An underappreciated but core part of metaethical expressivism is a thesis about moral practice’s function as a tool for coordination. Neil Sinclair’s Practical Expressivism brings this thesis to the fore and demonstrates that functions can help expressivists a lot. It does so by developing “practical expressivism,” a view with three core commitments. First, the metasemantic view that the semantic function of moral judgments is not to describe the world or express moral beliefs, but to express moral commitments (33). Second, the psychological view that moral commitments are (at least partially) “stable, reflectively endorsed, general attitudinally ascended states of approval and/or disapproval, including dispositions to avowal in moral discussion” (53). Third, the view that moral practice has the distinctive function as a coordination device that enables “human beings to work towards, test, refine, and sustain mutually beneficial patterns of action and reaction” (38). The third commitment, thereby, is extensively used to flesh out the others and to deal with problems for expressivism. While the idea to use functions in this way already figures in, for example, Simon Blackburn’s work, it has never before been given such an extensive and systematic treatment as in this book. Practical expressivism is, hence, the most sophisticated attempt to take functions in expressivism seriously.The book is divided into ten chapters, plus an interesting appendix comparing practical expressivism to other expressivist views. The first chapter starts with a (familiar) account of what metaethics is about and imposes two desiderata for adequacy on metaethical theories: first, to accommodate the characteristic features of and assumptions underlying moral practice; second, to fit with our wider views about the world.Chapters 2 and 3 flesh out practical expressivism’s core ideas and deal with initial problems, such as the moral attitude problem. Chapter 3 also offers a novel account of the expression relation, on which “to express a mental state type is to push that state into a public arena of discussion, as a candidate for acceptance or rejection by others” (65). Interestingly, the way this is fleshed out is similar to inferentialist accounts of assertion as being committed in certain ways that one can be challenged on and held accountable for, and that can license others to do certain things. So, here one can see some convergence between expressivism and inferentialism and that the former can appropriate resources from the latter. Chapter 3 also clarifies the quasi-realist project, which is an integral part of practical expressivism.The rest of the book argues that and how practical expressivism can accommodate some of the most important features of and assumptions underlying moral practice. Chapter 4 explains how practical expressivism accounts for moral disagreement, the role of reason-giving in moral disagreement, and how it captures that supervenience is a conceptual truth. Chapters 5 and 6 lay down the practical expressivists’ metasemantics for complex sentences and subsentential expressions. Chapter 7 explains how minimalism helps expressivists to allow normative judgments to be truth-apt and beliefs. Chapter 8 shows how practical expressivists can accommodate that (it is a conceptual truth that) moral truths are mind-independent and moral reasons categorical. Chapter 9 offers a systematic summary of the three main strategies used throughout the book and defends those strategies against the charge that their results fall short in some way. It also clarifies and criticizes so-called “presumptive” arguments for realism. Chapter 10 gives a brief conclusion and reflections on how further progress is to be made in metaethics.This book has a lot to offer. First, it offers an accessible introduction to some of the most central developments of the expressivist/quasi-realist program in the last twenty years. For example, among other things, the book helpfully clarifies the quasi-realist project, its explanatory burdens, and some of its core moves, provides a very insightful statement of the metasemantic dialectic between expressivists and their opponents, and offers a highly helpful account of why and how the metasemantic/semantic distinction makes orthodox semantics compatible with expressivism. Second, the book makes robust and original contributions to the expressivist/quasi-realist program (and, of course, some of the clarifications themselves are actually substantial contributions to metaethics). For example, the book offers the most fleshed-out and sophisticated version of Simon Blackburn’s suggestion on how to solve the Frege-Geach problem via “tree-tying commitments.”To my mind, though, the core contribution of this book is that it highlights and fleshes out the central role that functions can and should play for expressivists. Throughout the book, appeals to the function of moral practice are used to, among other things, explain what makes an attitude a moral attitude (49–55), when two mental states disagree (83–92) or are inconsistent (137–42), the conceptual truth of supervenience (97–100), moral mind-independence (202–6), and categoricity of moral reasons (215–16), why moral discourse takes predicative form (149–53), and more. All of these applications seem promising to me, and I think it is worthwhile to push the expressivist project further in this direction.While I am very sympathetic to Sinclair’s approach, though, in the rest of this review I want to focus on this contribution and raise some critical questions and concerns about how functions are understood and how they are used. First, Sinclair briefly explains (42–43) that he understands functions as systems-functions. However, it is not