Thought experiments are a mainstay in philosophy. We think about how to apply philosophical concepts to hypothetical cases in an effort to draw conclusions about their application more generally. To take a familiar example, one way we might learn that justified true belief isn’t knowledge involves first having an intuition that a subject has a justified true belief that doesn’t rise to the level of knowledge in a hypothetical case (a so-called Gettier case).About thirty years ago so-called negative experimental philosophers started to call this practice into question with the help of experimental results. According to them, we should distrust our intuitive verdicts about cases. Their reason: when we survey the folk, we find that variation in their intuitive verdicts about cases depends significantly on factors about their own psychology, which are irrelevant to the truth of the philosophical views at issue. They issued a challenge to explain why the intuitions philosophers rely on are any better off.One line of response to the challenge invokes a “perceptualist” account of intuition, which conceives of intuition as a sui generis mental state that works a lot like perceptual experience except it is intellectual. Broadly speaking, the response goes, the challenge fails because beliefs about hypothetical cases are only justified when they are produced in the right way. The right way typically relies on “perception-like” intuitions. Moreover, folk judgments about hypothetical cases tend not to be produced in the right way.Elijah Chudnoff defends a new response of this kind in his latest book, Forming Impressions. Crucial to his response is the claim that whether we are experts or not, we can have different, and indeed better, intuitions as we increasingly think about some issue. In the course of developing this response Chudnoff addresses a variety of issues, which span the psychology of expertise and modularity, the foundations of mathematics on the status of intuition, and early modern rationalist thought on philosophical method.The book has three parts. Part 1 provides a general theory of “expert impressions.” Chudnoff divides expertise into expertise in forming impressions and expertise in performing actions. Among kinds of expertise in forming impressions, there is expertise in forming intuitions in addition to the more familiar expertise in forming perceptual experiences. Chudnoff argues for the superiority of classifying expertise according to this picture against other taxonomies found in psychology before going on to explore how closely expert intuitions resemble expert perceptual experiences. Part 2 explains the place of impressions in epistemology. Here Chudnoff argues against phenomenal conservatism, or the principle that any experience as of p provides the subject having the experience with prima facie justification for believing that p. Chudnoff defends presentational conservatism instead. Only experiences as of p that have a certain phenomenology—“presentational phenomenology”—provide the subject having the experience with prima facie justification for believing that p. Chudnoff builds his case by comparing the two views’ treatments of experiences of occluded objects. The book culminates in a third and final part on philosophical methodology, which responds to the negative x-phi challenge.The account of intuition in Forming Impressions departs from perceptualist orthodoxy in two important respects. Perceptualists tend to claim that intuitions are a source of a priori justification. As a result, the judgments we make on their basis are a priori justified when justified at all. Chudnoff disagrees. In his account some intuitive judgments are a posteriori. What it takes for a judgment to count as intuitive is that it is formed by taking the contents of intuitions at face value. Moreover, when we take the contents of intuitions at face value, our perceptual experiences can sometimes play a substantial enough of a role in producing those intuitions that the resulting judgments are disqualified from counting as a priori. It may then turn out that many of the intuitive judgments arising from thought experiments in philosophy are a posteriori, pace standard rationalists (although it is worth noting that Chudnoff outlines a strategy for resisting this consequence—see especially 171–73).Second, and more significantly for the methodological aims of this book, the account of intuition departs from the standard assumption that intuitions are pretheoretic or naive. Chudnoff instead claims that what one intuits can change, and indeed improve, as a result of reasoning with concepts featured in the content of the intuition. Chudnoff develops this idea by applying it to cases of mathematical and philosophical intuition.In the philosophical case Chudnoff considers our intuitions about Thomson’s (1971) violinist. Part of what generates our intuition (assuming we share it) that the patient is morally permitted to detach themselves from the violinist is the argumentation surrounding the description of the case itself. Without the surrounding argumentation, we may have the intuition that detachment is not morally permissible. More generally, Chudnoff highlights various practices in argumentation, which he thinks can modify the contents of our intuitions: “drawing distinctions, clarifying the meanings of terms, evaluating analogies, highlighting logical form, engaging in dialectic, articulating principles, exploring models, [and] considering extreme cases” (204). In this