Making sense of the world: new essays on the philosophy of understanding 9780190469863, 0190469862

Making Sense of the World offers original work on the nature of understanding by a range of distinguished philosophers.

595 40 2MB

English Pages 262 [273] Year 2018

Report DMCA / Copyright


Polecaj historie

Making sense of the world: new essays on the philosophy of understanding
 9780190469863, 0190469862

Table of contents :
1. Introduction / Stephen R. Grimm --
2. The Unity of Understanding / John Bengson --
3. Understanding Persons and the Problem of Power / Remy Debes --
4. Religious Understanding, Naturalism, and Desire / Fiona Ellis --
5. Philosophy, Knowledge, and Understanding / Gordon Graham --
6. The Ethics of Understanding / Stephen R. Grimm --
7. Understanding and Structure / Allan Hazlett --
8. Aesthetic Understanding / Alison Hills --
9. Why (Study) the Humanities?: The View from Science / Jenann Ismael --
10. Understanding and Coming to Understand / Michael Lynch --
11. Against Understanding (as a Condition on Explanation) / Bradford Skow --
12. Understanding and Fluency / J. D. Trout.

Citation preview

Making Sense of the World

Making Sense of the World New Essays on the Philosophy of Understanding Edited by



3 Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and certain other countries. Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America. © Oxford University Press 2018 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by license, or under terms agreed with the appropriate reproduction rights organization. Inquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above. You must not circulate this work in any other form and you must impose this same condition on any acquirer. CIP data is on file at the Library of Congress ISBN 978–​0–​19–​046986–​3 9 8 7 6 5 4 3 2 1 Printed by Sheridan Books, Inc., United States of America


Contributors  1. Introduction—​S tephen R. Grimm  

vii 1

2. The Unity of Understanding—​J ohn Bengson  


3. Understanding Persons and the Problem of Power—​R emy Debes  


4. Religious Understanding, Naturalism, and Desire—​F iona Ellis  


5. Philosophy, Knowledge, and Understanding—​G ordon Graham   98 6. The Ethics of Understanding—​S tephen R. Grimm  


7. Understanding and Structure—​A llan Hazlett  


8. Aesthetic Understanding—​A lison Hills  


9. Why (Study) the Humanities?: The View from Science —​J enann Ismael  


10. Understanding and Coming to Understand —​M ichael Patrick Lynch  


11. Against Understanding (as a Condition on Explanation) —​B radford Skow  


12. Understanding and Fluency—​J. D. Trout  





John Bengson, Associate Professor of Philosophy, University of Wisconsin—Madison Remy Debes, Associate Professor of Philosophy, University of Memphis Fiona Ellis, Professor of Philosophy, University of Roehampton Gordon  Graham, Henry Luce III Professor of Philosophy and the Arts, Princeton Theological Seminary Stephen R. Grimm, Professor of Philosophy, Fordham University Allan Hazlett, Assistant Professor of Philosophy, Washington University, St. Louis Alison Hills, Professor of Philosophy, University of Oxford Jenann Ismael, Professor of Philosophy, University of Arizona Michael Patrick Lynch, Professor of Philosophy, University of Connecticut Bradford Skow, Associate Professor of Philosophy, MIT J. D. Trout, John and Mae Calamos Professor of Philosophy, Illinois Institute of Technology


Introduction Stephen R. Grimm

1.1. Introduction When we understand something, we are able to make sense of it. It could be some particular event that was puzzling, and is now clear. As in: “Why is he in such a good mood?” “Because he just received a job interview.” Or: “Why is the traffic moving so slowly?” “Because a tire is being changed by the side of the road, and there’s rubbernecking.” In these sorts of cases, making sense puts an end to puzzlement or lack of understanding, but on many occasions there is no puzzlement to begin with, as we automatically and easily make sense of things. When I see the baseball strike the window and the glass shatter, I do not pause while I try to work out why the glass shattered. I naturally and automatically attribute the shattering to the ball. Or again, when I hear a sentence spoken in English, for the most part I make sense of it without effort; only rarely do I need to work to understand what is being said. The nature of understanding, or of making sense, has received new attention among epistemologists, philosophers of science, and others for a few different reasons. For one thing, several epistemologists have recently argued that understanding represents a significantly different accomplishment than knowledge, and perhaps a more valuable one (see, e.g., Zagzebski 2001; Kvanvig 2003; Elgin 2006; Pritchard 2010). In the philosophy of science, the hope is that gaining clarity about understanding will help us to address longtime questions surrounding the nature of explanation (see, e.g., de Regt 2004; Lipton 2009; Strevens 2013).



In this introductory chapter, I  first expand on these debates in epistemology and the philosophy of science in §§1.2 and 1.3. In §1.4, I  then turn to types of understanding that have been relatively neglected in the recent literature, but that receive significant attention in this volume from a number of philosophers, including Jenann Ismael, Remy Debes, and Fiona Ellis. My goal in these sections is not to offer an exhaustive overview of recent work on understanding, nor to assimilate all of the contributions made by the essays in this volume. Instead, I try to offer a taste of some of the lively questions that are currently being addressed in the field, drawing on essays from this volume where relevant. In §1.5, I conclude by summarizing the chapters in this collection in a more systematic way.

1.2. Understanding in Epistemology One reason why many contemporary epistemologists have been drawn to the topic of understanding is that it seems to be satisfyingly “internal” in a way that knowledge allegedly is not, especially if understanding is construed, as above, as a kind of sense-​making. Note, for example, that a claim like “You think it makes sense to you, but it really doesn’t” sounds strange and perhaps incoherent. On the other hand, a claim like “You think you know, but you really don’t” is often perfectly apt. The difference seems to stem from the fact that you have a privileged perspective—​ an internal perspective, accessible by reflection alone—​ on whether you have managed to make sense of something in a way that you lack when it comes to knowledge. When something makes sense to you, for instance, there seems to be a sort of mental “click” that takes place. You suddenly “get” the thing in question. And how could you be mistaken about that? Pushing further on the distinction between understanding and knowledge, in this volume Alison Hills argues that understanding can be lucky in a way that knowledge cannot. Just by luck,1 you might come to learn about a certain causal relationship, and thereby acquire the right kind of “cognitive control” over the relationship that yields understanding. But, Hills argues, you would nonetheless not know the relationship obtains, because luck cancels out knowledge. This line of thinking also appears to reinforce the “internal” character of understanding or sense-​making, because it suggests that

1.   “By luck,” in the sense that you might easily have been mistaken about it.



what is important for understanding is how you manage the information you possess—​what you can do with it, and the sorts of patterns you can “see” within that information—​rather than how you came by that information in the first place. But there is also another side to the story. According to an opposing view in epistemology, understanding is not distinct from knowledge, but rather a species of it; what’s more, several philosophers argue that one typically cannot tell, just by reflection alone, whether one understands. This approach is represented in this volume by Allan Hazlett and Michael Lynch, and by some of my earlier work (e.g., Grimm 2006 and 2016). One of Hazlett’s examples nicely illustrates the idea: Consider someone who thinks that CO2 molecules are hotter than O2 molecules. She learns that the rising temperature is correlated with increased atmospheric CO2, and insightfully connects the hotness of CO2 with the changes in temperature. She thinks she understands why global mean temperature is rising—​it all makes sense to her; she has that distinctive feeling of “getting it”—​but she does not understand. Can she determine, by reflection alone, that she does not understand? Reflection is not what she needs—​what she needs is to have her mistaken picture of the connection between CO2 and temperature corrected, and in particular her false view that CO2 molecules are hotter than O2 molecules. But reflection will not disclose the falsity of that view. (Hazlett, this volume) Here we have a plausible case showing that from the first-​person point of view it can be very hard, perhaps impossible, to tell whether one understands. One might have felt the mental “click” described a moment ago—​the whole story might feel right—​but there might be nothing in the world that answers to the click. Apparently, then, understanding can often be illusory, just like apparent cases of knowledge. How then can we reconcile these conflicting intuitions about understanding? On the one hand, that there is a signature internal element that seems transparently accessible to the understander; on the other, that it can often be opaque, from the first-​person point of view, whether one really understands. A distinction might help (Grimm 2011). Let us say that some item—​ an event, say, or a subject matter—​subjectively makes sense to a person if it in some way “fits” or follows from other things the person takes to obtain



(cf. Riggs 2003). To objectively make sense of the item requires more, however. In addition to the subjective element just mentioned, it also requires that the putative explaining factors actually obtain, and whether or not these factors actually obtain will often be opaque to reflection alone—​they might appeal, for instance, to lawlike relationships or real dependencies in the world. It is therefore possible for someone to subjectively make sense of or understand some item without actually, or objectively, making sense of or understanding it. While I think this much seems plausible, one reason why understanding is an intriguing state is that this subjective sense of understanding is often such an achievement in its own right that we are naturally led to “focus on” it (Kvanvig 2003; 2009), rather than on any of the worldly relationships it might invoke. There is persisting tendency to suppose, for example, that conspiracy theorists, astrologers, and so on, have genuinely achieved their own understanding of the world, even though we would not dream of saying (at least, the epistemologists among us) that they have achieved their own knowledge of the world. A further reason why understanding continues to fascinate is that it seems to combine elements of rationalism and empiricism in a way that one does not find with ordinary instances of knowledge (perceptual knowledge of one’s environment, say). Along with the rationalist tradition, it appeals to notions of “grasping” or “seeing” which seem like special acts of the mind—​ acts that don’t just take in how things actually are, but how they must be, or how they might be. At the same time, however, along with the empiricist tradition, understanding typically tracks contingent relationships in the world. Thus understanding often hinges on discovering how the various elements of the world depend upon, and relate to, one another (cf. Lynch, this volume). Plausibly, a fully satisfying theory of understanding needs to combine both of these elements—​a rationalist element of grasping or seeing and the worldly relationship grasped or seen. The rationalist element is so intriguing, however, that it can often seem like the only part of the story, or at least the most important part.

1.3. Understanding in the Philosophy of Science Turning to issues in the philosophy of science, one prominent question is how to think about the relationship between understanding and explanation. Although the two notions are clearly intertwined, it is not easy to sketch out how the relationship works.



On the one hand, it seems like understanding is in some way normative for explanation (Turri 2015). Thus we often hear that we seek out explanations of various phenomena—​Brownian motion, the 2008 stock market collapse—​because we want to understand those phenomena. But then presumably explanations that fail to generate understanding are in some way lacking; they are not doing their job, so to speak. As a first pass, it is therefore tempting to think that we can evaluate the quality of an explanation based on whether it manages to yield understanding in real people, actual agents. On the other hand, it seems equally clear that real people, or actual agents, can be poorly situated in all sorts of ways, and can lack the cognitive tools required to appreciate good explanations. Thus a good explanation for the 2008 financial meltdown might appeal to obscure financial instruments like collateralized debt obligations or credit default swaps—​but this explanation might not help you in any way if its central concepts are alien to you, or if you lack the appropriate background information. Actual agents therefore can be, and often are, inapt judges of the quality of an explanation (see especially Bradford Skow’s critique in this volume). What’s more, and as J.  D. Trout has forcefully pointed out, an actual agent’s “sense” of understanding can be remarkably unreliable (Trout 2002; this volume). Thus an objectively poor explanation—​think again of conspiracy theories, or astrological predictions, etc.—​might make perfect sense to someone, and might lead to the sort of mental “click” described in the previous section, even though it provides no genuine understanding of the phenomena. A natural alternative would therefore be to turn to something like an “ideal agent” test for the goodness of an explanation. Thus we might say that a good explanation is one that would produce understanding in an agent possessing all of the relevant background knowledge, and lacking cognitive limitations. But now the concern is that because we have very little idea of what such an ideal agent might be like, we have little grip on how such an agent’s understanding could be normative for explanation. In short, since actual agents can be poor judges, and since ideal agents are shadowy creatures of the night, we should plausibly dispense with such tests altogether. Instead, we should try to evaluate explanations in terms of certain formal features, and not be concerned about their vague and uncertain connections to understanding. These worries, however, can be overblown. For we all think there are countless people who actually do understand things about the world—​indeed,



presumably we all take ourselves to understand many things about the world. But then we can ask: What is it that constitutes this understanding? What is it that the understander grasps? If we can give an informative account of the content of the grasp, then it is natural to think that a good explanation will be one that offers that sort of information, and a subpar explanation one that leaves that sort of information out. To illustrate the idea:  I  have argued elsewhere, following Jaegwon Kim (1994), that what understanders grasp are dependence relations, and that it is in virtue of grasping these relations that the understander can put his or her finger on why things are as they are, rather than otherwise (cf. Lynch, this volume). A good explanation will therefore be one that provides information about these relations, sometimes formally expressed in laws (which encapsulate these relations, often mathematically), but sometimes more informally expressed is causal stories. For instance, a good explanation for why the tides are running high will appeal to the proximity of the moon, and a good explanation for why people are refinancing their homes will appeal to the recent drop in interest rates. These explanations yield understanding—​I have argued, at least—​ because they provide information about the way in which the height of the tides depends upon its distance from the moon, and how borrowing patterns depend upon interest rates. They also allow us to see, by extension, what these phenomena do not depend on. Thus the explanations tell us, by their silence, that the height of the tides does not depend upon things like my personal distance from the coast, or my fondness for low tides. Similarly, they tell us that borrowing depends upon interest rates, but not upon how many fireflies will occupy my lawn this July. To achieve understanding, according to this model, an agent therefore needs to grasp these dependencies—​but of course all of us are not cognitively equipped to do this grasping, either because we lack the necessary concepts, or because getting a handle on the relationships exceeds our processing power. (Think of how only supercomputers can now crunch some of the data arising from science, and how this might spell a real limit to our understanding of the world.) Of course, this view about the relationship between explanation and understanding is a substantive one, and others might say (for instance) that a good explanation yields understanding because it shows why the phenomena had to happen. Alternatively, others might hold that a good explanation yields understanding because it reveals the mechanism that brought about the thing to be explained.



One benefit of the renewed interest in understanding in the philosophy of science, however, is that it might, and perhaps should, prompt advocates of these views to be more explicit about how exactly this explanatory information is supposed to generate understanding. What is it about learning mechanisms that is supposed to be so epistemically beneficial? If it is because mechanisms give us information about why things had to be a certain way (here one might think of the gears of a clock, forcing and turning things), then perhaps we can get information about why things had to be a certain way without appealing to mechanisms. Perhaps mechanisms are just one route toward this epistemic good among others. And so on. Theorizing about understanding and explanation in tandem therefore promises to breathe new life into debates about explanation in the philosophy of science that were arguably running out of steam at the turn of the 21st century (cf. Newton-​Smith 2000, 132).

1.4. Understanding without Grasping? According to the “grasp of dependence relations” view, understanding therefore involves locating a thing within a network of possibilities, because understanding turns on the ability to anticipate or predict how some phenomena would change (or fail to change) in light of other hypothetical changes. As noted earlier, Hills has nicely characterized this sort of grasping in terms of having “cognitive control.” In part what this means is that someone who understands is able to do something—​not necessarily intervene to make changes in the system (that would be difficult to do with many systems—​ think of planetary systems), but at least to make accurate judgments about what would happen, or would be the case, were other changes to occur. That is, the understander on this view is able to say how things would turn out if matters were different (cf. Woodward 2003). Because understanding and our sense of possibility therefore seem to be closely linked, it is tempting to think that our understanding of the world will grow or increase as our sense of possibility grows or increases (cf. le Bihan 2016). And this in turn suggests that there is a way of thinking about understanding that is not tied to making sense of things or figuring them out—​at least, if these two expressions imply a kind of manipulative control, or an ability to do things. Instead, it might involve a kind of receptivity to different ways the world might be, including different ways of living as a human being. The chapters by Jenann Ismael and Fiona Ellis in this volume both explore this ground. Ismael’s special focus is what we might call “humanistic understanding,”



and her view is that our understanding increases when we learn from great works of literature and art because they open up new possibilities for living, and help us to evaluate which of these possibilities are worth pursuing and which are not. In particular, Ismael claims that our sense of what makes for a good (or bad) way of life increases when we engage deeply with these works. This is a type of understanding that also seems importantly tied to wisdom, which might be understood as a knowledge of how to live well (cf. Grimm 2015). Ellis likewise writes of a kind of understanding, especially tied to religious experience, that comes from desiring God in the right way—​that is, apparently, in the way that will lead one farther along in one’s pilgrimage toward the divine. She also cites John Cottingham’s recent claim that there is a kind of understanding that consists in being properly attuned to the world, or open-​ mindedly receptive to it. Both Ellis’s and Cottingham’s views might therefore be grouped under the label of something like “understanding as proper orientation to the world.” These new approaches to understanding also raise several interesting questions, however. First, is there anything that might bind these various types of understanding together? That is, anything that might link the understanding that (plausibly) comes from grasping dependency relations, the understanding that (plausibly) grows as our sense of possibility grows, and the understanding that (perhaps!) arises from being properly oriented to, or attuned to, the world? Is the term “understanding” being used in unconnected ways in these different contexts, or is there a common thread that connects them? John Bengson addresses all of these questions and more in his insightful essay in this volume. Second, do these putative varieties of understanding all deserve to be thought of as “epistemic” goods, at least if we think of epistemic goods as roughly akin to “truth-​connected” goods? The answer seems to be yes for the grasp-​of-​ dependency-​relations view, because clearly we might have an accurate or inaccurate sense of how the various elements of a system depend upon one another, and there will clearly be accurate or inaccurate answers to “what if things had been different?” questions. Similarly, although more controversially, it seems like one could develop accurate and inaccurate views about which ways of life might be good or bad for a person, fulfilling or unfulfilling. But what might it mean for an orientation to the world to be “epistemically” apt or inapt? Perhaps with Heidegger one might say that more about the world is “disclosed” via an apt orientation than an inapt one—​hence one learns more about the world, but not in a way that can easily, or at all, be translated into propositional terms. But that is an idea, needless to say, that would need to be developed more fully.



1.5. A Guide to the Essays In this concluding section, I will now relay brief summaries of the chapters. While noting that understanding comes in a variety of forms, John Bengson argues for the unity of these forms, against the common tendency to view them as fundamentally heteronomous, or disunified. After identifying ten features of genuine understanding, which enable an argument for the existence of two distinct types of understanding, theoretical and practical, he poses a dilemma for theories that view them as disunified. Subsequently, he develops and defends a general account of understanding in terms of conceptions. What unifies diverse forms of understanding, on Bengson’s account, is a generic cognitive structure—​given by the notion of a noetic conception—​that is realized in different ways in various cases. In his chapter, Remy Debes points out that recent work in social epistemology suggests that those with power actively and passively hinder those without power from interpreting and communicating their experiences of suffering and persecution, thus obstructing their role in the production of knowledge about these experiences. This kind of “epistemic oppression” raises a puzzle about the nature and possibilities of interpersonal understanding, which he calls the problem of power. Put simply, if what counts as “knowledge” is regulated by those with power, then can empowered people ever genuinely understand oppressed people? Debes attempts to answer this question, which in turn leads to a new theory of what it means to understand other persons. Fiona Ellis takes up the image of a transformative experience through pilgrimage to capture what is at issue when the notion of religious understanding is introduced. She challenges the naturalist’s objection to the very idea of there being a journey in this sense, grants with John Cottingham that the transformation is moral and spiritual, and considers what it could mean for such understanding to be theoretical as well as practical. Further questions arise concerning the ‘fuel’ of this transformative journey, and she dialogues with Levinas’s claim that it is motivated by desire. She considers the merits of his position and concludes that it offers the shape for a model of religious understanding that can genuinely appeal to an expansive, i.e., nonscientistic, naturalist. Gordon Graham begins with the unfavorable comparison that has often been made between the substantial progress the natural and social sciences have made, and the interminable and inconclusive debates that philosophers engage in. His chapter goes on to examine the Humean attempt to make philosophy more scientific and the Lockean conception of philosophy as an “underlaborer” for the sciences. Graham argues instead for a conception of



philosophy as the pursuit of cognitive goals other than the kind of knowledge and explanation that are marks of the sciences. Stephen R. Grimm asks why the memorable French proverb “To understand all is to forgive all” should have even a ring of truth to it. After all, most of us—​probably all of us—​are less than admirable in various ways. We are shot through with vices such as vanity, cowardice, jealousy, and pride. But then why should coming to understand these elements of someone else lead to anything other than disapproval? In my chapter I review several different ways of thinking about understanding and argue that no one specific form of understanding addresses all the relevant cases neatly. If there is a legitimate connection between understanding and goods such as forgiveness or compassion, it seems to be a multifaceted one—​sometimes involving an understanding of the conditions that give rise to someone’s behavior, sometimes understanding the complexity of another’s situation, and sometimes involving an understanding of our common frailty. In his chapter, Allan Hazlett explores two claims about metaphysical structure: that “carving nature at the joints” is a valuable intellectual achievement and that understanding is constituted by a “grasp” of explanatory structure. He argues that explanatory understanding requires “carving nature at the joints,” but that neither “carving nature at the joints” nor understanding is plausibly seen as “the aim of belief ” or the “the aim of inquiry.” His chapter concludes with a discussion of the metaphysical preconditions for explanatory understanding through a discussion of the role of socially constructed properties in explanations. Alison Hills applies her well-​known account of understanding why to aesthetic understanding, by which she means understanding why a work of art is aesthetically valuable. She first develops some of her core ideas—​that understanding involves a kind of intellectual know-​how, that it is a matter of degree, and that it differs from knowledge. She then considers aspects of the account that may seem questionable when applied to aesthetics, notably that understanding involves explanation, and that that explanation can be articulated. Hills concludes by using her account to illuminate a particularly important activity: the appreciation of a work of art. Jenann Ismael’s chapter addresses the relationship between the humanistic and scientific visions of the human being, argues that the humanistic vision is not undermined by what science is teaching us about ourselves, and then turns to a discussion of the kind of understanding that the humanities provide. She claims that this understanding differs from the kind of understanding provided by the sciences, and that it is indispensable to human flourishing. Overall, Ismael argues, the humanities enrich our experience of the



world; educate the imagination; help us understand ourselves and other people; and teach us how to live, how to love, and how to feel. Michael Lynch notes that many philosophers take understanding to be a distinctive kind of knowledge that involves grasping dependency relations; moreover, they hold it to be particularly valuable. His chapter investigates and addresses two well-​known puzzles that arise from this conception: (1) the nature of understanding itself—​in particular, the nature of “grasping”; (2) the source of understanding’s distinctive value. According to Lynch, we can shed light on both puzzles by recognizing, first, the importance of the distinction between the act of coming to understand and the state of understanding; and, second, that coming to understand is a creative act. The idea that “explanation is that which produces understanding” is commonly accepted and often used to evaluate theories of explanation. But, Bradford Skow argues in his chapter, it cannot be used for this purpose. For, according to Skow, the claim either means that knowing the answer to the question why X is sufficient for understand why X—​in which case the claim is false; or it means that answering the question why X by performing the speech-​act of explaining invariably causes one’s audience to understand why X—​in which case the claim is useless, for theories of explanation aim only to say what it takes to be an answer a why-​question, not to say what it takes to provide an answer by performing the speech act of explaining. After defending these conclusions, Skow examines some prominent attempts to use the alleged connection between explanation and understanding to argue against one or another theory of explanation. Finally, J. D. Trout notes that many philosophers and psychologists appeal to a sense of understanding, typically a feeling invoked to explain people’s choices. “Understanding” in these cases seems loosely associated with properties like transparency (things we understand we can also introspect), or voluntary (cognitive) control (things we understand we can turn over in our mind). According to Trout, however, research on attention and memory shows that many candidate cases of understanding lack properties like transparency and voluntary control. In fact, “understanding” may denote an unprincipled stew of states, processes, capacities, and goals that are only occasionally present when philosophers, and ordinary folks, apply the term or concept. A unified account of understanding might be valuable, Trout argues, but understanding isn’t a natural kind or defined by a set of necessary and sufficient conditions. Any unity we find in understanding comes not from the involvement of common mechanisms across diverse cases, he concludes, but rather from messy cognitive activities in the common goal of pursuing the truth.



References de Regt, Henk. 2004. “Discussion Note: Making Sense of Understanding.” Philosophy of Science 71: 98–​109. Elgin, Catherine. 2006. “From Knowledge to Understanding.” In Epistemology Futures. Ed. Stephen Hetherington. New York: Oxford University Press. 199–​215. Grimm, Stephen R. 2006. “Is Understanding a Species of Knowledge?” British Journal for the Philosophy of Science 57: 515–​35. —​—​—​. 2011. “Understanding.” In The Routledge Companion to Epistemology. Ed. Sven Berneker and Duncan Pritchard. New York: Routledge. 84–​94. —​—​—​. 2015. “Wisdom.” Australasian Journal of Philosophy 93: 139–​54. —​—​—​. 2016. “Understanding and Transparency.” In Explaining Understanding: New Perspectives from Epistemology and the Philosophy of Science. Ed. Stephen R. Grimm, Christoph Baumberger, and Sabine Ammon. New York: Routledge. 212–​29. Kim, Jaegwon. 1994. “Explanatory Knowledge and Metaphysical Dependence.” Philosophical Issues 5: 51–​69. Kvanvig, Jonathan. 2003. The Value of Knowledge and the Pursuit of Understanding. New York: Cambridge University Press. —​—​—​. 2009. “The Value of Understanding.” In Epistemic Value. Ed. Adrian Haddock, Allan Millar, and Duncan Pritchard. New York: Oxford University Press. 95–​111. le Bihan, Soazig. 2016. “Enlightening Falsehoods:  A  Modal View of Scientific Understanding.” In Explaining Understanding: New Perspectives from Epistemology and the Philosophy of Science. Ed. Stephen R. Grimm, Christoph Baumberger, and Sabine Ammon. New York: Routledge. 111–​35. Lipton, Peter. 2009. “Understanding without Explanation.” In Scientific Understanding: Philosophical Perspectives. Ed. Henk de Regt, Sabina Leonelli and Kai Enger. Pittsburgh: University of Pittsburgh Press. 43–​63. Newton-​Smith, W. H. 2000. “Explanation.” In A Companion to the Philosophy of Science. Ed. W. H. Newton-​Smith. Oxford: Blackwell. 127–​33. Pritchard, Duncan. 2010. “Knowledge and Understanding.” In The Nature and Value of Knowledge: Three Investigations, by Duncan Pritchard, Alan Millar, and Adrian Haddock. New York: Oxford University Press. 3–​90. Riggs, Wayne. 2003. “Understanding ‘Virtue’ and the Virtue of Understanding.” In Intellectual Virtue: Perspectives from Ethics and Epistemology. Ed. Michael DePaul and Linda Zagzebski. New York: Oxford University Press. 203–​26. Strevens, Michael. 2013. “No Understanding without Explanation.” Studies in History and Philosophy of Science 44: 510–​15. Trout, J. D. 2002. “Scientific Explanation and the Sense of Understanding.” Philosophy of Science 69: 212–​33. Turri, John. 2015. “Understanding and the Norm of Explanation.” Philosophia 43.4: 1171–​75.



Woodward, James. 2003. Making Things Happen:  A  Theory of Causal Explanation. New York: Oxford University Press. Zagzebski, Linda. 2001. “Recovering Understanding.” In Knowledge, Truth, and Duty:  Essays on Epistemic Justification, Responsibility, and Virtue. Ed. Matthias Steup. New York: Oxford University Press. 235–​52.


The Unity of Understanding John Bengson

understands the why and how of things. The exemplary musician understands her instrument and how to play it. The exemplary athlete understands her sport, while the exemplary car mechanic understands engines, carburetors, and the tools she uses to manipulate them. And so on for the farmer, painter, designer, comedian, lover, conversationalist, businessperson, philanthropist, fitness planner, tour guide, moral exemplar, and myriad others who possess understanding of various kinds. The observation here should be familiar: understanding displays variety, coming in diverse forms. The question I wish to pursue is how these forms are related. Against the common tendency to see various forms or species of understanding as fundamentally distinct, perhaps so different as to warrant disbelief in a single, univocal genus to which they all belong, my aim will be to unearth unity amid diversity. I will focus throughout—​as a sort of case study—​on understanding in the realms of theory and practice.1 We seek to comprehend the world, to render intelligible what is, was, will be, and could or must be—​sometimes simply for the sake of illumination, other times in an effort to control our surroundings. This kind of theoretical understanding is paradigmatically manifested in scientific advances, which offer insight into, inter alia, laws, causes, mechanisms, and unifying principles. Yet this form of understanding also has a clear place in the ordinary course of things, lending meaning and context to our lives and enabling us to acquire some perspective on our human situation. THE EXEMPLARY SCIENTIST

1.   There may be other realms (e.g., the empathic), as discussed below. Using ‘understanding’ to cover multiple realms, as I do, leaves open that the term may be lexically ambiguous or designate a disjunctive kind. I will argue against both options below.

The Unity of Understanding


But our lives are not mere feats of contemplation. Being in the world involves engaging in a wide range of practical activities, both meaningful and mundane. At times, we navigate these activities perfectly well, but on other occasions our understanding fails and we are left confused or frustrated, unable to proceed effectively—​as when one eventually admits in the course of an effort at assembly, “I guess I don’t understand how to put the thingamajiggy together, after all.” In these ways, the phenomenon of practical understanding (or, sometimes, its absence) is manifest. Such understanding does not always show itself in action, as in a case where it goes unused. Yet its paradigm exercise remains skillful activity, which is importantly different from reflexive or instinctive behaviors, mechanical mimicries, and spurts of raw talent or mere knack. This is not because skilled action is more effective in comparison (it need not be),2 but rather because it is guided by the agent’s grasp of the action in a way that understanding-​less behavior—​even when overtly indistinguishable from expert performance—​is not. Investigation of the relation between theoretical understanding and practical understanding is related to, but distinct from, inquiry concerning the relation between knowledge-​that and knowledge-​how, as these are standardly conceived, namely, as the states or relations designated by the English expressions ┌x knows that p┐ and ┌x knows how to φ┐, respectively. Knowledge-​ that and knowledge-​how may offer specific examples of our categories. But our categories not only summon understanding, as opposed to knowledge (I return to this contrast below); in addition, our categories are more general, potentially including also, or instead, what is designated by a wide array of other expressions. In the case of practical understanding, the expressions may not include ‘how-​to’, nor any other infinitival clause headed by a question-​ word (e.g., ‘where-​to’), as when an accomplished dancer is said to ‘understand ballet’, a virtuoso musician to ‘understand counterpoint’, or a master carpenter to ‘understand that this tool (the one she is using, as opposed to another) is best’ for the task at hand. Likewise, the theoretical understanding of an exemplary scientist might be expressed without reverting to any ‘that’-​clause, as when a nuclear engineer is said to ‘understand fission’, a biologist to ‘understand why a particular species has certain traits but not others’, or a linguist to

2.   This was seen clearly by Aristotle (Metaphysics I.1): “With a view to action experience seems in no respect inferior to technê [skill].” Cp. Ryle (1949, 40): “there need be no visible or audible differences between an action done with skill and one done from sheer habit, blind impulse, or in a fit of absence of mind.”



‘understand Mycenean Greek’ or to ‘understand how to interpret Linear B’.3 The objects of the attitude at stake (or, in linguistic terms, the verb’s complement clauses’ designata) are broader than those associated with knowledge-​ how and knowledge-​that. The relation between understanding and knowledge of various sorts merits further scrutiny, and we will have an opportunity to discuss it below. The intention right now is not to argue that they are not intimately connected, but to avoid taking such a connection for granted, in a way that might bias our current investigation of the relation between the theoretical understanding of the “thinker,” whose paradigmatic manifestation is scientific advance, and the practical understanding of the “doer,” whose paradigmatic manifestation is skillful activity.4 In fact, it is only once we gain a better grasp of understanding in its various forms that we can productively address its connection to knowledge in its various forms—​or so I believe. Despite the significance we seem to place individually on both theoretical understanding and practical understanding, philosophers have tended to treat them as largely distinct. Much recent work on understanding in philosophy of science and epistemology has examined the former, focusing on such questions as whether and in what sense understanding is factive, transparent, or explanatory; how understanding interacts with skepticism; and why understanding is a (or the) cognitive aim of science. By contrast, practical understanding shows up primarily in discussions within philosophy of mind, action theory, and phenomenology, where the focus is instead on questions concerning its automaticity, (non)conceptuality, and

3.   It follows that reductions of knowledge-​that to knowledge-​how (e.g., Hartland-​Swann 1956)  or knowledge-​how to knowledge-​that (e.g., Stanley and Williamson 2001) do not by themselves entail reductions of one of our forms of understanding to the other; they do not secure the unity of understanding. And while nonreductionist positions—​such as Ryle’s dispositionalist anti-​intellectualism, or my own objectualist intellectualism (Bengson and Moffett 2011a)—​may be conjoined with bifurcation of the sort described below, they do not entail it (indeed, I will deny it). We should also bear in mind the possibility of nontheoretical knowledge-​that (see, e.g., Glick 2011, 412–​13) or theoretical knowledge-​ how (see, e.g., Lihoreau 2008, 281–​82); these might actually underwrite practical understanding and theoretical understanding, rather than the reverse. In sum: the know-​how/​know-​that debates are not irrelevant to our question, but they do not answer it. 4.   I say that φ (e.g., skillful activity) is the paradigmatic manifestation of a form of understanding K (e.g., practical understanding) if and only if what φ manifests (e.g., skill) is an instance of K, and for any x, x is an instance of K if and only if x is identical to what φ manifests (e.g., x is a skill) or x is a qualified version of—​i.e., approximates—​what φ manifests (e.g., x approximates skill). The intuitive idea is that φ serves as the ideal or fundamental case, by reference to which all other cases of the kind are understood.

The Unity of Understanding


embodiment—​as in Maurice Merleau-​Ponty’s declaration, “It is the body that understands.”5 The division is rarely explicitly addressed or acknowledged. But when it is, it is often treated not simply as a convenient division of labor, but as a compulsory division in nature. For example, Martin Heidegger claimed in Being and Time that practical skill possesses “its own kind of sight,” involving a proprietary type of nontheoretical cognition, which Heidegger labeled “circumspection.”6 Such a view is not specific to a particular period or tradition. Recent work in mainstream analytic philosophy of science has seen Peter Lipton similarly emphasize a division between theoretical understanding and practical understanding, the latter of which Lipton describes as “sui generis” (2009, 54). Jonathan Kvanvig goes further, influentially defending the epistemic value of understanding in such a way that practical understanding “gets left out,” and remarking confidently that this “is as it should be.”7 In the existing literature, then, we appear to be confronted with bifurcation, the idea that theoretical understanding and practical understanding are two very different phenomena, which can be analyzed, and comprehended, in isolation from one another. Perhaps they are tied together by some historical or etymological connection, but philosophically they are not of a piece. Understanding is thus in an important sense heteronomous, or disunified: like jade, which famously is either one of two disparate minerals, nephrite or jadeite, at bottom understanding is not one, but many. The common assumption of disunity would, if correct, have vast and deep implications for our conception of understanding and its nature, value, acquisition, dynamics, function, and extension. A few of these implications have already been noted (recall, for example, Kvanvig on its epistemic value and Merleau-​Ponty on the role of the body); several others will emerge in the subsequent discussion. If, however, my arguments in what follows are on the right track, one of the upshots is that basically all extant approaches to understanding—​which, as I shall explain, are committed to bifurcation—​are suspect. I will explore the possibility of bringing the practical and theoretical

5.   Merleau-​Ponty (1962, 142). The original French reads: “C’est le corps qui «comprend»” (1945, 169); cp. “Un mouvement est appris lorsque le corps l’a compris” (161). As the latter shows, Merleau-​Ponty sometimes but not always punctuates ‘comprendre’. 6.   Heidegger ([1926] 1962, 99). Such cognition is the epistemic backbone of the metaphysical project in Being and Time:  it is through, and only through, such cognition that the Being of work and equipment—​the “entities which we encounter as closest to us” (95)—​is disclosed. 7.   Kvanvig (2003, 186 and 190). Additional examples will be given below.



together so as to secure the unity of understanding. While I will acknowledge that there are important differences between them, my claim will be that theoretical understanding and practical understanding possess a common underlying nature, and I will seek to explain how this could be so, by identifying what it is that unites them.

2.1. Two Forms of Understanding I began with an intuitive contrast between understanding in the realms of theory and practice, which is meant to indicate what I mean by ‘theoretical understanding’ and ‘practical understanding’ without prejudging substantive questions about their proper philosophical treatment. Eventually I will propose a theory of understanding designed to reveal what they have in common and why this commonality is important. What I wish to do now is simply to suggest that (1) each is a form of genuine understanding, and (2) they are distinct (i.e., not identical). Let us consider each of these claims in turn. 2.1.1. Genuine Understanding: The U-​Profile It is possible to describe both thinkers and doers using locutions of the form ┌ x understands . . .┐. (Several possible examples were given above.) But this form of words does not invariably designate instances of genuine understanding.8 So, this essentially linguistic observation about the terms in which we sometimes describe thinkers and doers does not by itself offer an adequate basis for claim (1).9 What would substantiate that claim is an intuitive, and relatively theory-​independent, specification of core features of understanding (i.e., what the philosophy of understanding is the philosophy of ), together with reasons to think that they are shared by both scientific advance and

8.   See, e.g., Anderson’s (1986, 276) seminal discussion of evidentials. 9.   Dennis Stampe has impressed upon me the value of acknowledging that some of my descriptions of cases—​particularly in the realm of practice—​may sound unnatural to some readers, and of emphasizing that what matters is the phenomenon described. It suffices for present purposes that there are central cases of the phenomenon that call for the ordinary English descriptor ‘understanding’, and recognizable similarities between those cases and the others, which display the core features of genuine understanding described next. (One hypothesis—​also suggested to me by Stampe—​is that whether an utterance of an instance of the schema ┌x understands . . .┐ sounds felicitous to a speaker S depends not on understanding per se, but on S’s prior beliefs about  .  .  .’s susceptibility to being conceived in the manner explicated by the theory of understanding advanced in §2.4.)

The Unity of Understanding


skillful activity, our paradigms of theoretical understanding and practical understanding. Let’s begin, first, with a platitude:  to genuinely understand something is to grasp it—​whatever is understood—​in such a way that it makes sense to you. Among the principal characteristics of such grasp is, second, its status as a standing state, distinct from its exercises or effects (e.g., subsequent predictions or actions).10 Although, third, this state is psychological, or at least has a psychological component, genuine understanding is not merely psychological, but, fourth, is objective: more than a personal feeling of competence or comprehension, it involves genuinely grasping some portion of reality, and not simply enjoying a subjective sense of grasping it.11 Fifth, it is intelligent, in a way that mindless behavior and rote memorization—​however objective—​are not. Sixth, as shown by the possibility of mindful, nonrote engagement with what one does not yet understand, it is also robust, in a way that mere acquaintance with particular deeds or facts—​even if intelligent and objective—​is not. Genuine understanding also contrasts with confusion and muddle. It does this by, seventh, displaying coherence, rather than inconsistency or bare association (as in confusion), and, eighth, attaining order, of the sort absent from a miscellaneous list lacking arrangement or organization (as in muddle). A ninth core feature of genuine understanding is its multiple gradability:  it can be or become better, greater, deeper, stronger, or richer along various dimensions, for example, with respect to some of the features just mentioned, including orderliness, coherence, robustness, intelligence, and objectivity. Finally, tenth, possession of genuine understanding—​especially full understanding, but also even minimal understanding—​is a praiseworthy good. Attributing understanding to an individual is not merely to credit her with some kind of success (though it is of course that), but to compliment, or praise, her for it—​it is, as Plato said of epistēmē, to indicate that an individual is, at least in that respect and to some extent, “honorable and excellent.”12

10.   Throughout I use the term ‘state’ to designate standing properties or conditions, as opposed to (say) events. 11.   While the precise analysis of understanding’s objectivity is controversial, this should not obscure general agreement that understanding possesses this feature (this point is effectively made by Elgin 2007, 35). Similarly for the other core features of understanding identified in the main text (that it is a standing state, psychological, etc.). 12.   The modifier ‘in that respect’ is important: an evildoer who understands evil is not honorable and excellent all around.



