In Dear Prudence, Guy Fletcher explores the neglected topic of the meta-normativity of prudence. The main theme of the book is that many meta-normative views about prudence are structural counterparts to more familiar meta-ethical views about morality. Because of this similarity, Fletcher claims that importing some well-worn moves from the study of meta-morality can shed light on the brand-new study of meta-prudence, while at the same time studying meta-prudence can shed new light on old debates in meta-morality. Those interested in theoretical issues concerning prudence or in meta-normativity outside the confines of morality would be well served by studying this book.
The main aim of Dear Prudence is a ‘meta’ one: to defend the propriety and usefulness of studying the meta-normativity of prudence. Fletcher's primary opponents, therefore, are not so much those with particular views about prudence, but rather those who think prudence is meta-normatively uninteresting. United on that front are both philosophers who hold that prudence is unproblematically normative as well as those who hold that it is obviously non-normative. Sometimes Fletcher's opponents make such an assumption fairly explicitly, but more often than not they make the cagey move of implicitly committing to such an assumption by not discussing meta-prudence at all.
Thus, Fletcher makes an initial foray into the meta-normativity of prudence by canvassing some important questions that have analogs in the meta-ethics of morality. Those questions are:
1. Are there normative prudential reasons?
2. Are prudential judgements normative judgements?
3. Are these judgements motivationally special?
4. Is prudential discourse context-sensitive in significant ways?
5. Does prudential discourse involve the ascription of normative properties (in either a deflationary sense or a robust sense) and, if robust, what is the nature of prudential properties?
6. Is prudential talk partly (or wholly) expressive? (p. 3)
In light of this array of questions, don't be fooled by the title of the book, nor by Fletcher's rhetoric that he aims to give “a theory of various aspects of prudential discourse” (p. 4). Only 2, 3, 4, 5, and 6 are about prudential discourse per se, and even 5 appeals to discourse as a roundabout way of getting at the nature of normative properties. Thus, much of Fletcher's substantive theory goes beyond the language of prudence and into its metaphysics and epistemology.
While Fletcher's primary aim is to establish that it's worthwhile to ask these questions, his subsidiary aim is to defend particular answers to them. To that end, Fletcher explores 1 in the first chapter, where he defends the claim that there are indeed prudential reasons. In the second chapter, he defends an affirmative answer to 2. And in the third and fourth chapters, Fletcher answers 4 by arguing that prudential language features only a minimal form of context-sensitivity. As an answer to both 3 and 6, he defends, in the fifth chapter, a limited form of prudential judgement internalism. As can be seen, the overall thrust of these chapters is that the true meta-prudential theory looks much like a realist cognitivism familiar from the meta-ethics of morality. In the sixth chapter, Fletcher uses this realist view to undermine some familiar arguments for global error theory in a way that bears on 1. Finally, in the seventh chapter, he addresses 5 not so much by answering it but by defending the utility of a certain framing of the question inspired by debates in meta-morality.
To my mind, Fletcher is completely successful in accomplishing his primary task for the book. He does an admirable job laying out some of the major questions for the meta-normativity of prudence, and he displays how to fruitfully approach such questions in a manner familiar from the meta-ethics of morality. Moreover, he convincingly shows several specific ways in which taking the meta-normativity of prudence seriously can help with the meta-ethics of morality.
However, despite these merits, I think there are some shortcomings in the depth of his arguments and the breadth of conceptual resources on offer. This is understandable, however, in that Fletcher's main goal in this slender book is the ‘meta’ one of simply showing that exploring meta-prudence is important, as opposed to definitively defending a more substantive view. In the remainder of this review, I discuss the three areas where I find his ‘meta’ discussion most plausible and significant, yet where I also had most reservations about the details.
The first area concerns Fletcher's defense of the existence prudential reasons. In particular, Fletcher defends the thesis of Prudential Value Matters: “Evaluative prudential facts generate directive prudential facts (including facts about prudential normative reasons for action and for attitudes)” (p. 13). His primary positive argument, sensibly enough, is that this claim coheres with our intuitions – and with that I cannot agree more. Fletcher's main task, however, is defensive, and he recognizes that one of the strongest challenges to this thesis stems from Humeanism about reasons. Humeanism, in the broad, holds that reasons are only given by desires – and therefore they are not given by prudential value as such.
Fletcher rejects Humeanism, but he only considers the actual present desire version, which holds that the reasons an agent has at some time are given by their actual desires at that time. Fletcher rightly objects that this account gets the wrong results about reasons in cases where an agent could improve their lifetime wellbeing by taking certain actions but where they have no present desire to do so (sec. 5). In such cases, actual present desire Humeanism says that the agent has no reason to take the relevant actions that would improve their lifetime wellbeing.