respect, Chudnoff suggests, argumentation is supposed to supplement intuition in a way similar to how microscopes and telescopes supplement visual experience in science: both fundamentally alter the content of our experiences.There is a familiar concern for perceptualist accounts of intuitions like Chudnoff’s, which claim that intuitions have a special “presentational” phenomenology. Some struggle to locate experiences with such phenomenology, apart from perceptual experiences (Williamson 2007: 217). The point here is not to present this as a challenge for Chudnoff’s account (which, in fact, he has addressed elsewhere; see Chudnoff 2013, sec. 1.6). Still, readers sympathetic to this concern may find it hard on occasion to evaluate Chudnoff’s account. I’ll sketch an example.In Chudnoff’s view, what we can see (or intuit) can change as we develop relevant expertise. At the same time there is a crucial respect in which the two kinds of impressions are said to differ. Expert perceptual experiences can have high-level contents, which their novice counterparts cannot. For example, when I start out birdwatching, my perceptual experiences can’t represent properties such as being a superb fairywren. I see it as having certain more basic properties and go on to infer that what I am observing is a superb fairywren. But this changes if I develop expertise in birdwatching. An expert can see a bird as a superb fairywren. So long as the birdwatcher is a novice, however, they won’t be able to see a bird as a superb fairywren (this is part of what makes them a novice birdwatcher).Chudnoff then goes on to develop the intriguing proposal that intuitions do not resemble perceptual experiences in this respect. Instead, whatever contents can be accessed by an expert can also be accessed by the novice. Chudnoff illustrates his view by way of a mathematical example: specifically, how we represent in intuition that there is a curve that can touch every point on a plane. This claim strikes us as counterintuitive at first; after all, we cannot visualize a curve of this kind. However, Chudnoff claims, a trained mathematician can intuit its truth; moreover, so too can a beginner. An expert can show the beginner a series of constructions that increasingly approximate a space-filling curve so that they come to appreciate that the latter curve is possible after all. In the process the beginner is led to drop assumptions they made about curves initially (such as that all curves are visualizable). We reason our way to the intuition that there are space-filling curves.At this stage in the dialectic we might pause and ask whether the beginner’s guided judgment (or indeed the expert’s) is based on an intuition. When you go along with someone’s explanation for the existence of a space-filling curve, are you having an intuition in Chudnoff’s sense? You haven’t, presumably, had an intuition in the sense emphasized by dual-process theories in psychology; your judgment is not fast, automatic, or prereflective. You have, presumably, undergone a complex process of reasoning assisted by the visualization of a series of curves, but this is not quite what we are looking for. Instead you are supposed to affirm that you experienced a mental state with the right phenomenology—with features like “pushiness” and “forcefulness” (79–80). Here some readers may balk.Likely the book’s view of intuitions and their place in philosophical practice will be most helpful to readers already drawn to a perceptualist account of intuition. Other readers can also expect to profit from engaging with it, beyond simply reaping the benefits of getting a better understanding of a possible view. Take an alternative approach to the epistemology of thought experiments that looks to the imagination. For example, we judge—and in some cases come to know—that it is morally permissible for the patient to detach themselves from the violinist—by imagining the scenario Thomson describes and asking what is true in it. Plausibly, some of us are better at this task than others. They successfully stipulate the details explicitly mentioned in the description of the case and “fill in” further details appropriately. In fact, recent work by Amy Kind explores the view that we can be more or less skilled at imagining in various contexts, including the context of thought experimentation (see, e.g., Kind 2020). Parts of the rich framework developed in Forming Impressions may be reworked to suit an imagination-based view of thought experiments. Notably some of the central claims Chudnoff makes about intuition seem to carry over naturally to the imaginative exercises used in (philosophical) thought experimentation. Presumably we expect experts about a specific domain to be better at imagining scenarios about that domain. Trained philosophers could be a special case of this broader phenomenon. Moreover, the Chudnoff-style explanation of why philosophers are better at imagining Thomson’s case is satisfying (even if it is ultimately only part of the story). Background beliefs—which can be acquired from considering arguments and using the sorts of argumentative tools Chudnoff emphasizes—can have an impact on how we imagine scenarios, whether they concern philosophical concepts or not, in which case Chudnoff’s strategy may be of use to philosophers keen to defend the use of thought experiments in philosophy, regardless of any commitment to perceptualism about intuition.For helpful discussion, I am grateful to Nevin Climenhaga, Elijah Chudnoff, and Beau Madison Mount.