Let me emphasize that I  am not proposing to analyze genuine understanding in terms of this cluster of features. This is not yet an attempt at a theory of understanding, but a relatively uncontroversial, theory-​neutral description that identifies its core features, which jointly constitute understanding’s profile—​hereafter, the ‘U-​profile’. Later I will propose a theory of understanding that seeks to accommodate and explain the U-​profile. Here we may leave its constituent features at an intuitive level, at which their basic content, and their applicability to genuine understanding, should not be controversial. I take it to be obvious that in the realm of theory, thinkers sometimes satisfy the U-​profile. Perhaps less obvious, but equally real, is satisfaction of the U-​profile in the realm of practice. Recall that practical understanding is paradigmatically manifested by skillful activity. Skill, as I understand it, contrasts with various forms of ineptitude, such as clumsiness, sloppiness, and klutziness. It also requires know-​ how. If one is skilled at playing the piano, one knows how to play the piano. If one is skilled at chess, one knows how to play chess. If one is skilled at using a fretsaw, one knows how to use a fretsaw (see ­figure 2.1). But skill at an activity requires more than mere knowledge how to do that activity: after all, two individuals can both know how to play the piano, or play chess, or use a fretsaw, although one is a master who is skilled (e.g., skilled at piano-​playing) while the other, a novice, is not. By all reasonable standards, the latter is clumsy (e.g., clumsy at piano-​playing), or sloppy (i.e.,

Figure 2.1  A typical fretsaw

The Unity of Understanding


his openings are messy, midgame reckless, and endgame chaotic), or klutzy (i.e., he is prone to catch the blade, make haphazard cuts, and bang the frame against the wood), and does not yet possess the skill. It is not simply that the novice’s actions are less effective.13 Rather, the imperfection resides, first and foremost, in the novice’s grasp, which is limited and gives out quickly.14 It is a common refrain that if one is skilled, then one has what it takes to fluently navigate a wide range of different, even quite novel, scenarios or situations. John Campbell, focusing on skillful tool use, as opposed to mere “phrasebook” knowledge,15 finds here both systematicity and generality: [I]‌f you understand [in a practical way] how a tool works, there will be a certain systematicity [and generality] in your understanding of it. You will know how to use this tool in a wide variety of contexts, under various permutations of its intrinsic characteristics.16 The modulation of the pattern of use is systematic, in that the pattern of use covaries with variation in the standing properties of target and tool. And the modulation of the pattern of use is general, in that the same underlying sets of connections can be exercised in connection with endlessly many different tools.17 That skill requires know-​how but goes beyond it in these ways allows us to make good on the idea that the practical grasp of the doer—​whose paradigmatic manifestation is skill—​satisfies the U-​profile. Knowledge how to act is an objective, intelligent, gradable, standing psychological state. It also involves some

13.   Recall note 2. 14.   The thesis, which I am opposing, that skill and know-​how are equivalent is widely assumed and has recently been explicitly promoted by Stanley (2011, 5 and 11); see also Dickie (2012, 737). Cp. Stanley and Williamson (forthcoming, §1), who hold that skill is equivalent to some combination of knowledge how, when, where, whether, and the like (or dispositions thereto). Given that novices can also possess such knowledge (or dispositions thereto), the argument in the main text applies mutatis mutandis to this position as well. 15.   Campbell (2011, 170): “we can contrast someone who merely has grasped the use—​someone who only knows how to make the correct moves with the thing . . . —​from someone who is making [skillful] use of the thing.” 16.   Campbell (2011, 174–​75). Cp. Haugeland (1998, 199) on ‘complexity’ and ‘precision’; Annas (2011a, 101–​2) on how expertise is ‘dynamic’; and Fridland (2014, 2732) on experts’ ‘control’. 17.   Campbell (2011, 179). Campbell makes clear that the user’s understanding is what underlies and explains systematic and general patterns of use, and presumably would allow his characterizations to be extended so as to apply to practical understanding that is not primarily tool-​oriented.



degree of sense-​making (i.e., to some extent, the activity makes sense to one). But as the contrast between the clumsy novice (perhaps a confused muddler) and the master illustrates, know-​how is not always systematic and general, and thus it need not possess the sort of robustness, coherence, and orderliness—​and, hence, the kind of praiseworthiness—​characteristic of genuine skill.18 We are now in a position to see why it is reasonable to endorse claim (1): both the theoretical grasp of the thinker and the systematic, general practical grasp of the doer satisfy the U-​profile. In both cases, the grasp in question is a multiply-​gradable, standing psychological state, constituting a praiseworthy good, involving sense-​making that is more objective than a mere feeling of competence or comprehension, more intelligent than mindless behavior or rote memorization, more robust than simple acquaintance with particular deeds or facts, and possessing coherence and orderliness. Thus, I submit, the realms of theory and practice both afford instances of genuine understanding. 2.1.2. The Distinctness of Theoretical and Practical Understanding Once the U-​profile has been used to establish claim (1), one might deem it safe to return to locutions of the form ┌x understands . . .┐ in order to vindicate claim (2) through different ways of filling in ┌ . . .  ┐, or (in linguistic terms) distinct complement clauses of the verb: for example, ‘how’ complements versus ‘why’ complements or noun phrase complements.19 However, so far as I  can tell, there is no particular form of words in the English language that decisively marks practical understanding and theoretical understanding as such, or tracks the distinction between them. As noted at the outset, we may express the theoretical achievement of a linguist specializing in Mycenean Greek by saying that she ‘understands how to interpret Linear B’,

18.   Behaviors or processes that do not manifest a state that satisfies the U-​profile (e.g., robotic movements or subpersonal routines) are sometimes described in the language of ‘skill’. Such uses have a sort of derivative status, being felicitous only to the extent that their designata sufficiently resemble actions that manifest genuine skill, which (I am now arguing) satisfies the U-​profile. 19.   See again Kvanvig (2003, 188ff.), who focuses on the complement clauses, and moreover, claims that ‘understanding how’ picks out a kind of “understanding [which] is relevant more to practical purposes than to theoretical ones.” On the other side, it is common for epistemologists and philosophers of science, in the course of discussing theoretical understanding, to focus on ‘understanding why’; e.g., Pritchard (2010, 31) writes, “I want to take the paradigm usage of ‘understands’ to be in a statement like ‘I understand why such-​and-​such is the case’.” Cp. Hempel (1965, 334); Moravcsik (1979, 202); Kitcher (1989, 419); Grimm (2008); de Regt (2009, 588); Khalifa (2012; 2013a, 1155); Strevens (2013); and Hills (2016).

The Unity of Understanding


and the practical achievement of a virtuoso by saying that she ‘understands counterpoint’; similarly, we can express the practical understanding of a master carpenter by saying that she ‘understands why that tool is best’ for a given task. Notice, too, that in some cases one and the same form of words can do double duty: an accomplished performer and a scholarly expert might both be said to ‘understand ballet’ or ‘understand chord progression’. Of course, they understand it in different ways. The point is that those differences are not marked by or through the locutions alone. To catch sight of them, we must—​contra a simple linguistic or lexicological approach—​look beyond the attributions themselves. I propose we look to our paradigms. Theoretical understanding and practical understanding are reasonably viewed as distinct (not identical), at least insofar as they have distinct paradigmatic manifestations: scientific advance and skillful activity, respectively. This is not to suggest that their distinctness consists in their having different paradigmatic manifestations, but simply to register that they do, and that this provides reason to acknowledge the distinction itself. The content of those paradigms also suggests an initial characterization of the distinction. When an individual acts skillfully, her action is guided by her grasp of the activity itself, or of steps employed or tools utilized in its undertaking. A clown’s tripping and tumbling, which are a skillful imitation of clumsiness, manifest her practical understanding, which guides how she flails her limbs and tosses her head. Her understanding—​a standing state, “a factor which could not be separately recorded by a camera”20—​underlies and explains the successful, intentional execution of her action. Even when the clown is resting or absorbed in other activities, the state remains, poised to guide some such performance in one or another set of circumstances. In this way, practical understanding is action-​guiding, being centrally concerned with practical activity not merely in its topic or subject matter but in its function or character. By contrast, theoretical understanding centers on the culmination of theory. Kant at one point suggested that part of what is distinctive of theory—​whether physical, political, moral, etc.—​is its invocation of claims or “principles of a fairly general nature . . . abstracted from numerous conditions” ([1784] 1970, 61). Others have emphasized the importance to theory of laws, causes, mechanisms, unification, and dependence relations. From the

20.   Ryle (1949, 33); the clown example is his.



perspective of understanding, the interest of all of these is presumably that they shed light on, by playing roles in explanations of or related to, whatever is under investigation—​in other words, they help resolve queries (raised explicitly or implicitly) regarding the why, what, or how of what was, is, will, or might be. Whereas practical understanding is action-​g uiding, theoretical understanding is illuminating. In sum, I propose to treat the following conditions as distinguishing our two forms of understanding: x practically understands . . . only if for some activity φ related to . . . , x is in some state satisfying the U-​profile that is poised to underlie and explain the intentional execution of an intelligent action that contributes to φ-​ing (‘action-​guiding’). x theoretically understands . . . only if x is in some state satisfying the U-​profile that is poised to resolve explanation-​seeking questions of or relating to . . . (‘illuminating’).

2.2. The Notion of Unity To point to the existence of, and distinction between, theoretical understanding and practical understanding is not to suggest that the distinction is exhaustive. There may be outliers, that is, cases of genuine understanding that do not belong to either category: for example, a psychoanalyst’s empathic understanding of a patient, or a lover’s understanding of a beloved, is (perhaps) neither theoretical nor practical. Similarly, the distinction may not be exclusive, and it may fail to be sharp. Arguably, there are many mixed or borderline cases, which do not fit neatly into just one category: for example, Obama’s understanding of campaign strategy, which seems to involve a mixture of theoretical and practical elements. None of this, however, bears directly on the question of the unity of understanding in the sense in which I intend it. There are several different things one might mean by saying that understanding is ‘unified’, or that there is a ‘unity’ of theoretical and practical understanding. For example, one might mean that one form of understanding cannot be present in the absence of the other. (Compare the ancient thesis sometimes labeled the ‘unity of the virtues’, according to which—​on a popular interpretation—​possession of one virtue entails possession of all virtues.)21 Let us call this thesis copresence: 21.   For the interpretation, see Vlastos (1972).

The Unity of Understanding


copresence  One theoretically understands . . . if and only if one practically understands. . . .  There is some truth to this thesis, as theoretical understanding and practical understanding are often mutually reliant and cannot cleanly be disentangled. As stated, however, copresence is open to simple counterexamples. Presumably, an immobile but scholarly expert in neuroanatomy, kinesthetics, and dance history might have a theoretical understanding of ballet while lacking any practical understanding of ballet. Conversely, an accomplished but unscholarly dancer who grasps the practice might not be apprised of the theory.22 While there may be ways of tweaking or restricting copresence in order to handle such cases, I will not pursue these options.23 For I am not here invested in the idea that most or even all cases of understanding are entangled. copresence is not the unity thesis that I intend to pursue. Another thing one might mean by the ‘unity’ of practical understanding and theoretical understanding is that they are equally indispensable elements of a type of understanding that is, in some intuitive sense, full, complete, or best; each enriches one’s overall understanding, which is then ‘unified’. Let us call this thesis enrichment: enrichment  There is a type of understanding U, such that x has U with respect to . . . only if one both practically understands . . . and theoretically understands . . . , and U is superior, in respect of understanding, to either of these in isolation. This thesis requires, not entanglement (as in copresence), but rather that various forms of understanding can combine and aggregate so as to contribute jointly to improving an individual’s overall understanding. To the extent that we view the player-​coach or erudite artisan as possibly achieving an understanding of her craft superior to the understanding of a merely equivalently skilled player or artisan, we are committed to the truth of enrichment in at least some cases.

22.   Similarly for other activities:  for instance, a scholar of the Tour de France versus its star cyclist. Additional examples are described in §2.3. 23.   A modified version of copresence distinguishes Plato’s convergentist view (in the Republic) that full theoretical understanding just is full practical understanding from Aristotle’s divergentist view (in Posterior Analytics, Nicomachean Ethics, and Metaphysics A) that the two are substantially independent, even when full. See Cooper (1977, §§I–​II) and Annas (1981, 261ff.) for helpful discussion.



I am sympathetic to this thesis, and an important virtue of the theory of understanding I will develop below is that it both accommodates and explains it. However, enrichment is not my primary target. For this thesis does not speak to the question of what goes into, or makes up, theoretical understanding and practical understanding. Rather, it helps itself to both phenomena, leaving wholly unanswered the question of what makes each a form of understanding. So, although enrichment may in some sense locate unity in the interaction between theoretical and practical understanding, it does not find unity within them; it does not express what I mean by ‘the unity of understanding’. As I  use the term, to say that theoretical understanding and practical understanding are ‘unified’ is to say that it is one and the same thing that makes each a form of genuine understanding. This is not merely to say that they have something, even something very important, in common. Nor is the idea simply that both satisfy the U-​profile, or that they are species of the same genus (or determinates of the same determinable): understanding.24 Rather, the focus is on what makes both satisfy the U-​profile and belong to the same genus. In short, I wish to claim that they possess one and the same ground. I formulate the central idea as follows: ground  There is a type of nondisjunctive state, σ, in virtue of which the U-​profile is satisfied, such that: if x has practical understanding, then x is in σ1, and if x has theoretical understanding, then x is in σ2, and σ1 and σ2 are both instances of σ. This thesis asserts that instances of theoretical understanding and practical understanding qualify as instances of understanding—​as objective, intelligent, robust, multiply gradable, coherent, orderly, praiseworthy goods, centering on sense-​making—​because, or in virtue, of one and the same thing.25 In this sense, ground entails fundamental unity:  understanding is one, not many. Many philosophers have claimed or implied that ground is false, and that a disunified approach is required. I  believe this is the upshot of

24.   Of course, theoretical understanding and practical understanding do not satisfy the U-​profile in virtue of being species of understanding; rather, they are species of understanding in virtue of satisfying the U-​profile (or their being species of understanding just is their satisfying the U-​profile). 25.   I have chosen to employ ‘in virtue of ’, ‘because’, ‘makes’, and various other expressions associated with the metaphysics of ground, but as far as I can tell, various nongrounding ideologies (structure, essence, constitution, etc.) could be employed instead.

The Unity of Understanding


Heidegger’s claim, quoted above, that skill possesses “its own kind of sight,” Lipton’s assertion that practical understanding belongs to its own genus, and Kvanvig’s dismissal of the practical realm in his treatment of understanding’s value. Consider also the following remarks by Campbell: We can contrast a theoretical understanding of the causal properties of particular types of wood, for example, or different metals, such as iron or silver, with the understanding possessed by the carpenter or metalworker. The artisan’s grasp . . . is not a matter of having a detached picture . . . . It has to do rather with the . . . particular way in which he deals with various types of wood or how he uses different metals. His grasp . . . may consist entirely in his practical ability to respond suitably to their presence. . . . [It] is not an explicit, reflective grasp but consists in possession of suitable practical skills. (1994, 47–​48) Insofar as the “detached picture” or “reflective grasp” at the heart of theoretical understanding is not one and the same as the “practical ability to respond” in which practical understanding is said to “consist entirely,” what follows from Campbell’s remarks is the denial of ground. Yet, on the face of it, Campbell’s remarks appear to be eminently plausible; they may even seem innocuous. One wonders: is the fundamental disunity of understanding inevitable? I believe that despite the attraction of disunity, it is ultimately mistaken, and that there are good reasons to embrace the indicated unity thesis, ground. Before attempting to substantiate this thesis, however, it will prove instructive to address what are arguably the two most potent obstacles to unification.

2.3. Obstacles to Unification As explained above, mainstream philosophical theorizing displays a strong tendency, captured in Campbell’s remarks, to treat practical understanding and theoretical understanding in a bifurcated manner. It is fair to regard this as the orthodox position. One sign that a view is orthodox is widespread endorsement, often implicit, combined with systemic absence of explicit argument on its behalf. This is what we seem to find here. This of course does not imply that opposition to ground has no basis. Quite plausibly, it does. Arguably, a primary basis is the thought that the distinctive features of theoretical understanding and practical understanding—​illumination and



action-​g uidingness, respectively—​are to be accounted for in very different ways. Let me explain. A popular idea is that understanding is centrally linked to explanation—​ where this is neither mere description nor justification, and is canonically expressed by ‘because’-​statements (as in ┌p because q┐). This idea has enjoyed explicit endorsement by a variety of thinkers; for instance: Jaegwon Kim:  “[U]‌nderstanding, as Salmon puts it, results from ‘our ability to fashion explanations’. That is almost tautological.” (1994, 61, quoting Salmon 1984, 259)26 Peter Lipton:  “[W]‌e do not have a clear conception of understanding apart from whatever it is our explanations provide.” (2004, 23; cp. Khalifa 2013b) Jonathan Kvanvig:  “Understanding involves an already-​possessed awareness of the explanatory and other connections involved in the subject matter in question.” (2009, 99; see also 2003, 192ff.)27 Michael Strevens:  “Understanding a subject matter consists in grasping correct explanations of, using, or otherwise related to that subject matter.” (2010, 17) John Greco:  “To have an explanation is to be able to cite appropriate dependence relations,” and “understanding consists in a systematic knowledge of dependence relations.” (2014, 291–​92) We can formulate the proposed understanding-​explanation link as follows: explanationism  x understands . . . in virtue of bearing some cognitive relation to an explanation (or explanatorily relevant item) of or relating to . . .28 26.   See also Salmon’s (1998, 8–​9) discussion of various forms of nonscientific understanding, all of which he links to explanation or what Lipton (2009) terms “cognitive benefits of explanation.” 27.   Officially, Kvanvig requires grasp of explanatory connections “when they exist” (2009, 101), allowing that in some cases what is grasped are (say) probabilistic relations. In my terminology, these qualify as ‘explanatorily relevant’, and Kvanvig qualifies as an explanationist. 28.   In a moment I will discuss restricted versus unrestricted versions of this thesis. There are many further ways of elaborating explanationism, corresponding to the underlined expressions. Regarding the first, some explanationists will favor knowledge (Grimm 2006), while others will require only a certain sort of belief (Pritchard 2010, 2014), and still others will invoke an intellectual-​ability relation (see, e.g., Hills 2016; cp. Wilkenfeld 2013). Regarding the second, some will wish to privilege (say) laws or causes, while others will cast a broader net (see, e.g., Greco 2014 and Grimm 2014 on dependence relations), and still others will prefer to invoke not explanation per se but what Lipton (2009) calls the “cognitive benefits of explanation,” or what Kelp (2015) calls the “full account” of a phenomenon, or

The Unity of Understanding


Perhaps the foremost attraction of explanationism is that, insofar as explanations contribute to the resolution of inquiry, a tight understanding-​explanation link seems to account for the illuminating character of theoretical understanding. The attraction is fairly limited, however, since the specific version of this link promoted by explanationism (which invokes an in-​virtue-​of relation) is not required to generate such an account.29 Moreover, the attraction is inextricably linked to a potentially weighty implication. For it is difficult to see how many instances of practical understanding will obtain in virtue of the explanatory condition this thesis advances. Take, for example, the major league pitcher Phil Niekro’s understanding of the knuckleball (whose strange flutter is notorious), as displayed in and through his skillful pitching, which we may reasonably suppose was not accompanied by cognizance of the pitch’s explanation.30 Switching to the other side of the plate, there is substantial empirical research showing that expert batters are ignorant of the explanation of their success, which invokes the general fact—​of which they are unaware—​that successful batting requires repeatedly shifting one’s gaze ahead of the oncoming ball, anticipating (or predicting) where it will move next, as it hurtles toward one at speeds that make continuous visual tracking impracticable.31 Likewise, many carpenters show their skill at sanding by utilizing a stroke that does not break but rather cuts the wood’s fibers, although they are unaware that this is the explanation of why their stroke yields a smooth texture.32 As a final example, consider a self-​taught virtuoso musician, who may very well manifest impressive

what De Regt and Dieks (2005) and De Regt (2009) call an “intelligible scientific theory”; cp. Khalifa (2012; 2013a; 2013b). These details will not affect our discussion. 29.   I will describe my preferred alternative in §2.5.2. 30.   A skillful knuckleballer such as Niekro may have a practical grasp—​which satisfies the U-​profile—​ of the pitch (or how to throw it), and he may be able to describe it, but as indicated above, that is not yet to say that he can explain the pitch, e.g., why (or how) the grip affects the flutter. For example, if the knuckleball is the only pitch he has ever known, he may not know that the flutter depends on the grip. 31.   See Papineau (2013, 177–​78) and Brownstein and Michaelson (2016, 7–​8) for summaries of relevant research. 32.   Thanks to Farid Masrour for this example, and for the opportunity to peruse Aldren A. Watson’s beautiful and insightful Hand Tools: Their Ways and Workings. The book offers a glimpse into the understanding of a master craftsperson, and in so doing consistently emphasizes method—​procedure, or guide to action—​rather than explanation. Although the latter is also sometimes of interest, it does not guide the doing of what the doer does (cp. Ford 2017); it is not what is manifested in skillful activity. I return to the importance of method in §2.5.2.



understanding of—​hence, a state that satisfies the U-​profile with respect to—​the French horn, or of operatic signing or vocal chord progression (e.g., through a complex series of intentional modulations of her voice, resulting from subtle contractions of her pharynx, nuanced movements of her tongue and lips, minutely timed breaths, controlled inflections, diaphragmatic vibrato, and more), though she is unfamiliar with music theory and relevant sciences, and is incapable of providing any explanations of, using, or otherwise bearing on her music or its performance. Practical understanding is not, or at least not usually, explanatory in the way that explanationism maintains.33 It has seemed to many that what is paradigmatically manifested in skillful activity—​the practical understanding of the doer—​is centrally linked, not to explanation, but rather to practical ability, understood to include physical disposition, habit, or bodily activity. Recall Merleau-​Ponty’s remark about the body and Campbell’s invocation of the artisan’s dealings. Consider also: Alva Noë:  “Practical abilities amount to a type of understanding, one that we apply in our practices.” (2005, 285) Sean Kelly: “My bodily activity with respect to the object just is my way of understanding it.” (2002, 385) Hubert Dreyfus: “[S]‌kills are ‘stored,’ not as representations in the mind, but as dispositions to respond to the solicitations of situations in the world.” (2002, 367)34 Fred Dretske: “[I]‌n the case of all skilled actions, whether it be tying your shoelaces, playing a musical instrument, or dribbling a basketball—​the mind goes elsewhere while the body performs.” (1998) Robert Brandom: “[P]‌ractical understanding [is] a kind of adaptive attunement to the environment, the development of habits apt for successful coping with contingencies.” (2013, 112–​13)

33.   Cp. Khalifa (2013a, 1164 n. 18): “there is another kind of understanding-​how that is of a practical variety, e.g. Jimi understands how to play guitar. This is clearly not explanatory.” Jimi and other virtuosos may be able to describe or characterize their skills, but they need not (see, e.g., Fodor 1968, 633ff.); moreover, even when they can, their descriptions need not be accurate or effective; and, in any case, as noted above, a mere description is not an explanation. See also Hills (2015), who argues that although full moral virtue requires explanatory cognition, this is not always so for skill (whose link to practical understanding I defended in §2.1.1). 34.   Elsewhere Dreyfus makes clear that by ‘skill’ he means “the nonconceptual immediate intuitive understanding exhibited by experts” (2006, 43; emphasis added).

The Unity of Understanding


These and other theorists of practical understanding can be seen as pursuing an account of its action-​g uiding character in terms of practical ability. In so doing, they endorse something like the following thesis: abilitism  x understands . . . in virtue of bearing some practical ability relation to an activity (or action-​relevant item) of or relating to . . .35 Of course, many instances of theoretical understanding do not obtain in virtue of a thinker’s practical abilities. A physicist’s understanding of the knuckleball, for instance, need not be accompanied by any power to throw, catch, hit, or otherwise handle it. Similarly for the vision scientist’s understanding of successful batting. Theoretical understanding is not, or at least not usually, related to practical ability (disposition, habit, or bodily activity) in the way that abilitism maintains. Proponents of explanationism and abilitism may of course view these theses as restricted to theoretical understanding and practical understanding, respectively. The resulting position strings together the grounds invoked by these theses into a simple disjunction: one understands, when one does, either in virtue of explanatory cognition or in virtue of practical ability. Perhaps the most pressing objections to this dualist thesis are explanatory. First, it seems unable to accommodate and explain the thesis of enrichment, for it is not clear how two fundamentally disparate forms of understanding could interact in the way the thesis describes.36 Second, and perhaps more basically, dualism fails to explain why the U-​ profile is satisfied when and only when one of the disjuncts is. Above I argued that the constituent features of the U-​profile are possessed by instances of both theoretical understanding and practical understanding (recall §2.1.1). The dualist view is disbarred from providing a suitable answer to the question of why the features in the U-​profile are jointly instantiated when and only when one of the view’s preferred conditions obtains. The view does not merely drive an unsatisfying wedge through our conception of understanding; it also

35.   Those tempted to replace ‘  .  .  .  ’ with ‘how to φ’ are directed to §2.1.2. In a moment I  will discuss restricted versus unrestricted versions of abilitism, whose underlined expressions can be filled out in various ways. Some will wish to focus on the relations being-​able-​to, being-​disposed-​to, or being-​ counterfactually-​successful-​at, while others will posit another, possibly primitive or unique state (e.g., sensorimotor-​capacity-​to) or set of phenomena (e.g., a set of bodily movements); some will wish to privilege a single activity, while others will cast a broader net, allowing for diverse activities. 36.   I return to enrichment in §2.6.2.



leaves the fact that the U-​profile is satisfied, when and only when it is, seem both unprincipled and arbitrary. What is wanted is an explanation of the U-​ profile—​some insight into why its constituent features are jointly instantiated when and only when they are. Rather than facilitate insight, dualism courts mystery. To be sure, in some cases an unexplained disjunction is not mysterious but appropriate—​one example is jade. However, with jade we know that nephrite and jadeite offer distinct grounds for jade, and, furthermore, that they themselves are not grounded in one and the same thing. These two items of knowledge support, by entailing, the conclusion that jade is not fundamentally unified. We lack analogous knowledge, hence analogous support, in the present case. In this way, dualism about practical and theoretical understanding suffers from an epistemic, as well as explanatory, lacuna.37 A natural response to these concerns is to choose one disjunct, and to lift its restriction so that it is taken to apply to all cases of genuine understanding; those cases to which it cannot be applied are then classified as ‘other’.38 In fact, some proponents of an understanding-​explanation link have pursued this strategy, maintaining that all instances of understanding hold in virtue of a single condition, namely, explanatory cognition of the sort invoked by explanationism. This is the path de facto recommended by Linda Zagzebski when she writes: The person who has mastered a technê [skill] has a kind of understanding one cannot get any other way. He is able to explain features of the technê.39 Of course, a parallel move can be made by privileging practical ability, as in an unrestricted version of abilitism. 37.   The case of jade—​whereof there is a disjunctive genus—​should not be conflated with a case of lexical ambiguity, as with the noun ‘bank’. There is no reason to posit a genus (disjunctive or otherwise) of which both riverbanks and financial banks are species. That practical understanding and theoretical understanding both satisfy the U-​profile distinguishes the present case. For discussion of the case of jade, see Hacking (2007). 38.   Unrestricted versions of both theses cannot be endorsed simultaneously, as that would require equating explanatory cognition and practical abilities (which, as the examples above help to illustrate, is implausible). 39.   Zagzebski (2009, 143). See also Annas (2011b): “the skilled person can ‘give an account’ of what he does, which involves being able to explain why he is doing what he is doing” (20)—​e.g., in the case of a skilled electrician or plumber, “not just that you do the wiring or pipe-​laying such and such a way, but why” (19).

The Unity of Understanding


This imperialistic strategy faces three serious, interrelated difficulties: first, it chauvinistically excludes a set of central cases from the ranks of understanding; second, it incurs a significant justificatory debt, which it is not in a position to repay; third, it generates an unhelpful epicycle in addressing some of the deepest questions about understanding. After all, when promoting the imperialist strategy, the imperialist will inevitably omit, and thus be forced to try to discredit, a range of paradigm cases involving either skillful doers without explanatory cognition (recall the knuckleballer, the batter, the sander, and the virtuoso musician) or, instead, consummate thinkers without practical ability (recall the physicist and vision scientist). (That is the chauvinism.) Further, given that the U-​profile is satisfied in all of these cases, the imperialist’s gambit would appear to be unjustified—​or, worse, vulnerable to a host of mundane counterexamples that, again, given their satisfaction of the U-​profile, cannot simply be dismissed. (That is the unpaid debt.) Finally, since scientific advance and skillful activity are both significant achievements that share a range of important features, including those in the U-​profile, we will naturally be led to ask what they all have in common in virtue of which they do so. What makes it the case that the U-​profile is satisfied both in cases of scientific advance and in cases of skillful activity, even when the imperialist’s favored condition (explanatory cognition or, instead, practical ability) is missing? (This is the epicycle.) It may be objected that this misconstrues the imperialist’s ambition: namely, to effect unity by establishing that all instances of one form of understanding are really just instances of the other form of understanding in disguise. This is not indebted, epicycling chauvinism; it is exposé. However, this response underestimates the basic dichotomy implied by the imperialist’s strategy, one which has already cropped up in recent work on the nature of understanding. For example, in the course of defending an (increasingly popular) neo-​Aristotelian account of understanding in terms of knowledge regarding explanatory dependence-​relations, Greco unwittingly faces the choice between thinkers and doers and endorses the unhappy conclusion that a skilled agent, such as a “star gymnast,” who is ex hypothesi unaware of an explanation of her performance of a standing backflip, thereby “does not understand.”40 When prompted to choose, Greco privileges theoretical understanding. 40.   Greco (2014, 292); the same basic neo-​Aristotelian account is also endorsed by Grimm (2006; 2011; 2014). Greco focuses on understanding how the backflip is done, but his account implies that the star gymnast also does not understand the backflip, or how to do the backflip, or anything else that guides



But the choice is a false one. We should cast off imperialism, thereby removing any pressure to choose between thinkers and doers. By refraining from imperialistic schemes, we avoid the charge of chauvinism, the peril of unpaid debts, and the surfeit of epicycles. I draw two lessons from the foregoing considerations. The first is that, no matter how they are elaborated or developed, explanationism and abilitism ultimately obstruct unification. Nearly all extant theories of understanding accept (some version of ) at least one of these theses.41 But, as we have just seen, they imply either dualism, which denies the unity thesis’s conciliatory assertion of a common ground, or imperialism, which denies the unity thesis’s ecumenical embrace of all instances of the U-​profile. But it is important to recognize that although explanationism and abilitism (whether restricted or unrestricted) may each possess some initial appeal, neither is sacrosanct; both theses can be rejected. Further, and this is the second lesson, given the difficulties facing both dualism and imperialism, we have excellent reason to embrace this option—​and, in so doing, to resist the tendency toward bifurcation. Additional reason would be secured by ground. Importantly, while a proponent of this unity thesis should reject explanationism and abilitism, she need not view the tendency toward bifurcation as wholly irrational. To see this, consider the following thesis (which will play an important role below): incidentalism  There are a variety of species (forms) of understanding each of which is such that its members possess features that, while distinctive of, and perhaps essential to, the species, are incidental to the genus. This thesis implies that in order to appreciate the full richness and variety of understanding, features characteristic of particular forms of understanding are relevant; consequently, it makes good sense to explore such features, including those cited by explanationism and abilitism. The trouble arises when

her action. Greco’s discussion begins with the claim that a general understanding-​explanation link is among the “pre-​theoretical data” that function as constraints on an adequate account (287). What is needed, however, is a reason to believe that such a link is in fact a datum. 41.   Of the explanationists and abilitists quoted in the main text above, Heidegger ([1926] 1962), Merleau-​Ponty (1945, 1962), Campbell (1994), Dretske (1998), Dreyfus (2002; 2006), Kelly (2002), Kvanvig (2003), Noe (2005), and Lipton (2009) seem to favor dualism, whereas Zagzebski (2009), Strevens (2010), Annas (2011a,b), Greco (2014), and perhaps also Kim (1994) and Salmon (1984) appear to opt for imperialism. Pritchard (2010; 2014), Wilkenfeld (2013), and Kelp (2015) appear also to be explanationist-​imperialists, while Khalifa (2013a), Grimm (2011; 2014), and Hills (2015) seem to be sympathetic to dualism.

The Unity of Understanding


those features are treated not as candidate differentiae for particular forms of understanding, but as grounds for the understanding that takes those forms. (Compare the mistake of treating candidate differentiae for distinct species of the genus Rosa as what makes the members of those species roses.) In short, while incidentalism allows that such features merit attention, it implies that they are at the end of the day irrelevant to understanding per se—​as against bifurcation, and in tune with the unity thesis articulated in ground.

2.4. Noetic Conceptions In order to fully establish this unity thesis, it is not enough simply to criticize ideas that obstruct unification—​or even to do this in concert with a noncapitulating, rationalizing diagnosis of the tendency toward bifurcation. What is needed, in addition, is a positive account of understanding that vindicates unification. This is what I propose to do in the remainder of this chapter, by sketching a defense of this thesis through an account of understanding that positively identifies the state that makes ground true. The central idea of this account is that understanding, whether in the realm of theory or in the realm of practice, involves having a certain kind of conception of what is understood: what I call a ‘noetic’ conception. What unifies theoretical and practical understanding, on this approach, is not a self-​sufficient ingredient that is present in both, but a generic shape or structure—​given by the abstract notion of a noetic conception—​that is fulfilled in different ways in the two cases. I will pursue such unification in the next section. This section focuses on two preliminary questions: What is a conception? And what is it for a conception to be noetic? One’s conception of something is how one conceives or thinks, or is somehow inclined to think, of it.42 A conception has a tripartite basic structure, consisting of:

A mental state or attitude, conceiving (or conceiving-​of ) • An object, or what the attitude is of or about 43 • A content, or how the attitude characterizes its object •

42.   Conceptions and their potential theoretical importance have been emphasized by a number of philosophers working in a wide range of areas; see the citations in my (2015, 19 n. 40). While their views differ in the details, there is widespread agreement that conceptions are not concepts (the semantic values of terms such as ‘bachelor’ or ‘elm’): thus, for example, two individuals may possess the same concept while harboring very different conceptions of whatever satisfies that concept. 43.   In general, I understand the content of a state σ as what specifies the conditions in which σ is true, accurate, correct, or veridical.



For example, when one conceives of the environment as something sacred, there is the attitude of conceiving, the object of that attitude, the environment, and the content of that attitude, as something sacred. You have a conception of the environment that is different if and only if the attitude and object remain the same while the content differs (e.g., as something instrumental). This is just one illustration; examples abound. We have conceptions of ourselves, of gravity, of climate change, and of its causes and possible remedies, of how to build a fire or tie a reef knot, of the function of government, of the good life and what it takes to achieve it, and more. These and other conceptions underwrite, by structuring, a range of cognitive dispositions (e.g., to recognize, attend, notice, associate, imagine, articulate). Beyond this, they display a variety of philosophically significant properties. Five will play particularly important roles in what follows.44 The first two properties are correctness and completeness: A conception ξ of φ is correct to the extent that the content of ξ characterizes φ as having features F1 . . . Fn only if F1 . . . Fn are features of φ. A conception ξ of φ is complete to the extent that the content of ξ characterizes φ as having F1 . . . Fn if F1 . . . Fn are central features of φ. The notion of a central feature is context-​sensitive, selecting all and only what is needed to characterize what φ is in the manner that is determined by various aspects of that context. For example, what is central in a culinary context need not be identical to what is central in a botanical context; consequently, what it is for a conception of a given entity to be correct and complete in a botanical context (e.g., a conception of cocoa beans as, say, seeds produced by a small evergreen tree in the genus Theobroma and family Malvacae native to tropical regions of Central and South America) may diverge from what it is for a conception of that same entity to be correct and complete in a culinary context (e.g., a conception of them as required to make chocolate). If a conception is both correct and complete, then its content identifies the full range of its object’s central features and does not characterize the object as having any features it lacks. How do the features in the conception “hang together”? How does the conception itself “fit” with other relevant

44.   I will not here pursue a full theory of the attitude of conceiving or of the nature of its contents—​ their relations to other intensional entities, their logico-​semantical properties, their psychological roles, and so forth. Some may seek to reduce conceiving to a propositional attitude, such as belief, but in work now in progress I explain why my preferred theory is nonpropositional and nondoxastic.

The Unity of Understanding


conceptions? These questions point to the significance of two further properties of conceptions, corresponding to two types of coalescence: A conception ξ of φ is internally coalescent to the extent that the content of ξ identifies pertinent substantive connections, and pertinent features thereof, between φ’s central features. A conception ξ of φ is externally coalescent to the extent that ξ is rationally consistent with all other relevant conceptions (e.g., conceptions of φ’s central features).45 In the former case, the central features coalesce with one another; in the latter case, the conception coalesces with other conceptions. Pertinence and relevance, like centrality, are context-​sensitive; yet, neither pertinent connections nor relevant conceptions are themselves among φ’s central features. We can call the absence of internal coalescence miscellany, the paradigm of which is a conception whose content is a mere list. Take, for example, Charles Dickens’s character Bitzer’s conception of a horse as Quadruped. Graminivorous. Forty teeth, namely twenty-​four grinders, four eye-​teeth, and twelve incisive. Sheds coat in the spring; in marshy countries, sheds hoofs, too. Hoofs hard, but requiring to be shod with iron. Age known by marks in mouth.46 In Thomas Gradgrind’s schoolroom, where facts and definitions reign supreme, this zoological conception may well qualify—​let us suppose—​as fully correct and complete, identifying only those features possessed by horses and all of those features of horses that are central in that context. All the same, the young Bitzer’s conception is little more than a list, which says nothing about how, if at all, its elements are connected. Some such connections may not be pertinent in Gradgrind’s schoolroom, but others—​such as connections between gramnivorousness and possessing an abundance of grinders, or between marshy countries and shedding hoofs (which connections are of course not themselves among the central features of horses)—​presumably are. 45.   ‘Rationally’ consistent because, e.g., an exception should be made for those relevant conceptions that it would be rational for x to revise instead of ξ were x to become aware of the inconsistency. Thus external coalescence is a normative notion. I leave open how best to analyze the relevant type of consistency. 46.   Dickens (1854, 7). I am indebted to Ford (2011, §2.2.2), who invokes this passage, albeit in a different context.



That example reveals the importance of internal coalescence. Now consider external coalescence, whose function is to exclude a type of cognitive mix-​up that is compatible with a correct, complete, and internally coalescent conception. To see this, suppose that, while still in Gradgrind’s schoolroom, Bitzer’s conception of a horse was supplemented with pertinent connections among its elements, thereby precluding miscellany. Still, we can imagine that Bitzer possesses other relevant conceptions—​for example, a conception of hoofs, which (let’s suppose) he conceives incorrectly as being permanent—​ that conflict with his conception of a horse, as described above. Even if this would not affect the claim that Bitzer knows key facts about horses, it would certainly indicate that he remains somewhat mixed up about, and so does not fully understand, them.47 A fifth property invokes conceptual mastery (by contrast with minimal possession): A conception ξ is mastered by an individual x to the extent that x has mastery of the concepts (of the central features and pertinent connections), and their mode of combination, in the content of ξ.48 This property does not entail any of the previous four. It is possible to master each of the concepts in the content of a wholly incorrect or incomplete conception (e.g., the incorrect conception of the earth as having two moons). Equally, one could master each of the concepts in a conception that fails to be internally or externally coalescent. For example, Bitzer could master each and every concept in the content of his initial conception; yet, as we saw, that conception was little more than a list, hence not internally coalescent.49 The previous four properties are also not sufficient, individually or jointly, for a conception to qualify as mastered. Returning to Bitzer, let us remove the conflict between his conception of a horse and his conception of hoofs, so that he is no longer mixed up. Suppose, then, that his conception of a horse is correct, complete, internally coalescent, and externally coalescent. Compatibly 47.   The foregoing examples are not easily handled by extant theories of understanding, which tend to neglect one or both forms of coalescence (as in, e.g., theories that reduce understanding to true belief with explanatory content). 48.   I leave the reference to mode of combination implicit hereafter. 49.   As an example of mastery without external coalescence, consider an individual in an aesthetic context who has mastery of the concepts in the content of her conception of nature as merely instrumentally valuable, a conception that is not rationally consistent with her own conception of one of nature’s central features (in that context), natural beauty, as having final value.