clear that systems-functions can actually do (at least some of) Sinclair’s explanatory work. For example, to explain supervenience’s conceptual status, he argues that moral commitments must be constrained by it to play their coordinating function. This sounds like an explanation, but if “function” here is read as “systems-function,” it’s not clear that it is.A systems-function is a contribution that something makes to some capacity of a system of which it is a part. One prominent example (used to argue that systems-functions are not really functions) is this: doctors can use the heart’s thumping noises for diagnostic reasons, so one of the heart’s systems-functions is to make such noises, as they contribute to the doctor-heart systems’ capacity to make correct diagnoses. Now, let’s say that the heart must be able to contract to play this systems-function. Does this explain why the heart contracts? Not really. But then, it seems that the above account is not really an explanation of conceptual supervenience either.Would it help to say, as Sinclair does, that we are talking about the function that is constitutive of moral practice or moral commitments—that is, the function something has to play to be a moral practice or commitment? Even if this avoided the worries just raised, another problem with this suggestion is that it seems to have implausible implications in cases of malfunction. One prominent feature of systems-functions is that they do not allow malfunction: if something ceases to make the relevant contribution, it no longer has the relevant systems-function. But, surely, there are imaginable cases where moral practice malfunctions and really does not contribute to or even hinders the ability of “human beings to work towards, test, refine, and sustain mutually beneficial patterns of action and reaction” (38). If this was a systems-function constitutive of moral practice, the relevant practice is not moral practice in such cases. But, surely, moral practice should be able to malfunction.These worries could be avoided if functions were etiological functions: for example, if what explains the existence of (the mechanisms that produce) moral commitments is that they help us coordinate, that would explain why they have the features that allow them to play this function. And etiological functions do permit malfunctions. However, this reading ties the expressivists’ fate to empirical speculation about what explains the persistence of moral practice (which seems to be the reason Sinclair rejects it). I am not suggesting here, of course, that these sorts of worries cannot be overcome—I am optimistic that there is a notion that fits Sinclair’s explanatory purposes. But, without an account of this notion, his explanations rest on a promissory note.A second worry I have concerns the success of some of the function-based accounts themselves. For example, I am not sure whether Sinclair’s account of disagreement or inconsistency in terms of functions works. For disagreement, Sinclair offers the following account: “Two agents disagree-in-thought iff they have mental states that cannot collectively fulfil their constitutive functions” (89). Inconsistency is explained similarly (137–40). On first sight, though, this seems to overgenerate disagreement and inconsistency. Suppose the constitutive function of belief is to aim at truth, while desire’s constitutive function is to bring about the realization of its content. It seems as if this means that the belief that p and the desire that not p cannot collectively fulfill their constitutive functions, so they disagree (and are inconsistent) on Sinclair’s account. But neither do these states disagree, nor are they inconsistent.Maybe we can avoid this with different constitutive functions. Suppose the constitutive function of desires is not such that it is fulfilled when the desire’s content is realized, but only when agents are moved to do things that could bring about their content (I do not think this is Sinclair’s account, because moral commitments are supposed to be inconsistent due to their inconsistent contents, which fits more closely with the first reading). This deals with the above problem but raises the worry that Sinclair’s account undergenerates disagreement, because desires that disagree end up not disagreeing. After all, two desires that cannot be both realized might still be capable of moving even the same agent to do at least some of the things that could bring about their content.One way to evade these worries might be to restrict the account to mental states of the same type. However, at least for inconsistency Sinclair requires an account that can accommodate inconsistency between different mental state types. Otherwise, he could not explain how(1) If torture causes pain, then it is wrong.(2) Torture causes pain.(3) Torture is not wrong.form an inconsistent set of mental states. This is so because, on Sinclair’s account, (1) expresses a “tree-tying” commitment (135–37), (2) a representational belief, and (3) a moral commitment. So, Sinclair needs an account that works across mental state types.So, there is more work to be done for those who find the idea attractive that functions can substantially help expressivists. This, of course, is not a bad thing as it highlights fruitful avenues for further investigation, research, and exchange. Practical Expressivism is an outstanding book that excellently highlights these issues and pushes them further, and presents us with an attractive picture of expressivism that it is worth developing further. Anyone interested in metaethics can find a lot to learn and think about in this superb book, and I am sure it only marks the starting point in systematic thinking about the view it champions.