中文翻译：

---

形成印象：感知和直觉的专业知识


思想实验是哲学的支柱。我们思考如何将哲学概念应用到假设的案例中，以便更普遍地得出关于其应用的结论。举一个熟悉的例子，我们可能知道合理的真实信念不是知识的一种方式是，首先有一种直觉，即一个主体有一个合理的真实信念，但在假设的情况下（一个所谓的）没有上升到知识的水平。称为盖蒂尔案例）。大约三十年前，所谓的消极实验哲学家开始借助实验结果对这种做法提出质疑。他们认为，我们不应该相信我们对案件的直觉判断。他们的理由是：当我们调查人们时，我们发现他们对案件的直觉判断的变化很大程度上取决于他们自己的心理因素，而这些因素与所讨论的哲学观点的真实性无关。他们提出了一项挑战，以解释为什么哲学家所依赖的直觉会更好。对这一挑战的回应引用了对直觉的“感性主义”解释，该解释将直觉视为一种独特的心理状态，其作用与感性经验非常相似除非它是智力上的。一般来说，回应是，挑战失败了，因为对假设案例的信念只有在以正确的方式产生时才被证明是合理的。正确的方法通常依赖于“类似感知”的直觉。此外，民间对假设案例的判断往往不会以正确的方式产生。伊利亚·楚德诺夫在他的最新著作《形成印象》中为这种新的反应进行了辩护。 他的回应的关键在于，无论我们是否是专家，当我们越来越多地思考某些问题时，我们都可以有不同的、甚至更好的直觉。在提出这一回应的过程中，查德诺夫解决了各种问题，涵盖专业心理学和模块性、关于直觉地位的数学基础，以及关于哲学方法的早期现代理性主义思想。本书分为三个部分。第 1 部分提供了“专家印象”的一般理论。楚德诺夫将专业知识分为形成印象的专业知识和执行行动的专业知识。在形成印象的各种专业知识中，除了更熟悉的形成知觉经验的专业知识之外，还有形成直觉的专业知识。楚德诺夫论证了根据这张图对专业知识进行分类相对于心理学中发现的其他分类法的优越性，然后继续探索专家直觉与专家感知体验的相似程度。第二部分解释了印象在认识论中的地位。在这里，楚德诺夫反对现象保守主义，或者说，任何关于 p 的经验都为拥有该经验的主体提供了相信 p 的表面证据。相反，楚德诺夫为表象保守主义辩护。只有从 p 开始的具有某种现象学（“表象现象学”）的经验才能为拥有该经验的主体提供相信 p 的表面证据。楚德诺夫通过比较两种观点对被遮挡物体的体验的处理来建立他的案例。本书的第三部分也是最后一部分是关于哲学方法论的，它回应了消极的 x-phi 挑战。《形成印象》中对直觉的描述在两个重要方面背离了感性主义的正统观念。感性主义者倾向于声称直觉是先验论证的来源。因此，我们根据它们做出的判断在完全合理的情况下是先验合理的。楚德诺夫不同意。在他的叙述中，一些直觉判断是后验的。判断是否算直觉，是因为它是通过从表面上理解直觉的内容而形成的。此外，当我们从表面上理解直觉的内容时，我们的感知经验有时在产生这些直觉方面发挥了足够大的作用，以至于由此产生的判断不符合先验的条件。然后，结果可能是，哲学中的思想实验产生的许多直觉判断都是后验的、步调标准的理性主义者（尽管值得注意的是，楚德诺夫概述了抵制这种结果的策略——特别参见171-73）。 其次，就本书的方法论目标而言，更重要的是，对直觉的解释背离了直觉是前理论或幼稚的标准假设。相反，查德诺夫声称，通过用直觉内容中的概念进行推理，直觉可以改变甚至改进。查德诺夫通过将其应用于数学和哲学直觉的案例来发展这一想法。在哲学案例中，查德诺夫考虑了我们对汤姆森（1971）小提琴家的直觉。我们的直觉（假设我们也有同样的感觉）认为，病人在道德上是被允许与小提琴手分开的，部分原因是围绕病例本身描述的论证。 如果没有周围的论证，我们可能会直觉地认为超然在道德上是不允许的。更一般地说，楚德诺夫强调了论证中的各种实践，他认为这些实践可以修改我们直觉的内容：“区分、澄清术语的含义、评估类比、突出逻辑形式、参与辩证、阐明原则、探索模型，[和] 考虑极端情况”（204）。在这方面，查德诺夫认为，论证应该以类似于显微镜和望远镜补充科学中的视觉经验的方式来补充直觉：两者都从根本上改变了我们经验的内容。