The Unity of Understanding


with this, Bitzer might fail to master the concept quadruped, one of the concepts in the content of his conception of a horse, remaining genuinely open-​ minded about whether an individual animal who is born with only three legs but belongs to a species whose members normally are four-​legged qualifies as quadruped. Simply put, failure to attain mastery of relevant concepts may be directly responsible for deficiency of understanding, and it is for this reason that I include the fifth property.50 I will call a conception possessing all five of the properties explicated above to the greatest extent—​a conception that is fully correct, complete, internally and externally coalescent, and mastered—​fully noetic. I  will call a conception adequately noetic if it is not fully noetic, but nonetheless has each property to an adequate extent; and minimally noetic if it is neither fully nor adequately noetic, but has each property at least to some minimal extent. (We can assume that the thresholds for adequate and minimal noeticness are context-​dependent.) These possibilities, and in particular that a noetic conception is always noetic to a greater or lesser extent, should be borne in mind in what follows.51

2.5. Gaining ground I propose that a noetic conception is a plausible candidate for the state, σ, that makes ground true. To motivate this proposal, I  will argue that the U-​profile is satisfied in virtue of possession of a noetic conception (§2.5.1),

50.   Extant accounts of understanding do not impose any condition on conceptual mastery. As a result, they deliver the wrong verdict about an individual (such as Bitzer in the example just given) who satisfies those accounts, despite harboring conceptual deficiency, which undermines his or her understanding. See Fricker (2007, ch. 7)  for fascinating examples of the role of conceptual mastery in self-​understanding, and its ethical implications. For discussion of conceptual mastery itself, see Burge’s (1986) seminal discussion of grades of concept possession (e.g., partial versus full), which in turn is inspired by, inter alia, Frege’s distinctions between levels or sorts of mastery (e.g., “foggily” versus “clearly” or “sharply” possessing the concept number; see the excellent overview by Burge 1990, §§IV–​ VI; cp. Peacocke 1992, 29ff. and Bengson and Moffett 2007, 42–​43). These distinctions, and the possibilities described in the previous two paragraphs, can be explained by diverse theories; for example, on Bealer’s (1998) account, levels or sorts of mastery of a concept are explained by, roughly, the quality of intuitions one would have involving that concept in suitable cognitive conditions. 51.   I have focused on five properties that constitute noeticness. These properties may be related to various others discussed in recent work on understanding. For example, their instantiation by a conception may ground various intellectual abilities that “normally” accompany theoretical understanding (including, perhaps, those described by Hills 2016, §2; the qualifier ‘normally’ is hers). In addition, each property can come to be instantiated luckily or nonluckily (in various senses of ‘luck’). There is currently no consensus about the relation between understanding and various sorts of luck, and I will remain neutral on this matter here (but see the discussion of ‘resilience’ in my 2015).



and that noetic conceptions have important roles to play in the realms of both theory (§2.5.2) and practice (§2.5.3). 2.5.1. Noetic Conceptions and the U-​Profile The preceding section offered a general characterization of conceptions, explained their tripartite structure, and provided a definition of each of the properties required for a conception to qualify as noetic. The characterization and definitions are noncircular, informative, and (plausibly) modest, appealing to notions familiar from traditional debates (e.g., content, features, connections, consistency). This does not fit well with the hypothesis that they are covertly disjunctive, but suggests instead that a noetic conception is a type of nondisjunctive state. Arguably, one could not be in a state of this type without possessing genuine understanding. For possession of a noetic conception appears to entail, and make it the case, that all of the features in the U-​profile are present. A conception is a standing, psychological state. The correctness property ensures that it is, to some extent, objective. The completeness property, in conjunction with the mastery property, ensures that it is, to some extent, intelligent. The two coalescence properties ensure that it is, to some extent, robust, coherent, and orderly. Each of these properties can be possessed to a greater or lesser extent, rendering the state itself multiply gradable: assessable (e.g., as better, greater, deeper, stronger, or richer) with respect to the extent of correctness, completeness, internal coalescence, external coalescence, and mastery. If the state combines all five of these properties, then its subject conceives of something (the target) in a way that is not merely accurate, but at the same time displays cogency and organization amid complexity—​this is the hallmark of sense-​ making, and it founds a praiseworthy good. It follows that if one has a noetic conception, then one (or one’s state) thereby satisfies the U-​profile. 2.5.2. Theoretical Understanding: Illuminating Conceptions Earlier I  proposed that understanding in the realm of theory centers on illumination. Some conceptions have contents that identify laws, causes, mechanisms, unifying principles, dependence relations, and other items that are, or could be, appropriately invoked to answer explanation-​seeking questions. Such conceptions are illuminating conceptions, possessing the sort of information required to resolve queries regarding the why, what, or how

The Unity of Understanding


of things. Illuminating conceptions, when noetic, underwrite theoretical understanding. This connection between illuminating conceptions and theoretical understanding can be motivated by reflecting on, from one direction, what one has when one has a noetic conception, and, from the other direction, what it takes to achieve a state that both satisfies the U-​profile and is poised to resolve inquiry. Begin with the first direction:  if one has a noetic conception, then one has a mental state that contains information sufficient to fully characterize one’s topic (in ways relevant in that context). Now consider the other direction, concerning what it takes to be in a state that both satisfies the U-​profile and is poised to resolve relevant queries. If one lacks information sufficient to fully characterize one’s topic (in ways relevant in that context), then one is not in a state that satisfies the U-​profile, and one is not yet in a position to close investigation (in that context). These points suggest a convergence: one is in some state satisfying the U-​profile that is poised to resolve relevant queries just in case one has a conception that is both noetic and illuminating.52 Bitzer provides an illustration. Recall his initial, list-​like conception of discrete features of horses. At that initial stage, at which his conception is merely correct and complete, his state does not satisfy the U-​profile, as it is not yet robust, intelligent, coherent, and orderly. Suppose, however, that Bitzer’s conception of horses were to acquire the full range of properties constituting noeticness: consequently, it would not simply be correct and complete, but also internally coalescent, externally coalescent, and mastered. Then, and only then, would Bitzer attain a state that both satisfies the U-​profile and puts him in a position to resolve a wide range of explanation-​seeking queries regarding horses. Theoretical understanding goes lockstep with acquisition of a conception that is both noetic and illuminating. To see that a noetic illuminating conception is not merely sufficient but is also necessary for theoretical understanding, it may help to consider that mere knowledge-​that, even when abundant, is insufficient for the task. As we saw above (in §2.4), Bitzer may know key facts about horses, even while remaining confused or mixed up about them—​as when his conception is a mere list that conflicts with other relevant conceptions, lacking both internal and external coalescence—​and thus not genuinely understanding them. To be sure, in some cases, an individual who knows that p also has at least 52.   The connection between illuminating conceptions and theoretical understanding explains the attraction of explanationism, preserving the primary insight in that position while detaching it from its problematic implications. Both the attraction and the problems were discussed in §2.3.



some theoretical understanding of or relating to p.  But this connection is fairly weak, and there appear to be many exceptions, beyond the case of Bitzer. Intuitively, my highly pedestrian knowledge that 1 is prime does not merit classification as an instance of theoretical understanding. Nor does my knowledge that I am not currently experiencing severe pain, or that this (a flick of the wrist) is what is needed to make the ball spin. This is not, or at least not simply, because such knowledge may not qualify as theoretical, but because it may fall short of genuine understanding. Notice, after all, that a wide range of instances of knowledge-​that do not satisfy the U-​profile, as in cases of knowledge-​that stemming from rote memorization, spectatorial perceptual demonstration, or naive testimonial uptake. These points can be explained in terms of noetic conceptions. Knowledge-​ that—​even lots of knowledge-​that, including knowledge that p because q—​is compatible with an impoverished conception (or set of impoverished conceptions), which lacks many of the properties constituting noeticness. Since those properties ground satisfaction of the U-​profile, it follows that in a range of cases involving knowledge-​that, the U-​profile may go unsatisfied, leaving the knower without theoretical understanding, even when the knowledge-​that is theoretical. 2.5.3. Practical Understanding: Guiding Conceptions Let us now turn to practical understanding, the sort of understanding that is action-​g uiding. While such understanding might initially seem resistant to explanation by a view that privileges noetic conceptions, I will attempt to show that the present approach is well suited for the task. Earlier I noted that there is an important connection between practical understanding and know-​how (recall §2.1.1). Elsewhere I have argued (in a series of papers coauthored with Marc Moffett) for a view of know-​how that endorses the following two theses, which employ several notions—​‘conception’, ‘correct’, ‘complete’, and ‘mastery’—​explicated above: (α) To know how to φ is to master a (possibly implicit, possibly demonstrative) correct and complete conception of a method, or way, of φ-​ing, where a method of φ-​ing is constituted by a (possibly ordered, possibly singleton) set of action types, the execution of which is φ; (β) A conception ξ of a method of φ-​ing is correct and complete (in the relevant context) only if ξ is a guiding conception: the exercise of ξ could underlie and explain the successful, intentional execution of φ (viz., for one who masters it and acts on its basis).

The Unity of Understanding


I will not rehearse the arguments for this view here.53 What I wish to emphasize is that this account of knowing how entails that, and indeed explains why, practical understanding is centrally concerned with activity not merely in its topic or subject matter but in its function or character. For it follows from the conjunction of (α) and (β), together with the connection between practical understanding and know-​how, that an individual who has practical understanding will be in a state that is action-​g uiding, poised to underlie and explain the intentional execution of intelligent action. The account of know-​how in (α) invokes a conception that is correct, complete, and mastered; (β) tells us that such a conception is guiding. Although these theses uncover an intimate relation between practical understanding and guiding conceptions, they do not quite establish what is needed to secure unity, namely, a connection between practical understanding and noetic conceptions—​which are, in addition to being correct, complete, and mastered, also internally and externally coalescent. I propose that the latter properties enter in the transition from mere know-​how to skill. We have seen that if one has a noetic conception, then one thereby has a state that has the features constituting the U-​profile, including robustness, coherence, and orderliness (recall §2.5.1). We have also seen that in the practical realm, these three features correspond to the systematicity and generality of skill, which make skill more demanding than mere know-​how (recall §2.1.1). If skill requires a conception possessing internal and external coalescence, in addition to those features—​correctness, completeness, and mastery—​implied by the account of know-​how in (α), then the systematicity and generality of skill naturally follow. By way of illustration, consider a case in which your conception concerns the use of a tool, such as a fretsaw. Suppose that your conception is noetic:  it is correct, complete, internally coalescent, externally coalescent, and mastered. Because the conception is correct, complete, and mastered, then on the assumption that (α) is correct, it will imply knowledge how to use the saw; given (β), it will qualify as a guiding conception: it will be poised to underlie and explain intentional use of the saw. 53.   See Bengson and Moffett (2007; 2011a; 2011b) for elaboration and defense of both theses. On conceptions of methods, see especially our (2011a, §§5.2–​3), which offers a detailed explanation of the datum that know-​how is action-​g uiding. This datum is also discussed by Kumar (2011, §2), Cath (2015, 10 and 14–​15), and Santorio (2016, §4.1). A connection between know-​how and methods (albeit as referents of inferential rules) is also emphasized by Pavese (2015, esp. 13), and was a prominent theme in Ryle’s seminal discussion (see, e.g., Ryle 1945–​46, 4). I cannot engage the (vast and growing) literature on know-​how more fully here.



Because the conception is also internally coalescent, it identifies pertinent connections—​not always or primarily explanatory—​among its elements, including temporal, spatial, and normative connections between the steps within a given procedure or method of use (e.g., a turn of the blade should be followed by several short pull strokes), as well as connections between those steps and the steps in a variety of other methods (e.g., one is quicker than, or requires more dexterity than, or is more effective for maple than, another). Because, further, the conception is externally coalescent, it is rationally consistent with other relevant conceptions, including conceptions of blade, frame, and more generally saw and wood. Moreover, in a given instance of fretsawing, demonstrative conceptions of the particular blade, frame, and wood being used (and much else besides, including one’s own body) will be relevant. Given all this, it should be clear that in skilled agency, one’s conception of a method (or set of methods) for successfully using a fretsaw speaks not only to the specific situation in which one is utilizing that conception to work on a particular piece of wood with a particular fretsaw, but also to a wide range of other situations involving variations in the standing properties of this target and this tool, as when the moisture level increases or the blade’s teeth require sharpening, as well as in endlessly many cases involving different, perhaps novel targets (e.g., other pieces of wood) and tools (e.g., other fretsaws, coping saws, etc.). The conception displays both systematicity and generality.54 The upshot is that if it is true that skill requires possession of a noetic conception that is guiding, then it follows that skill requires not just know-​ how, but also systematicity and generality. And conversely, if it is true that skill requires know-​how plus systematicity and generality, it follows that skill requires a guiding conception that is noetic. In this way, there is a set of conditions for skill—​involving knowledge how (as construed in (α)), systematicity, and generality—​that links practical understanding to noetic conceptions.

54.   As this example illustrates, such systematicity and generality does not prohibit the conception from being situation-​specific. As stated in (α), the relevant conceptions may also be implicit and demonstrative (e.g., a four-​year old’s conception of a way of wiggling her ears—​viz., by doing this); in addition, their mastery may require a suite of abilities or kinesthetic properties. For another illustration, see the discussion of Heidegger’s hammering example in my (2015, 28–​30). See Chuard (2006, §3) for discussion of plausible constraints—​e.g., Discrimination, Context-​dependence, Attention, Location, and Inferential—​on demonstrative concepts. For the importance of implicit, demonstrative conceptions to knowledge how (e.g., some cases of knowledge how to wiggle one’s ears), see Bengson and Moffett (2007, esp. 51–​52).

The Unity of Understanding


2.6. Understanding in General I have argued that noetic conceptions forge unity between theoretical understanding and practical understanding, while accounting for the distinctive feature of each (à la incidentalism). This final section discusses what a conceptions-​based approach to the unity of understanding implies or suggests about understanding in general. 2.6.1. Intellectualism Considered in and of themselves, noetic conceptions are neutral between theory and practice. This is not to say that they resist categorization. For example, they are psychological states. Further, they are cognitive, as opposed to affective or sensory. Even further, they are intellectual, as opposed to (say) behavioral-​dispositional. Hence a conceptions-​based approach to the unity of understanding merits the label ‘intellectualism’.55 It might be objected that such intellectualism indicates that the proposed state is too close to the theoretical to serve as a neutral (non-​imperialistic) common ground, and that, in particular, it fails to do justice to nontheoretical forms of understanding, especially practical understanding, which is not an intellectual affair. But, first, the inference from intellectuality to theoreticality is a non sequitur. Even if the theoretical is intellectual, the intellectual need not thereby be theoretical. Reading Shakespeare, solving a sudoku puzzle, playing Jeopardy, counting silently to one hundred, and operating a sophisticated home theater are all plainly intellectual—​they directly engage the intellect—​but they are not, or at least need not be, theoretical. Second, as revealed by reflection on the complexity of many intelligent actions and the cognitive sophistication required for their skillful engagement, the idea that practical activities undertaken with understanding do not directly engage the intellect, in a way that implicates how doers conceive of their doings, both underappreciates and distorts their achievement.56

55.   In fact, I have chosen the Greek term ‘noetic’ because it connotes the intellect (but not necessarily theory). 56.   Recall the fretsaw example from §2.5.3. Elsewhere I  have argued that intellectual—​including conceptual—​failures undermine an agent’s know-​how even when corresponding abilities are in place (see the Salchow cases in Bengson and Moffett (2007, 46–​49; cp. 2011a), which also serve as counterexamples to abilitism). This provides independent motivation for an intellectualist approach and helps to defuse the charge of over-​intellectualization. For related discussion of that charge, see Bengson and Moffett (2007, 51–​55) and Bengson, Moffett, and Wright (2009).



2.6.2. Toward Enrichment While noetic conceptions are not in and of themselves specifically concerned with theory or with practice, they become so concerned when further conditions are met, as when, for example, they are illuminating or, instead, guiding. Or both—​as witnessed, arguably, in the player-​coach or erudite artisan, who grasp both the theory and the practice. This recalls the possibility of enrichment: enrichment  There is a type of understanding U, such that x has U with respect to . . . only if one both practically understands . . . and theoretically understands . . . , and U is superior to either of these in isolation. On the present approach, the relevant type of understanding, U, is achieved when one enjoys a noetic conception that is simultaneously illuminating and guiding. Importantly, such a state is not a gerrymandered conjunction of two fundamentally disparate types of understanding (as dualism would imply), or an unholy mixture of understanding plus something possibly ‘other’ (as imperialism would imply), but an integrated whole. It is a single conception, which is correct, complete, internally and externally coalescent, and mastered, poised to resolve inquiry regarding, and at the same time to guide the intentional execution of intelligent action concerning, whatever is understood. Success in this undertaking is no trivial accomplishment, but a dynamic process culminating in an epistemic ideal realized just when one is truly in command of some portion of reality. 2.6.3. Other Varietals What about other forms of understanding, besides those featured in the realms of theory and practice (and, as just discussed, in their combination)? Above I alluded to the diverse understandings of diverse exemplars, for instance, the farmer, painter, designer, comedian, lover, conversationalist, businessperson, philanthropist, fitness planner, tour guide, moral exemplar, and psychoanalyst. A conceptions-​based approach allows us to make sense of the differences between these understanders, while preserving the unity of their various achievements. Suitably generalized, the approach says, first, they are unified—​what makes them all forms of genuine understanding is one and the same—​insofar as they are all grounded in possession of a noetic conception (in keeping with a generalized version of GROUND), and, second, they are diverse insofar as each form has some distinctive

The Unity of Understanding


feature, over and above this, which serves as a criterion of individuation (again, per the thesis of incidentalism). There are at least two principled ways to distinguish various forms of understanding.57 The first privileges what is understood; hence it distinguishes forms of understanding by reference to the object of conception: for example, soils and seasons, forms and colors, fashions and fabrics, humor, people, norms, markets, charities, regimens and body-​types, locales, ways of living, and emotions and moods. While this method of individuation may be best in some cases, in many cases it does not deliver the most perspicuous taxonomies. A carpenter’s understanding of the wood with which she works (e.g., maple) and of the procedure by which it is dried have different objects (a concrete stuff versus a process or activity), though insofar as both are action-​ guiding they ought to be grouped together, as belonging to one and the same practical form of understanding. A narrow focus on the objects of conception also omits important distinctions: as noted above, a dancer and scholar might both understand ballet, albeit in very different ways—​the difference, of course, lies in how, not what, they understand, revealing an important limitation of an object-​oriented taxonomy.58 An alternative focal point is the content of conception, which might differ even between noetic conceptions of one and the same object. The content of a carpenter’s noetic conception of the maple she is handling may specify how to dry or cut it, whereas the content of a botanist’s conception of maple characterizes its phylogenetic properties, an artist’s highlights its aesthetic qualities, a firefighter’s indicates how it burns, a businessperson’s records its market value, and a tour guide’s focuses on its location and autumnal colors. This represents a second option for individuating forms of understanding. And it is often the one to favor: in a wide range of cases, the content is the difference that makes the difference. For example, it is the difference in the contents of the noetic conceptions of the carpenter, botanist, artist, firefighter, businessperson, and tour guide that explains why their understandings of maple are diverse: practical, theoretical, aesthetic, pyro, mercantile, and touristic, respectively.59 57.   My earlier objections to lexicology also cast doubt on the wisdom of citing linguistic forms, such as ‘understand why’, to mark the distinctions we seek. 58.   Several other examples of this sort were described above. 59.   More precisely, it is a function of the difference in content together with what qualifies as noetic in the context at hand, for example, at a botanical conference versus while fighting a forest fire (recall the various respects in which noeticness is context sensitive, as described in §2.4). Contents can of course differ in myriad further ways: for example, the content of a noetic conception of maple could be more or less abstract (e.g., as belonging to the ontological category material stuff ), specific (e.g., as



The idea of a content-​oriented taxonomy has general application. Take the case of understanding people: here we may countenance the possibility of diverse conceptions with diverse contents, concerning, inter alia, anatomical characteristics, cultural and historical properties, psychological qualities such as emotions and moods, or whatever makes a given individual beloved. Thus it is possible to have anatomical, cultural-​historical, empathic, or loving understanding of people. Here, again, the difference resides primarily in the content of the conception sought.60 That difference often implicates another. Achieving a conception with a particular content may in some cases be possible only once one has acquired a corresponding set of physical or mental characteristics. A particular guiding conception with a demonstrative content specifying a complex sequence of movements may require a certain suite of kinesthetic properties. A particular empathic conception with a content specifying the felt intensity of various emotions may require a certain suite of psychological properties (perhaps, e.g., one must have suffered emotional pain oneself in order to properly conceive of the emotional pain of others). Possession or acquisition of these properties may, in turn, implicate a certain course of experience. Here we may remember Mill’s emphasis on “experiments of living,” which he applauded for their capacity to generate “better taste and sense in human life,” including superior moral and self-​understanding—​through, I propose, improved conceptions of oneself, other people, and how to live well.61 that material—​thought while demonstrating maple), precise (e.g., as having flowers that are regular, pentamerous, and borne in racemes, corymbs, or umbels), and so forth. Further, in the case of practical understanding, both quality of method and quantity of methods matter. Such differences often do not secure interesting taxonomies, however. Differences in contents and objects interact in complex ways, thereby enabling complex forms of understanding and levels of understanding. One upshot is that a conceptions-​based approach accounts for the distinctions between familiar sorts of understanding while also making sense of why some of these distinctions are neither exclusive nor sharp (as noted at the outset of §2.2). 60.   Empathy has often been said to lie at the heart of the human sciences, and to make their cognitive goal fundamentally distinct from the cognitive goal of the natural sciences. (This idea was traditionally linked to a sharp division between verstehen and erklären.) But if the former aims at empathic understanding and the latter aims at theoretical understanding, a conceptions-​based approach implies that their goals are not fundamentally distinct, but have a common core—​namely, both strive for a noetic conception of their object. For discussion of the role of empathy in the “autonomy of historical understanding,” see Mink (1966, §§III–​IV). 61.   Mill (1859, ch. 3). I discuss the role of intuition in improvements of understanding in my (2015); the account is readily extended to the moral domain.

The Unity of Understanding


2.6.4. Comprehensive Understanding The foregoing remarks indicate how a conceptions-​based approach enables an extension of ground, beyond the realms of theory and practice, so as to secure a general unity thesis while safeguarding—​with the aid of incidentalism—​ the full richness and variety of understanding. The approach also allows a generalized version of enrichment, which recognizes the prospect that multiple forms of understanding (not just theoretical and practical) might combine and aggregate so as to contribute to increasingly superior forms of understanding. As more and more forms are introduced, it will of course become correspondingly more difficult to sustain a conception that is fully correct, complete, internally and externally coalescent, and mastered. Should one manage it, however, one will have arrived at a unified understanding that is not simply better, greater, deeper, stronger, and richer than what came before, but also more and more comprehensive, until at the limit—​which is perhaps unattainable—​everything is properly conceived.

Acknowledgments I am grateful to audiences at the Institut Jean Nicod, Hebrew University, the Midwest Epistemology Workshop at the University of Michigan, the Central APA in Chicago, and the Ranch Conference in Arizona, as well as to participants in graduate seminars on understanding at the University of Wisconsin–​Madison and Harvard University, for probing discussion. I received extremely helpful comments on earlier drafts from Tomas Bogardus, Paul Boghossian, Anton Ford, Ellen Fridland, Martha Gibson, Stephen Grimm, John Mackay, Blake Myers-​Schulz, Dilip Ninan, Alan Sidelle, Elliott Sober, Denny Stampe, and Daniel Wilkenfeld. I am deeply indebted to Marc Moffett for illuminating discussion at various stages in the chapter’s development, and to Anat Schechtman for extensive and insightful feedback, both critical and constructive. Work on this chapter was made possible through the support of a grant from the Varieties of Understanding Project at Fordham University and the John Templeton Foundation.

References Anderson, Lloyd B. 1986. “Evidentials, Paths of Change, and Mental Maps: Typologically Regular Asymmetries.” In Evidentiality: The Linguistic Coding of Epistemology. Ed. Wallace L. Chafe and Johanna Nichols. Norwood, NJ: Ablex. 273–​312.



Annas, Julia. 1981. Introduction to Plato’s Republic. Oxford: Oxford University Press. —​—​—​. 2011a. “Practical Expertise.” In Knowing How: Essays on Knowledge, Mind, and Action. Ed. John Bengson and Marc A. Moffett. New  York:  Oxford University Press. 101–​12. —​—​—​. 2011b. Intelligent Virtue. Oxford: Oxford University Press. Aristotle. (350 bce) 1908. Metaphysics. Trans. W. D. Ross and J. A. Smith. Oxford: Oxford University Press. Bealer, George. 1998. “A Theory of Concepts and Concept Possession.” Philosophical Issues 9: 261–​301. Bengson, John. 2015. “A Noetic Theory of Understanding and Intuition as Sense-​ Maker.” Inquiry 58: 633–​68. Bengson, John, and Marc A. Moffett. 2007. “Know-​How and Concept Possession.” Philosophical Studies 136: 31–​57. —​—​—​. 2011a. “Non-​propositional Intellectualism.” In Knowing How:  Essays on Knowledge, Mind, and Action. Ed. John Bengson and Marc A. Moffett. New York: Oxford University Press. 161–​95. —​—​—​. 2011b. “Two Conceptions of Mind and Action:  Knowing How and the Philosophical Theory of Intelligence.” In Knowing How:  Essays on Knowledge, Mind, and Action. Ed. John Bengson and Marc A. Moffett. New  York:  Oxford University Press. 3–​55. Bengson, John, Marc A. Moffett, and Jennifer C. Wright. 2009. “The Folk on Knowing How.” Philosophical Studies 142: 387–​401. Brandom, Robert. 2013. “From German Idealism to American Pragmatism—​and Back.” In Kant und die Philosophie in weltbürgerlicher Absicht. Ed. Stefano Bacin, Alfredo Ferrarin, Claudio La Rocca, and Margit Ruffing. Boston: de Gruyter. 107–​26. Brownstein, Michael, and Eliot Michaelson. 2016. “Doing without Believing.” Synthese 193: 2815–​36. Burge, Tyler. 1986. “Intellectual Norms and Foundations of Mind.” Journal of Philosophy 83: 697–​720. —​—​—​. 1990. “Frege on Sense and Linguistic Meaning.” In The Analytic Tradition. Ed. David Bell and Neil Cooper. London: Blackwell. 30–​60. Campbell, John. 1994. Past, Space, and Self. Cambridge, MA: MIT Press. —​—​—​. 2011. “Why Do Language Use and Tool Use Both Count as Manifestations of Intelligence?” In Tool Use and Causal Cognition. Ed. Teresa McCormack, Christoph Hoerl, and Stephen Butterfill. Oxford: Oxford University Press. 169–​82. Cath, Yuri. 2015. “Revisionary Intellectualism and Gettier.” Philosophical Studies 172: 7–​27. Chuard, Philippe. 2006. “Demonstrative Concepts without Re-​ identification.” Philosophical Studies 130: 153–​201. Cooper, John M. 1977. “The Psychology of Justice in Plato.” American Philosophical Quarterly 14: 151–​57. de Regt, Henk. W. 2009. “The Epistemic Value of Understanding.” Philosophy of Science 76: 585–​97.

The Unity of Understanding


de Regt, Henk W., and Dennis Dieks. 2005. “A Contextual Approach to Scientific Understanding.” Synthese 144: 137–​70. Dickens, Charles. 1854. Hard Times. London: Bradsbury & Evans. Dickie, Imogen. 2012. “Skill before Knowledge.” Philosophy and Phenomenological Research, 85: 737–​45. Dretske, Fred. 1998. “Where Is the Mind When the Body Performs?” Stanford Humanities Review 6.2. Dreyfus, Hubert. 2002. “Intelligence without Representation—​ Merleau-​ Ponty’s Critique of Mental Representation: The Relevance of Phenomenology to Scientific Explanation.” Phenomenology and the Cognitive Sciences 1: 367–​83. —​—​—​. 2006. “Overcoming the Myth of the Mental.” Topoi 26: 43–​49. Elgin, Catherine Z. 2007. “Understanding and the Facts.” Philosophical Studies 132: 33–​42. Fodor, Jerry A. 1968. “The Appeal to Tacit Knowledge in Psychological Explanation.” Journal of Philosophy 65: 627–​40. Ford, Anton. 2011. “Action and Generality.” In Essays on Anscombe’s “Intention”. Ed. Anton Ford, Jennifer Hornsby, and Frederick Stoutland. Cambridge, MA: Harvard University Press. 76–​104. —​—​—​. 2017. “The Representation of Action.” Royal Institute of Philosophy Supplement 80: 217–​33. Fricker, Miranda. 2007. Epistemic Injustice:  Power and the Ethics of Knowing. Oxford: Oxford University Press. Fridland, Ellen. 2014. “They’ve Lost Control:  Reflections on Skill.” Synthese 191: 2729–​50. Glick, Ephraim. 2011. “Two Methodologies for Evaluating Intellectualism.” Philosophy and Phenomenological Research 83: 398–​434. Greco, John. 2014. “Episteme:  Knowledge and Understanding.” In Virtues and Their Vices. Ed. Kevin Timpe and Craig A. Boyd. New York: Oxford University Press. 285–​301. Grimm, Stephen R. 2006. “Is Understanding a Species of Knowledge?” British Journal for the Philosophy of Science 57: 515–​35. —​—​—​. 2008. “Epistemic Goals and Epistemic Values.” Philosophy and Phenomenological Research 77: 725–​44. —​—​—​. 2011. “Understanding.” In The Routledge Companion to Epistemology. Ed. Sven Berneker and Duncan Pritchard. London: Routledge. 84–​94. —​ —​ —​ . 2014. “Understanding as Knowledge of Causes.” In Virtue Epistemology Naturalized:  Bridges between Virtue Epistemology and Philosophy of Science. Ed. Abrol Fairweather. Heidelberg: Springer. 329–​45. Hacking, Ian. 2007. “The Contingencies of Ambiguity.” Analysis 67: 269–​77. Hartland-​Swann, John. 1956. “The Logical Status of ‘Knowing That’.” Analysis 16: 111–​15. Haugeland, John. 1998. Having Thought: Essays in the Metaphysics of Mind. Cambridge, MA: Harvard University Press.



Heidegger, Martin. (1926) 1962. Being and Time. Trans. John Macquarrie and Edward Robinson. New York: Harper and Row. Hempel, Carl. 1965. Aspects of Scientific Explanation and Other Essays in the Philosophy of Science. New York: Free Press. Hills, Alison. 2015. “The Intellectuals and the Virtues.” Ethics 126: 7–​36. —​—​—​. 2016. “Understanding Why.” Noûs 50: 661–​88. Kant, Immanuel. (1784) 1970. “On the Common Saying ‘This May Be True in Theory, but It Does Not Apply in Practice.’” Trans. H. B. Nisbet. In Kant: Political Writings. Ed. Hans. S. Reiss. Cambridge: Cambridge University Press. 61–​92. Kelly, Sean Dorrance. 2002. “Merleau-​Ponty on the Body.” Ratio 15: 376–​91. Kelp, Christoph. 2015. “Understanding Phenomena.” Synthese 192: 3799–​816. Khalifa, Kareem. 2012. “Inaugurating Understanding or Repackaging Explanation?” Philosophy of Science 79: 15–​37. —​—​—​. 2013a. “Is Understanding Explanatory or Objectual?” Synthese 190: 1153–​71. —​—​—​. 2013b. “The Role of Explanation in Understanding.” British Journal for the Philosophy of Science 64: 161–​87. Kim, Jaegwon. 1994. “Explanatory Knowledge and Metaphysical Dependence.” Philosophical Topics 4: 51–​69. Kitcher, Philip. 1989. “Explanatory Unification and the Causal Structure of the World.” In Scientific Explanation. Ed. Philip Kitcher and Wesley C. Salmon. Minneapolis: University of Minnesota Press. 410–​505. Kumar, Victor. 2011. “In Support of Anti-​ intellectualism.” Philosophical Studies 152: 135–​54. Kvanvig, Jonathan. 2003. The Value of Knowledge and the Pursuit of Understanding. Cambridge: Cambridge University Press. —​—​—​. 2009. “The Value of Understanding.” In Epistemic Value. Ed. Adrian Haddock, Alan Millar, and Duncan Pritchard. Oxford: Oxford University Press. 95–​112. Lihoreau, Franck. 2008. “Knowledge-​How and Ability.” Grazer Philosophische Studien 77: 263–​305. Lipton, Peter. 2004. Inference to the Best Explanation. 2nd ed. London: Routledge. —​—​—​. 2009. “Explanation without Understanding.” In Scientific Understanding. Ed. Henk W. de Regt, Sabina Leonelli, and Kai Eigner. Pittsburgh: University of Pittsburgh Press. 43–​63. Merleau-​Ponty, Maurice. 1945. Le Phénoménologie de la Perception. Paris: Gallimard. —​—​—​. 1962. The Phenomenology of Perception. Trans. C. Smith. New York: Routledge & Kegan Paul. Mill, John Stuart. 1859. On Liberty. London: Longman, Roberts, & Green. Mink, Louis 1966. “The Autonomy of Historical Understanding.” History and Theory 5: 24–​47. Moravcsik, Julius M. 1979. “Understanding.” Dialectica 33: 201–​16. Noë, Alva. 2005. “Against Intellectualism.” Analysis 65: 278–​90.

The Unity of Understanding


Papineau, David. 2013. “In the Zone.” Royal Institute of Philosophy Supplement 73: 175–​96. Pavese, Carlotta. 2015. “Practical Senses.” Philosophers’ Imprint 15: 1–​25. Peacocke, Christopher. 1992. A Study of Concepts. Cambridge, MA: MIT Press. Pritchard, Duncan. 2010. “Knowledge, Understanding, and Epistemic Value.” Royal Institute of Philosophy Supplement 64: 19–​43. —​ —​ —​ . 2014. “Knowledge and Understanding.” In Virtue Epistemology Naturalized:  Bridges between Virtue Epistemology and Philosophy of Science. Ed. Abrol Fairweather. Heidelberg: Springer. 315–​27. Ryle, Gilbert. 1945-​ 46. “Knowing How and Knowing That.” Proceedings of the Aristotelian Society 46: 1–​16. —​—​—​. 1949. The Concept of Mind. Chicago: University of Chicago Press. Salmon, Wesley C. 1984. Scientific Explanation and the Causal Structure of the World. Princeton, NJ: Princeton University Press. —​—​—​. 1998. Causality and Explanation. Oxford: Oxford University Press. Santorio, Paolo. 2016. “Nonfactual Know-​How and the Boundaries of Semantics.” Philosophical Review 125: 35–​82. Stanley, Jason. 2011. Know How. Oxford: Oxford University Press. Stanley, Jason, and Timothy Williamson. 2001. “Knowing How.” Journal of Philosophy 98: 411–​44. —​—​—​. Forthcoming. “Skill.” Noûs. Strevens, Michael. 2010. “Varieties of Understanding.” Talk at the Pacific Division Meeting of the American Philosophical Association. Available online at http://​​research/​expln/​varund.shtml. —​ —​ —​ . 2013. “No Understanding without Explanation.” Studies in History and Philosophy of Science 44: 510–​15. Vlastos, Gregory. 1972. “The Unity of the Virtues in the Protagoras.” Review of Metaphysics 25: 415–​58. Wilkenfeld, Daniel. 2013. “Understanding as Representation Manipulability.” Synthese 190: 997–​1016. —​—​—​. 2009. On Epistemology. Belmont, CA: Wadsworth.


Understanding Persons and the Problem of Power Remy Debes

3.1. Introduction Oh, brethren, my way, My way’s cloudy, my way; Go send them angels down. ​Negro spiritual, “My Way’s Cloudy” In analytic philosophy, two problems are usually posed when it comes to understanding other people. One is the skeptical problem of other minds. The other is the problem of social cognition, often dubbed a problem of “mindreading.” The first wonders how we know that there are any minds besides our own. The second wonders how we know what is going on in those other minds assuming they exist. But this platitudinous framework has been upset by trending work in social epistemology, which has thrown into relief a third problem for understanding persons. Call it the problem of power. This problem can be expressed roughly as follows. Oppressed people face a double-​edged epistemic challenge. On the one hand, assuming they are aware of it, they must make sense of their oppression. On the other hand, social epistemology suggests they are often hindered from the means to do just this. In a variety of ways, those with power actively and passively hinder those without power from interpreting and communicating their experiences, thus obstructing their role in the “production” of knowledge about their own

Understanding Persons and the Problem of Power


oppression. The second hand of this problem has come to be called “epistemic” oppression. And it tends to work in two ways. First, oppressed people have had and continue to have their status as knowers called into question. Women, for example, have long endured both disparaging views of their intellect and the outright silencing of their opinions, judgments, outcries, and protests. Such silencing is sometimes active (e.g., actual prohibition of expression, as in the denial of the right to vote) and sometimes passive (e.g., the way a businessman might discuss key company initiatives in the locker room). Second, even when oppressed people do express themselves, these expressions are often not treated as knowledge. That is, oppressed people have had and continue to have the meaning fulness of their expressions challenged, either for being in some sense literally unintelligible or for being inappropriate. To take an obvious example, when hip-​hop first emerged it was widely discounted by mainstream white America as “not real music.” As late as 1996, the seminal trio De La Soul lamented this view of hip-​hop on their album Stakes is High with a striking sound bite from a white man who complains, “There’s no music in it. It’s just niggers talking.” It is not clear this has changed. Whites often still discount hip-​hop on the grounds of its being violent and profane—​despite the fact that such violence and profanity is often precisely the point, which is to say, the violence and profanity just is what many rap artists are trying to convey about the life they endure. In “My City,” the opening track of his newest album, Yo Gotti drawls: Just be patient, he gon fuck up, he a Memphis nigga Maybe I will, maybe I won’t Thugging Money don’t change that, fame can’t change that I’ve done shit to niggas in the streets I can’t take back So every second of my life I got a gun and I hate that But this Memphis, if you get caught without you gon regret that. Now, the primary concern of social epistemology has been to diagnose the nature of epistemic oppression, especially in these two respects—​i.e., the discounting of people as knowers and the discounting of what they express as knowledge. To this diagnosis there is sometimes added an analysis of the kind of harm that epistemic oppression constitutes (which hitherto has been



usually explained as some kind of threat to “epistemic agency”). But in the background of these analyses is a correlate problem for any general theory of understanding persons. After all, the “internal” problem of understanding just described (about how oppressed people make sense of their own oppression, i.e., for themselves) must be mirrored by an “external” one, namely, about how, or whether, such oppression can be understood by others, and, in particular, by those with power. Consider: Empowered people must not assume they can rely on (their) dominant ideologies to understand those without power.1 Nor can the empowered demand the oppressed to translate their experiences into terms consonant with those ideologies. Both these moves would only reify the relevant practices of epistemic oppression. Nor again can the empowered simply identify and abstain from practices of epistemic oppression. One does not “simply” think outside one’s ideology; that is oxymoronic. Moreover, it is a premise of the problem of power that sufficient conceptual resources are not yet in place for oppressed peoples to express directly (without translation) the meaning of their oppression. Thus we must wonder: Can empowered people ever genuinely understand oppressed people? Is it possible to theorize “successful” understanding in cases of unequal power without reifying existing modes of epistemic oppression? And what exactly is “understanding” in these kinds of cases? I propose working roughly backward through these questions, given that the first two seem to hang on an answer to the third. To be clear, my goal is not to analyze epistemic oppression, let alone the nature and causes of the dominant ideologies such oppression trades on. My goal is to marshal insights from existing analyses of epistemic oppression in order to motivate an inquiry into the nature of understanding persons, which inquiry, it seems, must be part of a comprehensive solution to the problem of power. And while I don’t pretend to offer a comprehensive solution here, I do hope to provide some traction.

3.2. Epistemic Oppression Yea, it makes me wanna holler And throw up both my hands

1.   The point is not that those in power can reach no kind of understanding whatsoever of the experience of oppression using their dominant ideologies (whether social, psychological, political, or otherwise). That would be a very strong claim. My point is only that for any given case the empowered cannot assume their dominant ideologies will be adequate for understanding experiences of oppression.

Understanding Persons and the Problem of Power


Crime is increasing Trigger happy policing Panic is spreading God know where we’re heading Oh, make me wanna holler They don’t understand ​Marvin Gaye, “Inner City Blues” The problem of power is not new. Something like it has been on the minds of sociologists at least since Clifford Geertz outed his own field in 1974 for its pretention to produce “walking miracles of empathy” (1974, 26–​ 45). And Geertz, for his part, credited the basic conceptual insight to the nineteenth-​century architect of hermeneutics, Wilhelm Dilthey. The problem of power is at the fore of W. E. B DuBois’s 1903 The Souls of Black Folk. And it is at least implicit in the stories of slave life told by DuBois’s near contemporary, Charles Chesnutt. It can be found in Marx and some of his heirs, like Simone Weil and Frantz Fanon. It is arguably part of what Mary Wollstonecraft was getting at in her 1792 A Vindication of the Rights of Woman. And in a sense, it just is the problem of Françoise de Graffigny’s 1747 Letters from a Peruvian Woman. I would not be shocked to hear of yet earlier sources. Despite this pedagogy, the problem of power has rarely been stated plainly. We thus already make progress here by dragging it into public. But also, and more to the point, as long as epistemic oppression persists in the world, then I assume there is sufficient cause not simply to think through the problem of power, but to rethink the problem—​that is, to actively search for new ways, new concepts, and new terms with which to express it. One might say that is precisely what the problem predicts of itself. So let us begin rethinking. In this section, I start by reviewing in more detail the nature of epistemic oppression. In the following two sections, I turn my attention directly to the nature of understanding persons. In the final section, I link the analysis of understanding persons back to epistemic oppression, thereby coming full circle to the problem of power. Gender, race, ability, sexual identity, class, education, clique, neighborhood:  these identities structure our most basic social relations. They are also ineliminably entangled in power dynamics, and for this reason become the substrate of various forms of inequality and oppression. Traditionally, the principal terms of analyzing such oppression have been social, political, and economic. Social epistemology, however, challenges this presumption.



The fundamental premise of social epistemology is that social relations can and do affect the production of knowledge. They shape the customs by which we exchange knowledge, whom we treat as a knower, the way our shared conceptual schemes are built or revised, and so on. But if this is correct, and given that social relations are power relations, then, as those like Miranda Fricker, Rae Langton, and Kristie Dotson (among many others) have shown, the production of knowledge can be the locus of inequality and oppression.2 Most obviously, oppressed peoples are subjected to various forms of exclusion from the production of knowledge, often referred to as “silencing.” In this case people on the losing side of power relations are treated as unreliable or totally incredible believers and witnesses, even with respect to their own experiences. Women, blacks, and gays, for example, have long suffered this kind of epistemic deprecation. When allowed in, they have been subject to derogatory epistemic stereotypes—​they are sources of “old wives’ tales,” “street smarts,” and “fashion sense.” But in matters of serious consequence, including their own oppression, their voices have been and remain subject to myriad forms of qualification, second-​g uessing, and rejection—​again, when they are not simply ignored. This species of epistemic oppression is sometimes called testimonial injustice. Less obvious are the ways in which social meaning itself is the locus of epistemic oppression. Consider: claims about the social world—​both first-​ order claims like “He’s gay” and second-​order claims like “Gay people are marginalized in the production of knowledge”—​must deploy innumerable concepts and descriptions. These concepts and descriptions are often the subject of disagreement, which, like any disagreement, we typically try to settle by appeal to “facts.” But of course, if we are talking about social facts, then at some level these facts depend on what we make of them—​what “we” understand “our” social world to be. In other words, social facts at some level depend on interpretation, especially “our” collective practices of interpretation.3 2.   The now seminal touchstone is Fricker (2007). Many other important voices might be mentioned, but see Langton (2000, 127–​45); Dotson (2014, 115–​138). Dotson’s piece is especially useful for its demonstration that epistemic oppression resists reduction to more familiar forms of social, economic, or political oppression. Most of these scholars trace the roots of theorizing about epistemic oppression to a range of earlier thinkers from Nancy Harstock, Edward Craig, Michael Foucault, Simon de Beauvoir, and even, in Dotson’s case, all the way to Plato. For an important extension to the case of children, see Burroughs and Tollefsen (2016). 3.   See also Fricker (1999, 191–​210); especially §IV, where she discusses the nature of the dependence relationship.