中文翻译：

---

实用表现主义

元伦理表现主义的一个未被充分认识但核心的部分是关于道德实践作为协调工具的功能的论文。尼尔·辛克莱 (Neil Sinclair) 的实用表现主义 (Practical Expressivism) 使这一论点脱颖而出，并证明函数可以为表现主义者提供很大帮助。它通过发展“实用表现主义”来做到这一点，这是一种具有三个核心承诺的观点。首先，元语义观点认为道德判断的语义功能不是描述世界或表达道德信仰，而是表达道德承诺（33）。其次，心理学观点认为道德承诺（至少部分）是“稳定的、反思性认可的、一般态度上升的赞同和/或不赞成状态，包括在道德讨论中公开承认的倾向”（53）。第三，认为道德实践作为一种协调机制具有独特的功能，使“人类能够朝着、测试、完善和维持互惠互利的行动和反应模式”（38）。因此，第三项承诺被广泛用于充实其他承诺并处理表现主义的问题。虽然以这种方式使用函数的想法已经出现在例如 Simon Blackburn 的著作中，但它以前从未像本书那样得到如此广泛和系统的处理。因此，实用表现主义是认真对待表现主义功能的最复杂尝试。本书分为十章，加上一个有趣的附录，将实用表现主义与其他表现主义观点进行了比较。第一章从对元伦理学是什么的（熟悉的）说明开始，并为元伦理学理论的充分性强加了两个必要条件：首先，适应道德实践的特征和假设；第二，为了适应我们对世界的更广泛的看法。第二章和第三章充实了实践表现主义的核心思想，并处理了最初的问题，比如道德态度问题。第 3 章还对表达关系提供了一种新颖的解释，据此“表达一种精神状态类型就是将该状态推向公共讨论领域，作为其他人接受或拒绝的候选者”（65）。有趣的是，这种充实的方式类似于推论主义者对断言的描述，即以某些方式承诺，人们可以对其提出质疑并对其负责，这可以许可他人做某些事情。因此，在这里可以看到表现主义和推理主义之间的某种趋同，前者可以从后者那里挪用资源。第 3 章还阐明了准现实主义项目，它是实践表现主义的一个组成部分。本书的其余部分论证了实践表现主义以及如何适应道德实践的一些最重要的特征和假设。第 4 章解释了实践表现主义如何解释道德分歧、给出理由在道德分歧中的作用，以及它如何捕捉到随附性是一个概念真理。第 5 章和第 6 章阐述了复杂句子和子句表达式的实用表达主义者的元语义学。第 7 章解释了极简主义如何帮助表现主义者允许规范判断成为真实的和信念。第 8 章展示了实践表现主义者如何适应（这是一个概念真理）道德真理是思想独立的，道德原因是绝对的。第 9 章对全书使用的三种主要策略进行了系统的总结，并针对这些策略的结果在某些方面存在不足的指控进行了辩护。它还澄清和批评了现实主义的所谓“推定”论点。第 10 章对如何在元伦理学上取得进一步的进展给出了一个简短的结论和思考。本书提供了很多内容。首先，它通俗易懂地介绍了过去二十年来表现主义/准现实主义程序的一些最重要的发展。例如，除其他事项外，这本书有助于澄清准现实主义项目、它的解释负担和它的一些核心步骤，对表现主义者和他们的对手之间的元语义辩证法提供了一个非常有见地的陈述，并提供了一个非常有用的解释为什么以及如何元语义/语义区别使正统语义与表现主义兼容。其次，这本书对表现主义/准现实主义计划做出了强有力的原创贡献（当然，一些澄清本身实际上是对元伦理学的实质性贡献）。例如，这本书提供了西蒙·布莱克本关于如何通过“捆绑承诺”解决弗雷格-吉奇问题的建议的最充实和最复杂的版本。不过，在我看来，这本书的核心贡献在于它强调并充实了功能可以而且应该为表现主义者发挥的核心作用。在整本书中，诉诸道德实践的功能被用来解释什么使态度成为道德态度 (49–55)，当两种精神状态不一致 (83–92) 或不一致 (137–42) )、随附性的概念真理 (97–100)、道德思想独立性 (202–6) 和道德原因的范畴性 (215–16)、为什么道德话语采用谓语形式 (149–53) 等等。所有这些应用对我来说似乎都很有前途，我认为将表现主义项目朝这个方向进一步推进是值得的。虽然我非常赞同辛克莱的方法，但是，在这篇评论的其余部分，我想重点关注这一贡献，并提出一些关于如何理解函数和如何使用函数的关键问题和担忧。