# Movements of the Mind  ## Metadata - Author: [[Wu, Wayne;]] - Full Title: Movements of the Mind - Category: #books - Summary: insert summary - My notes: - Summary: "Movements of the Mind" is a book that explores the structure of action and attention. The author argues that attention is a necessary component of action and that intention plays a crucial role in biasing attention. The book examines the relationship between intention and attention, as well as the impact of biases on action. It also discusses the role of memory in intention and how it regulates attentional phenomena. Overall, the book presents a theory of attention, intention, and action, providing a comprehensive understanding of these cognitive processes. - Source File: Movements of the Mind Theory by Wayne Wu.epub ## LLM Chats ## NotebookLM ## LLM Audio ## Highlights > Crucially, we must not infer the absence of agency from the presence of automaticity ([View Highlight](https://read.readwise.io/read/01hr4qn91s31gd5373s7k6bmq6)) > A *behavior space* describes behaviors available to a system at a time and context. The space is constituted by potential *couplings* of inputs to outputs, a causal linking of both. A system’s behavior is the instantiation of a specific coupling, a path in the space ([View Highlight](https://read.readwise.io/read/01hr4s0ebjrrg097e602y61h82)) > An agent’s acting is the agent’s responding in a specific way, guided by how she takes things given her biases. ([View Highlight](https://read.readwise.io/read/01hrpn42tz4d5wvvd0167fa1bj)) - Note: Wu's Basic Structure of Acting > Behavior spaces of minded systems have input takings that are *mental* phenomena with intentionality. *Taking things* is a term of art, referring to subject-level states, intentional mental phenomena such as perception, memory, imagination, and thought. > The behavior spaces within which action emerges are thus *psychological* behavior spaces, specifically *action* spaces. To be an action space, input-output couplings must meet a certain profile: the intentionality or content of the subject’s taking guides her response. To take a basic form of guidance, one’s visual experience identifies a location to respond to. The experience guides by providing spatial content that informs a movement. Guiding content explains why the response occurs as it does. How content precisely sets response will be relative to the action kind in question, say setting parameters for movement or encoding content in memory. ([View Highlight](https://read.readwise.io/read/01hr4s2hpf5kdhx5xc0wxz8az9)) > Second, a *causal* action space identifies what the agent can objectively do at a time yet is detached from the agent’s conception of her options. To capture *her* perspective, we focus on a *doxastic* action space largely delineated by her beliefs about her options. A human agent typically only *knows* about a proper subset of the actual causal possibilities ([View Highlight](https://read.readwise.io/read/01hr4y36pbht9rttxza7t5gbqf)) > A *normative* action space delineates the actions an agent ought to do, in the relevant sense of “ought.” Such spaces, defined by relevant normative requirements, might contain only one action as when Martin Luther protested, “Here I stand. I can do no other” ([View Highlight](https://read.readwise.io/read/01hr4y4dxgvg7bc3xqexs639s6)) > To uncover action’s structure, contrast action with *reflex*. An agent undergoing a reflex evinces behavior but not action. Action admits of the qualifications noted earlier. Reflexes, however, do not express rationality, grasp of reasons, skill, or expertise. They are not targets of normative assessment, are never intentional or done freely. They lie outside the set of actions. ([View Highlight](https://read.readwise.io/read/01hr4x4p8z4vwd8a6bbj2r26ve)) > I suggest that it is the *necessitation* of response. ([View Highlight](https://read.readwise.io/read/01hr4x5h231rkkrea2z6rr09e6)) > Jonathan Cohen notes: “the distinction between controlled and automatic processing is one of the most fundamental and long‐standing principles of cognitive psychology” (2017, 3 ([View Highlight](https://read.readwise.io/read/01hr4yqkc8f3h1qzdcfw1a50c2)) > Intentions explain why the agent acts as she does: a specific path is taken because it is what the agent intends to do. The content of the agent’s intention specifies a solution to the Selection Problem. Intention prioritizes one path over others in action space, specifically, the path corresponding to the kind of action intended. Consider a basic case: for visually guided acting, the content of intention sets what visual target, hence visual taking, guides response. The agent’s intention biases input processing and required output response. “Bias” refers to the sources that influence solving the Selection Problem while “biasing” refers to that source’s shifting priorities in action space ([Appendix 1.1](#oso-9780192866899-chapter-2-appendix-1) describes a classic connectionist implementation of similar ideas for the Stroop effect). ([View Highlight](https://read.readwise.io/read/01hrpnrb4ycms6fjhm9kdacgja)) > Attunement to guiding features is *attention* ([Chapter 