Understanding Persons and the Problem of Power


And yet, such interpretive practices have been and remain dominated by those in power; hence the need to speak not of our world, but of “our” world. “Black” “White” “Woman” “Man” “Married” “Gay” “Straight” “Normal” “Cool” “Weird”—​the meanings of concepts like these have long been normed to the interpretative practices of empowered social groups, who legislate, if only passively, the deployment and revision of existing concepts as well as the invention and codification of new ones. The result is that those on the losing side of power relations suffer a second kind of epistemic oppression, what Miranda Fricker has influentially dubbed “hermeneutic injustice.” One might wonder whether hermeneutic injustice by itself suffices to generate the problem of power for understanding persons—​indeed, whether they are mutually entailing. Fricker suggests as much in her own analysis, arguing that “the moment of hermeneutical injustice” occurs “only” when a person is hermeneutically marginalized through a “more or less doomed attempt on the part of the subject to render an experience intelligible, either to herself or to an interlocutor” (2007, 159). Now, bracket for a moment Fricker’s imbedded claim that hermeneutic injustice necessarily entails a failure of understanding. My question is this: Is hermeneutic injustice coextensive with the marginalization of persons engaged in an act of interpretation, and thus always and only a hermeneutic oppression of those on the losing side of power relations?4 There is reason to say no. As thinkers such as Charles Mills, Shannon Sullivan, and José Medina have argued, hermeneutic injustice also aptly describes the epistemology of “ignorance” that typically marks societies with trenchant social inequity (see Mills 1997; 2007, 11–​38; Sullivan 1997, 153–​72; Medina 2013). In these societies, dominant power identities (male, white, heterosexual, physically able) can become so much the norm that they are rendered essentially invisible. That is, social meaning is so deeply normed to dominant identities, that “normal” becomes synonymous with what is white, male, hetero, able, and so on. “Flesh” color Band-​Aids are made in white skin tones; nonhuman and even nonanimal characters like trains and trucks in children’s

4.   Even if it were, it is important to note, as Fricker herself does, that this would not excuse us from attending to the phenomena of silencing, which has a rather obvious entanglement with hermeneutic injustice. After all, one who is silenced faces a perverse challenge when it comes to resisting, revising, or creating new social meaning; and those with impoverished hermeneutic resources will struggle twice as hard to be recognized as credible, reliable knowers. See Fricker’s comments (2007, 159).



books are almost always masculine;5 television ads depict families with opposite-​sex parents; stairs are always the most direct way into a building; and so on.6 These may seem like mundane examples, but the mundaneness is precisely the point. It is the saturation, so to speak, of a dominant identity in a culture that renders it (or at least allows it to be) invisible. The consequence is a hermeneutic marginalization of the oppressors themselves, who are allowed to become ignorant about their own privileged identity. Indeed, they have a psychological stake in remaining ignorant—​that is, to resist efforts by oppressed groups to overturn, revise, or simply pluralize the conceptual landscape in ways that would allow the disempowered to express their oppression. For, of course, the success of such expression must involve somehow making visible the relevant dominant identities and thus the dominance itself, which visibility is necessarily threatening to these dominant identities. These identities are structured, if only unconsciously, on the privileges that come with being the invisible norm—​the unnoticed regulators of “the normal.”7 A final clarification about hermeneutic injustice is needed. The thesis of hermeneutic oppression does not entail, or even imply, hermeneutic bankruptcy. Oppression doesn’t create zombies, even if it sometimes seriously screws people up, epistemically speaking. Nor is the suggestion that oppressed people are some kind of interpretive simpletons, totally incapable of making sense of their oppression or totally unable to resist dominant forms of knowledge production. That would be insulting. And it would be insulting in part because it would constitute an act of testimonial injustice. It would be tantamount to claiming oppressed people cannot think outside dominant conceptual schemes and interpretive practices. Moreover, it would be false. As José Medina aptly notes, long before terms like “marital rape” or “domestic violence” gained currency, women were struggling to articulate their experiences of abuse, even if only in “fragmentary and precarious ways.”8 Likewise, 5.   This is especially obvious when the inanimate character has a plot-​neutral role, e.g., they don’t occupy a relevant familial role like “mother.” 6.   I’m grateful to my mother, Barbara Debes, for discussing this point. She is a lesbian, and when I asked her for intentionally mundane examples of how the social world in normed hetero, she replied, “What isn’t?”—​before giving me the example I use here (among many others). 7.   The fundamental point here about identity threat been traced to various important historical thinkers like Marx and Fanon. The first and arguably most poignant contemporary American treatment is Bell (1992). More recently, see Bernasconi (2007, 231–​39). The locus classicus of the conjoined point about privilege and its invisibility is McIntosh (1988). 8.   All quotations in this paragraph are from Medina (2013, 99–​100).

Understanding Persons and the Problem of Power


the intelligibility of same-​sex relations did not depend only on the emergence of labels like “same-​sex marriage” and “civil unions.” And so on for every kind of oppression. Medina’s point is that the struggle to make sense of and express one’s experience is already hermeneutic activity. Correspondingly, a social group’s hermeneutic capacity is not wholly dependent on having, as he puts it, “readily available terms and coined concepts.” In short, far from thinking hermeneutic injustice tells us something essential about those being oppressed, it tells us something about oppressors. Namely, oppressors have a responsibility to develop, as Medina puts it, “a hermeneutical sensibility” to attempts to communicate meaning outside normalized concepts and conventions. Those in power must become receptive to embryonic, even inchoate attempts to communicate what it’s like to live without power and privilege. In particular, they must be open minded about what seems to them—​from their already empowered point of view—​strange or foreign or uncomfortable in the expressions of those struggling to explain their experiences of oppression.9 In other words, empowered people must be ready to struggle in their own right. I endorse Medina’s dictate—​though, of course, it raises new questions. What exactly does such sensibility consist in? What exactly is required to satisfy a demand to develop it? And how does this demand and the corresponding duty of sensibility relate to other forms of epistemic oppression or other epistemic duties? These are urgent questions. And yet, they are not the questions I propose to chase up in the remainder of this chapter. For it seems to me that they skirt around the problem of power. At least, none of these questions strikes directly at the nature of understanding persons. Indeed, these questions arguably presuppose a conception of it. After all, unless we know what it is to understand another person, how can we decide whether we’ve successfully deployed a hermeneutic sensibility, let  alone decide what such sensibility consists in? To be clear, what I am suggesting is that the work on epistemic oppression would benefit from conceptual expansion. By focusing directly on the nature of understanding, and specifically what it is to understand persons, we might better dissect the problem of power as well as provide a resource for thinking about epistemic oppression generally.

9.   I borrow the specification of ‘embryonic’ and ‘inchoate’ from Medina (2013, 99). But for reasons that will clearer later, I think Medina’s dictate must be amended to include my own final point about the seemingly strange and foreign.



3.3. Understanding Persons: A First Take You know the reasons but still won’t ever know my life Kendrick AKA Compton’s human sacrifice ​Kendrick Lamar, “m.A.A.d city” Music is a crucible for thinking about the ways people have struggled to forge and express the meaning of their oppression. From Kendrick Lamar’s recent Good Kid m.A.A.d City, to Marvin Gaye’s seminal What’s Going On?, to the bars of the sorrow songs that DuBois famously positioned at the head of each chapter of the Souls of Black Folk, we hear not only the sounds of misery and anger, but also confusion. To be perfectly clear, in saying this I do not mean to suggest that I hear these sounds truly—​that is, that I take myself to be able to say what such music is really trying to express. Given what I shall go on to argue in this essay, it is philosophically important to reject any such intent. Instead, I venture. I venture that common to all this music is a note of tragic irony. They are songs and in some cases whole albums that express the experience of living in mean, marginalized, or mad conditions, while at the same time, often self-​consciously, calling into question their ability to do just that—​to make those experiences intelligible. Even DuBois, who at one point says that the meaning of the sorrow songs is “clear”—​that all sorrow songs express the hope that “sometimes, somewhere, men will judge men by their souls and not by their skins”—​even DuBois essentially marks their tragic irony when he tells us that these songs tell us not only of “death and suffering” but also “unvoiced longing toward a truer world, of misty wanderings and hidden ways.” In these songs, the slave “spoke to the world”; but, he adds, “Such a message is naturally veiled and half articulate” (Du Bois 1994, 162, 157, 159). These are depressing reflections, especially given that black American music is just one of countless possible examples of the intersection of oppression and art. But it also brings to the fore a crucial intuition: it is our experiences that matter. Our experiences, whether of oppression or love or loss or anything else, are the central focus of our efforts to establish understanding with others. So, let us try to leverage this intuition, by asking afresh: What is it to understand another person? Many things, it seems. We alternatively claim to understand the utterances, actions, beliefs, feelings, and so on, of other people. Sometimes we make even finer distinctions, as when we distinguish understanding what caused someone to do something from understanding her reasons for doing

Understanding Persons and the Problem of Power


it.10 Call this understanding-​about a person. But sometimes we mean something else entirely. Sometimes we mean quite literally that we understand her. That is, sometimes the object of our understanding is not a discrete utterance, action, belief, feeling, etc., of a person. It is the person, considered holistically, in a social or moral sense. It is the “you” that we face in our daily moral and social encounters, and which we sometimes speak of, third personally, as the “her” or “him” that we understand. Call this connotation of understanding persons, holistic understanding.11 How can we make sense of holistic understanding claims? First, note that such claims rarely purport anything like “total” understanding. “I totally understand you” simply doesn’t mean, in our daily exchanges, that we understand everything about a person. So too for other familiar expressions connoting holistic understanding such as, “Yeah, I  get you, man”; or “I feel you”; Or, to flip the direction of address to capture that cliché-​ but-​telling bit of lovers’ disputes and teenager angst, “Why can’t you understand me?!” Granted, eventually I must clarify exactly how “whole” holistic understanding has to be in order to count. And I will. But for now, the question is, in what sense do these claims purport to be holistic yet not totalistic? I suggest we render holistic claims as claims to understand personal perspective. On the one hand, it is commonplace to talk of people’s “outlook” or “point of view.” And most of us think of ourselves in similar terms. We think of ourselves as the locus of some distinctive point of view on the world—​or at least, on particular events in the world, including, especially, our own experiences. On the other hand, when we make holistic claims like, “Why can’t you 10.   These superficial distinctions can quickly become complicated, especially when we begin to ask—​as many have—​whether any of these varieties understanding in fact reduce to others. Donald Davidson, for example, argued influentially that a “primary reason” explains an action by being its cause. Similarly, Carl Hempel’s “covering law” model of explanation seems to invite us to account for understanding action by treating reasons as causes, if not by treating the relevant “law” as contingent and empirical. For an excellent discussion and vigorous challenge to these reductionist arguments, see Kim (2010, 105–​24). 11.   The idea of holistic understanding has obvious affinities with a broad array of existing disciplinary discussions, especially (1) those that aim to enhance our understanding of persons (including everything from critical race and queer theory to action theory), or (2) those that try to explain how we understand persons (especially social cognition theories like Simulation Theory, Theory-​Theory, and Direct Perception). And yet, neither of these existing bodies of literature addresses the specific questions I’ve posed. The first literature is useful and sometimes necessary to reach an understanding of persons or particular groups of persons; but that’s obviously not the same as saying what it is to understand persons. Meanwhile, the second literature is almost always framed in terms of what I dubbed “understanding-​about” persons. That is, the social cognition literature by and in large aims to explain how we understand discrete actions, utterances, behaviors, etc., of persons.



understand me?” or “I don’t really get her”—​it does seem to be a person’s point of view that we have in mind, that is, as what is to be understood. So there is something intuitively attractive here. Moreover, there is an obvious payoff in adopting this suggestion. By construing holistic understanding as a kind of perspectival understanding, we get immediate traction with the problem of power given that it is the experiences of oppressed people which partly motivate the problem. However, the suggestion needs to be defended. This requires saying a fair bit more about what perspective is, and what it means to understand someone’s perspective. I’ll do this first in a rough and ready way, and then, in the next section, in more detail. Even in the rough, the notion of perspective offers a double layer of content that promises the breadth and depth needed to make sense of the idea that it is sometimes the person—​in a holistic sense—​that we claim to understand. First, to reiterate, what one has a perspective on is, minimally, her experiences. And in the context of interpersonal understanding, these experiences are almost always thick kinds of things. Indeed, in most cases they will be more like little stories—​stories about being teased by your best friend at a party, or some betrayal of your love, or the struggle to be taken seriously by your parents. Better, because more precise, the experiences at issue will combine many smaller events and phenomena which, while they might be broken apart and made the object of discrete understanding-​ about claims, in everyday conversation we string together to make one thick experience. Second, and crucially, if we are talking about “perspective,” then it is not just the experience in and of itself, no matter how thick, that is at issue. It matters that it is someone’s experience—​indeed, someone particular: mine, yours, hers. In other words, when I claim to understand you, I don’t simply purport to understand the experience you had or have; I purport to understand the experience as yours. Or better, as you experience it. The double stress here, on both “as” and “you,” is intentional. To see the point, let me ask one more time (albeit still in the rough): What is it to understand a person’s perspective? The notion of perspective I’m developing might suggest that we answer this question in terms of a person’s “position.” That is, it might seem that what it is to understand a person’s perspective is to see as another person sees, e.g., to see someone’s experience as she sees it. In fact, I want to suggest that understanding a person’s perspective is more aptly described as seeing her experience as she sees it, where this change in emphasis implies the need for some further or special cognitive accomplishment.

Understanding Persons and the Problem of Power


Consider: On the first emphasis—​of seeing as she sees—​we are reminded of the familiar adage that “knowing” someone requires “walking a mile in his or her shoes.” And while it is beyond my scope here to defend a distinction between knowledge and understanding, the easy appeal to a connotation of “knowing” here is telling. For if understanding a person’s perspective only requires seeing as they see, there is an implication, or at least a risk of implying, that the mere accumulation of descriptive facts about a person’s experiences can suffice for understanding her. That is, there is a risk of implying that if I simply saw what she saw, heard what she heard, even had the beliefs she had—​this would suffice to understand her, in a holistic sense. But notice that if we accepted this, we actually seem to leave the person in question out of the understanding, in virtue of leaving her “take” on the facts out. That is, her interpretation of the facts suddenly seems inessential. And that would be bad irony. It seems confused, for example, for me to claim to understand you—​your perspective—​without referring to your interpretation of the facts. I take this point to be fundamental. Understanding a person requires being in contact with another person’s experiences not merely as she sees them, but as she sees them. Or to spin the metaphor another way, you can’t merely put on someone else’s shoes and walk around. You need a closer connection than that. As Atticus Finch famously told his daughter Scout, you need to “climb into his skin and walk around” (emphasis added).

3.4. Peculiar Perspectives Herein the longing of black men must have respect:  the rich and bitter depth of their experience, the unknown treasures of their inner life, the strange rendings of nature they have seen, may give the world new points of view and make their loving, living, and doing precious to all human hearts. — ​W . E .   B . D U B O I S ,

The Souls of Black Folk

Let me begin to rework my claims less metaphorically. Having a perspective shouldn’t be confused with merely having particular belief-​positions about substantive matters, political, moral, religious, and so on. Such beliefs inform a person’s perspective but are not identical to it—​not in the sense relevant to holistic understanding anyway. For example, your political point of view obviously depends on your political beliefs. Thus, without some knowledge of those beliefs it is unlikely I could understand your political perspective in a holistic sense. Still, if I know only your political beliefs, at most I will understand something about you. Holistic understanding claims, recall, purport to apply to particular persons—​to the particular you. By contrast, the content



of many beliefs is generalizable in the sense that for many beliefs you hold, anyone else might hold them.12 Thus we are driven back to thinking that the thinnest slice which makes for holistic understanding is an experience, even if that experience is only “what it is like” for a particular person to entertain a particular belief, political or otherwise. For only then does it make sense to introduce talk of “points of view” or “perspective” in the first place. And if my allowing for a “particular belief ” seems at risk of becoming too skinny to substantiate a holistic claim, keep in mind that beliefs are theory-​laden bits of cognition. They are always embedded in a network of other beliefs, memories, feelings, and so on. Thus, while the content of beliefs is in principle generalizable, what it is like for a particular person to entertain them is in practice always informed by a particular web of connections. Even a particular belief, then, is a potential inroad to a person’s particular perspective. And yet, this “what it’s like” talk is misleading. Having perspective, at least in an ethical context, can’t simply be to have experiences be “like” something.13 That risks conflating the notion of experience with perspective. For creatures like us, we can’t imagine merely having experiences without those experiences “being like” something. Perhaps ants or aliens with no consciousness can “merely have” experiences. But not us. So if we want a notion of perspective that is not already included in the notion of experience, we need to appeal to more than, or something different than, what experiences are “like.” I thus suggest the following: having a perspective is to have experiences mean something to you. That might seem question begging for the problem of power. But as I said much earlier, the problem of power doesn’t start with—​ and indeed is offended by—​the premise of interpretive simpletons. Moreover, I am not demanding thick, deep, or clear meaning. I am claiming that to have a perspective your experiences must simply have some significance, even if thin or opaque, and they must have this significance for you. By this I mean that you must have the in-​principle possibility of taking an interpretive stance on your experiences. You must be able to give or entertain a description of what a given experience means or meant to you.

12.   I’m grateful to Stephen Grimm for pointing out, about beliefs that contain indexicals like, “I am walking in Memphis,” that it would be suspicious to say that anyone else could hold that exact belief. 13.   My definition of perspective thus moves in a different direction from the way certain philosophers of mind or metaphysicians, like Lynne Rudder Baker, understand the concept. But I don’t think my direction is antagonistic to theirs. Case in point, I think Baker’s analysis and mine would fit together fairly neatly. See Baker (2013).

Understanding Persons and the Problem of Power


It is important that it suffices to be able to entertain a description. To say otherwise risks ruling out perforce not only young children and nonhuman animals, but also anyone who is merely unable to communicate, or communicate clearly, their interpretation. But a failure to communicate the significance of one’s experience at most entails a lack of expressive skills, conceptual repertoire, or self-​understanding, not the lack of perspective. In short, like having a therapist interpret your dreams, it suffices if you can recognize when a suggestion rings true and when it doesn’t; when the carrot you were holding in your dream represents a penis, or was, well, just a carrot. Being able to reject possible interpretations of your experience is enough.14 The foregoing thesis about the nature of perspective can be buttressed by shifting attention back to the concept of understanding. Consider:  Understanding a particular person’s perspective must involve more than grasping the generic significance of an experience. It is, after all, your perspective that’s at issue—​at least, if it’s really you I’m claiming to understand and not just a given experience as-​someone-​like-​you would suffer it. Sometimes we intend the latter, more generic claim. But typically we mean exactly the opposite. This is especially obviously when we are making demands to be understood by others. For example, when my wife is angry with me over some marital issue, and is pressing me to understand why she is angry, she would bring down the roof if I cut her off midexplanation by saying, “Yes, yes, I understand how wives can be angry about such things.” To rely on generic interpretations this way flies in the face of what satisfies us when we want to be understood—​that is, when we want our personal perspectives to be understood.15 Indeed, the problem of power emphasizes exactly this point. How irksome it must be to be black and be asked what such and such means to black people, as if there is but one black perspective. How irksome to be a gay and have someone say, “Oh I know just how you feel, some of my best friends are gay.” Instead, as DuBois emphasized in the case of black men, but could have said for all oppressed people, there is never one point of view 14.   One might wonder whether or not this excludes nonhuman animals. For the purposes of this chapter, I take no position on this further question. I only note that, it seems implausible if not ad hoc to insist that the only way one can confirm or reject the interpretation another is making of her experiences is through spoken language. So I  don’t see why this condition in itself should be thought to exclude nonhuman animals. 15.   Granted, sometimes our disputes are less complex. We don’t feel the need to be understood holistically—​only to have something about us understood, for example, some particular action. In this case, we might be content if others appreciate the legitimacy of our reasons.



the world must try to understand. There are many longings, many inner lives, many points of view. Of course, if you and I have a close relationship, or if I have the relevant background experiences (perhaps some of my best friends really are gay), then I might make some reliable guesses about the significance of your experience. In some cases, these guesses may even be more reliable than your own guesses inasmuch as they are or can be less occluded by the grip of bias and passion. Hence expressions like, “She knows me better than I know myself.” Nevertheless, they are still guesses. Like any interpretation of the significance of your experience, such guesses are always defeasible before the scales of your own eyes. Until your viewpoint is brought into the interpretive process, all interpretations are at best only hypotheses. Once again, to have a perspective is always to have a peculiar perspective. But there is one final, less obvious, though no less crucial implication of the foregoing analysis. Having a perspective is a dynamic aspect of being a person. This is because perspective is itself fluid. The experiences that are the object of perspective and which are what we ask others to understand (or seek to understand of them), are diachronic things, sometimes extending over considerable spans of time. Like the stories we tell or songs we sing to explain them, they are rarely defined by a single moment. And even if or when they do seem to be about a particular moment, they are nevertheless in principle varied by the extension of time it takes to tell or sing about them. Moreover, as we said, our experiences are always caught in a web. They connect in complicated ways to other experiences, beliefs, memories, hopes, and regrets. And these other connections are themselves in constant flux. Hence why we find ourselves revisiting the meaning of our experiences, and feel the pull when others demand us to do just that, even if we balk at actually doing so.

3.5. Understand-​ing Persons: Back to the Problem of Power The intelligence is defeated as soon as the expression of one’s thoughts is preceded, explicitly or implicitly, by the little word ‘we.’ —​Simone Weil, The Needs of the Soul It follows from the conclusions of the last section that we cannot, strictly speaking, “understand” someone’s perspective. In something like the way we

Understanding Persons and the Problem of Power


think of a person’s “personality,” a person’s perspective is impossible to grasp only in terms of generalities. Correspondingly, to understand another person can’t be a static cognitive achievement, as it can be, say, to understand why a house burned down after an electrical failure or to understand the phenomena of morning dew. Instead, when it comes to persons there is only a process of understand-​ing, corresponding to the dynamic and peculiar nature of perspective. Let me elaborate. If we begin to think about the conditions for successful understanding, notice how perverse it sounds to say you gave “appropriate” consideration to someone’s peculiar perspective if you treat it as static. Unless you mean something like, “This is how I understood him then”—​i.e., at a given time where you were engaged dynamically with him—​to treat a person’s perspective as static would mean you tried to understand someone’s perspective only in its generality (or in a generality). For example, this would be the case if you tried to interpret a black woman’s experience as a “human” would, or as a “woman” would, or as “black” person would. Such generalities have their uses, but in the context of understanding particular persons, such generalities entail making gross assumptions about experience, and thus cut against the very idea of perspective as something particular and fluid. Of course, we cannot avoid making some such assumptions. As social cognition theory tells us, there is a massive framing problem when it comes to interpersonal communication, the solution to which must involve “supplying” content at some level and to some degree. But the mere fact that we supply content is not the issue. The issue is, first, that the dynamic nature of perspective makes such content-​supply essentially suspect.1 Second, and following from the first, whatever else appropriate or successful understanding of perspective might require, it must involve confirmation by the person we claim to recognize. There must be some form of interpersonal communication by which she signals to us that we are “getting her right.” In other words, holistic understanding is essentially dialogic, even in a sense negotiated (albeit with the one-​to-​be-​understood enjoying a crucial epistemic privilege). This won’t sit well with old epistemologies that assume cognitive accomplishments like understanding are simply “in” the head. Nevertheless, my inquiry points the other way. Understanding persons seems to exist between persons—​ though, to be sure, sometimes we see no choice but to occupy both roles. Self-​ understanding is tough like that. This thesis is apt to raise some worries. Does my view entail we can’t legitimately claim to understand people we are not presently in dialogue



with? What about the dead, or fictional characters? And does it turn out that we never fully understand anyone else? After all, surely it is impossible to get fully inside someone else’s perspective. To say otherwise suggests undue credence in the power of imagination, but also belies my own point about the theory-​laden nature of perspective. For, if one’s experiences are networked to one’s other mental states in such a way that the significance of those experiences always depends on those other mental states, then how could we determine where to draw the line as we trace these connections outward? Isn’t it in principle possible that these connections eventually extend to the total set of one’s mental states? In short, isn’t holistic understanding really whole-​istic after all? I can legitimately claim to understand you only if I fully understand you? Call this last question the full understanding worry. There is not space here to address all these questions thoroughly. Thankfully, the first few worries seem to turn on exaggerations and confusions. For example, one would misspeak to say she understands, now, her dead lover. Either she means she understood him (1) in the past or “at that time,” or (2) she is speaking counterfactually: “Were he still with us, I would understand him.”16 The “past tense” case (1) is not in tension with my view, even if in practice it could be difficult to verify; e.g., it might be hard for the widow to verify now (to herself or anyone else) that she did then appropriately engage her husband in the kind of dialogue that would satisfy the dynamic conditions of understanding personal perspective. As for the “counterfactual” case (2), she is either appealing to the same evidence she would use in the past-​tense case together with an assumption that her dead husband is an unchanged person, or she is laying claim to some kind of ersatz dialogue; e.g., perhaps she discovered an unread letter from her husband that put her husband’s perspective into a new, easier-​to-​interpret light (“It is only now, having read his letter, that I understand him”). As for fictional characters, the worry is exaggerated. Only a boor would claim books, poems, movies, and the like are altogether beyond interrogation. Fictional characters are virtually always in a dialogue of some kind. Moreover, and more to the point, spoken discourse explicitly aimed 16.   I do not intend this claim to extend generally to all kinds of understanding. Plausibly there are kinds of understanding claims about past events that are perfectly legitimate when uttered in the present. My point here is that understanding persons is distinctive in this respect. Unless we assume an implied past-​ tense index or counterfactual, then it is not legitimate to claim to understand persons who can no longer verify the interpretations we make of them. This is just a consequence of my argument.

Understanding Persons and the Problem of Power


at holistic understanding has never been a condition of my analysis. It is only the paradigm. I’m thus happy to allow that fiction offers a grip for dynamic engagement. Of course, fictional characters cannot literally verify an interpretation we make of their perspective. So there can be no certainty of understanding them. But then again, isn’t this the correct result? Few things are more suspicious than claims to have hit upon the interpretation of a novel or movie or play, as if novels and movies and plays were not partly defined by the impossibility of precisely that. Their whole charm, or a considerable part of it anyway, depends on the fact that there is no end to how they can be interpreted. So it seems confused for anyone to claim that they “certainly” understand a fictional character. But now let me turn to the worry of full understanding, for this will lead back to the problem of power. The full-​understanding worry turns on a mistaken assumption that “full” should be conceived of descriptively, as something like “total” or “perfect” understanding. Yes, we say things like, “I totally get you.” But we rarely mean it literally, and we ought not to.17 Understanding a person’s perspective “fully” is or ought only to be a sufficiency claim. This is manifest in our everyday practice, where success is treated as understanding someone’s perspective “well enough.” Of course, one might now wonder how we decide what counts as “enough”? As it happens, however, we get a good start on this new question by simply fleshing out why holistic understanding should be normed to sufficiency instead of descriptive fullness. 3.5.1. The Moral Argument If a descriptively “perfect” grasp were the goal, who among us would really want to be fully understood? We are undoubtedly usually eager to have our individual perspectives taken into account, but this desire comes with the caveat that some distance is preserved. This is not just an issue of privacy. We cringe at the suggestion that we are not, to some degree, inscrutable in the eyes of others. Perhaps in those teenage moments of insecurity we are quick 17.   Epistemologists might wonder whether the present claim is meant to suggest that those like Catherine Elgin are correct in claiming that understanding isn’t factive in the way knowledge is. I happen to be largely persuaded by Elgin. However, strictly speaking, the present point is agnostic on this issue. I am merely highlighting the fact that the norms of success when it comes to understanding persons cannot be modeled on factive accuracy, even if factive accuracy proves to be one condition of success.



to allow our perspective to be amalgamated into claims of what “we” think, feel, desire. But most of us come (or come back) to a deep, if private, wariness of such claims, which seem to defeat our sense of self, intelligence, and agency. But also, the aim of perfect understanding seems a mischaracterization of the ethically proper mindset of the one-​trying-​to-​understand. To think we actually see the world exactly as someone else does, flies in the face of our basic concept of other-​regarding respect. Whatever the precise nature of such respect turns out to be, it must preserve the distinction between me and thee—​that is, of self and other. Yet this is precisely what a claim of perfect understanding, if taken literally, seems to efface. 3.5.2. The Epistemic Argument If one must dialogically negotiate her understanding with the person-​to-​ be-​understood, then plainly this understanding comes in degrees. For we allow that it can be judged as more or less accurate, namely, by the person-​ to-​be-​understood. And this is fitting, given that it is a commonplace among epistemologists to say that ‘understanding’ is marked partly by its coming in degrees—​as opposed, say, to some kinds of knowledge. For example, it would be strange to say one “sort of ” knows that the sun is rising. But it is perfectly ordinary to say one “sort of ” understands the sun’s rising (i.e., the physical explanation for why it does). Granted, there are theorists who favor reducing all talk of understanding to talk of knowledge or some specific species of knowledge talk, like know-​how (e.g., sort-​of-​understanding the sun’s rising would mean sort-​of-​knowing-​how it does). 18 But we can bracket this possibility here. Assuming I  must eventually say why a dialogically negotiated interpretation of personal perspective really is a species of “understanding,” then it is enough for now to note that successful interpretations are accomplishments that comes in degrees. 3.5.3. The Political Argument To norm holistic understanding to descriptive fullness would systematically favor socially empowered groups and harm disempowered ones, thus reifying both testimonial and hermeneutic forms of epistemic oppression. Empowered groups already enjoy certain epistemic privileges when it comes

18.   See, e.g. Stephen Grimm (2009, 84–​95) and Jason Stanley and Timothy Williamson (2001, 411–​44).

Understanding Persons and the Problem of Power


to self-​and interpersonal understanding. Dominant social groups trade on existing, “collectively” shared—​perhaps we should say, mainstream—​forms of social understanding to reach self-​and interpersonal understanding. Indeed, this privilege is simply another face of the problem of power, inasmuch as disempowered groups are pressed to conform to mainstream social understandings. These understandings may take myriad forms, but are especially obvious in the gamut of narrative tropes the mainstream uses to relate what it means to (1) occupy paradigm social roles (father, wife, teacher, doctor, bus driver, etc.); (2) experience a major life event, as well as to say an event is ‘major’ (job college graduation, promotion, birth of a child, retirement, etc.); or (3) possess life goods (job, house, gym membership, etc.). We rely on these narrative tropes to make sense of who we are, to ourselves and to others. And yet, as Miranda Fricker has argued, it is easy to entertain the idea that relations of unequal power “skew” these tropes, with the result that the powerful tend to have more widely accepted understandings of their experiences, “whereas the powerless are more likely to find themselves having some social experiences through a glass darkly, with at best ill-​fitting meanings to draw on in the effort to render them intelligible” (Fricker 2007, 148). The consequence is, if full understanding were set to a descriptive norm, then dominant groups would get the unearned privilege of better prospects for successful “full” understanding—​at least in the sense that they would more likely have the sense of being understood.19 Correspondingly, disempowered groups will be subject to disparaging judgments of their perspectives, which perspectives will inevitably be thought weird or simply unintelligible. This last remark might elicit a pause. On the one hand, the foregoing consequence hardly seems hypothetical, but rather, given all I  said at the outset, a description of the actual world we live in. On the other hand, I  said that everyday claims to understanding persons do not purport to be claims of descriptively full understanding, but instead claims of sufficiency. What, then, explains why disempowered groups are actually struggling for understanding? But the answer to this question is the point. So consider: Despite purporting only to sufficiency, our practices of understanding are nevertheless vulnerable to trenchant epistemic injustice. Even though our actual claims to

19.   One must be careful not to suggest that empowered people are in fact epistemically better off. On the one hand, as Stephen Grimm rightly pointed out to me, if the “accepted” tropes distort reality, it is not clear empowered people enjoy a genuine epistemic advantage. On the other hand, but relatedly, as thinkers like Medina (2003) have emphasized (and which I noted in §3.1), empowered peoples are prone to cultivating certain kinds of ignorance about the world around them and the people in it.



understanding persons don’t pretend to be more than, or demand more than, sufficiency, the way we go about deciding what is sufficient nevertheless trades on already accepted social tropes. For there are always two parts in the process of understanding persons. First, there is the actual back and forth of the interpretive dialectic. Second, there is the judgment, if only implicit, about whether that dialectic succeeded. That is, there is a judgment about whether I “got” your perspective, or you “got” mine. And when it comes to this second, normative part, this part is mired by the kinds of real epistemic inequities I canvassed at the start. Mainstream, empowered groups tend to count as sufficient only what can be rendered in existing descriptive tropes, which is to say, the tropes they accept. And it is the reality of this epistemic injustice that informs the political reason against norming understanding to descriptive fullness. Were we to defend descriptive fullness, we would reify in theory what is already an in-​practice injustice, and thus add insult to injury. We are at last in a position to draw some general conclusions about understanding persons and the problem of power. First, the theory of holistic understanding turns the problem of understanding others around to place the focus on the one-​to-​be-​understood, for it turns out that the only satisfactory account of understanding others is one that is dialogic and dynamic, so as to allow in the voice of the other whose perspective it is we are trying to understand. To put this crucial point another way, the dynamic nature of personal perspective is reflected normatively as a condition of understanding persons: To genuinely understand another person’s perspective, the other person must in principle be partner to the process. And not just in the sense of being, by definition, the necessary object of understanding. Rather, the other person must be engaged in the understanding—​in our understanding of her. And while what I’ve said here doesn’t yet offer an argument for a duty, moral or political, to understanding—​that is, a duty to understand or to try to understand other persons (I make that case elsewhere)20—​it does require anyone who claims to understand other persons, to engage them dynamically. For on my theory such inclusion is what it means to understand other persons. In the case of oppression, then, whenever those in power claim to understand those without power, even implicitly, they commit themselves, normatively, to dynamic inclusion of those without power and can be blamed for failing to do so. Correspondingly, they commit themselves to resisting epistemic

20.   See my “The Authority of Empathy (or How to Ground Sentimentalism)” (forthcoming).

Understanding Persons and the Problem of Power


oppression. And they can be blamed for failing to do so. This should be obvious for testimonial injustice, which is nothing other than the direct violation of dynamic inclusion. But it is no less true of hermeneutic injustice, which, in virtue of creating “we” tropes normed to those in power, also, albeit more subtly, violates dynamic inclusion. But wait, wasn’t the whole point of the problem of power, to question whether it was even possible for those in power to understand those without it? If so, what does it matter to say that those in power are obliged (insofar as they claim to understand those without power) to include them? Isn’t this just an exercise in futility? No. That is the import of rejecting descriptive fullness as a condition of success for understanding persons. Norming holistic understanding to sufficiency reveals a kind of red herring in the problem of power—​and a socially risky one at that. We never “perfectly” understand anyone, nor should we try. Thus we should neither treat perfect understanding as a condition of solving the problem of power, nor allow its impossibility to be an excuse for what is really required, namely, serious, earnest, vigilant efforts by those in power to build ongoing dialogues with those without power. Moreover, as the first conclusion showed us, successful dialogue requires individual awareness of and active resistance against epistemic oppression. Finally, success will be measured in individual moments corresponding to the peculiarity of perspective and the depth of awareness of the relevant parties. In other words, what counts as “enough” understanding for this person here of that one over there, can’t be assumed to be enough for that person over there of this one here. Similarly, what counts as “enough” understanding today can’t be assumed to be “enough” tomorrow. In short, the problem of power isn’t ultimately about whether white people can totally get black people. Or whether straight people can totally get gay people. Or whether men can totally get women. The problem is whether this white, straight, man can get this black, gay, woman . . . enough. Or more exactly, whether he can come to a point in discussion about her perspective that she says, is good enough.

Acknowledgments I am grateful to the John Templeton Foundation for its generous support of my research. I’m also grateful to the many participants of the “Varieties of Understanding” conference, where I  first presented some of the core ideas of this chapter. I owe a debt to the Liberty Fund, which over the years has provided me the opportunity to read carefully and thoroughly key historical



contributors to this debate like Du Bois, Chesnutt, and Graffigny. I want to thank Stephen Grimm for his constant encouragement and helpful feedback on the penultimate draft. I also want to thank my Memphis colleagues Luvell Anderson, Stephan Blatti, Verena Erelenbusch, John Tienson, and Deborah Tollefsen for invaluable conversations on various aspects of this chapter. I dedicate this chapter to my parents, especially my stepmother, for cultivating in me an abiding passion for social justice and a sense of duty to remain vigilantly critical of my own power and privilege as a white male.

References Baker, Lynne Rudder. 2013. Naturalism and the First-​ Person Perspective. New York: Oxford University Press. Bell, Derrick. 1992. Faces at the Bottom of the Well. New York: Basic Books. Bernasconi, Robert. 2007. “On Needing Not to Know and Forgetting What One Never Knew: The Epistemology of Ignorance in Fanon’s Critique of Sartre.” In Race and Epistemologies of Ignorance. Ed. Shannon Sullivan and Nancy Tuana. Albany: State University of New York Press. 231–​39. Burroughs, Michael, and Deborah Tollefsen. 2016. “Learning to Listen:  Epistemic Injustice and the Child.” Episteme 13.3: 359–​77. Debes, Remy. Forthcoming. “The Authority of Empathy (or How to Ground Sentimentalism).” In Ethical Sentimentalism. Ed. Remy Debes and Karsten Stueber. New York: Cambridge University Press. Dotson, Kristie. 2014. “Conceptualizing Epistemic Oppression.” Social Epistemology 28: 115–​38. Du Bois, W. E. B. 1994. The Souls of Black Folk. New York: Dover. Fricker, Miranda. 1999. “Epistemic Oppression and Epistemic Privilege.” Canadian Journal of Philosophy, Supplementary, 29: 191–​210. —​—​—​. 2007. Epistemic Injustice: Power and the Ethics of Knowing. New York: Oxford University Press. Geertz, Clifford. 1974. “‘From the Native’s Point of View’:  On the Nature of Anthropological Understanding.” Bulletin of the American Academy of Arts and Sciences 28: 26–​45. Grimm, Stephen R. 2009. “Understanding.” In The Routledge Companion to Epistemology. Ed. Sven Berneker and Duncan Pritchard. New York: Routledge. 84–​94. Kim, Jaegwon. 2010. “Reasons and the First Person.” In Essays in the Metaphysics of Mind. Oxford: Oxford University Press. Langton, Rae. 2000. “Feminism in Epistemology: Exclusion and Objectification.” In The Cambridge Companion to Feminism in Philosophy. Ed. Miranda Fricker and Jennifer Hornsby. Cambridge: Cambridge University Press. 127–​45.

Understanding Persons and the Problem of Power


McIntosh, Peggy. 1988. “White Privilege and Male Privilege: A Personal Account of Coming to See Correspondences through Work in Women’s Studies.” Working Paper 189, Wellesley Centers for Women, Wellesley, MA. Medina, José. 2013. The Epistemology of Resistance:  Gender and Racial Oppression, Epistemic Injustice, and Resistant Imaginations. New York: Oxford University Press. Mills, Charles. 1997. The Racial Contract. Ithaca, NY: Cornell University Press. —​—​—​. 2007. “White Ignorance.” In Race and Epistemologies of Ignorance. Ed. Shannon Sullivan and Nancy Tuana. Albany: State University of New York Press. 11–​38. Stanley, Jason, and Timothy Williamson. 2001. “Knowing How.” Journal of Philosophy 98: 411–​44. Sullivan, Shannon. 2007. “White Ignorance and Colonial Oppression: Or Why I Know So Little about Puerto Rico.” In Race and Epistemologies of Ignorance. Ed. Shannon Sullivan and Nancy Tuana. Albany: State University of New York Press. 153–​72.


Religious Understanding, Naturalism, and Desire Fiona Ellis

4.1. Introduction In a 1969 paper on religious understanding David Burrell claims that, in reflecting on the role understanding might play in matters of religion, the image of a pilgrimage spontaneously comes to mind. What we have here, he continues, “is an archetypal religious image of life: a journey with a goal quite out of the ordinary, freighted with expectation, carrying the promise of transformation. Yet the narration concerns itself not with the goal but with the way there” (Burrell 1969, 677). These claims encapsulate the themes with which I shall be concerned and gesture toward some of the conclusions I wish to draw. The idea that we are concerned with a goal quite out of the ordinary calls to mind the complaints of the naturalist when he takes issue with the “spooky” ontology of his “supernaturalist” opponent, insisting that we oppose “all forms of thought which assert the existence of a supernatural or transcendental Realm of Being and make knowledge of that realm of fundamental importance to human living.”1 We are to suppose that the envisaged goal is ruled out hereby, assuming that it involves reference to a transcendental Realm of Being, and that such a realm is forbidden. We shall see that the issues are rather more complex than the typical naturalist line suggests, and that there is scope for

1.   Randall (1944, 358). The naturalistic themes which pervade our current times are continuous with those with which this volume is concerned. See my (2014, ch. 1) for an overview.

Religious Understanding, Naturalism, and Desire


allowing that the “enchanted” (Burrell 1969, 678)  journey at issue can be defended on philosophical and, indeed, naturalistic, grounds. The notion of transformation—​personal transformation—​will be important to what follows, one question being what it could mean to describe religious understanding as transformative in this sense. Presumably we are concerned with a kind of moral or spiritual transformation, but how are these notions to be understood, and what is their relation to the religious, and, indeed, to a person’s religious understanding? It is in the context of raising this latter question that the relation between understanding and theory will come to the fore. Burrell’s image of a pilgrimage lends emphasis to the journey rather than its goal. This might be thought to sit awkwardly with talk of the goal’s extraordinary nature and our expectations in this regard, but it has important implications for how we think about understanding in this context, and there is a question about how it is motivated and sustained. My point of reference here will be Levinas’s claim that it is fueled by desire, and I end this chapter by considering how a position along these lines might contribute to the model of religious understanding it is my purpose to defend.