首先，Sinclair 简要解释 (42–43) 他将功能理解为系统功能。然而，系统功能是否真的可以（至少是部分）Sinclair 的解释工作尚不清楚。例如，为了解释随附的概念地位，他认为道德承诺必须受其约束才能发挥协调作用。这听起来像是一种解释，但如果这里的“功能”被理解为“系统功能”，则不清楚它是否是。系统功能是某物对它所属的系统的某些能力做出的贡献. 一个突出的例子（用于争论系统功能不是真正的功能）是这样的：医生可以出于诊断原因使用心脏的砰砰声，因此心脏的系统功能之一就是发出这样的声音，因为它们有助于医生心脏系统做出正确诊断的能力。现在，假设心脏必须能够收缩才能发挥这一系统功能。这是否解释了心脏收缩的原因？并不真地。但是，上面的解释似乎也不是对概念随附性的真正解释。像辛克莱那样说，我们正在谈论构成道德实践或道德承诺的功能——即，某种东西必须发挥的功能才能成为一种道德实践或承诺？即使这避免了刚才提出的担忧，这个建议的另一个问题是它在出现故障的情况下似乎具有难以置信的含义。系统功能的一个显着特征是它们不允许出现故障：如果某物停止做出相关贡献，它就不再具有相关的系统功能。但是，可以肯定的是，在可以想象的情况下，道德实践会出现故障，实际上不会促进甚至阻碍“人类努力、检验、完善和维持互惠互利的行动和反应模式”的能力 (38)。如果这是构成道德实践的系统功能，那么在这种情况下，相关实践就不是道德实践。但是，可以肯定的是，道德实践应该能够发生故障。如果功能是病因学功能，这些担忧就可以避免：例如，如果解释道德承诺（产生机制）的存在是因为它们帮助我们协调，这可以解释为什么他们具有允许他们发挥此功能的功能。病因学功能确实允许出现故障。然而，这种解读将表现主义者的命运与关于什么解释了道德实践的持久性的经验推测联系在一起（这似乎是辛克莱拒绝它的原因）。当然，我在这里并不是说无法克服这些担忧——我很乐观地认为有一个概念符合 Sinclair 的解释目的。但是，由于没有对这个概念的说明，他的解释依赖于期票。我担心的第二个问题是一些基于功能的说明本身的成功。例如，我不确定 Sinclair 关于职能方面的分歧或不一致的解释是否有效。对于不同意见，辛克莱提供了以下说明：“如果两个代理人的精神状态无法共同履行其构成功能，则他们在思想上存在分歧”（89）。不一致的解释类似（137-40）。然而，乍一看，这似乎过度产生了分歧和不一致。假设信念的构成功能是追求真理，而欲望的构成功能是实现其内容。这似乎意味着 p 的信念和非 p 的愿望不能共同实现它们的构成功能，因此它们在 Sinclair 的解释中不同意（并且不一致）。但这些状态既不不一致，也不矛盾。也许我们可以用不同的本构函数来避免这种情况。假设欲望的构成功能不是在欲望的内容实现时就实现了，但只有当代理人被感动去做可能带来他们的内容的事情时（我不认为这是辛克莱的说法，因为道德承诺由于内容不一致而被认为是不一致的，这更符合一读）。这解决了上述问题，但引起了人们的担忧，即 Sinclair 的解释不足以产生分歧，因为不同意的愿望最终不会不同意。毕竟，无法同时实现的两个愿望可能仍然能够促使同一个代理人去做至少一些可能带来其内容的事情。避免这些担忧的一种方法可能是将帐户限制在精神状态同类型的。然而，至少对于不一致，辛克莱需要一个能够适应不同精神状态类型之间的不一致的账户。否则，他无法解释（1）如果酷刑导致疼痛，那么它是错误的。（2）酷刑导致疼痛。（3）酷刑没有错。形成一组不一致的精神状态。之所以如此，是因为根据辛克莱的说法，(1) 表达了一种“绑树”承诺 (135-37)，(2) 一种代表性的信念，以及 (3) 一种道德承诺。因此，Sinclair 需要一个适用于各种心理状态类型的帐户。因此，对于那些认为功能可以极大地帮助表现主义者这一想法很有吸引力的人来说，还有更多的工作要做。这当然不是坏事，因为它突出了进一步调查、研究和交流的富有成果的途径。实用表现主义是一本杰出的书，它出色地突出了这些问题并将其进一步推进，并向我们展示了一幅值得进一步发展的表现主义的迷人图景。