2)](#oso-9780192866899-chapter-3). ([View Highlight](https://read.readwise.io/read/01hrpnsf2ke131dm1mkn9ck5xb)) > In intentionally acting, the agent expresses control. She can intentionally move objects, come to conclusions, and, in general, change the world. At the same time, action exhibits substantial automaticity. Consider recalling events from the previous evening’s soiree. You express control in remembering that specific event rather than last week’s party. Yet much of memory is automatic as when recalling a cutting remark overheard or the foul flavor of the appetizers. These memories just spring to mind ([Strawson 2003](#oso-9780192866899-bibliography-1-bibItem-359); [Mele 2009](#oso-9780192866899-bibliography-1-bibItem-240)). ([View Highlight](https://read.readwise.io/read/01hr6scsnc4g2qqj9cg8dz2dy2)) > An advance in the theory of agentive control can be secured if the three claims can be rendered consistent. They can. The problem is that (3) is generally read as distinguishing *kinds* of processes, yet given (1) and (2), intentional actions are a counterexample to this interpretation. The solution is then simple: reject dividing *processes* as automatic or controlled. Rather, automaticity and control in agency are to be explicated in terms of *features* of processes. Accordingly, a single process can exemplify both automaticity and control. Intentional action certainly does. ([View Highlight](https://read.readwise.io/read/01hr6she7g5ey70w1x786m3yjt)) - Note: The article is talking about how people do things on purpose. They say that there are two important things that people do when they act on purpose: automaticity and control. This means that sometimes we do things without thinking, and other times we have to make a plan and think about what we are doing. The article says that these two things are not different kinds of actions, but rather different features of the same action. This is important because it helps us understand how we control our actions. > Not only is the agent in control in intentionally acting, she is also the guide. ([View Highlight](https://read.readwise.io/read/01hr6tr6rpse52kaw5f19wmknd)) > That there are one–many relations between intention and execution establishes the *necessary automaticity of action*. The challenge of action-relevant selection is not discharged just by intending to act. If one intends to drink the coffee, the targeted mug presents many properties only some of which are relevant to guiding an appropriate reach-and-grasp movement. Thus, the color of the mug is not relevant but the location and contours of the handle are. Even then, given different ways to pick up the mug, action-relevant properties can be further subdivided. Intention cannot represent all these properties. ([View Highlight](https://read.readwise.io/read/01hrppke5zcy60zrrprpwsnwfe)) > Similarly, when the going gets tough, the agent can transition from Φ-ing automatically to doing it deliberately, bringing action under control precisely because she has to think about what to do, fine-tuning her intention. ([View Highlight](https://read.readwise.io/read/01hr6sv9fgb12q9z7wha38phtp)) > These further Selection Problems are not to be solved by deliberation or through conceptualization, at least completely (I called such problems *non-deliberative problems*; [Wu 2008](#oso-9780192866899-bibliography-1-bibItem-391)). A distance necessarily remains between the abstractness of the action representation in intention and the concrete details of the action performed. This distance entails the necessity of automaticity in the actions of finite minds, for not every property of the action done can be controlled since not every property can be represented in intention. The distance can be narrowed by further practical reasoning that fine-tunes the content of the intention to render it more specific. Still, a gap will always remain given limitations on what we can hold in mind. The intention can never specify all relevant details of action at the precise level of determinacy. ([View Highlight](https://read.readwise.io/read/01hrppp96k8zpgb54v4mczvkak)) > An input can also be biased in being randomly activated sufficient to guide a response, say due to noise in the neural basis of the input. This can lead to thoughts that randomly pop into one’s head or the feeling of a cell phone buzzing in one’s pocket when there is no cell phone. Spontaneous activity of the neural basis of thought or perception drives a response. Buridan’s donkey might find itself munching on the left bale precisely because spontaneous neural activity altered the perceptual representation of that bale to drive a response, automatically. Perhaps that bale suddenly seems closer, larger, more delectable, and now the donkey moves. Similarly, I suddenly reach for my cell phone when I hallucinate a tactile buzzing in my pocket ([View Highlight](https://read.readwise.io/read/01hr6v2qzjh6nzcyy53wf7m86z)) > That there are one–many relations between intention and execution establishes the *necessary automaticity of action*. The challenge of action-relevant selection is not discharged just by intending to act. If one intends to drink the