Certainly, this kind of simplistic Humeanism gets the wrong results here, but this is, well, just the simplest version of Humeanism. Many versions offer more sophisticated treatments of reasons. To pick just one: consider an idealized adviser view that makes no restriction to present desires of the agent. On this version, an agent has a reason for some action just in case their ideal adviser – taking into account the agent's desires over the course of their life – desires that their non-ideal counterpart take some action. It's plausible that this version of Humeanism avoids the implausible results of the actual present desire version noted by Fletcher. There's a good case to be made that someone's idealized adviser would have at least some desire that the person take actions that maximize their lifetime wellbeing. Of course, the ideal adviser might also desire that their counterpart do other actions as well, but there is at least a reason on this view for the subject to maximize their lifetime wellbeing. Yet, this reason comes from the idealized desires of the agent, not anything about prudential value per se.
Thus, I don't think Fletcher can argue on extensional grounds against Humeanism with respect to Prudential Value Matters. Rather, I think Fletcher is better served to focus on another strategy, which he briefly mentions. This is that even if Humeanism is extensionally correct about prudential reasons, it gets the explanation wrong in holding that desires generate all reasons and prudential value never does so itself (pp. 23–24). This is an approach that, to my mind, is much more promising – though, of course, explanatory considerations are much more vexed than extensional ones.
The second area concerns Fletcher's aim of showing that meta-prudence sheds light on meta-morality. Perhaps the most important point he makes in this regard is that moral error theory cannot quarantine off the normativity of prudence. As Fletcher notes, many error theorists seem to make an exception for prudence. One example is that J. L. Mackie and other error theorists hold that certain normative languages are better for us than others (p. 161). Another is that Richard Joyce holds that the error theorist still has reasons of self-interest to be nice to fellow human beings (p. 50). Against this mistaken tendency, Fletcher uses Prudential Value Matters – the claim that prudential value generates normative reasons – to develop a partners in crime argument against moral error theory. The basic idea is that prudence is normative if and only if morality is normative, and, given Prudential Value Matters, prudence is normative, hence morality is normative. To my mind, Fletcher convincingly shows that the error theorist must go whole hog and offer a completely global error theory, covering both prudence and morality. But the greater the error theory, the greater the argumentative burden.
However, Fletcher makes the further claim that this prudential partners in crime argument is more successful than the more common epistemological partners in crime argument. He suggests that this is because there is a greater prospect for epistemology to be reduced to non-normative ingredients than there is for prudence. The argument consists in a rough gesture toward a reduction of normative epistemology to non-normative notions of probability and true belief, followed by the claim that this sort of reduction seems more plausible than any reduction of normative prudential value to non-normative notions. Yet it seems to me that the prospect for reducing the one to non-normative goings on is no more or less plausible than the other. If the reductive epistemologist can work with true belief and probability, the reductive value theorist can work with non-normative substrates of wellbeing, such as pleasure and pain, or satisfied desires. Despite this issue, Fletcher's discussion usefully brings out that there are further partners in crime for morality than just epistemology. Prudence joins the syndicate as well.
The third area concerns the movement from the meta-ethics of morality to the meta-ethics of prudence. Fletcher holds that many apparently first-order discussions in prudence are better framed as trying to get at directly meta-normative considerations. This is an important point. Yet, I think Fletcher goes astray in his conception of how to directly approach meta-prudential considerations. He claims that prudential theorists should distinguish first-order questions about the substantive constituents of prudential value from second-order questions about what prudentially valuable properties are identical to (ch. 6, sec. 2). Thus, Fletcher holds that realists will claim that prudential value properties are identical either to non-natural properties or to natural but non-mental properties, whereas anti-realists will claim they are identical to mental properties.
Yet while I certainly accept the distinction between constituents of personal value and the nature of personal value itself, this distinction alone isn't going to do enough meta-prudential work. For one thing, it can't distinguish between realist and anti-realist views as Fletcher wants – at least not unless we smuggle in additional loaded concepts into the terms flanking the identity. Fletcher's example of an anti-realist view is one that holds that the property of good-for is identical to the property of fulfilling an actual desire. But, instead, they might think that the property good-for is generated by the property fulfilling an actual desire. In this case, they don't think the property of good-for is identical to any structured natural property – it's just identical, trivially, to itself. Hence, we can't use identity of the property alone to distinguish this anti-realist from even a realist non-naturalist.
Additionally, the appeal to property identity can't do the work of taxonomizing between different views within the realist and anti-realist camps. Consider two non-naturalist realists who both agree that good-for is not identical to any natural property. They might still disagree about what makes it the case why any particular constituent of wellbeing comes to have the property of good-for. One realist might think it's due to the non-natural essence of good-for. Another might think it's due to some normative law of metaphysics. Therefore, the distinction between whether or not good-for is identical to a non-normative property is not enough to distinguish them.
However, I do think Fletcher's basic contention that meta-prudence can fruitfully import resources from meta-morality is exactly right. It's just that there are more resources to import than property identity. When it comes to the above cases specifically, meta-ethicists can import notions of grounding, essence, and metaphysical laws – as some meta-ethicists are beginning to do. This is not to say that all such concepts are in good standing – but that something like them needs to be appealed to in addition to property identity.
In light of these qualms over some of the details of Dear Prudence, it's worth remembering that Fletcher is utterly successful in his primary goal of showing the importance of exploring the meta-normativity of prudence. Indeed, he prompted me to do a bit of exploration.