人们对像查德诺夫这样的直觉主义解释有一种常见的担忧，他们声称直觉具有特殊的“表象”现象学。除了感性体验之外，有些人还努力寻找这种现象学的体验（Williamson 2007：217）。这里的重点不是将此视为对 Chudnoff 帐户的挑战（事实上，他在其他地方已经解决过这个问题；参见 Chudnoff 2013，第 1.6 节）。尽管如此，同情这种担忧的读者有时可能会发现很难评估楚德诺夫的叙述。我将举一个例子。在楚德诺夫看来，随着我们发展相关专业知识，我们所看到的（或直觉的）可能会发生变化。同时，据说这两种印象在一个关键方面是不同的。专家的感知体验可以具有高水平的内容，而新手则不能。例如，当我开始观鸟时，我的感知体验并不能代表诸如成为优秀细尾鹩莺之类的属性。我认为它具有某些更基本的特性，并继续推断我所观察到的是一只极好的细尾鹩莺。 但如果我发展了观鸟方面的专业知识，情况就会改变。专家可以将一只鸟视为极好的细尾鹩莺。然而，只要观鸟者是新手，他们就无法将一只鸟视为出色的细尾鹩莺（这是使他们成为新手观鸟者的部分原因）。楚德诺夫接着提出了直觉所做的有趣提议。在这方面与感知体验不同。相反，专家可以访问的内容也可以被新手访问。楚德诺夫通过一个数学例子来阐述他的观点：具体来说，我们如何直观地表示存在一条可以触及平面上每个点的曲线。这种说法一开始让我们觉得违反直觉。毕竟，我们无法想象这种曲线。然而，查德诺夫声称，受过训练的数学家可以凭直觉知道其真相；此外，初学者也可以。专家可以向初学者展示一系列越来越接近空间填充曲线的结构，以便他们认识到后一条曲线毕竟是可能的。在这个过程中，初学者会放弃他们最初对曲线所做的假设（例如所有曲线都是可视化的）。我们推理出存在空间填充曲线的直觉。在辩证法的这个阶段，我们可能会停下来问一下初学者的指导判断（或者实际上是专家的判断）是否基于直觉。当你同意某人对空间填充曲线存在的解释时，你是否有楚德诺夫意义上的直觉？想必你还没有心理学双过程理论所强调的那种直觉；你的判断不是快速的、自动的或预先反思的。 想必您已经在一系列曲线可视化的帮助下经历了复杂的推理过程，但这并不是我们想要的。相反，你应该确认你经历了一种具有正确现象学的精神状态——具有“咄咄逼人”和“强制”等特征（79-80）。在这里，有些读者可能会犹豫不决。这本书对直觉的看法及其在哲学实践中的地位可能对那些已经被直觉主义的感性主义解释所吸引的读者最有帮助。其他读者也可以期望从参与其中获益，而不仅仅是获得更好地理解可能的观点的好处。对思想实验的认识论采取另一种方法，着眼于想象力。例如，通过想象汤姆森描述的场景并询问其中的真实情况，我们判断——并且在某些情况下了解到——病人与小提琴手分离在道德上是允许的。似乎我们中的一些人比其他人更擅长这项任务。他们成功地规定了案件描述中明确提到的细节，并适当地“填写”了更多细节。事实上，艾米·金德 (Amy Kind) 最近的工作探讨了这样一种观点，即我们在各种背景下或多或少都可以熟练地进行想象，包括思想实验的背景（例如，参见 Kind 2020）。 《形成印象》中开发的丰富框架的一部分可能会被重新设计，以适应基于想象力的思想实验观点。值得注意的是，查德诺夫关于直觉的一些核心主张似乎自然地延续到了（哲学）思想实验中使用的想象力练习中。据推测，我们期望特定领域的专家能够更好地想象该领域的场景。 受过训练的哲学家可能是这种更广泛现象的特例。此外，关于为什么哲学家更擅长想象汤姆森案例的查德诺夫式解释是令人满意的（即使它最终只是故事的一部分）。背景信念——可以通过考虑论证和使用查德诺夫强调的各种论证工具来获得——可以对我们如何想象场景产生影响，无论它们是否涉及哲学概念，在这种情况下，查德诺夫的策略可能对热衷于哲学的哲学家有用。捍卫思想实验在哲学中的使用，无论对直觉的感性主义有何承诺。对于有益的讨论，我感谢内文·克莱门哈加（Nevin Climenhaga）、伊利亚·楚德诺夫（Elijah Chudnoff）和博·麦迪逊·蒙特（Beau Madison Mount）。