In the burgeoning philosophical literature on conceptual engineering improving our concepts is typically portrayed as the hallmark activity of the field. However, Cappelen (2018) has challenged the idea that we can know how and why conceptual changes occur well enough to actively intervene in revising our concepts; the mechanisms of conceptual change are typically inscrutable to us. If the “inscrutability challenge” is correct, the practical aspect of conceptual engineering may seem to be undermined, but I argue that endorsing such pessimism would be a mistake. Even if the inscrutability challenge is correct, conceptual engineers often have good reasons to try to preserve existing concepts. I examine several cases where concept preservation is important and draw lessons about this activity for conceptual engineers. Conceptual engineering involves assessing concepts or representational devices, proposals to improve them, and efforts to implement these changes (Cappelen 2018). The practical effects of these concepts, particularly sub-optimal ones, are relevant considerations in each of these stages of the enterprise. But a problem for the second and third stages is the enormous empirical task involved in formulating plans for improving or “ameliorating”1 concepts and attempting to implement these plans. Cappelen (2018) argues that the mechanisms by which changes occur in linguistic and conceptual practice are typically inscrutable (ibid., 73) – we cannot figure out how and why they occur well enough to have confidence that we can actively intervene in changing 1 The use of the term “amelioration” in the conceptual engineering literature is due to Haslanger (2012). our concepts.2 I will follow Cappelen in calling this claim “Inscrutability.” Inscrutability is a surprising claim to hold if we want to place emphasis on conceptual engineering as an interesting and distinctive philosophical method. In other work, I have argued against Inscrutability, marshalling evidence that we can assess the empirical effects of our concepts and efforts to change them along a number of dimensions (Lindauer forthcoming, see also Nado forthcoming). But suppose Inscrutability is true. Does it follow that the practical aspect of conceptual engineering, arguably its distinctive aspect, is closed to us? I will argue that it does not, because in a range of cases, we have good empirical evidence that preserving our existing concepts, at least for some purposes, is preferable to revising them. Concept preservation is an active enterprise, one that draws our attention to kinds of practical considerations that sometimes lead us to attempt to improve our concepts, and often requires us to take action. I offer three cases in what follows and draw lessons from them for the role of concept preservation in conceptual engineering. 1. Preliminaries: Inscrutability and Lack of Control Before examining the cases, it will be useful to be a bit more specific about the claims under consideration. Cappelen distinguishes between Inscrutability, an epistemic claim, and Lack of Control, a metaphysical claim, which he argues follow from a commitment to metasemantic externalism.3 For metasemantic externalists, the meanings of our words can be influenced by a 2 As I elaborate below, inscrutability and lack of control over conceptual change, as Cappelen notes (pp. 72-74), are distinct, but lack of control follows from inscrutability, and Cappelen puts a great deal of emphasis on inscrutability when defending lack of control (p. 74). 3 More precisely, he argues that they follow from a set of externalist assumptions (pp. 72-73). whole range of factors external to or “outside the heads” (Burge 1979) of speakers: such as past baptismal acts (Kripke 1980) and actions intended to determine meanings, information sources from the past, statements of other people and especially experts, and how they are used over time. I will not question any of these claims, or how they are supposed to imply Inscrutability and Lack of Control. It is clear enough how metasemantic externalism could be thought to imply Inscrutability. Rather than concepts, Cappelen focuses on the meanings of terms as subject to change in conceptual engineering.4 To effectively change the meaning