coffee, the targeted mug presents many properties only some of which are relevant to guiding an appropriate reach-and-grasp movement. Thus, the color of the mug is not relevant but the location and contours of the handle are. Even then, given different ways to pick up the mug, action-relevant properties can be further subdivided. Intention cannot represent all these properties. ([View Highlight](https://read.readwise.io/read/01hr6szdafy9n6td5yywve5jgp)) > We can explain passive action, understood technically, as constituted by action capacities which are activated fully independent of intention. Passive actions, like reflexes, are fully automatic, yet are distinct from pure reflexes in being solutions to the Selection Problem. Such passivity includes certain forms of mind wandering and daydreaming but also pathologies of action, such as passivity phenomena in schizophrenia, the hearing of voices or the experience of inserted thoughts (see Cho and Wu 2013 for an overview of mechanisms of auditory hallucination) ([View Highlight](https://read.readwise.io/read/01hr6va8e0cw2mst5327mxx03j)) > The virtuoso learned through practice and instruction, and in doing so she acquired both a capacity to act and, through intense practice, a skill. The shift from indirect to direct intention correlates with a shift from control to automaticity for certain features of the action. At the beginning, the specification of subactions is part of the agent’s conception of what it is to play the violin. In intending to do those subactions, her intention directly engaged extant action capacities present at the time of learning. Definitionally, these subparts, things she could directly do, were controlled in being explicitly intended. This allowed her to learn something more complicated based on things she already knew how to do. ([View Highlight](https://read.readwise.io/read/01hr6wk1h70e999wge8bnfppvd)) > the story of attention onto the structure of action in three modes: vigilance, atten*tion*, and atten*ding*. Of these, atten*tion* has explanatory priority, but start with the most complex of the phenomena, *attending* as an action, say looking for something. Looking is conceptualized as overt attention, a direction of the eye toward visual targets. Attending as action involves a coupling of input to output, here a visual experience of a target to an eye movement. The agent moves his eyes toward a target guided by his visual experience of it. He looks. The dynamics of looking were vividly demonstrated in Yarbus’ experiment ([Section 1.6](#oso-9780192866899-chapter-2-div1-6)). In action space, attending is constituted by a coupling of input and output, ([View Highlight](https://read.readwise.io/read/01hrpqg29d55t9a20cq1ep9vbh)) > Attention in the sense of that which is guiding response is a *component of action*, and is present in every action. ([View Highlight](https://read.readwise.io/read/01hrpqpb4rj42ezavr66ga0vkr)) > All three attentional phenomena are explicated in relation to the Selection Problem. To track the subtle distinction between attentional guidance *in* action and attending *as* action, I refer to the former, reluctantly, as *attention* and the latter as *attending.* The term *attention* is unfortunate since it suggests a state whereas guiding is a process, so I would prefer to refer to it as attending as well, thus courting ambiguity. What we have are attending*-in-action* (henceforth: *attention*) and attending*-as-action* (e.g. looking). Since we are explaining action in general, attent*ion* is prior in the order of explanation relative to attend*ing*, a type of action. As attention is present in all actions as the basis of guidance, this chapter explains what attention is. ([View Highlight](https://read.readwise.io/read/01hrpqsh0w4s5vjsfhm3vp86n7)) > We, the folk, do reliably track attentional phenomena because we have a distinctive form of access to our intentional agency. We can know about our intentional actions without observation ([Anscombe 1957](#oso-9780192866899-bibliography-1-bibItem-14)). We can intentionally attend to something to act on it, and we can accordingly know what we “pay attention” to, that is, what we are acting on. Privileged access provides an initial justification for affirming William James’ claim that we know what attention is: > Everyone knows what attention is. It is the taking possession by the mind, in clear and vivid form, of one out of what seem several simultaneously possible objects or trains of thought. Focalization, concentration, of consciousness are of its essence. It implies withdrawal from some things in order to deal effectively with others. ([James 1890](#oso-9780192866899-bibliography-1-bibItem-190), 403) ([View Highlight](https://read.readwise.io/read/01hrpr90kss3sb2ch9g8rsmv90)) > To explain what attention in action is, I draw on David Marr’s theory of explanation for psychological phenomena ([Marr 1982](#oso-9780192866899-bibliography-1-bibItem-233), ch. 1). Marr outlined a schema for answering questions of the form “What is *X*?” An answer is structured by three levels: > (a) the *computational level* that specifies what the phenomenon is for, its