is that it often feels unnecessarily difficult to figure out exactly what those ideas are, and how they all fit together. Dickie's formulations can feel somewhat longwinded and cumbersome, in a way that sometimes obscures the precise point she's trying to make. And I often found myself lost in the dialectic, wondering why this is being discussed here, and why it matters to the overall picture (though perhaps that merely reflects my own inadequacies rather than the book's). That's a shame because the book contains much of value.

I now discuss Dickie's account of reference-determination for perceptual demonstratives, focussing on her account of how perception justifies perceptual demonstrative belief. Chapter 3 explicates a notion of justification applicable to behaviour. The behaviour of a subject S is strongly justified by intention iff, for some Φ:Footnote ⁵

1. (A) S intends to Φ.

2. (B) That intention determines how S behaves.

3. (C) S is skilled in Φ-ing, so that the resulting behaviour is a reliable means towards S Φ-ing.

Chapter 4 uses strong justification in an account of how ordinary perception justifies perceptual demonstrative beliefs formed on the basis of it, e.g., my belief that that is a cup, looking at a cup-like entity on the table before me. Given Reference and Justification, this account of how perception justifies perceptual demonstrative beliefs is central to Dickie's account of reference-determination for them. I will argue, however, that Dickie slides unwarrantedly from strong justification of belief-forming behaviour to justification of the beliefs resulting from it. I close by suggesting a way to fix this problem, by adopting a richer conception of the mind's ‘need to represent things outside itself’ that Dickie claims motivates belief-formation in response to perception.

In chapter 4 (esp. 4.3), Dickie argues that (at least in typical, paradigm cases) perceptual demonstrative <F(that)>-beliefs are justified because they're strongly justified. However, Dickie isn't concerned with intentions at this point, but with another kind of motivational state she calls ‘needs’ (introduced above). So Dickie modifies the above definition of strong justification by intention as follows. The behaviour of a subject S is strongly justified by need iff, for some Φ:Footnote ⁶

1. (A) S needs to Φ.

2. (B) That need determines how S behaves.

3. (C) S is skilled at satisfying the need, so that S's behaviour is a reliable means towards S Φ-ing.

With a bit of simplification and omission of (for present purposes irrelevant) detail, the instances of (A)–(C) relevant to the justification of perceptual demonstrative <F(that)>-beliefs are:

1. (A*) S needs to represent external things.

2. (B*) That need determines that S forms a body of <F(that)>-beliefs in response to perceptual processing.

3. (C*) S is skilled at representing external things, so that S forming a body of <F(that)>-beliefs in response to perceptual processing is a reliable means towards S representing external things.

From (A*)–(C*) and the definition of strong justification by need, Dickie concludes:Footnote ⁷

1. (D*) S's body of <F(that)>-beliefs formed in response to perceptual processing are strongly justified because they are formed by S implementing a skilful strategy that reliably leads to fulfilment of S's need to represent external things.

However, (D*) does not follow. What follows is not that S's <F(that)>-beliefs themselves are strongly justified. It follows only that S's forming a body of beliefs in response to perceptual input is strongly justified. That process or activity of belief-formation in response to perceptual input is what's strongly justified by the need to represent, not the beliefs that result from it; at least, not without further argument. So there is a missing step in Dickie's argument. She needs an argument from (a) belief-forming behaviour that is a skilful means towards satisfaction of need, to (b) epistemic justification of the beliefs that result from that behaviour. What is that argument?

One suggestion is that I'm being overly literal here. Dickie's discussion is couched in terms of strong justification of a belief-forming activity. But we could perhaps reformulate in terms of strong justification of believing itself, rather than merely of an activity leading to belief. We could then replace (A*)–(C*) with:

1. (A!) S needs to represent external things.

2. (B!) That need determines that S believes <F(that)> in response to perceptual processing.

3. (C!) S is skilled at representing external things, so that S believing <F(that)> in response to perceptual processing is a reliable means towards S representing external things.

(A!)–(C!) and the definition of strong justification entail:

1. (D!) S's believing <F(that)> in response to perceptual processing is strongly justified because it results from S implementing a skilful strategy that reliably leads to fulfilment of S's need to represent external things.

We now have a valid argument that uses the definition of strong justification to conclude that S's perceptual demonstrative beliefs formed in response to perceptual input are (at least in typical cases) strongly justified. However, this modified approach over-generates justification.

Any old perceptual demonstrative belief formed in response to perceptual input can be seen as part of a skilful strategy for representing external things. Suppose there's an orange on the table before me, viewing conditions are normal, and no funny business is afoot. One way for me to satisfy my need to represent external things is by forming the perceptual demonstrative belief <Orange(that)>. A different way is by believing one of the following instead: <¬Orange(that)> or <Blue(that)>. Each of these beliefs represents an external thing, namely the orange to which my perceptual demonstrative <that> refers. Yet only the first should be justified by my perceiving the (orange) orange. Although the need to represent external things is satisfied by any belief I form about the orange, regardless of what properties I predicate of it, justified belief is more demanding. If skilful satisfaction of the need to represent external things sufficed for justification of the resulting beliefs, then any perceptual demonstrative belief formed on the basis of perception would count as justified, irrespective of whether one believed the object to be as perception presents it; justification would require only that one be skilled at forming beliefs in response to perception, not that one be skilled at matching character of belief with character of perception. This alternative account of how perception justifies perceptual demonstrative belief therefore over-generates justification.

A potential solution is close at hand: the need to represent external things is really the need to represent external things as they really are; it is the need to represent truly. This resolves the problem because satisfaction of this need in response to perceptual input does indeed require that how objects are perceptually presented matches how one believes them to be. What this suggests is that, if strong justification by need is to explain reference for perceptual demonstratives, the need to represent cannot be the need merely to be engaged in representation that Dickie's presentation suggests it to be. The goal of belief-formation in response to perception is not just to have any old cognitive perspective on what's out there. The goal is an accurate mental representation of our surroundings. Moreover, accurate mental representation occupies a practical role that mere representation of external things does not: actions guided by accurate representation are ceteris paribus more likely than those guided by mere representation to satisfy the agent's preferences and desires. For example, if I desire an orange and I represent that thing on the table – which, as it happens, is an orange – as an orange, I am ceteris paribus more likely to reach for it and thereby satisfy my desire, than if I represent it as something other than an orange. The practical utility of accurate representation may thus be used to explain why we possess a need to accurately represent, whereas it is unclear what practical use is served by a mere need to represent external things without regard for accuracy, or why we should possess such a thing.