of a term, we would need to understand the mechanisms of reference change – how events in the past such as introductory events, the statements of experts, communicative chains, and perhaps various forms of activism and public policy can fit together in a causal process that changes a term’s meaning.5 For Cappelen, no cohesive understanding of such a causal process is available to us. But if this understanding is not available––meaning change is inscrutable––then we can’t deliberately aim to bring about such changes with any confidence. Of course, even if we did understand the full process by which we could change the meaning of a term, we might still lack the means of implementing it – the activists may disagree with us, or we may not be able to pass the relevant policies, or whatever. Lack of Control, a metaphysical point about what we are capable of doing, follows from Inscrutability, an epistemic point about what we are capable of knowing or understanding, but not the reverse. In this paper I focus on the kind of Lack of Control that follows from the Inscrutability that Cappelen argues we are subject to in relation to the mechanisms of conceptual engineering; 4 He continues to use the term ‘conceptual engineering’ largely to maintain continuity with the literature and the self-descriptions of other theorists (see, e.g., pp. 3-4). 5 Cappelen also holds that the current intensions of terms are inscrutable for us (pp. 73-74), but this claim, as I point out below, doesn’t bear on my argument. the mechanisms by which concepts or the meanings of terms change. Here is not the place to settle a debate regarding concepts, and I will often refer to concepts rather than terms and their meanings for ease of exposition.6 By presenting and discussing a set of realistic (if not perfectly real) examples, my hope is to persuade my audience that concept preservation is not subject to the same worries that concept improvement is. That is, the Inscrutability challenge does not apply, or not in as significant a way, to preservation projects, and Lack of Control is similarly mitigated as a concern. I will argue that this suggests that the almost singular emphasis on concept improvement in discussions of conceptual engineering may be misleading, and that conceptual engineers will often have an important role to play in concept preservation instead. 2. Case 1: The Bad President, ‘Citizen,’ and Concept Preservation For the first case, I’ll ask you to imagine that a democratic country has elected a bigoted authoritarian president. Suppose that prior to the groundswell of support for this president from his supporters, the concept or representational device ‘citizen’ had progressed substantially. In the more distant past, the country embraced an ethnonationalist, hierarchical concept of citizenship according to which the “real citizens” were understood to be white, whereas nonwhite people counted at best as second-class citizens. But due to the hard work of social movements, legal decisions, and many other events, this ethnonationalist concept was replaced by7 a more egalitarian citizen concept. Predictably, this progress met with ongoing resistance, 6 Other preferred cognitive or linguistic items may also be substituted in without changing anything important for the points that I make here. 7 We may instead view this as reforming the prior concept. Nothing important turns on this choice either. and the new president’s support is partly explained by the resistance to this change. He promises people who prefer the old concept to bring their country back to the “good old days,” getting rid of the egalitarian concept in favor of the older ethnonationalist one. Now a set of conceptual engineers thinking about citizenship in that country get together and ask themselves, what should we do about this proposed change to the concept? It is not an amelioration or improvement by their lights – quite the opposite. Allowing the society to go back to the old concept would undo the progress made. In such a case, the conceptual engineers, being practically-minded philosophers of citizenship, should not only abstain from supporting the president’s desired change but actively resist it. They should, in other words, prevent moral backslide8 