function and logic; > (b) the *algorithmic level* that specifies the relevant representations and algorithms for the function, > (c) the *implementation level* that specifies the hardware that realizes the algorithms, say specific circuits in the brain. > It is at the computational level that philosophical inquiry most directly engages. ([View Highlight](https://read.readwise.io/read/01hrpry0q3hzhej50m5ykbtje6)) - Note: nice summary of Marr > if *S* selects *X* to guide action, then *S* attends to *X*. > Call this the *Jamesian Condition on Attention*. ([View Highlight](https://read.readwise.io/read/01hrrw6fcwkt2thapz9nr2ghw6)) - Note: The Jamesian Condition > In a slogan: *attention is guidance in action*. When an input, the subject’s taking *X* rather than other targets, guides response, the mind thereby takes possession of *X* rather than others in order to deal with it. Guidance, selection, and coupling are explicated within the structure of action. We thus merge Jamesian and empirical conceptions of attention with the structure. That is, we link the psychological theory of attention to the metaphysical theory of action yielding a metaphysical theory of attention in action. The confluence of well-grounded folk psychology, empirical psychology, neuroscience, and the philosophy of action provides strong grounds for treating selection for action as the right theory of attention ([View Highlight](https://read.readwise.io/read/01hrpsf6973pesk4d4pp4jzxmm)) > *Attention is not a mechanism modulating neural activity; rather it, as a subject-level activity of guiding response, is constituted by specific neural modulations.* > Attention has a causal impact by guiding response. That is its primary causal role, part of an agent’s acting. Readers with a passing familiarity with the empirical literature on attention, however, will have come across the metaphor of visual attention as a spotlight. On this common conception, attention is like a tool that one, or one’s brain, can *deploy*, highlighting a target. In the science of attention, attention is said to modulate psychological and neural processing. That is, attention is a causal factor that scientists appeal to in order to explain other biological phenomena. I disagree, for attention is not a cause or explanans as the spotlight metaphor implies. Attention is an explanandum, what we are trying to explain. We must push back on this causal conception of attention because it distorts our understanding of attention at multiple levels of analysis, including philosophical. ([View Highlight](https://read.readwise.io/read/01hrrvzp08ytz2cnr2ka5cgasm)) > The solution is to demonstrate that there is a *single* characterization of attention accepted (typically implicitly) by all scientists of attention who treat standard experimental paradigms as valid, namely that expressed by the Jamesian Condition. ([View Highlight](https://read.readwise.io/read/01hrrw3s85rbd18qahm9331k3b)) - Note: "if S selects X to guide action, then S attends to X." From 2.2 > I focus on visual attending as an action, hence as having the basic structure: an input guides an output where that input is atten*tion*. ([View Highlight](https://read.readwise.io/read/01hrrwdjzpj7m5exhf1bzjbfa0)) > The sign of mundane attentional capture is automatic *orienting*, the movement of the body to bring a sensory organ to an optimal position relative to the stimulus, say fixating on a visible stimulus by shifting one’s eyes, turning one’s head to track a sound or haptically exploring an object carefully with one’s fingertips. This is *overt* perceptual attending. ([View Highlight](https://read.readwise.io/read/01hrrwf3xg43c6dhb9q2xka65v)) > Introspection suggests that when we switch from verbal source *A* to *B* and back, there is a difference in how the sources strike us. When attended, the words in *A* are clear while we do not, in some sense, hear *B* with the same auditory acuity, a fact that can lead to significant social awkwardness. This is reversed when we voluntarily switch to *B*. Extensive behavioral studies in the 1950s and 1960s bore this out (reviewed in [Wu 2014](#oso-9780192866899-bibliography-1-bibItem-396), ch. 1; see also [Pashler 1998](#oso-9780192866899-bibliography-1-bibItem-288), ch. 1). These gave rise to one of the first major debates about attention: is sensory selection early or late in processing? For example, subjects show a decrement in their ability to identify the words voiced in the unattended channel though they access basic features such as the voice’s pitch. Attending involves changes in sensory phenomenology and accessibility rooted in shifts in sensory processing. Covert auditory attention as experienced at a party provides a familiar case where we can change auditory attunement at will, flipping back and forth between auditory streams and sharpening our access to attended sounds at the expense of others. ([View Highlight](https://read.readwise.io/read/01hrrwrd1p1x1jksvqq9h3h8f9)) > 4.6 First-Personal Access to Intentional Action ([View Highlight](https://read.readwise.io/read/01hr6zdpndrnd6f19bx9trjwq0)) - Tags: [[toread]] > y goal in this section focuses on explaining