4.2. Naturalism and the Extraordinary Philosophers have declared the truth of naturalism,2 and they vastly outnumber those for whom the position is “nothing short of a philosophical disaster.”3 It has become clear, however, that the nature of this thing we should be endorsing is either entirely obscure or so obviously contestable that the received wisdom is immediately undermined. According to one version of the contestable picture, we should all be reductive materialists or materialist naturalists. As Thomas Nagel puts it, “among the scientists and philosophers who do express views about the natural order as a whole, reductive materialism is widely assumed to be the only serious possibility” (Nagel 2012, 13). Reductive materialism admits of various characterizations, ranging from David Armstrong’s “the world contains nothing but the entities recognised by physics” (Armstrong 1980, 156) to John Dupré’s more recent “if there is only material stuff in the world (no spooky stuff ), then the properties of stuff

2.   Thus, we are told in a recent collection that “[n]‌aturalism has become a slogan in the name of which the vast majority of work in analytic philosophy is pursued (de Caro and Macarthur 2004, 2). 3.   Richard Bernstein gives voice to this complaint in his (1995, 58).



must ultimately explain everything” (Dupré in Nagel 2012). As per my original paper, the reference is to Dupre’s NDPR review of Nagel. The strong reductionist takes the measure of nature to be physics. A more moderate naturalism defines this measure with reference to a broader conception of science (Why just physics? How on earth could that explain everything? And what reason could be given for insisting upon this restriction?); and an even more moderate position challenges the assumption that the offending restriction can be lifted only in terms which are themselves restricted by science (Why just science? How on earth could that explain everything? And what reason could be given for insisting upon this restriction?). It is in the context of giving expression to these latter complaints that we find John McDowell recommending that we “discourag[e]‌this dazzlement by science” which leads us to suppose that “genuine truth is restricted to what can be validated by their [sciences’] methods” (McDowell 2002, 295). It should go without saying that this is not a rejection of science. McDowell’s naturalism breaks out of the offending scientistic strictures, and it promises to accommodate the idea that there are values in the natural world which make normative demands upon us. It is a one-​world position which eschews reference to a second, supernatural, realm of being, this being one obvious reason for retaining the “naturalist” label. As James Griffin has put it in the context of articulating a similar “expansive naturalist” outlook, “[v]‌alues do not need any world except the ordinary world around us . . . . An other-​worldly realm of values just produces unnecessary problems about what it could possibly be and how we could learn about it” (Griffin 1996, 44). We can note Griffin’s description of the natural world around us as “ordinary,” “ordinary” in this context being the logical complement of “supernatural” in the offending otherworldly sense. We might also compare this ordinary world with the “larger” universe of Robert Solomon’s “naturalised spirituality” (Solomon 2002), which, we are told, involves “the beauty and majesty of everything from mountain ranges, deserts, and rain forests to the exquisite details in the design of an ordinary mosquito . . . our grandest passions, love in particular  .  .  .  our sense of humanity and camaraderie” (Solomon 2002, xv–​xvi). Quite extraordinary in one clear-​enough sense, albeit with no hint of a second, supernatural realm. As Solomon puts it, it is “this universe, this world, this nature, not supernature” (Solomon 2002, xv). Talk of nature versus supernature brings a theistic dimension to the discussion, and we are to suppose that the expansive naturalist’s position must be shorn of any reference to God. After all, it involves rejecting a second, supernatural, realm, and God is not reducible to nature. McDowell himself

Religious Understanding, Naturalism, and Desire


stops short of proclaiming the truth of atheism, using the imagery of darkness to refer to that which exceeds the limits of his more relaxed conception of nature. It is here that we are returned to his own preferred conception of the meaning of ‘natural’, namely, “not supernatural (not occult, not magical),” and he adds: There is no need for me to take a stand on whether everything is natural in that sense (thereby, among other things, giving needless offence to people who think respect for modern science is compatible with a kind of religious belief that preserves room for the supernatural). (McDowell 2008, 218) The imagery of darkness would suit the apophatic theologian, but McDowell is no such thing—​intentionally at least—​and seems to be of the mind that this mysterious dimension, such as it is, could have no bearing upon nature and our natural human being. It is, after all, “occult” and “magical,” and it calls to mind the “rampant platonism” he criticizes elsewhere, with its implication that our lives are “mysteriously split, somehow taking place both in nature and in some alien realm—​in ‘Plato’s Heaven’ perhaps.”4 McDowell’s criticisms of the supernatural are like those of the scientific naturalist when he objects to the values which form part of McDowell’s ontology, the implication being that they too are “occult” and “magical.” Thus, we find Peter Railton complaining of the “worrisome ontological expansion” which ensues if we exceed his own preferred social-​scientific terms, to which he adds that we can explain everything that needs to be explained about value without making this problematic move—​the relevant entities just produce unnecessary problems about what they could possibly be and how we could learn about them. The refrain is getting familiar.5

4.   McDowell (1998, 77). Rampant Platonism is said to “picture human beings as partly in nature and partly outside it” (McDowell 1994, 77) and has a difficulty in explaining how we can latch onto, e.g., the requirements of morality by virtue of being natural beings. We are forced to conclude that we ourselves are partly supernatural: we have a “foothold in the animal kingdom and a mysterious separate involvement in an extra-​natural world” (McDowell 1994, 78). 5.   Witness Railton:  “[i]‌f for example, moral facts are identified with—​or otherwise reducible to—​ natural facts, then there is no special mystery about what sort of thing they are, or how we come to have knowledge of them, refer to them, and so on . . . . If moral facts are identical with—​or otherwise reducible to—​natural facts, then cognitivism may be possible without worrisome ontological expansion” (Railton 1993, 280). This is all said in the context of an extended discussion with David Wiggins, whose position counts as an expansive form of naturalism in the sense at issue here. Much of ch. 2 of my



What does this have to do with God? Well, the expansive naturalist’s conception of “religious reality” (to coin a phrase) turns it into something occult and alien—​i.e., supernatural/​extraordinary in the pejorative sense. But what if this conception can be challenged in the way that he challenges the scientific naturalist’s understanding of his own preferred conception of value? The expansive naturalist could try to block this parallel by objecting that the two cases are completely different—​God is “occult” and “alien” in a way that value is not. Yet this response simply begs the question against the possibility of an alternative framework—​one which challenges the assumption that God must be viewed in these pejorative terms and hence, that nature must exclude him. Does this not lead to a form of pantheism? After all, the naturalist is committed to a one-​world position, and I seem to be suggesting that God can be accommodated within such a framework. The theist is no pantheist, but he denies that God sits alongside nature to create the unnecessary problems to which Griffin refers—​as if it is a matter of adding an irrelevant and alien realm to the beauty and majesty of the ordinary world.6 Rather, it is this world which is God-​involving, for he is actively present in all things. So God is rescued from “some alien realm,” but he is not reducible to the world in which he is present, for he is its source and sustainer, and, as such, to be distinguished from anything within it, or indeed, beyond it.7 The theist claims further that we can enter into loving fellowship with this actively present God, that our receptivity in this context brings personal transformation, and that this transformation is expressed most significantly at the level of morality. Indeed, Levinas goes so far as to suggest God, Value, and Nature is devoted to a not unsympathetic consideration of Railton’s position and the question of how their disagreement might be adjudicated. 6.   It is the assumption that this model of God is mandatory, which makes a pantheist response seem so tempting in the first place. Thus Karl Rahner: “That God really does not exist who operates and functions as an individual existent alongside other existents, and who would thus as it were be a member of the larger household of all reality. Anyone in search of such a God is searching for a false God. Both atheism and a more naïve form of theism labour under the same false notion of God, only the former denies it while the latter believes it can make sense of it” (1978, 63). Brad S. Gregory offers a riveting account of where the offending line of thought goes wrong in his paper “No Room for God?” (2008, 495–​519). 7.   Compare Herbert McCabe:  “If God is whatever answers our question, how come everything? Then evidently he is not to be included amongst everything. God cannot be a thing, an existent among others. It is not possible that God and the universe should add up to two” (“Creation,” in 1987, 6).

Religious Understanding, Naturalism, and Desire


that being moral is the only way of relating authentically to God, anxious as he is to deflate the cognitive pretensions of those for whom God is a mere object of theory.8 Levinas offers a salutary reminder of our cognitive limitations in this context, and it will become clear that I am not suggesting that there is an easy route to theism, nor that the difficulties it confronts are not genuine. My point at this stage is simply to argue that the typical naturalist’s conception of these difficulties can be challenged, and that once we move beyond the unduly restrictive parameters of scientism, there is scope for questioning the assumption that an expansion in the direction of God is bound to be philosophically disastrous. If the expansive naturalist has shown us anything at all, it is that the limits of nature are entirely unclear. I began this section with the naturalist’s objection to the very idea of there being an authentic form of religious understanding, the complaint being that its object is too extraordinary to be part of our ontology. The scientism which underlies some versions of this complaint can be challenged, but there are “expansive” conceptions of naturalism which transcend this limitation to bring a richer world into view—​one which involves beauty, goodness, and love, for example. God seems to be another matter, and expansive naturalists like McDowell consign him to the “region of darkness,” denying that this realm could have any bearing upon nature or our natural human being. So there is continued allegiance to the idea that reference to God would have to be extraordinary in the pejorative sense, and this, of course, is what the scientific naturalist says of the expansive naturalist’s conception of the moral. The expansive naturalist quells this latter charge with an accusation of scientism, but we can question his own analogously pejorative conception of God. After all, there are good theological reasons for denying that God and nature are to be opposed in the manner he assumes, although the darkness imagery has a point. This imagery is embraced by Levinas, and the emphasis he places on morality (rather than theory) has important implications for what it means for understanding to be religious as opposed to, say, moral. It also raises the issue of how we are to interpret the force of the term ‘understanding’ in this context. It is to these questions that I now turn.

8.   Hence we are told that to know God is to know what is to be done (Levinas 1990, 17), and that “there can be no ‘knowledge’ of God separated from the relationship with men” (Levinas 1969, 78).



4.3. Cottingham on Religious Understanding In his paper “Transcending Science:  Humane Models of Religious Understanding,” John Cottingham distinguishes between two ways of comprehending religious understanding, such understanding referring to a “certain mode or manner of understanding the world.”9 More specifically, we are concerned with what it is to relate to the world religiously or what it is to understand things in a religious way (Cottingham 2015, 4). The first, disputed, position is that being religious involves “subscribing to an explanatory hypothesis about the origins of the cosmos and our human nature” (4), so that “understanding the world religiously is . . . an attempt to dissect and analyse and explain it in the manner of modern science” (5). The second position corresponds to Cottingham’s preferred new model of religious understanding, and it involves taking seriously the idea that “understanding the world religiously is not an attempt to dissect and analyse it and explain it in the manner of modern science . . . but rather a mode of engagement with, or connection with, reality as a whole” (5). This mode of engagement “cannot be achieved by the critical scrutiny of the intellect alone, but requires a process of attunement . . . a moral and spiritual opening of the self to the presence of the divine” (5). For Cottingham then, a nonscientific approach to religious understanding is a matter of denying that the understanding it involves—​of God and/​ or nature—​is scientific, and this is just another way of saying that religious understanding is not the same as scientific understanding, and that it must be explained on its own terms. Explaining it on its own terms is said to involve viewing it as a mode of engagement with, or connection with, reality as a whole—​one in which the self becomes morally and spiritually open to the presence of the divine as she embarks upon a path of moral and spiritual change (Cottingham 2015, 5–​6). The idea is familiar from Burrell’s talk of a journey which carries the promise of personal transformation, and Cottingham notes that this approach takes us away from classical theorizing about religion and closer to traditional religious thought and practice. It involves the kind of awareness which enables one to see the world transfigured, so that it is irradiated with meaning and value, and the human 9.   This paper was written as a contribution to the New Models of Religious Understanding project, a subproject of Varieties of Understanding (Cottingham 2015).

Religious Understanding, Naturalism, and Desire


subject, caught up in that mystery, is unmistakably called on not to be any longer a spectator, a mere “tourist,” but to respond, to be a morally responsive agent, part of a cosmos that is diaphanous, transparent to the divine. (Cottingham 2015, 7) Cottingham raises a potential objection to such an approach, namely, that even if we grant that religious understanding involves the kind of awareness described, we still need classical natural theology if we are to “characterize the object of such awareness and the content of such belief,” and “give a proper theoretical account of the nature of the divine” (7). It is quickly made clear that these ambitions are futile. They are futile because the object of theological enquiry “stands outside our human frame of meaning and reference, the properties of the divine qua divine [being] in a radically different category from any of the properties that characterise the empirical world” (8). However, these limitations are not fatal, for we can relate to God at a nontheoretical level, this involving religious understanding in Cottingham’s preferred sense. God is not “grasped” at this level of understanding, but there are “traces” of his presence to be discerned “in the beauty of the natural world and the compelling power of our moral sensibilities” (10). Cottingham ends with a supposed paradox, namely, “that our understanding of existence can be transformed by using a framework of interpretation whose structure does not function like an explanatory theory” (Cottingham 2015, 10). It is left unclear why this should be paradoxical, but the point seems to be that our understanding is transformed by an interpretative framework whose object cannot be pinned down in the manner demanded by explanatory theism. We are told that the paradox is solved not by further theorizing but by action, for religious understanding is “inseparable from moral action and spiritual practice,” and “cannot come about by abstract theorizing” (10).

4.4. Some Questions We have been given two conceptions of religious understanding, one theoretical, the other practical. The theoretical mode is said to correspond to that of the natural theologian or explanatory theist and it involves approaching God in the manner of a scientist. Cottingham objects that, on such an approach, we end up relating to something other than God—​a “mere idol of our own construction,” as he puts it (Cottingham 2015, 10). The practical mode captures the understanding of the ordinary believer, it is inseparable from “moral action and spiritual practice,” and it is at this level that we are said to relate to



God. Cottingham doubts that these modes can coexist, and implies that theory must be rejected—​understandably so, given that its object “stands outside our human frame of meaning and reference,” and given that “the properties of the divine qua divine are in a radically different category from any of the properties that characterize the empirical world.” We could be forgiven for thinking that Cottingham has reverted to the framework it is my own purpose to transcend, namely, one according to which God is consigned to an “alien realm” which stands opposed to anything within the natural world. But we know, of course, that this is not his position, for he insists that there are traces of the divine to be found within the natural world, and that we relate to them when we discern nature’s beauty and feel the pull of morality. This much is compatible with the claim that the properties of the divine qua divine are in a radically different category from those that characterize nature, for it could be a way of capturing the idea that God cannot be known in himself, and that we can know him only in so far as he is revealed in the world—​revealed at the level of the relevant “traces,” if Cottingham is to be believed. Yet this does not preclude the possibility of these traces being discernable to theory, and Cottingham seems to concede such a possibility, for he tells us that God “can never be brought fully within the grasp of the human mind.”10 So it is not ruled out that theory and practice are on a par in this context, nor, indeed, that God remains utterly inscrutable either way. After all, it is a familiar from Solomon—​and many others—​that morality, beauty, and spirituality have nothing to do with God.11 Cottingham exhorts us to relinquish “the epistemology of control” for “an epistemology of receptivity,” lest we “clos[e]‌ourselves off from allowing the evidence to become manifest to us.” Hence, we are told that “[t]he question is not ‘Can I, while scrutinizing the data and remaining detached and fully in control, satisfy myself of the rational acceptability of belief in God?’ ”; but rather . . . ‘How can I embark on a path of moral and spiritual change which might open me to a deeper awareness of something that I now glimpse only faintly.’ We must be ‘open, yielding, receptive listeners’ rather than ‘detached, critical evaluators’ ” (Cottingham 2015, 6). We can agree that an epistemology of receptivity is required if this is just another way of saying that religious understanding must involve being open to God.12 However, this does

10.   My italics. Compare Descartes: we can touch God but not comprehend him. 11.   See, for example, Comte-​Sponville (2007), and Harris (2014). 12.   The scientist must, of course, be analogously receptive to his subject matter.

Religious Understanding, Naturalism, and Desire


not rule out the possibility of being theoretically receptive, unless there are reasons for thinking that there is something about theory per se which makes it inhospitable to God. It seems plausible to suppose also that we must operate as detached, critical evaluators in such a context. At least, this is so if being detached and critical is a way of protecting against distortion and prejudice—​ being detached in this sense is not the same as being detached from one’s subject matter. Cottingham has the latter scenario in mind when he talks of being closed off from allowing the evidence to become manifest to us, and we can relate this point to something Levinas says in the context of criticizing what he elsewhere describes as the empty “theological arithmetic” (Levinas 1990, 274)  which reduces God to something that can be adequately grasped (and controlled) in thought: We have been reproached for ignoring theology; and we do not contest the necessity of a recovery, at least, the necessity of choosing an opportunity for a recovery of these themes. We think, however, that theological recuperation comes after the glimpse of holiness, which is primary. Levinas 1998a, ix) The glimpse of holiness seems primary for Cottingham, and some of what he says suggests that it is inimical to theory. It will be remembered, however, that his target is a particular theoretical approach to God, and although it is not spelled out in any detail, it is said to involve subscribing to an explanatory hypothesis about the origin of nature, giving a proper theoretical account of the nature of the divine, satisfying oneself of the rational acceptability of belief in God, and, in all of this, approaching the world in the manner of the analyzing and controlling modern scientist. We must consider then what this could really mean.

4.5. Explanatory Theism and Natural Theology Here’s one way of thinking about the kind of approach under dispute which is implicit in some of what Cottingham says and explicit in the response of his scientist.13 Explanatory theism involves putting forward a theory or hypothesis in terms of which to explain some observational data—​the natural world

13.   I spell out further details and assumptions of this approach in my “God and Other Minds” (2010, 331–​51).



or some aspect of it—​where it is not ruled out that the data is better explained in some other, nontheistic, way. (Compare the way in which theories within science are treated in a similarly provisional manner.) It follows that the relevant data must be characterized in terms which are neutral with respect to the theory’s truth—​which means, in the case at hand, that the natural world must be characterized in terms which do not make reference to God, the God hypothesis being just one possible way of explaining it. Add to this that theism has been superseded and that it has been superseded by science—​the scientist has no need of that hypothesis—​and it follows that atheism must be our starting point, that theism is for the scientifically ignorant, and that religious understanding in Cottingham’s sense must be explained away. We are returned to the standpoint of the scientific naturalist, and it likewise encapsulates the attitude of Michael J. Buckley’s natural theologian as spelled out in his study of modern atheism (Buckley 1987). Such a figure tackles the question of God in abstraction “from any common religious tradition and the experiences it involves” (Buckley 1987, 348), i.e., he ignores religious understanding in Cottingham’s sense, and he is said to be informed about God “from the outside,” i.e., from outside the perspective of one who so understands. We are told also that his approach turns the question of God into a purely philosophical question (Buckley (1987, 348). More specifically, it involves “apply[ing] to the philosophers for philosophic information,”14 this information is to be found in the natural world, and it constitutes the putative evidence for God’s existence. The philosophers in question take the natural evidence to fall short, insisting that there are good philosophical reasons for insisting upon this negative conclusion. After all, nature can be adequately comprehended in non-​God-​involving terms, and a move in the direction of God does no more than to introduce a further and highly problematic realm of being which explains precisely nothing.15 The framework is familiar, it presupposes that nature contains no trace of God, and we are encouraged to suppose that this is the philosophically respectable position. Granted then that natural theology and/​or explanatory theism 14.   Buckley (1987, 342). The phrase comes from Étienne Gilson’s Elements of Christian Philosophy (1960, 33). 15.   This natural theological approach corresponds to what Paul Tillich calls “cosmological-​scientific” in his “The Two Types of Philosophy of Religion” (1946). As John E. Smith puts it: “the point of departure for the cosmological-​scientific type is . . . the world of limited things and processes as they are known both through ordinary experience and the precisely formulated knowledge of the natural sciences. This way of approach, often called the ‘way from Nature to God,’ begins the quest with a world of fact beyond the self, although this world is often said to include man as well” (Smith 1958, 929).

Religious Understanding, Naturalism, and Desire


are to be comprehended along these lines, we can agree with Cottingham that they can be of no assistance in helping us to comprehend the object of religious understanding in his preferred sense, which, of course, is equivalent to saying that they can tell us nothing about God, assuming that it is at this level of understanding that God is revealed. The problem with the theoretical approach under dispute is that it is rigged in favor of atheism. God becomes an explanatory posit whose credentials are undermined on scientific and philosophical grounds, these grounds dictating that the natural evidence for his existence falls short. It does indeed on the assumption that nature contains no trace of God, and if we accept this assumption, then religious understanding is ruled out in both of the senses with which Cottingham is concerned. However, this atheistic framework can be questioned, as indeed it is questioned by Cottingham when he claims that there are divine traces within nature to which we respond when we understand things religiously in his preferred sense. Yet he seems to agree with our natural theologian that these traces elude theory, and that there can be no theorizing about God. Hence the seemingly antitheoretical approach. This antipathy toward theory is on target as far as the disputed approach is concerned. However, if taken to rule out the possibility of theorizing about God in any sense at all, then we play into the hands of one who believes that religious understanding has nothing to do with how things really are. It also blocks the possibility of saying anything about what is at issue—​or what could be at issue—​when we respond to reality in the relevant religious sense.16 The secular expansive naturalist agrees with Cottingham that the natural world is ‘irradiated with meaning and value’, but struggles to see how this could have anything to do with God. This response threatens to consign God to the ‘alien realm’ which makes atheism so tempting in the first place, and although I  have given reasons for questioning such a framework—​good reasons if it is scientistically motivated—​it would be philosophically and theologically naïve to suppose that we have here an instant route to God. Indeed, it is in the nature of the case that there can be no such thing, lest God be pinned down and reduced to something within the world.17 Not for nothing the imagery of

16.   Compare Sarah Coakley, who, in her 2012 Gifford Lectures Sacrifice Regained:  Evolution, Cooperation, and God talks of the importance of reclaiming the notion of natural theology. She claims that it needs to go on and that without it “Christian theology and philosophy of religion [are] disturbingly empty-​handed, or at the best merely defensive and inward looking, in the face of mammoth incursions on its former territory by secular science and philosophy.” 17.   See Moser (2008) for some important thoughts in this context.



darkness and the idea that we are pilgrims in this context—​grappling toward something we can barely understand, caught up in mystery, yielding, and receptive. Burrell tells us that the narration here concerns itself not with the goal but with the way there. This rightly discourages a crude means-​end conception of the relation between journey and goal, although it would be misguided—​and analogously problematic—​to suggest that the goal in question drops out as irrelevant. After all, the relevant journey has a telos, even if we acknowledge, as we must, that it is a telos in progress which involves the aforementioned darkness. Indeed, we might borrow here from Talbot Brewer, who, in the context of articulating his own preferred and more general conception of understanding, lends emphasis to the sense in which it involves reflective activity rather than being an achieved state—​something that we do rather than something that we have.18 We are told also that it “fuel[s]‌its own dialectical advance” in the “active pursuit of its own deepening” (Brewer 2011, 308), and that its real value “seems to lie in the running actualisation of that understanding itself, both in one’s own life and in the lives of others whose view of nature is deepened and extended by one’s inquiries” (Brewer 2011, 308). We are returned to the question of nature’s limits as our ever-​deepening understanding throws up a level of putative evidence which exceeds the parameters of Buckley’s natural theologian, namely, that which belongs to one who understands the world religiously in Cottingham’s sense. The secular expansive naturalist struggles to appreciate this level of understanding, for he has a difficulty with the idea that nature can be transparent to the Divine, and hence, that God is actively present in this sense. However, he has no worries about allowing that it is irradiated with meaning and value, and may even confess to being spiritually aware—​in a nontheistic sense, of course. We must clearly do more than simply reassert, in the face of his skepticism, that nature is God-​involving and that religious understanding is exactly what Cottingham says. Indeed, he has given us good reason for refusing this route 18.   Brewer (2011, 305). Hence, and in the context of describing the value of understanding, Brewer claims that we arrive at the correct interpretation of his position “only when we take seriously the gerundival form of the term ‘understanding.’ Just as the value of possessing hearing lies in this capacity’s actualization in hearing (for instance, in hearing music or in hearing another human being), so too the value of possessing understanding lies in its actualization in understanding or appreciating various things (for instance in understanding a book one is reading or appreciating a natural phenomenon to which one is attending). The final value of understanding comes to light only when we conceive of understanding not as an achieved state that can be attributed to people while they are mentally inactive, but as a complete or unimpeded actualization of that achieved state in active appreciation of the world in which our lives unfold” (2011, 306–​7).

Religious Understanding, Naturalism, and Desire


in the present context, namely, that the limits of nature cannot be taken for granted, and that we cannot assume that we are remotely clear about where these limits are to be drawn. It follows that the distinction between what is natural and what is supernatural is analogously problematic—​which, of course, can work to the natural theologian’s advantage if there is a case for saying that the world which is irradiated with meaning and value is already charged with the grandeur of God. The case for saying this in the present context is that we have had reason to question the assumption that God is an irrelevant cosmic excess, and to consider what it could really mean to be morally and spiritually aware, i.e., to understand religiously in Cottingham’s sense. All of this constitutes “philosophic information” which is relevant to the question of God and which can be appreciated by the nonbeliever provided that she approaches the issue with honesty, impartiality, and open-​mindedness.19 One is informed about God “from the outside” in this respect, but this is not the same as being closed off from God in the manner of Buckley’s natural theologian, for one is encouraged, though not required, to consider the theistic significance of the experiences and actions under dispute, one possible outcome being that things are seen in a different, and perhaps God-​ involving light. As Sarah Coakley puts it in the context of making a related point, something may be “newly and surprisingly revealed.”

4.6. Understanding and Desire Burrell’s image of a pilgrimage lends emphasis to the way rather than the goal. I have suggested that these are inextricably bound up together, but it is a point in favor of Burrell’s preferred emphasis that it discourages us from thinking that religious understanding can be “had” in the manner of a commodity—​ as if we are motivated in this context simply by the prospect of such possession. The relevant way is expressed at the level of moral action and spiritual practice, and it goes without saying that morality and spirituality resist being “had” in the offending sense.20 Our theoretical endeavors in this context have a similarly open-​ended trajectory as we struggle to comprehend what 19.   Hence Brewer’s “virtue epistemological” approach according to which “[e]‌xcellence in theoretical reflection requires courage, perseverance, carefulness, impartiality, unflinching honesty, and openness to the views and criticisms of others” (Brewer 2011). 20.   Compare Mark Johnston’s insightful account of Eve’s disobedience: “it was driven by her anxious hope that ‘wisdom,’ a correct conception of good and evil, and hence the knowledge of how to live, could be got from a tree, off the shelf as it were—​as if it were something fixed and complete like an ideal commodity, as if it were something that could be possessed by human beings” (2009, 168).



these notions really mean, our conclusions generating further questions and insights which themselves contribute to our ongoing understanding. This is what Brewer is getting at when he talks of understanding “fuel[ing]” its own dialectical advance in the “active pursuit of its own deepening.” Or as he puts it elsewhere, “the activity of understanding [is] drawn into an unending dialectical extension of itself ” (Brewer 2011, 313). Levinas takes this endless activity to be fueled by desire, his primary focus being our moral interactions with others. However, this desire is to be distinguished from the hunger which “calls for food,” having nothing to do with wanting and getting things for oneself.21 Rather, it tends interminably “towards something else entirely, toward the absolutely other” (Levinas 1969, 33), and it is insatiable in this respect. The naturalist’s “alien realm” springs to mind, but Levinas has no truck with “worlds behind the scenes,”22 making it clear that the object of this desire is something to which we are eminently capable of relating (Levinas 1969, 38). However, it is a mode of relating which is not an “aiming” (Levinas 1998b, 222), and its object is said to arouse and sustain the desire rather than to satisfy it: “desire is an aspiration that the Desirable animates; it originates from its ‘object’; it is revelation” (Levinas 1969, 62). The object of this desire is God, and we know that Levinas shares Cottingham’s concerns about the possibility of relating authentically in such a context. The turn to desire is intended to bypass the relevant difficulties, moving us beyond the idols to which we are confined on the offending approaches. First, given that the desire in question does not call for food, it is ruled out that God is at the service of egoistic consciousness—​something to be “aimed” at in this controlling sense in order to be grasped and “consumed.”23 Second, God’s resistance to such treatment explains the desire’s insatiability—​the subject cannot be “filled up” in this respect and this is why her desire does not cease. Third, however, this is not to be read as a limitation, for there is no call for food in this context, the desire in question “desir[ing]’ 21.   Thus we are told that the activity it involves is “of a different order than that of affectivity or the hedonic activity by which the desirable is invested, attained, and identified as an object of need” (Levinas 1998b, 221). 22.   See Levinas (1991, 8). In the final paragraph of this work Levinas tells us that it “does not seek to restore any ruined concept” and he refers with relief to “the death of a certain god inhabiting the world behind the scenes” (1991, 185). 23.   Compare: “The sky calls for a gaze other than that of a vision that is already an aiming and proceeds from need and to the pursuit of things. It calls for eyes purified of covetousness, a gaze other than that of the hunter with his ruse, awaiting the capture” (Levinas 1998b, 163).

Religious Understanding, Naturalism, and Desire


beyond anything that can simply complete it” (Levinas 1969, 34). The desiring subject’s approach to God “increases the distance” in this respect, guaranteeing that God’s transcendence is preserved. It serves equally, however, to secure the most intimate connection, for she is motivated to be moral in this context, and it is at this level of interaction that her desire for God is truly expressed (Levinas 1969, 78). The subject is said to be “brought to [her] final reality” in this manner (Levinas 1969, 178). Levinas sums up the position as follows: [T]‌he desirable (or God) must remain separated within desire: near, yet different—​which is, moreover, the very meaning of the word “holy.” . . . The desirable separates itself from the relation to desire it called forth, and, through this separation or this holiness, the desirable remains a third person, a He at the base of the Thou. He does not fill me with good, but compels me to goodness, better than the good to be received . . . . God is torn out of the objectivity of presence. . . . He is no longer an object or an interlocutor in a dialogue. His distancing or his transcendence turns into my responsibility. (1998b, 223–​24) Levinas has told us already that the glimpse of holiness is primary, the context there being the possibility of theological recuperation. Such a glimpse is crucial to what he is saying here, his aim being to clarify what it could mean for the dimension of the holy to be truly revealed. The idea that God must remain separated within desire is intended to guarantee his transcendence or difference: “there is neither vision nor aiming here,” God being “transcendent to the point of absence.” Rather, this desire-​involving movement—​which is itself a revelation—​is refracted back onto the person before whom I am morally responsible, this being the only way of guaranteeing that I relate truly to God. It is in this sense that the “referral to the other is an awakening to nearness” (Levinas 1998b, 223). We are reminded of Cottingham’s human subject who, caught up in God’s mystery, is “called on not to be any longer a spectator, a mere ‘tourist,’ ” but to respond, to be a morally responsive agent. Levinas avoids the suspect supernaturalism which locates God in a second, supernatural, world, agreeing with the naturalist that our focus must be upon this universe, this world, this nature. The idea that God is “transcendent to the point of absence” suggests a further concession to naturalism, for it implies that God has been squeezed out of the picture altogether. This impression is further corroborated with the emphasis Levinas places upon our status as moral agents, and it would be easy to conclude that the religious quest has



been reduced without remainder to the moral. It should be clear from what has been said, however, that Levinas’s conception of morality is irreducibly God-​involving, and that what he seeks to eliminate is not God himself but the various idols which stand in the way of genuine revelation. A paradigm perhaps of what it could mean for God to be actively present in things.24 Levinas offers a way forward for one who believes that suspect supernaturalism is a thinly disguised form of atheism. He veers close to atheism in one clear-​enough sense, but this atheism, such as it is, promises a more authentic relation to God, and the position as a whole suggests a way of defending a new and more satisfactory conception of religious understanding. It appeals in particular because the model at which we arrive involves a rejection of the very things which make religious frameworks seem so problematic and superfluous in the first place. Small wonder then that there is a knife-​edge between this position and that of the secular expansive naturalist, and I have noted already that there is the potential here for something to be newly and surprisingly revealed. A glimpse of holiness perhaps, and we know from before that this is the way to theological recuperation as far as Levinas is concerned.

4.7. Conclusions I began this chapter with Burrell’s image of a pilgrimage with a “goal quite out of the ordinary, freighted with expectation, carrying the promise of transformation,” this being his way of capturing what is at issue when the notion of religious understanding is introduced. I  have challenged the naturalist’s objection to the very idea of there being a journey in this sense, granted with Cottingham that the transformation at issue here is moral and spiritual, and offered these thoughts and more to illustrate what it could mean for understanding in this context to assume a more theoretical shape. There are ways of theorizing about God which seem doomed to miss the target, when, for example, it is assumed that nature and God must be held in permanent opposition. However, theory is not bound to take this problematic form, and we relinquish it at the cost of being unable to question the offending assumptions and misunderstandings, including, of course, those to which we ourselves are prone.

24.   Compare Simone Weil:  “[t]‌he real aim is not to see God in all things; it is that God through us should see the things we see. God has got to be on the side of the subject and not on that of the object during all those intervals of time, when, forsaking the contemplation of the light, we imitate the descending movement of God so as to turn ourselves towards the world” (1956, 358).

Religious Understanding, Naturalism, and Desire


Further questions arose concerning the “fuel” of this transformative journey, Levinas’s response being that it is motivated by desire. I laid out the merits of his position, and concluded that it offers the shape for a model of religious understanding which could genuinely appeal to the secular expansive naturalist. It is such a figure, after all, who rightly insists that supernaturalism and reductive naturalism are two sides of the same problematic coin, and that we do not need any world except the ordinary world around us.

Acknowledgments This work was supported by funding from the John Templeton Foundation as part of the Varieties of Understanding project. Many thanks to Stephen Grimm and his team for selecting and supporting the Heythrop College subproject New Models of Religious Understanding. I benefited greatly from discussion with John Cottingham, Anna Abram, Marilynne Robinson, Michael Barnes, Michael Inwood, and Sarah Coakley.

References Armstrong, D. M. 1980. “Naturalism, Materialism, and First Philosophy.” In The Nature of Mind and Other Essays. Ed. D. M. Armstrong. St Lucia: University of Queensland Press. 149–​65 Bernstein, Richard. 1995. “Whatever Happened to Naturalism?” Proceedings and Addresses of the American Philosophical Association 69: 57–​76. Brewer, Talbot. 2011. The Retrieval of Ethics. Oxford: Oxford University Press. Buckley, Michael J. 1987. At the Origins of Modern Atheism. New Haven:  Yale University Press. Burrell, David. 1969. “Religious Life and Understanding” Review of Metaphysics 22: 676–​99. Coakley, Sarah. 2012. “Reconceiving ‘Natural Theology’: Meaning, Sacrifice and God.” Gifford Lectures: Sacrifice Regained: Evolution, Cooperation, and God. http://​www. faith-​​2012/​05/​sarah-​coakley-​2012-​gifford-​lectures.html. Comte-​Sponville, André. 2007. The Book of Atheist Spirituality. Trans. Nancy Huston. London: Bantam Press. Cottingham, John. 2015. “Transcending Science:  Humane Models of Religious Understanding.” Paper for New Models of Religious Understanding, part of Varieties of Understanding. De Caro, Mario, and David Macarthur. 2004. “Introduction: The Nature of Naturalism.” In Naturalism in Question. Ed. Mario De Caro and David Macarthur. Cambridge, MA: Harvard University Press. 1–​17



Dupré, John. 2012. Notre Dame Philosophical Reviews. Thomas Nagel, Mind and Cosmos:  Why the Materialist Neo-​Darwinian Conception of Nature Is Almost Certainly False. October 29. Ellis, Fiona. 2010 “God and Other Minds.” Religious Studies 46: 331–​51. —​—​—​. 2014. God, Value, and Nature. Oxford: Oxford University Press. —​—​—​. 2018. New Models of Religious Understanding. Oxford: Oxford University Press. Gilson, Étienne. 1960. Elements of Christian Philosophy. New York: Doubleday. Gregory, Brad S. 2008. “No Room for God? History, Science, Metaphysics, and the Study of Religion.” History and Theory 47: 495–​519. Griffin, James. 1996. Value Judgement: Improving our Ethical Beliefs. Oxford: Oxford University Press. Harris, Sam. 2014. Waking Up:  Searching for Spirituality without Religion. London: Bantam Press. Johnston, Mark. 2009. Saving God: Religion after Idolatry. Princeton, NJ: Princeton University Press. Levinas, Emmanuel. 1969. Totality and Infinity. Trans. Alphonso Lingis. Pittsburgh: Duquesne University Press. —​—​—​. 1990. Difficult Freedom: Essays on Judaism. Trans. Seán Hand. Baltimore: John Hopkins University Press. —​—​—​. 1991. Otherwise Than Being or Beyond Essence. Trans. Alphonso Lingis. Dordrecht: Kluwer. —​—​—​. 1998a. Of God Who Comes to Mind. Trans. Bettina Bergo. Stanford, CA: Stanford University Press. —​—​—​. 1998b. “A God ‘Transcendent to the Point of Absence.’” In God, Death, and Time. Trans. Bettina Bergo. Stanford, CA: Stanford University Press. 219–​24 McCabe, Herbert. 1987. “Creation.” In God Matters. London: Continuum Press. McDowell, John. 1994. Mind and World. Cambridge, MA: Harvard University Press. —​—​—​. 1998. “Two Sorts of Naturalism.” In Mind, Value, and Reality. Ed. John McDowell. Cambridge, MA: Harvard University Press. 167–​97 —​—​—​. 2002. “Response to Charles Lamore.” In Reading McDowell on Mind and World. Ed. Nicholas H. Smith. London: Routledge. 294–​96 —​—​—​. 2008. “Reply to Fink.” In John McDowell: Experience, Norm, and Nature. Ed. Jakob Lindgaard. Oxford: Blackwell. 214–​19 Moser, Paul. 2008. The Elusive God:  Reorienting Religious Epistemology. New York: Cambridge University Press. Nagel, Thomas. 2012. Mind and Cosmos:  Why the Materialist Neo-​ Darwinian Conception of Nature Is Almost Certainly False. New York: Oxford University Press. Rahner, Karl. 1978. Foundations of Christian Faith:  An Introduction to the Idea of Christianity. Trans. William V. Dych. London: Darton, Longman and Todd. Railton, Peter. 1993. “What the Non-​cognitive Helps Us to See the Naturalist Must Help Us to Explain.” In Reality, Representation, and Projection. Ed. John Haldane and Crispin Wright. Oxford: Oxford University Press. 279–​97.

Religious Understanding, Naturalism, and Desire


Randall, John Herman, Jr. 1944. “Epilogue: The Nature of Naturalism.” In Naturalism and the Human Spirit. Ed. Yervant H. Krikorian. New York: Columbia University Press. 354–​82. Smith, John E. 1958. “The Present Status of Natural Theology.” Journal of Philosophy 55: 925–​36. Solomon, Robert. 2002. Spirituality for the Skeptic:  The Thoughtful Love of Life. New York: Oxford University Press. Tillich, Paul. 1946. “The Two Types of Philosophy of Religion.” Union Seminary Quarterly Review 1: 3–​13. Weil, Simone. 1956. The Notebooks of Simone Weil. Vol. 2. Trans. Arthur Wills. London: Routledge & Kegan Paul.


Philosophy, Knowledge, and Understanding Gordon Graham

5.1. Progress in Philosophy What is the aim of philosophy as an intellectual activity? Can it achieve its end? Does it ever do so? The purpose of this chapter is to offer answers to these questions. But it is necessary to begin by saying something about why they arise. After more than two thousand years of thought and argument, it appears that none of the central issues in philosophy has ever been resolved. Philosophical disagreement about, for instance, the relation of mind to reality, the role of freedom and determinism in human action, the rational basis of ethical judgment, is interminable, and insofar as there is any advance in the exploration of these issues, its effect is always to increase rather than diminish their problematic character. In short, philosophical problems are perennial. In this respect, philosophy contrasts sharply with the natural and social sciences. Leaving aside occasional anxieties about “whiggism” in the history of science.1 it is widely agreed that, at least since the seventeenth century, physics and biology have progressed enormously, and produced scientific explanations of natural phenomena that are far more satisfactory than anything known to the ancient or medieval worlds. Such progress, it is true, is much less widely conceded with respect to the social sciences. Yet here too there is an important contrast to be drawn with philosophy, because there has clearly

1.   Herbert Butterfield’s work The Whig Interpretation of History (1931) had a huge influence on the writing of political history, but also raised anxiety among some historians of science. See Weinberg (2015).

Philosophy, Knowledge, and Understanding


been methodological progress. The gathering and processing of social, economic, and psychological data is hugely advanced on what it was, and these shared and successful methods make collaborative research (and research assistance) possible within the social sciences in a way that has no counterpart in philosophy. The absence of progress, and the contrast with other disciplines, raises a doubt about philosophy as an intellectual activity, but this doubt is not new. In the late nineteenth century, Henry Jones, Professor of Moral Philosophy at St Andrews, published an article in Mind that gave eloquent expression to the very same thought. At the present moment as philosophers would themselves acknowledge, there is no theory that either obtains or deserves unquestioning confidence.  .  .  .  Unprejudiced observers  .  .  . who contrast the long catalogue of defeats sustained by the philosophers and the shattered condition of their ranks today, with the solid and advancing conquests of the natural sciences, have very naturally concluded that philosophy is seeking by a doubtful method an unattainable goal. ( Jones 1893, 160) “The sad picture which Hume drew of philosophy in his day,” he goes on to say, “represents with much faithfulness its condition in our own,” and he quotes Hume in the introduction to the Treatise: Principles taken upon trust, consequences lamely deduced from them, want of coherence in the parts, and of evidence in the whole—​these are everywhere to be met with in the systems of the most eminent philosophers, and seem to have drawn disgrace upon philosophy itself. Nor is there required such profound knowledge to discover the present imperfect condition of the sciences, but even the rabble without doors may judge from the noise and clamour which they hear, that all goes not well within. (Hume [1739] 1978, xviii). Jones compares “the present moment” at which he was writing, to the moment at which Hume was writing, 150 years previously. As far as the general point goes, though, “the present moment” could just as easily be the twenty-​ first century. A somewhat dispiriting conclusion thus seems inescapable; the condition of philosophy as an intellectual discipline has not advanced much, if at all, in over 250 years.



Actually, 250 years is an arbitrary number in this respect. Jones refers to Hume because Hume’s Treatise set itself to pursue philosophical questions in a new style and manner that, or so Hume believed, would make real progress. A  century and a half later when Jones was writing, this ambition was still alive, at least to some extent. That is why he reflects on the period that has intervened. But it is quite possible to extend the same reflection to the whole history of philosophy and construe its effort, going all the way back to Plato possibly, as being no less inconsequential as far as definite “results” are concerned. Even if this is indeed the case, however, we ought not to infer from it that over this long period, things have stood still. There have been obvious changes and developments in philosophy. Very evidently, contemporary philosophers no longer pursue their subject in the same way that Plato did. Still, it is neither false nor unintelligible for some of them to be regarded (or regard themselves) as Platonists, just as the Cambridge Platonists were in the seventeenth century, and the Neoplatonists in the fourth century. Such identifications are only possible, though, if the differences between Plato’s time and these other periods, however evident, are not fundamental. By contrast, over the same period change in, for instance, military technology has rendered the trireme quite useless from the point of view of modern naval engagement. Between Hume’s time and ours, there is one important difference nonetheless. In the eighteenth century, “science” and “philosophy” were not distinguished as they commonly are today, and indeed it was the rise of just this distinction in the century after Hume that generated the intense concern of Jones and others with the state of philosophy. To him, as to many philosophers, it seemed a matter of shame and embarrassment that their chosen subject was treading water while the natural and social sciences were moving steadily ahead. One consequence was the altered character of the “rabble without.” In Hume’s world this could be taken refer chiefly to those who knew very little about academic inquiry, but by the late nineteenth century those outside philosophy and looking in, with disdain, included natural and social scientists of great intellectual sophistication and considerable accomplishment. From their perspective, it was impossible not to wonder at the endlessness of debates in which philosophers engage, to question whether, since they never seemed to get anywhere, they had any point, and to suggest that, truth be told, nothing would be lost if they were abandoned. Jones, then, had real cause for concern and despite the passage of a further century and more since his article was published, natural and social scientists can plausibly entertain the same thoughts today.