by doing what they can to preserve the concept that is currently in place. Notably, this could involve preventing moving back to a previously held concept, as in this case, or preventing a brand new concept from being introduced whose effects would be worse than retaining the present one. One lesson we can take from this case has already been stated – conceptual engineering must not only involve re-engineering but also, in some cases, preserving the concepts that we have upon assessing the practical context that we are in. Indeed, insofar as we accept the metaphorical language of “engineering” concepts at all, the case suggests that we have good reason to include preserving concepts as a part of conceptual engineering, just as preserving buildings and bridges is a part of architectural engineering.9 Even in other cases where we are 8 The present case involves moral backsliding, but we might also refer to a broader phenomenon of normative backsliding, where the practical effects of new or reintroduced concepts would be normatively worse than preserving the current concept. The sense of worseness may then be broadened to include negative epistemic or aesthetic as well as moral effects. 9 I am grateful to Mark Pinder for suggesting that I use the analogy to architectural engineering to draw out my points in this paper, an analogy that I return to in Section 5. Hume (1751/1998) notably likens laws to houses, whose broad fun

中文翻译：

---

作为概念保存的概念工程

在有关概念工程的新兴哲学文献中，改进我们的概念通常被描述为该领域的标志性活动。然而，Cappelen (2018) 挑战了我们可以知道概念变化如何以及为什么发生足以积极干预修改我们的概念的想法；概念变化的机制对我们来说通常是难以理解的。如果“难以理解的挑战”是正确的，那么概念工程的实践方面似乎会受到破坏，但我认为支持这种悲观主义将是一个错误。即使难以理解的挑战是正确的，概念工程师通常也有充分的理由尝试保留现有的概念。我研究了几个概念保存很重要的案例，并为概念工程师吸取了有关此活动的经验。概念工程涉及评估概念或代表性设备、改进它们的建议以及实施这些更改的努力（Cappelen 2018）。这些概念的实际效果，尤其是次优概念，是企业各个阶段的相关考虑因素。但第二和第三阶段的问题是制定改进或“改进”1 概念的计划并尝试实施这些计划所涉及的大量经验任务。Cappelen (2018) 认为语言和概念实践中发生变化的机制通常是难以理解的（同上，73) – 我们无法弄清楚它们如何以及为什么发生得足够好，以至于有信心我们可以积极干预改变 1 在概念工程文献中使用“改善”一词是由于 Haslanger (2012)。我们的概念。2 我将跟随 Cappelen 将这种说法称为“不可思议”。如果我们想把重点放在概念工程作为一种有趣且独特的哲学方法上，那么不可思议性是一个令人惊讶的主张。在其他工作中，我反对 Inscrutability，整理证据表明我们可以评估我们的概念的经验影响以及在多个维度上改变它们的努力（Lindauer 即将推出，另见 Nado 即将推出）。但假设不可测度是真的。是否遵循概念工程的实践方面，可以说是其独特的方面，对我们关闭了吗？我会争辩说它不会，因为在一系列情况下，我们有很好的经验证据表明，至少出于某些目的，保留我们现有的概念比修改它们更可取。概念保存是一项积极的事业，它将我们的注意力吸引到各种实际考虑上，这些考虑有时会导致我们尝试改进我们的概念，并且经常需要我们采取行动。我在下面提供了三个案例，并从中吸取了概念保存在概念工程中的作用。1. 初步：难以理解和缺乏控制 在审查案件之前，对所考虑的索赔更具体一点是有用的。Cappelen 区分了不可测度（一种认识论主张）和缺乏控制（一种形而上学主张），他认为这源于对元语义外在主义的承诺。 3 对于元语义外在论者来说，我们的话的含义可能受到 2 正如我在下面详述的那样，如卡佩伦所指出的那样，难以理解和缺乏对概念变化的控制（第 72-74 页） )，是不同的，但缺乏控制源于难以理解，而卡佩伦在为缺乏控制辩护时非常强调不可理解性 (p. 74)。3 更准确地说，他认为它们是从一组外部主义假设得出的（第 72-73 页）。演讲者外部或“头脑之外”（Burge 1979）的所有因素：例如过去的洗礼行为（Kripke 1980）和旨在确定意义的行为、过去的信息来源、其他人特别是专家的陈述，以及随着时间的推移如何使用它们。我不会质疑任何这些说法，或者它们应该如何暗示难以理解和缺乏控制。很明显，元语义外在主义如何被认为暗示了不可测度。Cappelen 关注的不是概念，而是术语的含义，因为它们会在概念工程中发生变化。 4 为了有效地改变术语的含义，我们需要了解指称变化的机制——过去的事件，例如介绍性事件，专家的陈述、交流链，也许还有各种形式的激进主义和公共政策可以组合在一起，形成一个改变术语含义的因果过程。5 对于卡佩伦来说，我们无法对这种因果过程进行统一的理解。但是，如果没有这种理解——意味着变化是高深莫测的——那么我们就不能有信心地刻意去实现这样的变化。当然，即使我们了解了我们可以改变一个术语的整个过程，但我们可能仍然缺乏执行它的手段——积极分子可能不同意我们，或者我们可能无法通过相关政策，或者其他什么. 缺乏控制，一个关于我们能够做什么的形而上学观点，来自不可测度，一个关于我们能够知道或理解什么的认识论点，但不是相反。在这篇论文中，我关注的是卡佩伦认为我们在概念工程机制方面受到的不可测度所导致的那种缺乏控制；4 他继续使用“概念工程”一词，主​​要是为了保持与其他理论家的文献和自我描述的连续性（例如，参见第 3-4 页）。5 Cappelen 还认为，术语的当前内涵对我们来说是难以理解的（第 73-74 页），但是正如我在下面指出的那样，这种说法与我的论点无关。概念或术语含义变化的机制。这里不是解决关于概念的争论的地方，为了便于说明，我经常会提到概念而不是术语及其含义。 6 通过展示和讨论一组现实的（如果不是完全真实的）例子，我的希望是说服我的听众，概念保留不会受到与概念改进相同的担忧。也就是说，不可测度挑战不适用于保护项目，或者在很大程度上不适用于保护项目，并且缺乏控制同样作为一个问题得到缓解。我会争辩说，这表明在概念工程的讨论中几乎单一地强调概念改进可能会产生误导，并且概念工程师通常会在概念保存方面发挥重要作用。2. 案例 1：坏总统、“公民”和观念保留 对于第一个案例，我让您想象一个民主国家选出了一位偏执的专制总统。假设在他的支持者对这位总统的支持高涨之前，“公民”的概念或表征手段已经取得了实质性的进展。在更遥远的过去，该国接受了一种民族主义的、等级分明的公民身份概念，根据该概念，“真正的公民”被理解为白人，而非白人则最多被视为二等公民。但由于社会运动、法律决定和许多其他事件的辛勤工作，这个民族主义概念被更平等的公民概念所取代。可以预见的是，这一进展遇到了持续的阻力，6 其他首选的认知或语言项目也可能被替代，而不会改变我在这里提出的观点的任何重要内容。7 我们可以将其视为对先前概念的改革。也没有什么重要的事情会影响这个选择。新总统的支持部分是由于对这一变化的抵制。他向喜欢旧观念的人们承诺，让他们的国家回到“过去的美好时光”，摆脱平等主义观念，转而支持更古老的民族主义观念。现在，一群思考该国公民身份的概念工程师聚在一起问自己，我们应该如何处理这一对概念的拟议更改？他们的灯光并不是改善或改进——恰恰相反。允许社会回到旧的概念会破坏已经取得的进步。在这种情况下，概念工程师作为具有实践意识的公民哲学家，不仅应该放弃支持总统所期望的变革，而且应该积极抵制它。换言之，他们应该尽其所能保护现有的概念，从而防止道德倒退8。值得注意的是，这可能涉及防止回到以前持有的概念，如在这种情况下，或防止引入一个全新的概念，其效果会比保留现有概念更糟糕。然后，糟糕的感觉可能会扩大到包括消极的认知或审美以及道德影响。9 我很感谢 Mark Pinder 建议我使用建筑工程的类比来阐述我在本文中的观点，我在第 5 节中回到了这个类比。 Hume (1751/1998) 特别将法律比作房屋，其广泛的乐趣