this common ground regarding access: that the agent’s reliably accurate reporting (or her judging, taking it to be the case, etc.) that she is Φ-ing bears the conjunction of familiar features attributed to practical knowledge. Such accessing is non-observational, authoritative, and distinctively first-personal. The results of access provide evidence of knowledge, and accessing what we are doing is an intentional action we can explain. The structure of action shows that these features co-occur because the agent’s intention in action is the agent’s cognitive attention to her action that informs her report of what she is doing. ([View Highlight](https://read.readwise.io/read/01hrp929m33mcajv6t5st1ew8z)) > Access draws on practical remembering, namely the updating of intention by fine-tuning. ([View Highlight](https://read.readwise.io/read/01hrp92gysnywfmw267rppyh4s)) > Of course, if we are raising questions about introspective accuracy, one side cannot simply claim that its report is the correct one. Which report is accurate, the naive response or the proposed correction? In support of the latter, the biology of the visual system explains why clarity decreases away from the fovea: spatial resolution drops off as one moves away from the fovea, the size of visual neurons’ receptive fields increases in the periphery leading to less acute discrimination, the cortex magnifies representations of central vision at the expense of processing in the periphery, and there is a psychophysically measurable drop in behavioral performance in discriminating peripheral detail. If we take visual consciousness to have a neural basis, then the nature of our experience should cohere with the biology. The loss of acuity makes sense given the corresponding differential distribution of neural representational resources. So, we have independent grounds to accept the probity of the correction. ([View Highlight](https://read.readwise.io/read/01hr6zw2vftctbf01t3q838yqv)) > Indeed, a savvy subject might retort that her introspection was accurate: visual experience is spatially acute throughout the visual scene. After all, normal visual experience is temporally extended, served by multiple saccadic eye movements, one to three times a second. In extended experience, much of the visual world is spatially acute across time precisely because we foveate much of the visual scene over time. Wasn’t that the point of asking the subject to look around, to assess experience over time? Schwitzgebel considers fixations over time but only to explain how the subject is mistaken in concluding that everything is “all clear *simultaneously*” (2008, 255; my emphasis). Yet the savvy subject can aver that she is not getting it wrong relative to the task instruction which was this: “Look around a bit. Consider your visual experience as you do this.” When looking around a bit, why isn’t it accurate to say that one’s view is largely clear? ([View Highlight](https://read.readwise.io/read/01hr6zxyxbz23kek0zjbnr5hhc)) > by focusing your attention on the phenomenal quality of [an experience], you can come to know something about your current experience. Philosophers generally agree on this much. ([Gertler 2012](#oso-9780192866899-bibliography-1-bibItem-151)) ([View Highlight](https://read.readwise.io/read/01hr705yp59yjcydan1wvbcd4r)) > When we attend to our own mental states, it feels like that is just what we are doing: focusing our internal attention on something that is there for us to discern. (Lycan 2004) ([View Highlight](https://read.readwise.io/read/01hr70agtaw6jq44q05rx7bxzg)) - Tags: [[quote]] > Many other philosophers endorse what we might call an *inner spotlight model* where we can directly attend to the phenomenal properties of experience (Chalmers 2003; Lycan 2004; Levin 2006; Papineau 2006, 2002; Carruthers 2011; Balog 2012; Gertler 2012; Horgan 2012). Yet no empirical evidence for the existence of the type of attention postulated is presented, and little is done to investigate the nature of introspective attention. This is a striking and puzzling lacuna. What follows fills this gap, drawing on the theory of attention ([Chapter 2)](#oso-9780192866899-chapter-3). ([View Highlight](https://read.readwise.io/read/01hr709n8cg9raegssam080n8b)) > Expanding the basic structure in [Section 4.6](#oso-9780192866899-chapter-5-div1-6) by drawing on Evans’ insight, I suggest an empirical hypothesis: the mind generates accurate introspective judgments regarding perceptual experience by drawing on the capacities deployed in generating accurate perceptual reports including *perceptual* attention. ([View Highlight](https://read.readwise.io/read/01hr70h0wqphexxg5j1bmhrmhz)) > Figure 7.1 An object with feature *T* is visually experienced. The guiding feature *T* informs, through being seen, a conceptual response, the application of the visual concept *T* in a report: “There is a *T.