Philosophy, Knowledge, and Understanding


5.2. Philosophy’s Inconclusiveness Unwelcome though it may be to philosophers, skepticism about the value of philosophy rests on what appears to be a salient and solid matter of fact. The sciences, both natural and social, are indeed marked by progressive development. Carefully designed, repeatable experiments and/​or statistical information rigorously gathered and processed have played an important part in building a common body of knowledge to which competing explanations must answer. By this means, some of the competitors are rightly rejected, and over time they are discarded as live explanatory options. Hypotheses that survive this test for a period, provide important stepping stones in the formulation of explanatory theories that are ever more comprehensive in their scope. It may be true, as Thomas Kuhn famously argued in The Structure of Scientific Revolutions, that in the most striking cases scientific progress is neither smooth nor uniform. An accumulating burden of awkward observations initially prompts amendments and adjustments that are then incorporated into widely accepted theoretical frameworks. This process of adjustment, though, can go only so far. When a kind of evidential and explanatory “breaking point” is reached, even the best-​established theories are finally abandoned. The point to emphasize, however, is that if the Kuhnian picture is indeed a more accurate account of major change in scientific understanding, it is no less a mark of progress. Whether it happens little by little within “normal science” or in a grand revolutionary overthrow, scientific theories do get refuted. Consequently, even the most venerable and widely accepted of them can in time become matters of interest only to historians of ideas. However great their historical significance was and remains, they are no longer viable options for the practitioners within whose scientific discipline they once claimed ascendency. As the recurrence of “Platonism” alluded to earlier shows, the situation with respect to philosophy is quite different. Though contemporary philosophers do sometime speak of “results,” and refer to philosophical contentions as “theories,” the fact is that they never build on these “results” in the sense of taking them as fixed. Nor do they regard any “theory,” however limited the support it may continue to command, as finally refuted in such a way that we can “move on.” On the contrary, it is not merely standard practice, but a characteristic method in philosophy to dispute and debate foundations. This attack on foundations results both in raising doubts about purported “facts,” even when they are generally regarded as “obvious,” and in the revival of “theories” long thought to be indefensible. The history of dualism in the



philosophy of mind is a striking example. This philosophical inclination—​to uncover and then question the suppositions that lie behind observations and explanations—​necessarily denies to the conclusions philosophers reach, the status of “known.” In short, from the point of view of philosophical inquiry, no serious thought or contention can be dismissed as mere error, or discounted as something long since refuted. That is why no substantial philosophical position is ever completely superseded by another, as Aristotelian physics may be said to have been superseded by Newtonian mechanics. Philosophy has a history, certainly, but it is very rare for philosophical contentions to become matters of purely antiquarian interest. To put the point simply: contemporary philosophers read Aristotle’s Metaphysics for its philosophical import. No contemporary physicist reads the Physics in this way, and that difference may explain why most philosophical debates so easily strike scientists as idle. It seems then that no matter how much time elapses, or how much intellectual effort is expended, there is something importantly inconclusive about philosophical debate. It is hardly surprising therefore, in Jones’s words, that nonphilosophers “have naturally concluded that philosophy is seeking by a doubtful method an unattainable goal,” and philosophers have not infrequently been haunted by something of the same thought. Is there any satisfactory response to these allegations or suspicions? In the history of philosophy there seem to me to be three positions that recur with some regularity. The first, which takes its cue from Hume, suggests that philosophy should relinquish its traditional methods in favor of, or at least in alliance with, the empirical sciences. The second response, of which John Locke might be thought to be the author, is less radical, and accords philosophy a subsidiary role as an important contributor to these sciences. The third response, in which the influence of Kant (among others) is detectable, is that the problem of philosophy’s inconclusiveness is misconceived. It only arises when we erroneously regard philosophy as some sort of science, instead of according it a distinctive integrity of its own. The next few sections will explore the respective merits of these in turn.

5.3. Making Philosophy a Science Famously, Hume’s purpose in the Treatise was to remedy the unhappy condition that he describes in its preface, by adopting a new method modeled on Francis Bacon and Isaac Newton (though Hume does not explicitly mention Newton), namely the method of “observation and experiment.” One hundred

Philosophy, Knowledge, and Understanding


years later, John Stuart Mill, in his Examination of Sir William Hamilton’s Philosophy, was pursuing the same strategy. In the twenty-​first century, it is easy to find philosophers who think that many of the traditional problems of moral philosophy and human agency have to be rethought in the light of results generated by the investigations of contemporary moral psychology. In many respects this is just a modern version of what Hume aspired to. It engages in systematic “observation and experiment” with the aim of recording the actual attitudes and motivations of moral agents, as opposed to the motivations and attitudes that moral philosophy has commonly imputed to them, or thinks they ought to have. Moral psychology of this kind, in fact, can be thought of as one aspect of a much more ambitious program—​namely, naturalizing ethics, aesthetics, epistemology, and so on, by investigating the traditional problems of philosophy empirically. It is now common to find books and papers that draw on paleoanthropology, sociobiology. and evolutionary psychology to illuminate issues in the philosophy of mind and emotion, and experiments employing the power of MRI (magnetic resonance imaging) in neurophysiology to uncover the biological basis of aesthetic preferences. Hume’s strategy, in other words, has found new advocates, and even an eighteenth-​century term—​“science of mind”—​can plausibly be revived to capture this important new movement of thought. Occasionally, enthusiasm for empirical inquiry has generated the proposal that philosophy should simply be abandoned. For a brief period in the mid-​ twentieth century, political science, for instance, seemed set to displace political philosophy completely. Such displacement, however, has always proved temporary, and for the most part the current turn to empirical psychology, evolutionary biology, and neurophysiology rests upon a more modest contention; philosophers will make far greater headway with the problems that interest them if they pursue them collaboratively with these (and other) sciences. It is a claim that philosophers can be found making no less than psychologists and neurophysiologists, and many philosophers who believe in the potential rewards of such collaboration will add that it is only an unhelpful and unhealthy obsession with intellectual boundaries that could make philosophers resist this. There is a perspective from which this view of the matter seems so sensible, there cannot be much reason to reject it. Yet, as my references to Hume and Mill suggest, it is a salutary fact that we have been here before. As the ambition of the eighteenth-​century science of mind unfolded into the nineteenth century, the adoption of new methods, for the most part empirical, reshaped intellectual inquiry in the express hope that by throwing better light on old



and seemingly intractable problems, these new methods would secure real advances in knowledge and understanding. An especially striking instance of this can be found in the work of one of Mill’s disciples—​Alexander Bain. Widely regarded as a key figure in the early history of scientific psychology, Bain founded the journal Mind and contributed many of the most important papers that appeared in its early volumes. In an essay entitled “Associationist Controversies,” Bain expressly advocated a methodological change of the kind under discussion. We are, at the moment, in the midst of a conflict of views as to the priority of Metaphysics and Psychology. If indeed the two are closely identified as some suppose, there is no conflict; there is in fact, but one study. If, on the other hand, there are two subjects, each ought to be carried on apart for a certain length, before they can either confirm or weaken each other. I believe that in strictness, a disinterested Psychology should come first in order, and that, after going on a little way in amassing the facts, it should revise its fundamental assumptions.  .  .  .  I  do not see any mode of attaining a correct Metaphysics until Psychology has at least made some way upon a provisional Metaphysics. (Bain 1903, 38) Bain is here advocating an alliance between modern psychology and traditional metaphysics, though his own faith was pinned principally on the former, for which, in fact, he commandeered the older label “science of mind.” Elsewhere, in an address to the Anthropological Section of the British Association meeting in Aberdeen (where Bain held the Regius Chair of Logic between 1860 and 1880), he advocates an even grander alliance between philosophy, the science of mind, anthropology, and zoology. It was the spirit of the age, and similar proposals can be found in other subject areas, notably economics and politics. For better or worse, the outcome of this program was not in fact a newfound unity of the sciences. On the contrary, it was disciplinary differentiation. Psychology, economics, politics, linguistics, sociology, and anthropology quite rapidly gathered strength and significance, often as subdivisions within departments of philosophy, and eventually went their own way. In many instances, they explicitly addressed topics that had hitherto been the exclusive province of philosophy, while developing new interests of their own. Importantly for my purposes, however, logic, metaphysics, philosophical psychology, and moral philosophy were not subsumed within these new studies

Philosophy, Knowledge, and Understanding


and disciplines. Rather, they all reasserted themselves as distinctive intellectual inquiries whose key concerns still required to be addressed directly. Why was this? The answer (in my estimation) is to be found in the normative dimension to these questions, a dimension that is indispensable in philosophy. Logic provides a simple illustration of the point. One function of the mind is to reason, and there is much to be learned about how human beings actually reason from close observation and carefully constructed experiments. Logic, though, which we may here take to be the rules of reasoning, is concerned with evaluating and not merely describing patterns of reasoning. How people do reason is one question, and how they ought to reason is another. Human beings not infrequently reason fallaciously. Careful empirical observation may reveal recurrent patterns of reasoning, of which, perhaps, logic has to take some account. But since faulty reasoning may arise from systematic errors to which human beings are prone, no such patterns can capture logical validity. In short, the science of mind applied to reason, however sophisticated it may be, cannot generate principles by which valid inferences are to be distinguished from logical fallacies. The point applies not only to logic. Kant makes exactly the same observation about aesthetics, when he comments (in a footnote to the first Critique) on what he regards as Alexander Baumgarten’s “abortive” attempt to make aesthetics a science by grounding criteria of taste in generalizations about human preferences. Such an endeavor, Kant thinks, is “fruitless” because “the criteria are, as regards their chief sources, merely empirical, and consequently can never serve as determinate a priori laws by which our judgment of taste must be directed. On the contrary, our judgment is the proper test for the correctness of the rules” (Kant [1781] 1929, 66). Developments in economics over the course of the nineteenth century exhibit the same trajectory. Utilitarianism as expounded by Bentham aimed to tell us what we ought to prefer and by what criterion forms of social organization should be judged good or bad. Mill’s attempt to refine the crude doctrine he inherited from Bentham had the same purpose, but it pushed things along on a direction that eventually produced sophisticated mathematical modeling of “preference satisfaction.” At this point, welfare economics came into its own. Yet, whatever its merits, it did so only by relinquishing Bentham’s original aim. Recorded preferences, however widespread, cannot tell us what people ought to prefer.2 2.   Mill famously argues that that just as the proof of something’s being visible is that people actually see it, so the proof that something is desirable must lie in the fact that people actually desire it. This



The same observation can be made about the associationist psychology inaugurated by Hume and advocated by Bain. Even if we suppose that the Treatise of Human Nature successfully describes the “habits” characteristic of the human mind, this tells us nothing, as Hume himself acknowledged, about how we ought to think. I think the same point will be found to be true of moral philosophy. While there is no doubt that the empirical experiments and surveys of contemporary moral psychology have thrown up intriguing facts about human action and cast significant doubt on, for instance, common assumptions people make about altruism, these discoveries do nothing to illuminate the philosophical question about egoism and altruism that has occupied moral philosophers, namely, “How can the needs of others generate reasons for me to act in their interests?” As observed earlier, the interminable nature of philosophical debate has generated three different responses. I conclude from this brief intellectual history that it would not be wise to pin much hope on the first of them—​the contention that philosophers should replace, or at any rate supplement, their traditional methods of inquiry with those of the empirical sciences. Let us then consider the second response—​that philosophy can best make a contribution to the growth of human knowledge by acting as ‘under-​labourer’ to the empirical sciences.

5.4. Philosophy as the Handmaid of Science The term “under-​labourer” is Locke’s. Here is the relevant passage from his Essay Concerning Human Understanding. The commonwealth of learning is not at this time without master-​ builders, whose mighty designs, in advancing the sciences, will leave lasting monuments to the admiration of posterity: but everyone must not hope to be a Boyle  .  .  . and in an age that produces such masters as . . . the incomparable Mr. Newton, . . . it is ambition enough to be employed as an under-​labourer in clearing the ground a little,

is utilitarianism’s best-​known attempt to bridge the descriptive/​normative divide. It has prompted a lot of debate. I side with those who hold that it rests upon a crucial ambiguity in the word “desirable,” meaning both “able to be desired,” which empirical generalization can confirm, and “worthy of being desired,” which it can’t.

Philosophy, Knowledge, and Understanding


and removing some of the rubbish that lies in the way to knowledge. ([1689] 1975, 9–​10) Since in the seventeenth century the terms were virtually interchangeable, Locke could as easily have called Newton and Boyle philosophers as scientists. Accordingly, the distinction is not one in which he shows much interest, but the line Locke draws between the “master” and the “under-​labourer” has sometimes been employed to articulate the relationship between philosophy and science. Thomas Reid gestures in this direction in his Inquiry into the Human Mind on the Principles of Common Sense. With Hume’s Treatise of Human Nature in mind, probably, he remarks that It is genius, and not the want of it, that adulterates philosophy, and fills it with error and false theory. A creative imagination disdains the mean offices of digging for a foundation, of removing rubbish, and carrying materials; leaving these servile employments to the drudges in science, it plans a design, and raises a fabric. Invention supplies materials where they are wanting, and fancy adds colouring, and every befitting ornament. The work pleases the eye, and wants nothing but solidity and a good foundation. It seems even to vie with the works of nature, till some succeeding architect blows it into rubbish. (Reid [1764] 1997, 15) Reid is here identifying the cause of the huge discrepancy he sees between the physical and the human sciences in his own time. That our philosophy concerning the mind and its faculties, is but in a very low state, may reasonably be conjectured, even by those who have never narrowly examined it. Are there any principles with regard to the mind, settled with that perspicuity and evidence, which attends the principles of mechanics, astronomy and optics? These are really sciences, built upon laws of nature which universally obtain. What is discovered in them is no longer matter of dispute: future ages may add to it, but till the course of nature be changed, what is already established can never be overturned. But when we turn our attention inward, and consider the phænomena of human thoughts, opinions and perceptions, and endeavour to trace them to the general laws and the first principles of our constitution, we are immediately involved in darkness and perplexity. (Reid [1764] 1997, 16)



Like Locke, Reid uses the terms “philosophy” and “science” more or less interchangeably and the general point he is making is that careful analysis of concepts and close scrutiny of fact is less glamourous, but far more likely to be productive than a rush to theorizing. This relatively modest contention, however, falls considerably short of a much more ambitious idea that gained currency during the era of “philosophy as conceptual analysis.” The most profitable task for philosophers, some of the proponents of conceptual analysis thought, was to do the preliminary linguistic and analytical work that natural and social scientists would need if they were to proceed more effectively with their empirical investigations. This implies, in effect, that philosophy should play the role of “handmaiden” to science as it was once thought to be philosophia ancilla theologiae—​ handmaiden to theology.3 Hume had a word of warning in this latter case. Too close an alliance with theology, he thought, was dangerous because “the same fires, which were kindled for heretics, will serve also for the destruction of philosophers” (Hume [1757] 1993, 166). Times change, of course, and contemporary philosophical heretics are unlikely to be burned. Hume’s broader contention, nevertheless, remains valid; philosophy in the service of theology puts its own autonomy at risk. A similar danger, it can be argued, arises from too close an alliance with science. The conceptual clarity and logical rigor that philosophical analysis achieves must ultimately count for very little, if the intellectual significance of this achievement is dependent on the viability of a scientific hypothesis that scientists could at any moment disprove and hence discard. While Reid’s claim that “what is already established can never be overturned” is strictly true, it may mislead us into forgetting that even the most widely accepted scientific theories can turn out not to have been “established” after all. There is a second drawback to the idea of philosophia ancilla scientiae. Neither natural nor social scientists have ever shown much inclination to hand any part of their intellectual agenda over to philosophers, still less to suspend their own inquiries until the results of that conceptual analysis is complete. There is good reason for their reluctance. For a short spell in the 1960s historians thought they should turn to philosophers for clarification on the concept of historical explanation as a valuable preliminary to formulating more satisfactory explanations of historical events. They quickly found 3.   The medievals who used this expression would not have taken it to imply that philosophy was thereby subservient to theology. They were conscious that as “ancilla” to the Lord, Mary the mother of Jesus had been accorded an indispensable role in the redemption of the world.

Philosophy, Knowledge, and Understanding


that the philosophers simply raised further difficulties and, if anything, left the historians more uncertain than before about what was and was not a good explanation. This brief episode illustrates a very important point, in fact. The interminable philosophical disagreements that Jones describes are not simply an unfortunate outcome of defective methods. Rather they reflect something structural about the trajectory of philosophical thought. While history generally eschews generalization and seeks to describe and explain the particular, philosophy moves in the other direction—​from the particular to the abstract. The effect of this is the uncovering (not invention) of complexities, which runs counter to establishing agreed “results.” To concur on the adequacy of an explanation depends upon identifying, and giving special weight to, a single determinant, and thereby simplifying the potential range of complicating factors. In short, the style of thinking characteristic of philosophy makes it especially unsuited to serve as an underlaborer to the exact sciences.

5.5. The Normativity of Philosophy From the point of view of the third proposal outlined at the start, the conclusions of the last few sections are only to be expected. Since it is a mistake, this third response suggests, to try to cast philosophy in the mold of the empirical sciences, its failure to look like a science is not really a failure at all. In “The Nature of Philosophy,” the paper with which we began, Jones pursues this idea, and makes the following important points. First, we should not confuse real development with measurable progress. In the natural sciences, at any rate, progress is the mark of development, but this is not true in the arts. There is no doubt that between the medieval and modern periods, Western classical music developed very considerably. But it adds nothing to describe this development as “progressive” and is seriously misleading if it causes any one to suppose that the existence of later works makes earlier works musically less valuable or interesting, and hence no longer worth playing. Second, as Jones puts it, Hume’s “rabble without was wrong to infer from the “noise and clamour” that “all goes not well within.” On the contrary, debate is a sign of vitality, and the greater the vigor, the greater the vitality. The silence of consensus, far from being a welcome outcome, would be a sure sign that philosophical activity is at an end. Third, one scientist can build upon the results of another because scientific inquiry is importantly impersonal. Other people cannot do my philosophical thinking for me any more than they could write my poems.



As the analogies with music and poetry suggest, Jones invites us to compare philosophy with the arts rather than the sciences. The comparison seems to me illuminating, and yet there appears to be a crucial disanalogy. Philosophy, it seems, unlike music or poetry, is a truth-​seeking activity. That is to say, it strives to arrive at true beliefs about the nature of reality, the validity of moral obligation, the basis of human motivation, the criteria of genuine knowledge, the possibility of action, and so on. Since music and poetry are essentially exercises in imagination they need not “track” the truth. It is this relation to truth that puts philosophy, unlike music, in the same company as the sciences. If this is correct, then it seems we have simply circled back to the puzzle with which we began—​philosophy’s apparent failure ever to arrive at the truth. The solution, I shall suggest, lies in acknowledging that, while philosophy has cognitive aims that music does not, the cognitive value it seeks is neither knowledge nor explanation. “Did the Almighty,” says Lessing, “holding in his right hand Truth, and in his left Search after Truth, deign to proffer me the one I might prefer, in all humility but without hesitation I should request Search after Truth.” So writes Mill’s bête noir, Sir William Hamilton (Hamilton 1853, 40). Hamilton’s point is that philosophy is a form of activity, and since knowing the truth is essentially passive, philosophy must be the search for truth rather than its possession. There is something right, and yet something not right about this. To begin with, if philosophy is indeed the search for truth, mustn’t this search culminate in its discovery? A  “search” that never produces what is sought seems both frustrating and pointless. The objection sounds obvious, though its force relies on a certain kind of example. If I am engaged in a legal search to discover the true ownership of a property, then enjoying the search is small consolation for never finding out, and undertaking a search in the belief that I will never actually find out would indeed be very odd. This example is clear enough, but to generalize from it in a certain way leads to an importantly mistaken way of thinking. The following generalization sounds equally obvious. Intellectual inquiry is the search for knowledge. The proper object of knowledge is truth. Consequently, truth is necessarily the aim of intellectual inquiry. It seems to follow that if philosophy is indeed an intellectual inquiry, then pace Hamilton, it cannot rest content with searching for truth, but at some point must actually find it. So the problem of philosophy’s inconclusiveness returns.

Philosophy, Knowledge, and Understanding


The plausibility of this argument, of course, rests upon the truth of its premises. Is it true that intellectual inquiry is the search for knowledge (or truth, for that matter)? The answer is that knowledge has its place in intellectual inquiry, but is only very rarely its terminus ad quem. The straightforward search for knowledge is more characteristic of other kinds of investigation—​medical diagnosis, criminal trials, market research, investigative journalism, family genealogy, for example. Intellectual inquiries like history and the natural sciences are certainly “cognitive” pursuits, but their distinguishing mark lies in what they aim to do with the knowledge that they seek, or have already secured. In these disciplines the terminus ad quem is a different cognitive value—​explanation, usually—​and knowledge is valuable chiefly as a means to it. We are inclined to think otherwise because “knowledge” is so easily assumed to be primarily a matter of establishing facts and communicating information. This is not wrong, exactly, but as Wittgenstein points out in the Philosophical Investigations the word ‘know’ carries other connotations. The grammar of the word “know” is evidently closely related to the grammar of the words “can,” “is able to.” But also closely related to that of the word “understand.” (To have mastered a technique.) (§150) But there is also this use of the word “know”: we say “Now I know!”—​and similarly, “Now I can do it!” and “Now I understand.” (§151) In the light of this observation we can return to the problem of interminable philosophical debate. We might agree that interminable and inconclusive debate is incompatible with the establishment of fact, and hence with the production of knowledge. But is it also incompatible with achieving understanding? If what philosophical thinking offers is not knowledge of propositional or demonstrable truths, but understanding, then its search is for something different, the sort of knowledge reflected in another sentence Wittgenstein expressly connects with philosophy—​“Now I know my way around.” No one can seriously doubt that the natural sciences of astronomy, chemistry, physics, genetics, and so on, have increased human knowledge beyond anything that previous ages could have imagined. We cannot conclude from this alone, though, that these sciences are to be accorded high cognitive value, since plainly we can have genuine knowledge of trivial truths, propositions that are no more valuable for being true. It follows that knowledge is not necessarily valuable in itself, but derives its value from the wider purposes within which it is set. In the case of intellectual inquiry, these are the purposes of



explanation and understanding. If the search for truth (to use Hamilton’s language) never produces demonstrable truth, it might nevertheless result in greater understanding. In this case, it is not pointless in the way that the example of a fruitless legal search would lead us to suppose. To think of philosophy as a search for understanding rather than the pursuit of knowledge resolves the “problem” of its inconclusiveness in the following way. Philosophy’s tendency to uncover complexities and thereby multiply rather than minimize the questions with which it begins, does militate against the establishment of firm conclusions that can be expected to command the rational assent of philosophers in general. But this is compatible with generating greater understanding. If Wittgenstein is right that the declaration “Now I  know my way around” can signal philosophical accomplishment, then it seems right to say that what I know my way around is the depth and complexity of the philosophical issue in which I am interested and with which I am grappling. To pursue the analogy with knowledge of place a little further, if there is no terminus ad quem in philosophical inquiry, philosophical understanding nevertheless enables us to distinguish between dead ends and promising pathways. Knowledge in this sense cannot be encapsulated in the kind of propositional knowledge that serves well in medical science, criminal justice, and political commentary. This fact serves to demarcate it, but does nothing to show that philosophical inquiry is cognitively worthless. It only shows how mistaken it is to suppose that the value of philosophy must be modeled on the empirical sciences that have made impressive progress with respect to knowledge and explanation.

5.6. Philosophy and Personal “Quest” It seems, then, that the right response to the “problem” of philosophy’s inconclusiveness is to reject the comparisons that make it problematic. This conclusion can be strengthened by the observation that rejecting the scientific parallel as inappropriate has a further advantage. It accords with another important feature of philosophical inquiry that necessarily stands in the way of philosophy’s being progressive—​the first-​person nature of philosophical understanding. I have to grapple with philosophy’s problems and possibilities for myself, and cannot trade, as I can in medicine or astronomy, on someone else having successfully grappled with them before me. Philosophical writings from times past, and contemporary philosophical discussion can undoubtedly illuminate those problems and present new possibilities. That is the point of studying philosophy. Familiarizing myself with the long, rich tradition of

Philosophy, Knowledge, and Understanding


inquiry that stretches back so many centuries, and mastering the techniques of reason that have emerged over that time, serves to deepen and enrich my own understanding. But the philosophical works I read cannot be a repository of answers or solutions. I need to ponder the problems and explore the possibilities for myself. This emphasis on the necessity of personal engagement sometimes prompts the thought that philosophy is not a form of intellectual inquiry at all, but rather a personal quest of some kind. Professional philosophers rightly reject the use of the expression “my philosophy” as having any connection with their activity, and regard it as signaling an amateurishly mistaken conception of their subject, one that reduces philosophy properly so-​called to the affirmation of aphorisms or homely sayings. For this reason, and despite the arguments rehearsed in this chapter, there is likely to be a residual resistance to any suggestion that philosophy is not in the truth business. Only if we continue to make truth central to philosophical inquiry, the supposition is, can we distinguish it from a common tendency (as Michael Oakeshott once put it) to “seek in philosophy what wiser men would look for in a gospel, some guidance as to le prix de choses, some convincing proof that there is nothing degrading in one’s being alive” (Oakeshott 1933, 1). But is it the pursuit of truth that prevents philosophy from being converted into gospel? Since truth is claimed by many to be a merit of their “gospel,” it doesn’t seem that truth seeking will adequately mark the difference. In another place, however, Oakeshott more fully elaborates this properly philosophical alternative to “gospel.” Philosophy springs from a certain bent of mind which, though different in character, is as much a natural gift as an aptitude for mathematics or a genius for music. Philosophical speculation requires little in the way of knowledge of the world and is, in comparison with some other intellectual pursuits, independent of book learning . . . . [T]‌he impulse to think systematically is, at bottom, nothing more than the conscientious pursuit of what is for every philosopher the end to be achieved. The passion for clearness and simplicity, the determination not to be satisfied with anything inconsequent, the refusal to relieve one element of experience at the cost of another, are the motives of all philosophical thinking. (Oakeshott 1960, xiii–​xv) The values of clarity, simplicity, coherence, and comprehensiveness are no less cognitive than truth and knowledge. Unlike truth and knowledge, though,



which can often be secured without much difficulty, pursuing these intellectual goals is very demanding, too demanding in fact for most people. To generate activity, therefore, they must, as Oakeshott observes, not only be values, but also be ends toward which our inquiries are actually directed. That is to say, it is not enough to recognize these as cognitive values; they must also be intellectually motivating. Accordingly, cognitive values such as clarity and coherence necessarily have a double nature. “Getting clear” is both an intellectual accomplishment and a personal achievement. On the one hand, arriving at a philosophically significant thought is different from a personal achievement like winning a race. It is something that can be communicated and understood by others. On the other hand, it is not a “result” of inquiry, a piece of information such as can be found in a dictionary, detachable from the thinker to the benefit of all. Philosophers engage with the work of other philosophers in a fashion somewhat analogous to the way in which musicians engage with the music that others compose. Being able to play the music of Bach or Beethoven is a musical accomplishment related to, but different from playing my own compositions. This observation, it is worth adding, is wholly consonant with recognizing that no composition of mine is ever likely to match theirs. What is the aim of philosophy as an intellectual activity? Can it achieve its end? Does it ever do so? These are the questions with which we began, and these are the answers we have reached. The aim of philosophy is to attain certain cognitive goals with respect to questions generated by the most fundamental features of human existence and experience. These cognitive goals do not include knowledge for its own sake, or the explanation of observed regularities. That is why it is a mistake to treat philosophy as some sort of “science” and bewail its inconclusiveness. Rather, the goals of philosophy are clarity, simplicity, coherence, comprehensiveness in thinking about questions whose intricacy and complexity have become ever more evident over the time that human beings have been thinking about them. This shows that there is point in formulating and communicating to others the outcome of such thinking. But no work of philosophy can ever eliminate the necessity for readers to think through these questions for themselves.

References Bain, Alexander. 1903. Dissertations on Leading Philosophical Topics. London: Longman Green. Butterfield, Herbert. 1931. The Whig Interpretation of History. London: G. Bell.

Philosophy, Knowledge, and Understanding


Hamilton, Sir William. 1853. Discussions on Philosophy and Literature, Education and University Reform. London: Longman Green. Hume, David. (1739) 1978. A Treatise of Human Nature. Ed. Lewis Amherst Selby-​ Bigge, and Peter H. Nidditch. Oxford: Clarendon Press. —​—​—​. (1757) 1993. The Natural History of Religion. Ed. John C.  A. Gaskin. Oxford: World Classics. Jones, Henry. 1893. “The Nature and Aims of Philosophy.” Mind 2: 160–​73. Kant, Immanuel. (1787) 1929. Critique of Pure Reason. Ed. and trans. Norman Kemp Smith. London: Macmillan. Kuhn, Thomas. 1970. The Structure of Scientific Revolutions. 2nd ed. Chicago: University of Chicago Press. Locke, John. (1689) 1975. An Essay Concerning Human Understanding. Ed. Peter H. Nidditch. Oxford: Clarendon Press. Mill, John Stuart. 1865. Examination of Sir William Hamilton’s Philosophy. London: Macmillan. Oakeshott, Michael. 1933. Experience and Its Modes. Cambridge:  Cambridge University Press. —​—​—​. 1946. “Introduction to Hobbes’ Leviathan.” In Hobbes on Civil Association. Oxford: Basil Blackwell, 1960. Reid, Thomas. (1764) 1997. An Inquiry into the Human Mind upon the Principles of Common Sense. Ed. Derek R. Brookes. Edinburgh: Edinburgh University Press. Weinberg, Steven. 2015. “Eye on the Present: The Whig History of Science.” New York Review of Books 62: 82–​84. Wittgenstein, Ludwig. (1953) 2009. Philosophical Investigations. 4th ed. Trans. G. E. M. Anscombe, Peter M. S. Hacker, and Joachim Schulte. Chichester: Wiley-​Blackwell.


The Ethics of Understanding Stephen R. Grimm

6.1. Introduction The mission statement of the United States’ renowned Fulbright Program tells us that two of its central goals are: • To increase mutual understanding between the people of the United States and the people of other countries. • To assist in the development of friendly, sympathetic, and peaceful relations between the United States and other countries of the world.1 The second of these goals, moreover, seems to build on the first. While we might seek mutual understanding for its own sake, the suggestion is that we should also, and perhaps more importantly, seek it because it will lead to friendlier and more peaceful relations with others. The Fulbright Program is also hardly alone in taking mutual understanding as a core aim. A quick Google search reveals dozens of other organizations, especially international or interreligious organizations, that likewise view mutual understanding as not just an inherently worthy goal, but one worth pursuing because of its connection to other moral and political benefits.2

1.   http://​www.fulbright-​​AboutUs/​MissionStatement/​tabid/​73/​Default.aspx. 2.   E.g., “mutual understanding” also appears in the mission statements of the Red Cross, the Council on American-​Islamic Relations, the US Bureau of Educational and Cultural Affairs, the Malaika Foundation, the Japan Society, the Asian Society, and so on.

The Ethics of Understanding


But if it is true that understanding helps to promote goods such as peace, love, and friendship, this should strike us as a deeply puzzling fact. After all, most of us—​probably all of us—​are less than admirable in various ways. We are shot through with vices such as vanity, cowardice, jealousy, and pride. We might sacrifice for others when it is convenient, but often not when the costs run deep or when they might take away some special comfort. In short, it is not difficult to sympathize with Kant’s claim in the Groundwork that at the bottom of many (all?) of our actions we find the “dear self ” smiling back up at us—​that is, a self primarily concerned with its own status or comfort, rather than with a large-​minded concern for others (Kant (1785) 2012, 10). But if that is the case, then why should coming to understand these facts about other people, or about other groups, lead to anything other than disapproval? Perhaps you are thinking that this is too uncharitable, naturally calling to mind the admirable people in your life who are driven not simply by self-​ interest, but by something larger or more noble. Still, the puzzle remains, because the claim we are considering is not this limited one: that better understanding the admirable will lead to goods such as peace, love, and friendship. Instead, we are looking at something more unrestricted: that better understanding anyone—​even, presumably, the selfish, vain, and vicious—​will lead to goods such as peace, love, and friendship. It is this view, I submit, that we still need to make better sense of. A slightly different and perhaps more evocative way to try to get on a grip on our puzzle is by asking why the lovely French proverb—​ Tout comprendre, c’est tout pardoner or, in English To understand all is to forgive all —​should have even have a ring of truth to it. For again, why should understanding others in all their limitations and faults lead us closer to forgiveness, and not toward a more negative attitude instead? In this chapter I will try to show that if there is a legitimate connection between understanding and goods such as forgiveness (or friendship, or love), it is a multifaceted one—​sometimes involving an understanding of the conditions that give rise to someone’s behavior, sometimes an understanding of the complexity of another’s situation, and sometimes involving an understanding of our common frailty. I will nonetheless try to spell out where these different



kinds of understanding earn their special place, as well as where they are limited in various ways. In addition, we should note that the connection between understanding and goods such as forgiveness, peace, and love is sometimes characterized in a descriptive way and sometimes in a normative way. That is, sometimes it is claimed that understanding others will lead to forgiveness, peace, love, etc., and sometimes that understanding should lead to these goods. In this chapter, however, I will mainly focus on the normative issue, trying to identify what it is about understanding (in its various forms) that properly brings about these goods. Perhaps the most controversial claim I will make is that while understanding is compatible with blaming or sanctioning someone else, it is not compatible with what I shall call “angry blame,” or blame laced with resentment or bitterness. Contrary to recent advocates of angry blame such as R. Jay Wallace (2011), Susan Wolf (2011), and Victoria McGeer (2013), I  will argue that angry blame is never appropriate for the sort of imperfect and frail creatures that we are. This leaves me in the unenviable position of defending what Wolf has called “wimpy blame”—​a hard task, according to Wolf and others, because while resentment and bitterness might be undesirable emotions in many ways, they nevertheless are said to play an important role in our moral lives. If I am right, however, not only can we get by without these negative emotions, but we should do so, once we properly understand our common limitations.

6.2. Conditions In thinking about different ways in which understanding should prompt compassion and forgiveness, perhaps the most natural place to begin is with cases where understanding reveals that someone is not truly responsible for his or her actions. To illustrate, consider the following passage from the Buddhist monk Thich Nhat Hanh. Understanding and love aren’t two separate things, they’re just one. Suppose your son wakes up one morning and sees it is already quite late. He decides to wake up his younger sister, to give her enough time to eat breakfast before going to school. It happens that she is grouchy and instead of saying, “Thank you for waking me up,” she says “Shut up! Leave me alone!” and kicks him. He will probably get angry, thinking, “I woke her up nicely, why did she kick me?” . . . But

The Ethics of Understanding


then he remembers that during the night his sister coughed a lot, and he realizes she must be sick. Maybe she has a cold, maybe that is why she behaved in such a mean way. He is not angry anymore. At that moment there is budh [the root of the word “Buddha”] in him. He understands, he is awake. When you understand, you cannot help but love. (Hanh 2005, 15–​16) We might therefore say that, for Hanh, understanding leads to forgiveness because it helps us to see that “conditions” or circumstances were responsible for someone’s bad behavior, rather than a faulty character or ill will.3 More generally, it is this sort of understanding that often leads people to forgive the faults of those who have had a wretched upbringing, or who have been beaten down by life in various ways. It would have been a miracle, it is often said, if someone subject to those conditions had turned out differently. A proper understanding of his or her circumstances should therefore lead us to recognize this fact and absolve the person of blame. The main problem with this attempt to link understanding and forgiveness, however, is that while it might hold for the sort of examples just described, it does not seem to cover a wide enough range of cases. In particular, it does not seem to account for the large number of cases where our faulty character presumably is to blame. To take a small example, familiar to academics: suppose you sometimes underline passages in library books, or write small comments in the margin. I take it that in this case you are at fault—​this is something you shouldn’t be doing, lest the books become unusable for others—​but in the normal course of things there are no conditions compelling you to mark up the books. Your actions simply betray the tendency, we can suppose, to put yourself above others on occasion, or try to except yourself from rules you expect others to obey. This pattern seems to apply to more serious cases as well. Thus Bernie Madoff presumably did not cheat on his customers and friends because of conditions, but rather because of greed, and Tiger Woods was not unfaithful to his wife because of conditions, but rather because of lust or intemperance. 3.   We also find this idea in the Buddhist thinker Śāntideva: “We are not angry at bile and other such sources of great suffering. We are angry, however, at those with minds. But they’re all incited by conditions”; for analysis and sources, see Bommarito (2011). Martha Nussbaum claims that a similar idea can be found in the Stoics, especially Seneca. She writes: “Seneca says that the wise person is not surprised at the omnipresence of aggression and injustice, ‘since he has examined thoroughly the circumstances of human life’ (condicio humanae vitae). Circumstances, then, and not innate propensities, are at the origins of vice” (Nussbaum 1993, 100).



In short, although conditions or circumstances sometimes best explain fault, at other times it is the usual parade of vices. For our purposes, the important point is that the mantra to understand all is to forgive all should presumably hold in these “noncompelled” cases too—​in those moments where we do wrong because of our deficient character, rather than because of conditions. We therefore need to search further in order to appreciate how understanding might lead to forgiveness in a more wide-​ranging way.

6.3. Simulations A second promising way to link understanding and forgiveness begins by supposing that to understand another person is to “take up” or assume his or her perspective, in the way that simulationists such as Jane Heal, Robert Gordon, and Alvin Goldman have described.4 On this view, in order to understand another person’s actions I need to imaginatively assume his or her own beliefs and desires while temporarily bracketing my own. For instance, in light of imaginatively taking on (a) your desire to get milk and (b) your belief that Joe’s Market is the best place nearby to buy milk, I will “see” or appreciate that your trip to Joe’s is the most sensible course of action—​the fitting or apt thing to do. If I can appreciate how this is the most sensible course of action from your perspective, however, then presumably I should not condemn you for your action, but take a much more favorable attitude instead. More generally, suppose that X is the wrong thing to do but you mistakenly believe that it is right or appropriate thing to do. Well, then, what are you supposed to do in this case? Something other than X? As Holly Smith notes, For [such a person] to have done the objectively right act would have been for them to do what they believed to be wrong. Such an act would necessarily have stemmed from a worse configuration of desires. (1983, 559) It is therefore appealing to think that when we understand others in the sense of “taking on” or assuming their perspectives, we will appropriately steer away from judgments of blame and toward judgments at least of acceptance, if not

4.   For representative examples of their views in this area, see Heal (1986), Gordon (2000), and Goldman (2006).

The Ethics of Understanding


of full love and support. If we add to this line of thinking the popular idea that our beliefs are not under our voluntary control, the case seems to be stronger still. After all, if we act according to our beliefs about how it is best to act, and our beliefs about how it is best to act are not under our meaningful control, then attitudes such as blame and resentment simply seem out of place. The basic problem with this general approach, however, is that even if we grant that beliefs are not under our direct voluntary control,5 they seem to be enough under our control to warrant judgments of fault and blame. Return again to the Bernie Madoff case. Even though his sons resolutely denied that they knew anything about their father’s crimes, the common response, shared by the district attorney, was that, well, they certainly should have known. That is, they should have managed their beliefs in a way that they did not turn a blind eye toward conspicuous evidence of fraud, or downplayed the importance of this evidence. We therefore do not arrive at universal forgiveness simply by noting that given a particular person’s configuration of beliefs, X was the only reasonable thing to do. Instead, we often take one another—​ legitimately, it seems—​to be responsible for the particular beliefs we happen to have, and by extension for the actions that flow from them. Imaginatively taking on or assuming the beliefs of another—​understanding “from the inside” what lies behind their actions—​therefore does not automatically or appropriately yield goods such as forgiveness and love, nor does it eliminate blame and culpability from the picture.

6.4. Moral Knowledge One plausible upshot of the previous section is there is an epistemic condition on blameworthiness:  that is, for someone to be blameworthy for X it is necessary that he or she either actually knew or should have known that X was wrong.6 I  will now suggest that the epistemic condition has special relevance for our project because it reveals a new and different way in which understanding can and should lead away from blame and toward forgiveness and acceptance. In particular, the more we appreciate that moral knowledge

5.   I cannot, after all, now believe that I am the mayor of Cleveland, even if you offered me a briefcase full of money to form this belief (cf. Alston 1989). 6.   Or perhaps it is better to think of this as a doxastic condition: thus the idea would be that the person either believed or should have believed that X was wrong (cf. Goldberg, forthcoming). My claim opposes Elizabeth Harman (2015; n.d.). Harman claims that Arpaly (2003) and Markovits (2010) likewise deny the epistemic condition.



was hard to come by in a given situation, the more we should steer away from blame.7 I will also argue that while this thought is correct and important to note, it is also too limited to make complete sense of the idea that to understand all is to forgive all. To see this, consider an area where uncertainty and second-​g uessing reigns: parenting. The main reason, I take it, is that when it comes to parenting a number of subtle and difficult-​to-​weigh considerations are morally important:  there is the need to promote autonomy in one’s children, for instance, but also to instill discipline. The need to cultivate proper respect for one’s country, one’s elders, perhaps God or one’s religious traditions, but also to promote free-​thinking and experimentation. The need to allow a place for friends in one’s child’s life, but also a need to make sure that the friendships do not become manipulative or destructive. Finally, and to bring things into the twenty-​first century, there are constant judgments to be made about the harms or benefits of technology in their lives. Parenting is a never-​ending attempt to juggle these goods, with no clear rules in sight. One can walk into a local bookstore at any time and find experts confidently advocating every side of an issue. Tiger moms and hippie parents both have their heroes. Until you become a parent, it is therefore often tempting to judge, or even to condemn, other parents for their actions. Once parenthood finds you, this tendency markedly, and I think appropriately, decreases. In coming to appreciate how hard it is to choose rightly in these areas, and in particular how hard moral knowledge is to come by, the grounds for blame seem to fade away. This pattern seems again typical. Thus former American presidents are typically very cautious about judging sitting presidents, generals about judging other generals, head coaches about judging other head coaches, and so on. In all of these cases, it is common to hear thoughts along the lines of: “I would never judge him/​her, because I know how hard the job is.” Of course, this is compatible with the plausible claim that there are certain things that everyone should know from a moral point of view, and hence that a failure to know them always brings blame in its wake. Perhaps your conscience tells you these things, or maybe they have been “written on our hearts” by God, as St. Paul says, or perhaps they are just in some sense “obvious,” so everyone is on the hook for them. We might think here of the natural 7.   For more on the difficulty of acquiring moral knowledge, and hence (it is claimed) on the prudence of trusting authorities on moral questions, see Jones (1999).