*” ([View Highlight](https://read.readwise.io/read/01hr71hbshjpc9662n7qp8s6pn)) > Some might find it puzzling that the same form of attention can guide both perceptual and introspective report. Yet we deploy the same form of attention for many different responses. Consider all the different bodily movements that can be made based on the same form of perceptual attention: jabs, reaches, kicks, grasps. Similarly, many different mental responses can be made guided by the same perceptual attention: imagining, encoding, reasoning about, etc. Thus, seeing the location of an object, can guide a myriad bodily and mental responses to it. The same form of visual attention can serve a myriad responses, including introspective report. Crucially, on this view, talk of *introspective* attention does not identify a distinct type of attention that differs from perceptual attention (e.g. one internally rather than externally directed). Rather, it specifies a different *deployment* of perceptual attention, namely for introspective response. Introspective attention is, here, perceptual attention for introspective response in report or judgment. The result is an Evansian attention model of introspective judgment that involves the redeployment of basic attentional capacities exercised in observational judgment (cf. Byrne 2018). Unlike raw appeals to a form of introspective attention we do not understand, the model presented here draws on psychological capacities that we do understand. ([View Highlight](https://read.readwise.io/read/01hr71pqzd1a20e88z6g6wb528)) > In contexts where perceptual judgments of *X* tend to be accurate in sensory modality *S*, then simple introspecting that one *S*-experiences *X* will also tend to be accurate. > Thus, in contexts where one is reliably accurate in visually reporting the shape of an object, one will tend to be reliably accurate in introspectively reporting *seeing* that shape. In both cases, the report is based on visually attending to the shape. ([View Highlight](https://read.readwise.io/read/01hr72351w2f3p9by4bt3bj7vn)) > Introspection of peripheral blur is clearly not univocal across subjects, and that should give us pause. Indeed, this variability is among the data to be explained. After all, given a common biological basis explaining the loss of peripheral visual acuity, we might expect that sameness or similarity of biology will ground sameness of peripheral visual phenomenology. ([View Highlight](https://read.readwise.io/read/01hrp9vyetxa5xazd15pj34ge2)) > Detecting blur at the optometrist’s office draws on two forms of attention, perceptual attention to (say) the letter “E” on the chart given one lens power, and cognitive attention to, namely memory of, past perceptual experience of the same “E” given another lens power. That is, we compare the current letter experienced with one lens against the maintained memory of the letter experienced with another lens. ([View Highlight](https://read.readwise.io/read/01hrp9xg535gt71vygn8zkwaqr)) > Accordingly, when introspecting veridical and hallucinatory experience, visual representational content guides the application of the same phenomenological concept in both cases via the capacity to apply the requisite observational concept, *RED*: it visually seems to me red. Thus, perceptual and introspective reliability can come apart in that perception is not reliably accurate in hallucination but introspection is. When the subject renders a perceptual report in hallucination, the reliability of the attentional link between experiential content of red and the corresponding judgment that there is something red is offset by the unreliability of perception which is, by hypothesis, disconnected from the world. Resulting perceptual judgments based on hallucinations are reliably wrong. In introspective judgment, however, the unreliability of perception need not offset the reliability of attention. The content of the hallucination links to introspective judgment, and arguably, that judgment about visual experience will be correct, so long as one uses more cautious concepts: “It visually seems to me as if *X*” rather than “I see *X*.” ([View Highlight](https://read.readwise.io/read/01hrpaq8kh50ynjvqy66rmp9k5)) > Representationalists find this specific relationalist account of hallucination absurd (it is not obligatory for relationalists to deny phenomenology in hallucination). Their response is partly driven by confidence that introspection clearly (infallibly) reveals phenomenology in hallucination. How could one possibly be wrong about how things (phenomenally) seem while hallucinating? Yet this standard response about introspection begs the current question about introspective reliability in cases of perceptual unreliability. What is at issue is not that introspecting subjects will say and firmly believe that hallucination has phenomenology. The relationalist can agree to that. Rather, the question is whether these judgments are accurate. That is not obviously so. ([View Highlight](https://read.readwise.io/read/01hrpaw0wm40fvs83s8wkrzyt0)) > Thus, even if the relationalist concedes the presence of a common factor, the action capacity normally deployed in making a perceptual judgment, the requirement for non-question begging conceptualization of this factor correctly leaves the metaphysical question about sensory phenomenology open. ([View Highlight](https://read.readwise.io/read/01hrpb1dnf1q3v0ztsdx25wdnp))