The Ethics of Understanding


law tradition in ethics, where thoughtful reflection on human capacities and powers is enough to reveal both a list of moral goods and a list of moral rules. But even if we grant the appeal of these views and allow that certain moral principles are known (or knowable) by all, a common thread in these traditions is that the more one descends into particular cases, the more difficult it is to know what to do. As Aquinas puts it, “Practical reason . . . is busied with contingent matters, about which human actions are concerned:  and consequently, although there is necessity in the general principles, the more we descend to matters of detail, the more frequently we encounter defects” (Summa Theologica I.94.4). That is, the more we become sensitive to the complexities of a situation, the more difficult it often is to know how to act appropriately. I thus submit that a deeper understanding of, or appreciation for, the complexities of a person’s station in life will therefore often, and appropriately, lead us toward compassion and away from blame or condemnation. We can thus see another sense in which the French proverb rings true. At the same time, we once again bump up against the limitations of a proposed account, because this new account likewise seems only to apply to a limited number of cases—​namely, cases in which moral knowledge is particularly difficult to come by. The trivial case of scribbling in the margins would again qualify here, but it would also apply, I  take it, to more serious cases involving lying, infidelity, theft, and murder—​cases, in particular, in which there are no competing deep moral goods on the scene, and hence cases where knowledge of the wrongness of the act is not difficult to obtain. It therefore still leaves us unable to explain how understanding should lead to compassion and forgiveness in those cases where moral knowledge is readily available but fault still occurs. In the remaining sections I will argue that blame, sanction, punishment, and so on are entirely appropriate in these cases, and that a better understanding of a person’s actions does not fittingly lead to forgiveness along these lines. More specifically, understanding in these cases does not fittingly lead to forgiveness taken as exoneration. There is, however, another crucial sense of forgiveness that is relevant to these cases, and that is tied to increased understanding: namely, forgiveness taken as dissolving anger or resentment, and in particular as dissolving the anger or resentment that often accompany judgments of blame.8

8.   For more on this distinction between types of forgiveness, see Murphy (2003) and Hughes (2015).



In order to appreciate this point, I  suggest that still another variety of understanding is required—​not, as we have canvassed earlier, either (a) an understanding of the conditions or circumstances that sometimes compel action, or (b) an understanding of the complexities of a situation that make moral knowledge hard to come by. Rather, in the remainder of the chapter I will focus on a new dimension of understanding, namely (c) an understanding of our own frailties and faults as human beings. In particular, I will argue that while blame, sanctions, punishment, and so on might be appropriate in many cases, it is hard to see how “angry blame,” or blame charged with resentment, ever is—​at least, when one is armed with the appropriate understanding. Since this stance against angry blame is currently unpopular in philosophy—​opposed by Susan Wolf, R. Jay Wallace, and Victoria McGeer among others—​we will also consider the case in favor of angry blame.9

6.5. Blame, Criticism, and Angry Blame Let’s now ask: Is it ever appropriate to feel bitterness or resentment toward someone we are blaming? To make things clearer, suppose that the blame is directed against someone who has wronged us. Imagine your friend has promised to pick you up at the train station at 3:50 p.m. As you step out of the train and into a steady rain, your friend is nowhere to be seen. Soon it is 4:00 p.m., and then 4:15, and you are now the only person on the platform. Suppose it also turns out that your friend “just forgot” to pick you up. He was playing an addictive new game on his phone and simply lost track of time. When he appears at 4:35 full of apologies, it would surely be natural not just to blame him but also to resent him for his behavior. Suppose in this case that blame is not only natural, but also appropriate—​ there were no conditions to exculpate your friend, and no moral complexities that made it hard for him to know he should have picked you up. According to many philosophers, resentment too would not only be natural but also apt or fitting. For some, what resentment signals in cases like this is a proper sense of self-​respect. Resentment is therefore appropriate because it tracks harm against something that is truly valuable (see, e.g., Franklin 2013). What’s more, a number of authors have argued that blame without resentment or

9.   For a review of the literature on blame and discussion of central topics, see Coates and Tognazzini (2013; 2014) and Strabbing (forthcoming).

The Ethics of Understanding


anger is not really blame at all—​that is, that to drain blame of resentment is in some sense to miss the essence of blame (see, e.g., Wallace 2011). They then add that a world without blame would be morally deficient in some way, or at least less morally good than a world with angry blame. In addition to helping us track morally important things in the way just described, resentment is therefore valuable because the threat of it helps to keep people in line—​for you will, plausibly, be less likely to commit a wrong if you know that anger and resentment are waiting for you, rather than a more emotionally neutral form of punishment. What we need to ask now, however, is whether it is possible for blame to fill these tracking and deterrent roles while shedding the elements of anger and resentment. In the remainder of this section I will argue that this is not only possible, but desirable, because angry blame is at bottom irrational. More exactly, it is irrational because it is grounded in a false presupposition—​one that an appropriate understanding of one’s situation should first reveal, and then lead us to eliminate. To get a better sense of how blame could shed itself of anger and resentment while still playing its central role in our lives, compare the related but distinct notion of criticism. Clearly criticism plays a pervasive role in our lives:  we regularly and perhaps inevitably critique painters, singers, cooks, politicians, and of course other philosophers. It also plays a very valuable role, for we need to keep track of how people are faring relative to standards of good singing, good cooking, and so on—​to pick out or flag the truly excellent among us, and to distinguish them from the mediocre or just plain poor. Call this approach to criticism, for lack of a better description, “plain criticism” or “standards-​based criticism”—​it is criticism in light of commonly accepted standards, perhaps objective ones. At the same time, I  take it that there is also an “ugly” side to criticism that often accompanies standards-​based criticism. It accompanies standards-​ based criticism so regularly, in fact, that it has helped to give all criticism a bad name. Here I’m thinking of the sort of criticism that does not just evaluate how someone is faring relative to a standard but instead tries to establish that the other person is somehow “beneath one” in relation to the standard, or in some way has ideas above her station. We might call this “pecking-​order criticism” because it is significantly concerned with asserting one’s status relative to others. To see the sort of thing I have in mind, note that a very natural response to criticism along the lines of “Well, this really isn’t a very good painting,” or “To be honest, her singing is pretty dreadful” is: “I’d like to see you do better



yourself !” or “Who are you to talk? You’re a godawful singer.” A  response along these lines is almost automatic among children in particular, who instinctively interpret any criticism as indicating that the critic could do better than the person being criticized. It is clearly possible, however, though not easy, to criticize someone else in a way that is free of any implication of “I could do this better than you,” or a desire to diminish another person’s status relative to one’s own. Standards-​ based criticism therefore does not necessarily imply or bring with it pecking-​ order criticism. That said, our usual practice of criticism is so saturated with ego that it is often fair to suppose, as a default, that pecking-​order criticism is taking place. That is one reason why criticism is so hard to take properly as well as to give properly. The moment we suppose that someone is not just assessing us relative to a neutral standard but is trying to “lord it over us,” we feel upset and defensive. A truly skilled critic will therefore try to deliver her evaluation with the ego removed, or without trying to suggest that she could do it better. All that said, suppose we were now trying to give a scholarly account of the nature of criticism. Wouldn’t it be easy to imagine someone, at first blush, claiming that criticism without attention to pecking order wasn’t really criticism at all—​that somehow it was of the essence of criticism to try to emphasize one’s status relative to others? Although this would be a natural response, it nonetheless would represent a confusion between (a) what usually goes on, psychologically, in less than admirable people like ourselves, and (b) the proper and almost indispensable role that criticism plays in our lives. We might say that pecking-​order criticism is a corruption of what criticism could and should be, and that there’s little or nothing of moral worth in the pecking-​order concerns that as a matter of fact, but sadly, infuse our practices. Turning back now to blame, I would like to make two claims. First, that taking the anger or resentment out of blame does not “leave the blame out of blame” (Wallace 2011, 349). Alternatively, it does not leave us with what Susan Wolf has called mere “blame for wusses” or “wimpy blame” (Wolf 2011, 336)—​which is presumably to say, blame not worth having. Instead, the anger or bitterness that typically accompanies blame is an optional add-​on, much like pecking-​order criticism is an optional add-​on to standards-​based criticism. Second, and more importantly for our purposes, the justification for shedding anger from our practice of blame is even stronger than the justification for shedding pecking-​order concerns from our practice of criticism, because the anger is characteristically based on the false presupposition that we ourselves are somehow exempt from the fault. A proper understanding of

The Ethics of Understanding


our own frailties should help to reveal this irrationality, I will now argue, and thus should help take the rancor out of blame. Consider again the case of the delayed train pickup. As you wait and stew and then wait and stew some more, part of what gives rise to the resentment and anger, I take it, are thoughts such as the following: “How could he have been so thoughtless! Or so stupid! If I were in his shoes I never would have forgotten to pick him up!” What a little reflection should show, however, is that this is probably not true! Perhaps, given the right amount of weariness or distraction, you would indeed have done the same thing. Or even if you are particularly fastidious and unerring about this specific thing—​that is, even if it is true that you never miss pickups, for whatever reason—​it is still highly plausible that you engage in other thoughtless behavior from time to time, behavior that fails to respect people in the way they deserve. Realizing or appreciating this should then lead us to retract the “I never would have,” and with the retraction of the “I never would have,” I submit, this particular brand of angry blame should, rightfully, fade away. Or again, suppose you discover your spouse cheating on you, and you are filled with anger and resentment. “How could she do this to me! How could she treat me like this! I never would have done something like this to her—​ something so hurtful, so humiliating!” But again, even if you are right about this particular type of hurtful behavior, is it really the case that you have never failed in comparable ways? Assuming a negative answer to this question—​ assuming, that is, that you have been thoughtless, or hurtful, or perhaps given into temptation, against your better principles or judgment—​then this should help you to realize how it is possible, even when there is nothing like malice or ill will in the picture. My suggestion then is that resentment arises from incorrectly identifying ourselves as somehow exceptional or immune from the relevant fault. Appreciating how we are liable to similar failures should therefore show us that we are not exceptional or immune and should therefore, if we are clear-​ sighted, undercut the source of the resentment. The account here is thus in line with the idea that resentment is naturally undercut by a thought along the lines of, “There but for the grace of God go I” (cf. Watson 1987; 2004, 245–​46). That is to say, since I have shown myself to be weak, frail, selfish in areas X, Y, and Z, I am fortunate that I have managed to keep my head above water in the related but distinct areas X*, Y*, and Z*. I should therefore curtail my resentment toward those who have failed in these other parts of life. Or again, it naturally calls to mind the biblical



injunction, “Let he who is without sin cast the first stone.” For again, the attitude of the stone thrower is presumably something along the lines of: “How dare he . . .” or “How disgraceful . . . I would never . . . !” Calling to mind our own frailties, or having them called to mind by others ( Jesus, for instance, in the Gospel stories), makes us confront the irrationality of these stances on their face, and should make the one attitude—​of bitterness or resentment or “righteousness indignation”—​naturally fade away. The account here is also similar to, but importantly different than, a sentiment of Dietrich Bonhoeffer’s, the heroic Lutheran pastor whose opposition to the Nazis led to his imprisonment and execution. Thus Bonhoeffer writes in The Cost of Discipleship: Judging others makes us blind, whereas love is illuminating. By judging others we blind ourselves to our own evil and to the grace which others are just as entitled to as we are. ([1937] 1995, 185) Bonhoeffer’s thought is similar to the one developed here because it too links judging (blaming, resenting, lack of forgiveness) with a blindness to our own evil. Our accounts differ in that while Bonhoeffer takes the judging to be the cause of the blindness, I have argued that things are the other way around. It is our blindness to own frailty that leads to the judging, I have suggested, and it is an appreciation or proper understanding of that frailty that should make the judging dissolve in time. Finally, note that we have been dwelling here on cases of the morally frail resenting the morally frail. But consider instead the case of the fully admirable or virtuous judging the frail. What would, or should, their attitudes be like? To my ear at least, it sounds extremely odd to say that the fully virtuous person would be resentful or bitter toward the weak. Instead, it seems much more natural to suppose that the fully virtuous person would feel a kind of pity toward the weak—​a kind of lament, perhaps, for their ongoing struggles to do well. Yet if this is right, then it is hard to see where resentment or bitterness or rancor could ever have a legitimate place in the moral life. If we are frail, then our attitude toward fellow frailty should apparently be one of sympathetic compassion. If we are fully virtuous, our attitude should again arguably be compassionate, though now drawn more, it seems, from pity (in the noncondescending sense of “pity,” admittedly rare today). In either case, resentment seem to represent a corruption not only what our blaming practices could be, but what they should be, were understanding more present.

The Ethics of Understanding


In a general sense, we might therefore say that a proper understanding of “the human condition”—​of our human plight, as Schopenhauer might say—​should lead us to be more compassionate and gentle with one another. In these areas, the problem is not that moral knowledge is difficult to come by, but rather that it is difficult to live by. Once we see that we are subject to these same struggles, our resentment, grounded in a false sense of our own exceptionalism, should properly fade away.

6.6. Objections Before turning to close, in this section can now consider three natural objections to our account. The first objection is that I  haven’t taken resentment’s value as a deterrent seriously enough. Surely—​the thought goes—​people are more apt to stay morally in line, or to honor their obligations, if they can expect that their misconduct will be met not just with impartial sanctions or punishments—​ blame for wimps—​but with some deeply personal resentment. While this may be true, however, the normative implications of the claim are hard to see. After all, sanctions accompanied by seething hatred would presumably have an even stronger deterrent effect, and sanctions accompanied by corporal punishment still more. Why then stop with resentment when one could have still stronger deterrent effects? In addition, if we are evaluating practices mainly in terms of their forward-​looking benefits, then it is hard to see how blame accompanied by resentment and anger might fare better than the more impartial blame recommended here. As residents of the Middle East can attest, lingering resentment and anger is not generally conducive to well-​being. The second objection is that the proposal developed here is “psychologically unrealistic” (McGeer 2013, 163). As Victoria McGeer notes, the view I have argued for is in keeping with the general project of trying to “sanitize” or “civilize” blame of its unappealing emotional elements, such as bitterness, rancor, and resentment. According to McGeer, however, this is so outside the reach of actual human beings that it should not even belong among our moral ideals. She writes: Is it possible for creatures like us to engage in such a process under the burden of perceived injury without feeling or expressing angry emotions, however mild these may be? My response to this is brief. Perhaps it is possible, at least for the saints among us. But however



much we may suppose this is a preferable state of affairs, I doubt it is practically available to the common run of humanity. And this must pose a serious challenge to any philosophical theory that recommends it. (McGeer 2013, 181) In other words, blame without resentment is just not the sort of thing ordinary human beings are capable of. Since ought implies can, it is also not among our obligations. I suggest, however, that blame without resentment is more attainable than McGeer claims. First off, recall our earlier discussion of criticism, and the distinction between standards-​based criticism and pecking-​order criticism. Here too, I take it, the inclination to intermix the two, so that every case of criticism is infused with an element of pecking-​order criticism, is extremely strong. Indeed, for children and adolescents it might be all but inevitable. As we mature and see the drawbacks of pecking-​order criticism, however, it is clearly possible to move away from it, and to strive for a more “civilized” or “sanitized” form of criticism—​standards-​based criticism—​even if is contrary to inclination. Further, although blaming without resentment is admittedly difficult, it is not fair to say that it is possible only for the saints among us. I write this not long after a deeply troubled white gunman took the lives of nine members of a black church in Charleston, South Carolina. Although the act was tragic and depraved, the response on behalf of the victims’ families was inspiring, as one-​ by-​one they approached the captured shooter and forgave him for the crime. Consider the following testimony from a family representative for Ethel Lance, a seventy-​year-​old grandmother who was killed in the shootings: You took something very precious away from me. I will never talk to her ever again. I will never be able to hold her again. But I forgive you and have mercy on your soul. You hurt me. You hurt a lot of people, but I forgive you.10 Note well what forgiveness involved here: not acquittal, because the family did not cease to hold the shooter blameworthy or at fault or deserving of punishment (cf. Nussbaum 1993, 124). Instead it seemed to be forgiveness tied 10.   https://​​news/​p ost-​nation/​w p/​2015/​06/​19/​hate-​wont-​win-​the-​ powerful-​words-​delivered-​to-​dylann-​roof-​by-​victims-​relatives/​.

The Ethics of Understanding


to a decision to let go of hatred or hostility toward the offender; in other words, a decision to acknowledge the blame or the fault, but to drop the vindictive emotions. Note too, and crucially in terms of McGeer’s concern, that while the families in Charleston were heroic, they were not, it seems fair to say, saints. They were simply behaving in a way that they felt called to behave as Christians. It is nonetheless a model that many, Christian or not, justifiably viewed with admiration—​indeed, a model worthy not just of respect but emulation.11 A final objection is that my view fails to recognize a difference between big, heinous, stomach-​churning faults and more minor faults. Perhaps we have all thoughtlessly forgotten to pick up someone from the train station, or perhaps we have all done selfish little things like writing in the margins of a library book. Very few of us, however, have ever done something like wantonly betray a friend, or murder an innocent person, or commit treason against our country. So, perhaps if someone were to say, of some class of wrongs—​those that involve the taking of innocent life, say—​that he or she would never do such a thing . . . perhaps that person is right! Hence perhaps there is no internal incoherence in the area that might undermine someone’s resentment or bitterness with respect to these graver faults. In reply I suggest that although this objection is ultimately mistaken, it nonetheless points to something correct and worth noting. The flaw lies in the confident assertions about what we would “never” do. Thus if Stanley Milgram’s famous experiments at Yale, or the Stanford Prison Experiment, has taught us anything, it is that odds are, had we been in these situations, we too would have done something previously unthinkable.12 We are therefore likely much more capable of even grave wrongs than we usually imagine. That said, the element of truth in the objection is that there are almost certainly some things we would, in fact, never do—​or perhaps better, some ways of acting that are beyond the large majority of human beings. Here I am 11.   Julia Driver has suggested in conversation that forgiveness in the Charleston case was in some ways atypical, because the offender was already captured, and therefore justice in one sense had already been done. Had the offender not been captured, angry blame could have been crucial to sustaining the psychological resolve to track down the offender, especially if the search was prolonged (say, ongoing for years). In response, while I acknowledge the psychological usefulness of angry blame in many cases, again I am not sure what follows. Suppose it is true of a given person that he/​she could only pursue justice over the course of years by thinking of the offender as “scum” or “less than human” or “a monster who doesn’t deserve to live.” Would those attitudes be appropriate? I do not think that they would. The psychological utility of various states therefore needs to be distinguished from their warrant. 12.   For more on the Milgram and Stanford prison experiments, see Doris and Stich (2014).



thinking of sadists or psychopaths who don’t just do terrible things, but revel or delight in them. For two fictional but all-​too-​plausible examples, think of depraved characters like King Joffrey or Ramsey Bolton in The Game of Thrones, characters who only appear happy when they are making others suffer. In these cases there doesn’t seem to relatable moral weakness or the like at play, but something much darker. Regarding such cases of sadism, there are two ways we might reply. First, we might say that such cases truly do merit our anger and resentment—​ or, at the very least, they do not merit responses such as compassion and gentleness—​but that this is not a problem for the thesis that “to understand all is to forgive all,” because actions like this simply fall outside the scope of our understanding. Perhaps we genuinely cannot see—​not now, and not even while bearing in mind Milgram-​style cases that call our boundaries into question—​how human life could mean so little, or even nothing, to another person. Second, we might say that in extraordinary cases such as this it is not even clear that blame is appropriate, much less resentment or bitterness, because something has gone so badly awry that it is not even clear human agency is at play, rather than mere animal instinct. Put in our earlier categories, it is not even clear that the person can be blamed because he or she “knew or should have known better,” because it seems like conscience has no voice at all in these cases, or that the moral law has not even been written on these hearts. I will not take a stand on either response here, but I think both are plausible.

6.7. Conclusion In closing, I  will briefly suggest a few different ways in which these reflections might open the door to a more robust “ethics of understanding.” For one thing, it would be interesting to consider how understanding, in the various forms explored above, might take its place alongside more conventional virtues such as courage and temperance. If what I have said here is correct, understanding seems to be crucial to both (a) determining whether someone is a fitting subject of blame to begin with, and (b) mitigating the feelings of anger and resentment that naturally accompany our judgments of blame. It is therefore hardly an optional add-​on for the virtuous person, yet in the virtue ethics literature it seems to have received little or no attention. Finally, more work needs to be done to distinguish the “resentful anger” I have focused on above from the sort of “righteous anger” that plausibly does have a place in the virtuous life. Perhaps righteous anger takes its legitimacy

The Ethics of Understanding


from the fact that it is more neutral or less self-​implicating—​perhaps one only properly feels righteous anger on behalf of someone else, or perhaps on behalf of a cause or principle that one feels important or worthwhile—​and it might be possible to cultivate this in a way that does not have the poisonous effects associated with resentment. Of course, articulating the subtle distinctions between righteous anger and resentful anger will likely be a very challenging task. But it seems like a project well worth pursuing, if we are to have a better understanding of our moral lives.

References Alston, William. 1989. “The Deontological Conception of Epistemic Justification.” In his Epistemic Justification: Essays in the Theory of Knowledge. Ithaca, NY: Cornell University Press. 115–​52. Arpaly, Nomy. 2003. Unprincipled Virtue. New York: Oxford University Press. Bommarito, Nicholas. 2011. “Bile & Bodhisattvas:  Śāntideva on Justified Anger.” Journal of Buddhist Ethics 18: 357–​81. Bonhoeffer, Dietrich. (1937) 1995. The Cost of Discipleship. New York: Touchstone. Coates, D. Justin, and Neal Tognazzini. 2013. “The Contours of Blame.” In Blame: Its Nature and Norms. Ed. D. Justin Coates and Neal Tognazzini. New York: Oxford University Press. 3–​26. —​—​—​. 2014. “Blame.” In The Stanford Encyclopedia of Philosophy. Summer 2014 ed. Ed. Edward N. Zalta. http://​​archives/​sum2014/​entries/​blame/​. Doris, John, and Stephen Stich. 2014. “Moral Psychology: Empirical Approaches.” In The Stanford Encyclopedia of Philosophy. Summer 2014 ed. Ed. Edward N. Zalta. http://​​archives/​fall2014/​entries/​moral-​psych-​emp/​. Franklin, Christopher. 2013. “Valuing Blame.” In Blame: Its Nature and Norms. Ed. D. Justin Coates and Neal Tognazzini. New York: Oxford University Press. 207–​23. Goldberg, Sanford. Forthcoming. “Should Have Known.” Synthese. Published online 30 January 2015. Goldman, Alvin. 2006. Simulating Minds: The Philosophy, Psychology, and Neuroscience of Mindreading. New York: Oxford University Press. Gordon, Robert. 2000. “Simulation and the Explanation of Action.” In Empathy and Agency: The Problem of Understanding in the Human Sciences. Ed. Hans Kogler and Karsten Stueber. Boulder, CO: Westview Press. 62–​82. Hanh, Thich Nhat. 2005. Being Peace. 2nd ed. Berkeley, CA: Parallax Press. Harman, Elizabeth. 2015. “The Irrelevance of Moral Uncertainty.” In Oxford Studies in Metaethics 10. Ed. Russ Shafer-​Landau. New York: Oxford University Press. 53–​79. —​—​—​. N.d. “Ethics Is Hard. What Follows?” Heal, Jane. 1986. “Replication and Functionalism.” In Language, Mind, and Logic. Ed. Jeremy Butterfield. New York: Cambridge University Press. 135–​50.



Hughes, Paul. 2015. “Forgiveness.” In The Stanford Encyclopedia of Philosophy. Ed. Edward N. Zalta. Spring 2015 ed. http://​​archives/​spr2015/​ entries/​forgiveness/​. Jones, Karen. 1999. “Second-​Hand Moral Knowledge.” Journal of Philosophy 96: 55–​78. Kant, Immanuel. (1785) 2012. Groundwork of the Metaphysics of Morals. Trans. Mary Gregor. 2nd ed. New York: Cambridge University Press. Markovits, Julia. 2010. “Acting for the Right Reasons.” Philosophical Review 119: 201–​42. McGeer, Victoria. 2013. “Civilizing Blame.” In Blame: Its Nature and Norms. Ed. D. Justin Coates and Neal Tognazzini. New York: Oxford University Press. 162–​88. Murphy, Jeffrie. 2003. Getting Even:  Forgiveness and Its Limits. New  York:  Oxford University Press. Nussbaum, Martha. 1993. “Equity and Mercy.” Philosophy and Public Affairs 22: 83–​125. Smith, Holly. 1983. “Culpable Ignorance.” Philosophical Review 92: 543–​71. Strabbing, Jada. Forthcoming. “Responsibility and Judgment.” Philosophy and Phenomenological Research. Wallace, R. Jay. 2011. “Dispassionate Opprobrium:  On Blame and the Reactive Sentiments.” In Reasons and Recognition: Essays on the Philosophy of T. M. Scanlon. Ed. R. Jay Wallace, Rhul Kumar, and Samuel Freeman. New York: Oxford University Press. 348–​72. Watson, Gary. 1987. “Responsibility and the Limits of Evil: Variations on a Strawsonian Theme.” In Responsibility, Character, and the Emotions: Essays in Moral Psychology. Ed. F. Schoeman. New York: Cambridge University Press. 256–​86. —​—​—​. 2004. Agency and Answerability. Oxford: Clarendon Press. Wolf, Susan. 2011. “Blame, Italian Style.” In Reasons and Recognition:  Essays on the Philosophy of T. M. Scanlon. Ed. R. Jay Wallace, Rhul Kumar, and Samuel Freeman. New York: Oxford University Press. 332–​47.


Understanding and Structure Allan Hazlett

7.1. Introduction In the Phaedrus, Socrates sympathetically describes the ability “to cut up each kind according to its species along its natural joints, and to try not to splinter any part, as a bad butcher might do.” (265e)1 For example, you might think that the periodic table of the elements is more true to the world’s chemical structure than the ancient quadripartite system of earth, air, fire, and water. Examples like this suggest that “carving nature at the joints,” whatever exactly this expression metaphorically represents, is a valuable intellectual achievement. However, many have wondered whether there exists structure of the sort that Plato’s metaphor requires—​whether there are “natural joints” along which kinds might be cut up. And, just as important, they have wondered where—​i.e., in what domains of inquiry—​such structure exists. Prior to the more or less epistemological questions of the nature and value of “carving nature at the joints,” there is a metaphysical question about whether and where nature has “joints” at which to be “carved.” In contemporary epistemology, many argue that understanding is (at least partly) constituted by a “grasp” of explanatory structure. For example, understanding why global mean temperature is increasing will be (at least partly) constituted by a “grasp” of the explanation of the fact that global mean temperature is increasing—​to understand this, you would need to see how the rise in CO2 is connected to the rise in temperature, e.g., by appreciating how the “greenhouse effect” works. Only in this case would you genuinely 1.   Alexander Nehamas and Paul Woodruff ’s translation, from Plato 1997.



understand why global mean temperature is increasing. In any event, contemporary epistemologists generally agree that understanding is a valuable intellectual achievement. But just as we can ask whether and where nature has “joints” at which to be “carved,” we can ask whether and where there exists explanatory structure to be “grasped.” This chapter explores that question—​ with a large part of the discussion devoted to clarifying the nature and value of understanding, and the relationship between understanding and metaphysical structure. Before going any further, we must attempt to cash out the metaphor of “grasping” explanatory structure. It seems to me that the best way to do this is by appeal to the notion of representation, which is at least less metaphorical than the notion of “grasping.” Here is what I propose as an articulation of the idea that understanding is (at least) partly constituted by a “grasp” of explanatory structure: The representation view: Understanding why p is a species of correctly representing E as the explanation of the fact that p. A few points of clarification. First, although the present formulation speaks only of “understanding why p,” something analogous could be said of other kinds of understanding that intuitively involve “grasping” explanatory structure, e.g., understanding how something happened or how something works or understanding some phenomenon, process, or event. In any event, on the representation view, understanding why p is a species of correct—​i.e., accurate—​representation.2 Second, the present formulation is neutral on the context-​sensitivity of explanation attributions. It is consistent with the idea that different things could count as “the explanation of the fact that p” in different contexts, depending on, for example, assumed background conditions or interests. However, the present formulation is also consistent with the idea that explanation attributions are not context-​sensitive.3

2.   Cf. Grimm and Ahlstrom-​Vij 2013. We can now see one reason why the metaphor of “grasping” was nonideal: literally grasping something, e.g., getting a hold of it with your hands, is not a way of representing it. Similarly, Linda Zagzebski’s (1996) evocative phrase “cognitive contact with reality,” which suggests correct representation, is potentially misleading: being in contact with something is not a way of representing it. Acquaintance emerges as a hard case. 3.   In all this, the present formulation treats the relationship between understanding and explanation as analogous to the relationship between propositional knowledge and truth (cf. §7.2.1).

Understanding and Structure


Third, although the present formulation implies that there is a unique (perhaps contextually determined) explanation of any understood fact, the spirit of the representation view would survive an amendment of “the explanation” to “an explanation.” The leading idea behind the representation view is that understanding is a species of correct representation. Fourth, “explanation” is ambiguous. In a metaphysical sense, explanation is a relation between one set of events, facts, or truths (the explanans), on the one hand, and another set of events, facts, or truths (the explanandum), on the other. In this sense, for example, the fire explains the smoke. In a representational sense, explanation is a relation between someone or something that represents the explanans, on the one hand, and the explanandum, on the other. In this sense, for example, the forest ranger explains the smoke, by appeal to the fire. In the present formulation, “explanation” is used in its metaphysical sense. Fifth, the present formulation is meant to be neutral on the nature of explanation. In particular, it is meant to be neutral on the possibility of noncausal explanations. Consider, for example, the question of why lying is wrong, which does not obviously suggest an answer in causal terms. In any event, in what follows I  assume that the representation view is true.4

7.2. Consequences of the Representation View This section explores some consequences of the representation view (§7.1), by way of clarifying and articulating what it says. 7.2.1. Understanding and Knowledge The representation view (§7.1) is neutral on the question of whether understanding is a species of knowledge (unless knowledge is coextensive with correct representation). The representation view is consistent with the following two platitudes in the epistemology of understanding: (1) Understanding why p is distinct from knowing that p. (2) No amount of propositional knowledge is sufficient for understanding.

4.   For defenses of the representation view, see Trout 2002; Grimm 2011a; 2014; Strevens 2013; Khalifa 2013.



The truth of these platitudes has led some to conclude that understanding is not a species of propositional knowledge (cf. Zagzebski 2001; Kvanvig 2003). But that conclusion is not entailed by (1)  and (2). Understanding might be constituted by propositional knowledge plus something that is not propositional knowledge, in which case understanding would be a species of propositional knowledge, whilst (1) and (2) remained true. In any event, the representation view is consistent with, but does not entail, that understanding why p is knowing that p because q, that understanding why p is knowing why p, and that understanding why p is knowing the explanation of the fact that p. It seems obvious that propositional knowledge is at least a cause of understanding, even if it is not a constituent of understanding. Someone who is propositionally ignorant about climate change, atmospheric CO2 levels, astronomy, the nature of heat, and the like, is not a likely candidate for coming to understand why global mean temperature is increasing. A necessary step on the road to understanding is the acquisition of propositional knowledge about the thing you are trying to understand. (This is occasionally missed by epistemologists and philosophers of education who are dismissive of the value of learning “mere” facts.) Of particular importance here are counterfactual (propositional) knowledge—​e.g., that were there less CO2 in the atmosphere, less heat would be trapped—​and general (propositional) knowledge—​e.g., that increases in atmospheric CO2 have generally coincided with increases in mean global temperature. 7.2.2. Understanding Is Factive Knowing that p requires that it is true that p; in this sense knowledge is factive. Understanding is factive in what superficially looks like the same sense: understanding why p requires that it is true that p. You cannot understand why p unless it is the case that p.5 No one understands why Nixon killed Kissinger, because Nixon did not kill Kissinger. However, given the representation view (7.1), understanding is factive in a different sense, and one more analogous to the sense in which knowledge is factive. Understanding, like propositional knowledge, is a species of correct representation. Just as knowing that p requires correctly believing that p, which 5.   Compare: knowing that p often precedes understanding why p (although these two things sometimes come together, as when we come to know something only when we come to understand how it could be so).

Understanding and Structure


entails that p, understanding why p requires correctly representing E as the explanation of the fact that p, which entails that E is the explanation of the fact that p. Just as you can only know truths, you can only understand real explanatory structure. In cases of knowledge, the person who knows must represent (as true) a true proposition; in cases of understanding, the person who understands must represent (as real) real explanatory structure.6 Understanding requires a correct, and not a mistaken, representation of the relevant explanatory structure. Consider someone who connects increases in CO2 to increases in temperature via the principle that CO2 molecules are hotter than naturally occurring O2 molecules. This person does not understand why global mean is increasing; she misunderstands the rise in temperature. We do naturally speak of someone’s having a “mistaken understanding” of something. But compare the claim that some falsehood was true “for all I  knew” or that some falsehood is true “for all they know.” And compare the sense of “knowledge” employed by scholars in many disciplines, on which “knowledge” refers (roughly) to an entrenched belief system, as in: “[t]‌raditional Diné (Navajo) knowledge teaches that there are two great systems that are responsible for all processes and existence on earth.” (Semken and Morgan 1997, 109) Epistemologists employ a different sense of “knowledge,” on which knowledge is factive.7 Just as we often usefully speak of false knowledge, we often usefully speak of mistaken understanding. But in whatever sense knowledge cannot be false, understanding cannot be mistaken. You might object that understanding is compatible with having many false beliefs about the thing understood (cf. Kvanvig 2003; Elgin 2007; 2009). Maria might understand how whales evolved from land mammals, even if she falsely thinks that whale sharks are whales. But this is orthogonal to the present point, which implies that Maria must correctly represent the explanatory

6.   It seems to me this goes a long way toward explaining the appeal of propositional knowledge and understanding. Contrast the present conception of propositional knowledge with Richard Rorty’s (1991) view that “knowledge is, like truth, simply a compliment paid to the beliefs we think so well justified that, for the moment, further justification is not needed.” (24) What Rorty calls knowledge is not a species of correct representation—​and for this reason it is valuable in a quite different way than knowledge on the present conception. It is one thing to want beliefs that have been paid a compliment by your community, and that to that extent put you in solidarity with said community; it is quite another thing to want to represent things correctly. 7.   Hazlett 2010; 2012; cf. Goldman 2002, 183–​85 and Kusch 2009, 72–​73. The difference between these two senses of “knowledge” is related to, though not the same as, the Wittgensteinian distinction between knowledge and certainty: the former consists of conclusions grounded on evidence and arguments; the latter provides an ungrounded “background” or “framework” for our thinking and inquiring.



structure relevant to whale evolution. Believing that whale sharks are whales does not entail that one does not correctly represent the connections between land mammals and whales. Understanding is factive in the sense that it is a species of correct representation, not in the sense that those who understand are omniscient about their areas of understanding. (We should concede that understanding can be better or worse, and that although Maria understands mammalian evolution well enough to count as understanding, full stop, she would understand better were she not mistaken in her belief about whale sharks.)8 You might object that it is possible to understand a false theory, such as astrology, or the theory of phlogiston. But this is consistent with the factivity of understanding, for what one understands, in such cases, is the false theory, of which one has a correct representation (cf. Elgin 2007). There is a crucial difference between the astrologer, who does not understand human affairs, in virtue of wrongly representing their connection to astronomical phenomena, and the sociologist of astrology, who does understand the astrologer’s theory, in virtue of correctly representing how that theory works. And there is the same crucial difference between the phlogiston theorist, who does not understand combustion, in virtue of wrongly representing how it works, and the historian of science, who does understand the phlogiston theorist’s theory, by correctly representing the principles and components of that theory.9 You might object that it is possible to understand something by using a model or approximation that is, strictly speaking, false (cf. Elgin 1996, 122–​29; 2004; 2007). So, for example, it is possible to understand molecular bonding by picturing molecules as composites of extended and colored spheres. But molecules are not composites of extended and colored spheres. However, this is consistent with the factivity of understanding, for in this case you do not wrongly represent the physical structure of molecules. You correctly represent the physical structure of molecules, by intentionally and knowingly using an 8.   It also seems (cf. Hazlett forthcoming) that understanding attributions are contrastive: Maria might understand whale evolution, relative to someone entirely ignorant of whale evolution, but not relative to an expert evolutionary biologist. 9.   We can thus note an ambiguity in saying, for example, that someone “understands physics.” We could mean that she understands the phenomena that physicists attempt to explain—​e.g., the existence of matter, the structure of the universe—​or we could mean that she understands the theories that physicists propose in their attempts to explain said phenomena. A skeptic who thought physics was bunk could understand physics in the second sense, but not in the first. This ambiguity arises in virtue of the fact that the names for some topics, disciplines, and subject matters are ambiguous as between certain phenomena (in the broadest possible sense of “phenomena”) and our representations of those phenomena; consider “biology,” “chemistry,” “history,” etc.

Understanding and Structure


incorrect representation of them—​in this case, a model that represents them as extended and colored spheres. What do I mean by saying that you “use” an incorrect representation? Compare:  a detective shows you a drawing of a criminal suspect, and you say that the person you saw looked like that, but with a moustache. You correctly represent the subject, by intentionally and knowingly using an incorrect representation of him—​in this case, the drawing of him without a moustache. The case of understanding through the use of models and approximations is relevantly similar to the case of understanding through the use of metaphor. According to David Foster Wallace, Roger Federer’s forehand “is a great liquid whip.” This metaphor can enable us to understand Federer’s forehand better: his forehand is a whip, delicate and fast, but it is a liquid whip, with none of the clunky solidity of a nonliquid tennis racket, and it is a great liquid whip, one whose force and power seems to extend meters beyond Federer’s visible arm-​and-​racket, both as it uncoils in windup and as it guides the ball across the net. As above, we correctly represent Federer’s forehand by intentionally and knowingly using an incorrect representation of it—​in this case, the sentence “Federer’s forehand is a great liquid whip.” Is a correct representation that uses, in the present sense, an incorrect representation worse, in some way, than a correct representation that doesn’t? Is understanding through the use of models and approximations, or through the use of metaphor, somehow less of an intellectual achievement, for example, than understanding by other means? This is unclear. There seems no better description of Federer’s forehand than that it is a great liquid whip. In any event, it seems possible that, at least in some cases, understanding through the use of incorrect representations may be the only kind of understanding available, either absolutely or for creatures like us. But none of this speaks against the factivity of understanding, as understood here, which is analogous to the factivity of knowledge. 7.2.3. Understanding Is Threatened by Skepticism At the turn of the century, Linda Zagzebski (2001) lamented the fact that understanding was unpopular in epistemology. Her argument is, in part, responsible for the current epistemological popularity of understanding. She argued that the epistemological preoccupation with Cartesian skepticism explained the unpopularity of understanding, since only propositional knowledge, and not understanding, is threatened by Cartesian skepticism. The reason for this, so the argument goes, is that understanding is



transparent to reflection in a way that propositional knowledge isn’t: you cannot in general determine, by reflection alone, whether you possess propositional knowledge (this is one of the lessons of Cartesian skepticism), but you can in general determine, by reflection alone, whether you possess understanding. However, given the representation view (§7.1), it is not plausible that you can in general determine, by reflection alone, whether you possess understanding. Consider someone who thinks that CO2 molecules are hotter than O2 molecules. She learns that the rising temperature is correlated with increased atmospheric CO2, and insightfully connects the hotness of CO2 with the changes in temperature. She thinks she understands why global mean temperature is rising—​it all makes sense to her; she has that distinctive feeling of “getting it”—​but she does not understand. Can she determine, by reflection alone, that she does not understand? Reflection is not what she needs—​what she needs is to have her mistaken picture of the connection between CO2 and temperature corrected, and in particular her false view that CO2 molecules are hotter than O2 molecules. But reflection will not disclose the falsity of that view. Just as we can imagine a malicious demon who aims to deceive us about the truth, we can imagine a malicious demon who aims to deceive us about explanatory structure. Descartes’s demon succeeds by causing his victim to believe propositions that aren’t really true. Our demon succeeds by causing his victim to represent explanatory structure that isn’t real. Here’s how we can imagine this deception working. Using our world as a template, and focusing on explanatory connections between events for the sake of simplicity, for every event x apparently explained by event y, the demon constructs an undetectable event z, and makes it the case that z, and not y, explains x, while leaving the appearances intact. Appearances of explanations, in the demon’s world, are thus systematically illusory. Our counterparts, in such a world, do not understand why anything happens, but things seem to them exactly as things seem to us. Most philosophers think the possibility that we are systematically deceived about the truth is a “remote,” and not a “live,” possibility. Is the possibility that we are systematically deceived about explanatory structure likewise “remote” and not “live”? One possibility seems more plausible in the case of structure than in the case of truth: that there is no structure, either in general or in particular domains, and thus no possibility of understanding, either in general or in those particular domains. We’ll return to this idea below (§7.4).

Understanding and Structure


7.2.4. Understanding Is Apt to Be Transmitted It is a platitude in contemporary epistemology that understanding cannot be transmitted through testimony—​or, at least, that the transmission of understanding through testimony is somehow more difficult or more problematic than the transmission of propositional knowledge through testimony. Given the representation view (§7.1), however, it is unclear why this would be so. Let me explain. There are ways of thinking about understanding that would make it difficult to see how understanding could be transmitted through testimony. Consider: The feeling view: Understanding is a species of feeling; in particular, it is constituted by the distinctive feeling of insight, which characterizes the “Aha!” or “Eureka!” moment in inquiry. The acquaintance view: Understanding is a species of acquaintance (e.g., direct awareness or historical familiarity) with the thing understood. If either of these views is true, then it is hard to see how understanding could be transmitted through testimony. Feelings do not seem to be the kind of things that can be transmitted through testimony—​I can describe and explain my feelings to you, but I cannot transmit them to you, in the way that I can transmit my beliefs to you. Even if you come to feel as I do, it will not be on the basis of my feeling as I do, but because of the operation of sympathy. In the same way, acquaintance does not seem to be the kind of thing that can be transmitted through testimony—​no matter how much I say about Mexico City (with which I am acquainted), and no matter how much you learn from me, you will not become acquainted with Mexico City on the basis of my testimony. Feelings and acquaintance, unlike belief, cannot be transmitted through testimony. Locke writes that “we may as rationally hope to see with other Mens Eyes, as to know by other Mens Understandings.” (Essay Concerning Human Understanding, I.iv.23) The comparison suggests not just the disvalue of testimonial knowledge, but its impossibility. His reference in the same passage to “[t]‌he floating of other mens Opinions in our brains” suggests that there is something metaphysically suspect about the whole idea of transmitting knowledge from one person to another—​that at best someone else’s beliefs might take up residence in me, without really being mine. This suggestion



has found few adherents when it comes to propositional knowledge, but, as I mentioned above, there is a contemporary consensus that testimonial understanding is difficult, problematic, or impossible. The reason, it seems to me, may derive form the appeal of the feeling view and the acquaintance view. Just as other people’s eyes cannot be in my head, other people’s feelings and relations of acquaintance cannot be mine. But if these are essential to understanding, then understanding is likewise not capable of being transmitted from person to person. The problem is not that there is merely something unappealing about testimonial understanding—​it is that the very idea is metaphysically problematic. For it is not just that there is something unappealing about, for example, being testimonially acquainted with Mexico City—​it is that there is simply no such thing as being testimonially acquainted with Mexico City. As formulated, these feeling view and the acquaintance view are compatible with the representation view, so long as there are species of feeling and species of acquaintance that are also species of representation. But the representation view obviously entails neither of these views. And it does not sit well with them either. The representation of explanatory structure intuitively requires neither a feeling of insight nor acquaintance with said structure. Once we start thinking of understanding as a species of representation, it is much more natural to think of the feeling of insight as something that sometimes, but not always, accompanies the acquisition of understanding, and it is much more natural to think of acquaintance with the thing understood as one, but not the only possible, basis for understanding. However, if the representation view is true, and neither the feeling view nor the acquaintance view is true, then it is mysterious why there would be anything problematic about the transmission of understanding through testimony, by contrast with the transmission of propositional knowledge through testimony. Both understanding and propositional knowledge are species of correct representation: what then would prevent you from representing such-​and-​such content on the basis of the fact that I represent such-​ and-​such content? Representations can generally be copied: I can believe what you believe, draw what you draw, assert what you assert, and so on.10 Why then can I not represent E as the explanation of the fact that p, on the 10.   However, I cannot see what you see. But I can believe the representational content of your episodes of seeing, so long as your descriptions of them are adequate. So, it seems to me, although the experiential aspect of sense perceptions cannot be transmitted through testimony, their representational aspect can be.

Understanding and Structure


basis of the fact that you represent E as the explanation of the fact that p? Given the representation view, the burden is on the defender of the view that understanding cannot be transmitted through testimony to argue for this and explain why it would be so.

7.3. Understanding Requires “Carving Nature at the Joints” I now want to spell out for another consequence of the representation view:  that understanding requires “carving nature at the joints” (§7.1)—​at least given a plausible cashing out of that metaphor. This will shed further light on the nature and value of understanding, and set the stage for a discussion of where understanding is possible (§7.4). 7.3.1. Explanatory Structure and “Natural Joints” We encountered two kinds of “structure” above (§7.1)—​the kind of structure implied by Plato’s metaphor of “carving nature at the joints” and the kind of structure implied by the contemporary epistemological metaphor of “grasping” explanations (which I  called “explanatory structure”). In this section I will spell out a connection between these two kinds of structure. Ted Sider (2009) writes that: The world has an objective structure. . . . [C]‌ommunities that choose the wrong groupings may get at the truth, but they nevertheless fail badly in their attempt to understand the world. . . . There is more to be discovered [than the truth], more that is mandatory for inquirers to think about. The world has objective streaks in it; it has structure. . . . You can state truths if you don’t speak in terms of this structure, but you miss out; you are deficient along one of the main axes of cognitive success. (398–​99) Sider’s account is motivated by the idea that some, but not all, categories pick out properties that are, in a difficult-​to-​articulate sense, part of the world’s structure. Compare the property of being hydrogen with the property of being hydrium, where something is hydrium iff it is either hydrogen on or before 1 January 2016 or helium after 1 January 2016. You might think that the property of being hydrogen, but not the property of being hydrium, is part of



the world’s structure.11 Others might put this by saying that the property of being hydrogen, but not the property of being hydrium, is natural or sparse, that hydrogen, but not hydrium, is a natural kind, or that being hydrogen, but not being hydrium, is a real property (cf. Quine 1969; Armstrong 1978; Lewis 1983; 1986; Sider 2009; 2011; Mellor 2013). As David Lewis puts it, natural properties are those “whose sharing makes for resemblance, and the ones relevant to causal powers,” (1983, 13) by contrast with those nonnatural properties that are “as gruesomely gerrymandered . . . as you please. They pay no heed to the . . . joints, but carve things up every which way.” (1986, 59) We should adopt these three criteria, with a slight amendment, in line with our neutrality on the possibility of noncausal explanations (§7.1). Let’s say that jointy properties are those such that (1) individuals that share jointy properties resemble one another, (2)  jointy properties partly constitute explanations, and (3)  the boundaries of jointy properties are not arbitrary (by contrast with the sort of properties exemplified by the property of being hydrium).12 On this characterization, two related points of clarification. First, Lewis (1983; 1986) assumes that natural properties must be qualitative and intrinsic; the present characterization makes no such assumption. Second, the present characterization does not assume that jointy properties provide the ultimate or fundamental ground of the explanatory powers of individuals. Given these two points, note well that the present characterization allows for the possibility of nonfundamental (“higher-​level”) jointy properties and for the possibility of dispositional jointy properties. This characterization of jointy properties leaves open the question of whether jointy properties are found only in nature (as opposed to artificial spaces), of whether they are all and only those properties posited by the natural sciences (as opposed to the social sciences or the humanities), and of whether the properties posited by the natural sciences are indeed jointy. (This is why we are stuck with the awkward “jointy” as opposed to the elegant “natural.”) Jointy properties partly constitute explanations. This criterion can be illustrated with an example. Why did the Hindenburg explode? Because it was

11.   Nothing I say here depends on the assumption that hydrogen, but not hydrium, is part of the world’s structure—​this is just an example used for the purposes of illustration. Any example of one property that is, and one property that is not, part of the world’s structure would do. 12.   I focus here on properties for the sake of simplicity; our discussion could be broadened to include individuals and logical operations (cf. Sider 2011, chs. 6, 9–​10). I’ve also ignored the issue of whether “naturalness” comes in degrees (Lewis 1983; 1986, 60–​61; Sider 2011, 128–​36; Mellor 2013).

Understanding and Structure


filled with hydrogen (which is flammable). It did not explode because it was filled with hydrium (which was flammable prior to 1 January 2016). The property of being hydrogen partly constitutes the explanation of the Hindenburg’s exploding; the property of being hydrium does not. The date does not have the kind of place in the explanation of the Hindenburg’s explosion that it would have were hydrium a part of the explanation of that event:  the reason the Hindenburg exploded was because it was filled with hydrogen, not because it was filled prior to 1 January 2016 with hydrium. Understanding is a species of correctly representing explanatory structure (§7.1). But given the connection between explanations and jointy properties, correctly representing explanatory structure requires thinking in terms of categories that pick out jointy properties. Correctly representing E as the explanation of the fact that p requires identifying those jointy properties that (partly) constitute E. Therefore, given the representation view, understanding requires thinking in terms of categories that pick out jointy properties. To understand why the Hindenburg exploded, you must know that it was filled with hydrogen. Someone who did not know this would not have “grasped” the explanatory connection between the contents of the Hindenburg and its catching fire. So I think Sider is right to say that choosing “the wrong groupings” can prevent you from understanding something. The Hindenburg was filled with hydrium; that is a truth to be known. But without the category hydrogen, you cannot understand why the Hindenburg exploded. Correctly representing explanatory structure requires thinking in terms of categories that pick out jointy properties—​and, therefore, understanding requires “carving nature at the joints.” 7.3.2. “Carving Nature at the Joints” Is Optional However, the significance of “carving nature at the joints” should not be overstated. Sider (2011) writes that “joint-​carving thought . . . is a constitutive aim of the practice of forming beliefs, as constitutive as the more commonly recognized aim of truth.” (61) I shall argue that this is implausible. Assume that propositions are structured individuals that have properties as constituents, and define a jointy proposition as one such that all the properties that constitute it are jointy.13 And just as a true belief is a belief in a true proposition, let us define a jointy belief as a belief in a jointy proposition. 13.   Once the jointiness of individuals and logical relations are taken into account (see footnote 12), we can say that a jointy proposition is one such that all of its constituents are jointy. Note that the definition of a jointy proposition would be different if properties are sets of possibilia; see Lewis 1986, 61.



In the case of belief, on the most plausible reading of “constitutive aim,” a constitutive aim is a normative standard that flows from the essential nature of an attitude or practice (Sider 2011, 61). Supposing truth is a constitutive aim of belief, is jointiness also a constitutive aim of belief ?14 It will be instructive to consider the evidence that people offer for thinking that truth is a constitutive aim of belief: If in full consciousness I could acquire a ‘belief ’ irrespective of its truth, it is unclear that before the event I could seriously think of it as a belief, i.e. as something purporting to represent reality. (Williams 1973, 148) In order for a propositional attitude to be an attitude of belief, it cannot represent itself as wholly unaccountable to truth or evidence. . . . A self-​ representation of certain of one’s attitudes as ‘aiming at’ truth is partially constitutive of belief. (Railton 1997, 57–​59) Truth is not an optional end for first-​personal doxastic deliberation. [W]‌ithin the first-​person deliberative perspective  .  .  .  the question whether to believe that p seems to collapse into the question whether p is true . . . . [O]ne cannot settle on an answer to the question whether to believe that p without taking oneself to have answered the question whether p is true. (Shah 2003, 447) These phenomena, so the argument goes, suggest that truth is a constitutive aim of belief. But no analogous phenomena exist that might suggest the same relationship between belief and jointiness. You can acquire, in full consciousness, a belief irrespective of whether it is jointy (e.g., I now believe that this glass of water contains many atoms of hydrium), you can represent your beliefs to yourself as not jointy (e.g., my belief just described), and forming jointy beliefs is an optional end for first-​personal doxastic deliberation (e.g., I can ask myself whether to believe that this glass of petrol contains any atoms of hydrium, while explicitly setting aside the question of whether such a belief would be jointy). To put all this another way, considerations of truth and falsity are the “right kind of reasons” for forming, revising, and sustaining beliefs, whereas considerations of jointiness are the “wrong kind of reasons” for forming, revising, and sustaining beliefs (cf. Hieronymi 2005; Shah 2006). That a proposition is nonjointy is not the kind of reason on the basis of which you 14.   I speak here of the constitutive aim of belief rather than (following Sider) the constitutive aim of the practice of forming beliefs. Nothing will hinge on this simplification.

Understanding and Structure


could disbelieve it or stop believing it. It is, by contrast, the kind of reason on the basis of which you could decide not to inquire about whether it was true, or the kind of reason on the basis of which you could desire not to find out whether it was true. And to put this yet another way, although believing that p commits you to the truth of the proposition that p, it does not commit you to the jointiness of the proposition that p (cf. Hieronymi 2005). Jointiness is not a constitutive aim of belief, not in the way that truth seems to be. Evaluating beliefs and believers vis-​à-​vis the standard of jointiness is one way we may evaluate them, but this standard does not flow from the essential nature of belief. This suggestion isn’t decisive; perhaps there is some alternative argument that jointiness is a constitutive aim of belief. But it is unclear what that argument would look like. Sider (2011) also maintains that: The goal of inquiry is not merely to believe truly (or to know). Achieving the goal of inquiry requires that one’s belief state reflect the world, which in addition to lack of error requires one to think of the world in its terms, to carve the world at the joints. (61) And that thinking about structure is “mandatory for inquirers.” (2009, 398) I shall argue that this, too, is implausible. What does it mean to say that something is “the goal of inquiry”? We could understand this as a psychological claim, to the effect that inquiry essentially requires a desire or intention to form jointy beliefs. But this is implausible. You can engage in inquiry without having any such desire or intention. I can inquire about whether this glass of petrol contains any hydrium atoms, for example, by investigating the date and the chemical composition of petrol. Moreover, the goal of forming jointy beliefs is optional for inquiry, in a way that the goal of forming true beliefs isn’t. You can’t coherently claim to be inquiring unless you’re trying to arrive at true beliefs. “Inquiry” that doesn’t seek truth is no inquiry at all, but rather pretended or feigned inquiry. (This is not to say that we might have good reason to pretend or feign inquiry; consider detectives seeking legally admissible evidence or a quality control inspector doing routine checks.) Inquiry that doesn’t seek jointy beliefs, by contrast, whatever you might way to say about it, is still inquiry. The alternative to that psychological reading of “the goal of inquiry” is a normative reading, on which jointiness is a normative standard that flows from the essential nature of inquiry. (In this case “the goal of inquiry” is the “constitutive aim” of inquiry, given how we understood “constitutive aim,”



above.) But this too is implausible. My inquiry about whether this glass of petrol contains any hydrium atoms, for example, ought not be evaluated by this standard—​I did not set out to form a jointy belief, so why would my failure to do so count against my efforts? Suppose, to strengthen the intuition here, that I had good reason to seek knowledge of a nonjointy proposition: a prize was offered for those who could locate samples of hydrium. Someone who had set out to “carve nature at the joints,” and failed, would deserve some kind of negative evaluation, but someone who for good reason makes no such attempt should not be evaluated as having failed to do something that she did not set out to do. Some instances of inquiry do seem, in some sense, to have the goal of “carving nature at the joints”—​think of a chemist who wants to discover the molecular composition of some substance—​but other instances of inquiry seem indifferent to “carving nature at the joints.” Sometimes philosophers talk of “aims” and “goals” as a way of expressing allegiance to certain values. Sider (2011) says that “it’s better to think and speak in joint-​carving terms.” (61) Could we understand the idea that jointy belief is an “aim of belief ” or “the goal of inquiry” in evaluative terms? There does not seem to be the right kind of connection between belief or inquiry, on the one hand, and the value of jointy belief, on the other. Think of the various goods that belief and inquiry can yield: profit, for example, or pleasure. Profitable and pleasurable belief and inquiry are better than costly and painful belief and inquiry, but this is not aptly articulated by saying that profit and pleasure are “aims of belief ” or “the goals of inquiry.”15 The same, it seems to me, when it comes to jointy belief.16

15.   However, jointiness plausibly provides one species of significance, where significant questions contrast with those that are trivial or uninteresting (Baril 2010; Grimm 2011b; Treanor 2014): significant questions are those answered by a jointy proposition. 16.   Consider Sider’s (2011) idea that, of people who think of the world in terms of “grue”-​like predicates, “it is almost irresistible to describe these people as making a mistake,” consisting in the fact that “they’ve got the wrong concepts,” and so are “carving up the world incorrectly.” (2) This is implausible for reasons similar to those articulated above: just as belief and inquiry do not plausibly “aim” at “carving nature at the joints,” neither does mere concept possession. False beliefs count as mistaken because beliefs are representations of truth—​to believe that p is to represent the proposition that p as true. But instances of concept possession are not representations at all—​and certainly not representations of jointiness. I possess the concept grue, but I am not thereby mistaken about anything. If I believed that grue picked out a jointy property, then I would be mistaken—​but that would be the familiar mistake of false belief. However, there is an alternative sense of “mistake,” on which to make a mistake is just to miss out on something good. “It was a mistake for you not to try the spring rolls,” I say, after tasting them. Here “to make a mistake” means (something like) to do something imprudent or foolish. Perhaps, then, in this sense, a community that chooses non-​joint-​carving concepts is making a mistake—​they

Understanding and Structure


7.3.3. Understanding Is Optional In the same way that “carving nature at the joints” is optional (§7.3.2), understanding is optional. Let me explain. What I mean is that understanding is not plausibly an “aim of belief ” or “the goal of inquiry.” The reasons for this are of the same kind as those given above, concerning “carving nature at the joints.” First, none of the things that make us think truth is an “aim of belief ” have analogs when it comes to understanding. You can acquire, in full consciousness, a belief irrespective of its having any connection to understanding, you can represent your beliefs to yourself as having no connection to understanding, and acquiring understanding is an optional end for first-​personal doxastic deliberation. There is no reason to think that understanding is a normative standard that flows from the essential nature of belief. Second, on neither a psychological nor a normative reading of “the goal of inquiry” is understanding plausibly “the goal of inquiry.” You can engage in inquiry without desiring or intending to acquire understanding, and in some such cases, a negative evaluation of your inquiry on the grounds that you did not acquire understanding would not be appropriate. Again, talk of “aims” and “goals” may be code here for the expression of evaluative allegiance, e.g., allegiance to the value of understanding. That understanding is not an “aim of belief ” or “the goal of inquiry” does not, on my view, imply that it is not valuable. Nor does it imply that believers and inquirers ought not seek understanding, nor that they ought not generally or always seek understanding. What it implies, I think, is that the value of understanding does not flow from the essential nature of belief and inquiry. However, it is important that, if talk of “aims” and “goals” is evaluative, then the idea that understanding is an “aim of belief ” or “the goal of inquiry” can do nothing to explain or justify the idea that understanding is valuable.

7.4. When Is Understanding Possible? Let’s now return to our original question of whether and where there exists explanatory structure to be “grasped.” (§7.1) Given the representation view, this is the question of whether and where there exist explanations to be represented. Where there is no explanatory structure, understanding is impossible.

are imprudently or foolishly passing up the opportunity to do something valuable, namely, to “carve nature at the joints.”



And given the connection between explanations and jointy properties (§7.3.1), explanations exist only where jointy properties exist. Where there are no jointy properties, therefore, understanding is impossible. Given the representation view, this is not surprising. Think of the analogous claims about propositional knowledge. Propositional knowledge is a species of correct representation; more precisely, it is a species of correct belief. A belief that p represents the proposition that p as true, and thus a belief that p is correct if and only if it is true that p. Where there is no truth, therefore, propositional knowledge is impossible. Consider domains that we think of as nonfactual, e.g., those comprising questions of mere taste or preference. Your favorite color is blue; mine is green. There is no question of who is right here—​and so no possibility of correct belief, and so no possibility of knowledge. In the same way, when it comes to domains in which we think there is no explanatory structure, including those in which we think there are no jointy properties, there will be no possibility of the correct representation of explanations, and so no possibility of understanding. You would be right to point out that this conclusion must be qualified. For the representation view was formulated (§7.1) so as to cover all and only cases of understanding why something is the case. I  said above that it could be extended to cover all cases of understanding that intuitively involve “grasping” explanatory structure. And it seems to me that this is a broad and significant epistemological category. But there are certainly important and legitimate uses of the word “understanding” that refer to something else, e.g., as Homer tells us, “a friend with an understanding heart is no whit worse than a brother,” which has no obvious implications vis-​à-​vis explanatory structure. So the impossibility of the present species of understanding does not per se rule out the possibility of other species. (That said, I  will continue to use “understanding” to refer to the present species of understanding.) It is also important to keep in mind that understanding is optional (§7.3.3). If understanding is impossible, it does not mean that our beliefs are doomed not to fulfill their “aim” or that our inquiries are doomed not to achieve their “goal.” The impossibility of understanding would not be an intellectual disaster. I have spoken throughout (e.g., §7.2.2, §7.3.1) of the “existence” and “reality” of explanatory structure (e.g., explanations). But the metaphysical status of such structure is controversial. You might think, for example, that many of the categories we commonly employ pick out socially constructed properties. If the possibility of understanding requires the existence of jointy properties,

Understanding and Structure


does this mean that our thinking that employs such categories cannot amount to understanding? Race properties—​the properties picked out by our racial categories—​are taken by many academics to be socially constructed, and this view is relatively widespread among philosophers of race (see, e.g., Outlaw 1996; Mills 1998; Haslanger 2000; Mallon 2004; Glasgow 2009). What does it mean to say that race properties are socially constructed? First, it means that our racial categories do not pick out properties that exist independent of social practices and institutions—​as they would on a biological realist view according to which racial categories are biological categories. Our common-​sense system of racial classification looks nothing like the genetic reality as understood by contemporary biology and anthropology.17 Second, it means that racial categories are dependent on, and are relative to, social practices and institutions. Different societies employ different systems of racial classification, which determine the race properties that exist for a given society. Here I shall assume social constructivism about race properties, but what I have to say about understanding and socially constructed properties does not depend on this assumption. Race properties seem like constituents of the kind of explanations the representation of which can amount to understanding. For example, suppose implicit racial bias leads a hiring committee to prefer hiring a less qualified white candidate to hiring a more qualified black candidate (call her “Maya”). In such a case, the following seems like a good (at least partial) account of what happened: Maya didn’t get the job because she was black. And someone who represented her being black as the explanation of Maya’s not getting the job would, it seems to me, correctly represent (at least part of ) the explanation of her not getting the job. And someone might understand why she didn’t get the job (at least partly) in virtue of representing her being black as the explanation of her not getting the job. I argued that understanding requires thinking in terms of categories that pick out jointy properties (§7.3.1). Is this compatible with the idea that someone might understand why Maya didn’t get the job by representing her being black as the explanation of her not getting the job? If it is, race properties are jointy properties. Is this compatible with their being socially constructed? 17.   So it is not denied that there exists an underlying genetic reality, studied by biology and anthropology; what is denied is that that reality includes race properties. Consider the uncontroversially socially constructed property of being cool (Haslanger 1990). There may very well be genetic traits shared by the cool—​they are, after all, more athletic than the uncool—​but this does not mean that the cool comprise a natural kind. Similarly, there may very well be genetic traits shared by members of particular races, but this does not mean that racial categories are biological categories.



I think so. Socially constructed race properties meet the three criteria for jointiness (§7.3.1). First, black people resemble one another—​not necessarily genetically (as for the biological realist), but in terms of social status, lived experience, and history. Second, blackness partly constitutes explanations: witness Maya’s not getting the job, which was causally explained by her being black. Third, the boundaries of the property of being black is not arbitrary—​ its boundaries are determined by real social practices and institutions. Socially constructed properties, at least in this case, are jointy properties. Socially constructed explanatory structure, therefore, is real enough to provide for the possibility of understanding. That a domain comprises social constructions is therefore no barrier to the possibility of understanding in that domain. However, consider two people who both represent Maya’s being black as the explanation of her not getting the job. The first is a social constructivist about race—​in classifying Maya as black, she intends to classify her in terms of social status, lived experience, and history, and in representing her being black as the explanation of her not getting the job, she intends to connect this status, experience, and history with that event. The second person is a biological realist about race—​in classifying Maya as black, she intends to classify her in genetic terms, and in representing her being black as the explanation of her not getting the job, she intends to connect her genetic properties with that event. It seems to me that both of these people may understand why Maya didn’t get the job, but that, given the assumption that race properties are socially constructed, the first person understands why Maya didn’t get the job better than the second person. The first person has a better understanding of why Maya didn’t get the job than does the second person. This, in at least two senses. First, the social constructivist’s understanding is fuller or more complete than that of the biological realist—​she understands more than the second person does. Someone who knows both that x happened because y happened and that y happened because z happened seems to know more about why x happened than someone who merely knows that x happened because y happened. Similarly, the social constructivist seems to know more about why Maya didn’t get the job than the biological realist. Knowing more, in this way, seems to make for a fuller or more complete understanding. Second, the biological realist’s understanding of why Maya didn’t get the job contains a mistake, an error, that is no part of the social constructivist’s account.18

18.   It is sometimes said that understanding comes in degrees. That, it seems to me, is different from the idea that some understandings are better than others in virtue of being more complete or less mistaken. My (faster, more reliable, better looking) car is better than your car, but not in virtue of being a car to a greater degree than your car.

Understanding and Structure


Socially constructed explanatory structure is real enough to provide for the possibility of understanding—​but those who recognize the socially constructed nature of such structure will acquire a better understanding of the relevant phenomena than those who do not. It seems to me that something similar can probably be said about other kinds of “constructed” structure—​e.g., fundamental physical properties, as they are understood by a constructive empiricist in the philosophy of science, or properties of things in the phenomenal world, as they are understood by a transcendental idealist, or properties of sets, as they are understood by a constructivist in the philosophy of mathematics.19 In any event, my proposal is that understanding will require the existence of properties that meet the three Lewisian criteria (§7.3.1): they must make for resemblance, constitute explanations, and not be arbitrary. If “constructed” properties have these features, then there are “joints” to be carved, and so explanatory structure to be “grasped.”

7.5. Conclusion Having assumed the representation view (§§7.1–​2), on which understanding is a species of correctly representing explanatory structure, I argued that correctly representing explanatory structure requires thinking in terms of categories that pick out jointy properties (§7.3). I then considered the question of where explanatory structure is to be found, and argued that the fact that a domain comprises social constructions does not per se preclude the possibility of explanatory structure, and therefore understanding, in that domain (§7.4). Ancient philosophers took seriously the possibility that the metaphysical conditions for the possibility of knowledge are never met. Sextus Empiricus sympathetically discusses several pre-​Socratic philosophers—​including Heraclitus, Protagoras, and Parmenides—​who suggest this. Nothing like this is plausible when it comes to our contemporary conception of propositional knowledge, but things are less clear when it comes to understanding. We are familiar with situations in which it seems like understanding is impossible because there is nothing to be understood. Abraham Lincoln and Charles Darwin were born on the same day. Some people might be tempted to look for some meaning or

19.   The case of mathematics is complicated by the question of whether mathematical understanding involves explanations at all; see Lipton (2011) and Shapiro (2011) for discussion.



significance to this, some explanation of why these two important people were born on the same day. But nothing explains why Lincoln and Darwin were born on the same day—​it’s just a coincidence. Conspiracy theorists seek understanding where there is none to be found—​they look for explanatory structure where it does not exist. Against this, you might point out that a complete explanation of the chains of events leading up to Lincoln and Darwin’s births would provide an explanation of why Lincoln and Darwin were born on the same day. Perhaps this is right, but in other cases this is less plausible. I’m made to understand that there is (at least typically) no explanation of why a radioactive atom decays at the precise moment that it decays—​this event is genuinely random. Nothing explains why this atom decayed at such-​and-​such a time, as opposed to a moment later. And so there is no possibility of understanding why this atom decayed at such-​and-​such a time, as opposed to a moment later. The question of whether and where explanatory structure exists (§7.4) is the question of the extent of this type of situation. Suppose that, in some domains but not in others, there is no explanatory structure. The present view—​that the possibility of understanding requires the existence of explanatory structure—​enables us to contrast these two sorts of domains, by saying that the former, but not the latter, are domains in which understanding is possible. Suppose, alternatively, that there is no explanatory structure in any domain. Does this threaten the present view? You might object that understanding is clearly possible, and so the conditions for the possibility of understanding must be met, and so the possibility of understanding must not require the existence of explanatory structure. However, I think what is clear is that we want understanding. But it is not clear that, in this instance, we can get what we want. And the present view articulates what it is that we want: we want to represent real explanations—​to “grasp” explanatory structure and to “carve nature at the joints”—​even if this is impossible.

Acknowledgments I presented versions of this chapter at the University of California, Riverside, in November of 2015, at the workshop “Non-​Alethic Aims of Inquiry” at the University of St Andrews in October of 2014, at the American Philosophical Association Pacific Division Meeting in Seattle in April of 2012, at the University of Cambridge Serious Metaphysics Group in February of 2012, and at the conference “Knowledge, Understanding, and Wisdom” in Bled, Slovenia, in June of 2011. For valuable feedback, thanks to my audiences on those occasions, as well as to Anne Baril, Stephen Grimm, Kareem Khalifa, Ted Sider, and Iain Thompson.

Understanding and Structure


References Armstrong, David M. 1978. Universals and Scientific Realism. Cambridge: Cambridge University Press. Baril, A. 2010. “A Eudaimonist Approach to the Problem of Significance.” Acta Analytica 25: 215–​41. Elgin, Catherine Z. 1996. Considered Judgment. Princeton, NJ:  Princeton University Press. —​—​—​. 2004. “True Enough.” Philosophical Issues 14: 113–​31. —​—​—​. 2007. “Understanding and the Facts.” Philosophical Studies 132: 333–​42. —​—​—​. 2009. “Is Understanding Factive?” In Epistemic Value. Ed. Adrian Haddock, Alan Millar, and Duncan Pritchard. Oxford: Oxford University Press. 322–​30. Glasgow, Joshua. 2009. A Theory of Race. New York: Routledge. Goldman, Alvin I. 2002. Pathways to Knowledge: Private and Public. New York: Oxford University Press. Grimm, Stephen R. 2011a. “Understanding.” In The Routledge Handbook of Epistemology. Ed. Sven Bernecker and Duncan Pritchard. New York: Routledge. 84–​94. —​—​—​. 2011b. “What Is Interesting?” Logos and Episteme 2: 515–​42. —​ —​ —​ . 2014. “Understanding as Knowledge of Causes.” In Virtue Epistemology Naturalized:  Bridges between Virtue Epistemology and Philosophy of Science. Ed. Abrol Fairweather. Heidelberg: Springer. 239–​45. Grimm, Stephen R., and Kristoffer Ahlstrom-​Vij. 2013. “Getting It Right.” Philosophical Studies 166: 329–​47. Haslanger, Sally. 1990. “Ontology and Social Construction.” Philosophical Topics 23: 95–​124. —​—​—​. 2000. “Gender and Race: (What) Are They? (What) Do We Want Them to Be?” Noûs 34: 31–​55. Hazlett, Allan. 2010. “The Myth of Factive Verbs.” Philosophy and Phenomenological Research 80: 497–​522. —​—​—​. 2012. “Factive Presupposition and the Truth Condition on Knowledge.” Acta Analytica 27: 461–​78. ———. Forthcoming. “Moorean Pragmatics, Social Comparisons, and Common Knowledge.” In New Essays on Entitlement. Ed. Nikolaj J. L. L. Pedersen and Peter Graham. New York: Oxford University Press. Hieronymi, Pamela. 2005. “The Wrong Kind of Reason.” Journal of Philosophy 102: 437–​57. Khalifa, Kareem. 2013. “Understanding, Grasping, and Luck.” Episteme 10: 1–​17. Kusch, Martin. 2009. Knowledge by Agreement:  The Programme of Communitarian Epistemology. New York: Oxford University Press. Kvanvig, Jonathan. 2003. The Value of Knowledge and the Pursuit of Understanding. New York: Cambridge University Press. Lewis, David. 1983. “New Work for a Theory of Universals.” Australasian Journal of Philosophy 61: 343–​77.



—​—​—​. 1986. On the Plurality of Worlds. Oxford: Blackwell. Lipton, Peter. 2011. “Mathematical Understanding.” In Meaning in Mathematics. Ed. John Polkinghorne. New York: Oxford University Press. 49–​54. Mallon, Ron. 2004. “Passing, Traveling, and Reality:  Social Construction and the Metaphysics of Race.” Noûs 38: 644–​73. Mellor, David H. 2013. “Nature’s Joints: A Realistic Defence of Natural Properties.” In Classifying Reality. Ed. David S. Oderberg. Oxford: Blackwell. 23–​40. Mills, Charles. 1998. Blackness Visible:  Essays on Philosophy and Race. Ithaca, NY: Cornell University Press. Outlaw, Lucius T. 1996. On Race and Philosophy. New York: Routledge. Plato. 1997. Complete Works. Ed. John M. Cooper and D. S. Hutchinson. Indianapolis: Hackett. Quine, Willard V. O. 1969. “Natural Kinds.” In Ontological Relativity and Other Essays. New York: Columbia University Press. 114–​38. Railton, Peter. 1997. “On the Hypothetical and Non-​hypothetical in Reasoning about Belief and Action.” In Ethics and Practical Reason. Ed. Garrett Cullity and Berys Gaut. New York: Oxford University Press. 53–​79. Rorty, Richard M. 1991. “Solidarity or Objectivity?” In Objectivity, Relativism, and Truth. New York: Cambridge University Press. 21–​34. Semken, Steven C., and Frank Morgan. 1997. “Navajo Pedagogy and Earth Systems.” Journal of Geoscience Education 45: 109–​12. Shah, Nishi. 2003. “How Truth Governs Belief.” Philosophical Review 112: 447–​82. —​—​—​. 2006. “A New Argument for Evidentialism.” Philosophical Quarterly 56: 481–​96. Shapiro, Stewart. 2011. “Addendum on Peter Lipton’s ‘Mathematical Understanding.’” In Meaning in Mathematics. Ed. John Polkinghorne. New York: Oxford University Press. 55–​59. Sider, Theodore. 2009. “Ontological Realism.” In Metametaphysics:  New Essays on the Foundations of Ontology. Ed. David J. Chalmers, David Manley, and Ryan Wasserman. New York: Oxford University Press. 384–​423. —​—​—​. 2011. Writing the Book of the World. New York: Oxford University Press. Strevens, Michael. 2013. “No Understanding without Explanation.” Studies in History and Philosophy of Science 44: 510–​15. Treanor, Nick. 2014. “Trivial Truths and the Aim of Inquiry.” Philosophy and Phenomenological Research 89: 552–​59. Trout, J. D. 2002. “Scientific Explanation and the Sense of Understanding.” Philosophy of Science 69: 212–​33. Williams, Bernard. 1973. “Deciding to Believe.” In Problems of the Self. New York: Cambridge University Press. 136–​51. Zagzebski, Linda. 1996. Virtues of the Mind: An Inquiry into the Nature of Virtue and the Ethical Foundations of Knowing. New York: Cambridge University Press. —​—​—​. 2001. “Recovering Understanding.” In Knowledge, Truth, and Duty:  Essays on Epistemic Justification, Responsibility, and Virtue. Ed. Matthias Steup. New York: Oxford University Press. 235–​51.


Aesthetic Understanding Alison Hills

8.1. Introduction There are many different kinds of understanding that have a role in aesthetics. One is the idea of understanding a work of art itself. There is much to be said about what that understanding might consist in, what background knowledge might be required for it. (Do you need to know what the artist intended? Or the historical context in which the art was produced? Or the tradition into which the work falls?) Another is understanding an artist. (Do you need to know all her work? All the work of her contemporaries? Of her predecessors?). But these are not my topics here. Instead, I am interested in what it means to understand why a particular work of art has aesthetic (or artistic) value, or not, for instance that A has merit, or is beautiful, or valuable. For the remainder of the chapter, this is what I will mean by aesthetic understanding. In other work I  have given an account of moral understanding and of understanding why in general, and here I want to see to what extent this can generate an account of aesthetic understanding that is independently plausible, and that has an important role in aesthetics. I will begin by sketching an account of aesthetic understanding that has some similarities with knowledge, but also some striking differences. Aesthetic understanding, unlike aesthetic knowledge, requires a grasp of the reasons that make the claim in question true. What does this grasp consist in? I will explain it as a kind of know-​how, or set of abilities, including the ability to give explanations of value in this and in other cases. Aesthetic understanding



of this kind is important in several ways, but here I will concentrate on just one: its role in a key aesthetic activity, appreciation. The resulting conception of aesthetic understanding, and its role in aesthetics, is attractive. But at the same time it faces some serious objections, notably, that it requires that aesthetic value have an explanation (where this is understood as an account that can be expressed propositionally); and that aesthetic understanding be articulate. Both of these might be disputed, and I will discuss both issues in some detail. But before I begin, some preliminaries. I will be talking here about truth in aesthetics, and of aesthetic knowledge. I  assume that this is consistent with a wide variety of different views in meta-​aesthetics: not only a straightforward realism about aesthetic value, but also the more plausible versions of subjectivism, expressivism, or relativism that allow for a notion of aesthetic truth. And I assume too that they will allow for the possibility of aesthetic knowledge, whereby one’s aesthetic judgments are true and are well formed. I will not here take a stand on what is required for knowledge: perhaps they need to be justified, or responsive to evidence, or reliably formed; but I will assume a standard, and so hopefully uncontroversial account of knowledge where if you know that p, your belief that p is true, is not made on the basis of “luck” or on “defeated” evidence, and this knowledge can be passed on relatively easily by means of testimony. I  will argue that aesthetic understanding is rather different from aesthetic knowledge, understood in this standard way. I will be interested here in judgments about the value of works of art. Of course, they can be valuable in many different ways:  they can be expensive, perhaps they can have moral value, perhaps other kinds too. But I will be interested in aesthetic value. It is not easy to define what this is, except that it is their value qua work of art (and so is not the same as their price or moral value, for instance, which they do not have qua artwork). I will assume that in practice we have a sufficiently good grasp as to whether a claim about a work’s value is a claim about aesthetic value or not. Sometimes a distinction is drawn between a work’s aesthetic value and its artistic value, where artistic value but not aesthetic value can differ between two perceptually indistinguishable objects; others prefer to extend the concept of aesthetic value to account for this sort of case. As far as possible, I do not take a stand on this, or between theories of aesthetic value (e.g., whether it is the object’s formal qualities that ground its value; or its place in an art-​historical tradition, or some combination of these); I will try to give an account of aesthetic understanding that is neutral on these questions.

Aesthetic Understanding


8.2. Aesthetic Understanding I will call an “aesthetic proposition”, a proposition concerning aesthetic value, for instance that Middlemarch is an exceptionally good novel; that Citizen Kane an excellent film; and so on. By aesthetic understanding, I mean understanding why p (some aesthetic proposition) is true, for instance, understanding why Middlemarch is an exceptionally good novel and Citizen Kane an excellent film.1 If you understand why Middlemarch is good, you must believe that it is good, and you must have some view as to the reasons why (for instance, that it is good because it is such a morally serious and humane book). If it is not good after all, or your explanation of why it is good is not right, you cannot understand why Middlemarch is good. Aesthetic understanding, in this sense, is factive. And it is not transparent: you can have more (or less) aesthetic understanding than you think you have. So far, aesthetic understanding is rather similar to aesthetic knowledge: if you know that Middlemarch is exceptionally good, it must really be good, and you can know more (or less) than you think that you do. But understanding is not the same as knowledge.2 Understanding requires a grasp of the reasons why p is true: why Middlemarch is good and Citizen Kane excellent. What is this grasp? I think it should be considered a kind of intellectual know-​how. Just as if you physically grasp something, you have it under your control and can manipulate it at will; so if you grasp the relationship between two propositions (p, and its explanation, q), you have that relationship under your control and can manipulate it. Thus you can draw conclusions on the basis of the reasons why they are the case, and give an explanation of your conclusion; both in this, and in related cases. This know how involves a number of abilities, which allow you to treat q as the reason why p, not merely believe or know that q is the reason why p. If you understand why p (and q is why p), then in the right sort of circumstances you can successfully:

1.   My conception of aesthetic understanding here parallels an account I give elsewhere of moral understanding (Hills 2009; 2010) and understanding why in general (Hills 2015a). 2.   Kitcher (2002), Woodward (2003a, 179), Lipton (2004, 30)  Grimm (2006; 2012)  and Khalifa (2012) argue that understanding is a kind of knowledge. Zagzebski (2001), Kvanvig (2003), Pritchard (2010a; 2010b) and Newman (2014) argue that it is distinct, but on different grounds in each case. See Hills 2015a for more details.



(1) Follow an explanation of why p given by someone else (2) Explain why p in your own words (3) Draw the conclusion that p (or that probably p) from the information that q (4) Draw the conclusion that p′ (or that probably p′) from the information that q′ (where p′ and q′ are similar to but not identical to p and q) (5) Given the information that p, give the right explanation, q (6) Given the information that p′, give the right explanation, q′ I will call having this set of abilities essential to aesthetic understanding having “cognitive control” over the relationship between p and q, the reasons why p. This theory raises a number of questions, which I will now begin to address.

8.3. What Is an Explanation of an Aesthetic Proposition? Given the close link between understanding and explanation, what kind of explanation do you need to be able to give of why some aesthetic proposition is true to count as understanding why it is? I think that a variety of views about explanation in aesthetics are consistent with my account of aesthetic understanding, and I am inclined to allow many different types of explanation to count as acceptable. For instance, can the explanations offered make use of distinctively aesthetic concepts, or must a good explanation ultimately use only nonaesthetic concepts? It is very natural to explain why a work of art has aesthetic value using “thick” aesthetic terms, such as: it is graceful, or bold, or exciting; and correspondingly to explain why another lacks value by saying it is (for instance) clumsy, lifeless, or dull. Perhaps works of art are clumsy or bold in virtue of some particular natural (nonaesthetic) properties. It is a very intuitive thought that beauty, and other kinds of aesthetic value, do not and cannot “float free” of the nonaesthetic properties of an object; quite the opposite: they are dependent on them.3 In that sense, nonaesthetic properties are more fundamental (more “natural”) than aesthetic ones. But even if that is right, it does not follow that an explanation of aesthetic value can only make use of the most basic fundamental properties. In general, I do not think that understanding why p requires you to explain p “all the way down”: if q is the explanation of p, and r is the explanation of q, you

3.   Zangwill (2014); a classic discussion is Sibley (1959; 1965).

Aesthetic Understanding


can understand why p by grasping the relationship between p and q; you do not need to grasp the relationship between q and r (that is a separate matter, required for understanding why q). It follows that explanations using thick aesthetic terms are perfectly acceptable, whether or not they pick out fundamental properties. There is a debate in general about the role of principles or laws in explanations. According to some theories of explanation, explanations must use only strict laws (without exceptions); according to others ceteris paribus laws are acceptable; yet others will allow fully particularist explanations that do not make use of any kinds of law at all.4 This has significant implications for the very possibility of aesthetic understanding, because it is a matter of controversy whether there are any lawlike aesthetic principles that could play a role in aesthetic explanation. For instance, on the one hand, Beardsley thought that there were: he claimed that objects have aesthetic merit either by being unified or intense or complex (Beardsley 1958, ch. 11). On the other hand, many have thought that judgments of aesthetic value are “anomalous,” that is, that they cannot be supported by lawlike generalization. Sibley, for instance, is well known for denying that there are or could be any valid “principles of taste” (Sibley 1959). One might well worry that if Sibley is right and there are no valid aesthetic generalizations, and if explanations quite generally require lawlike generalizations, there can be no valid aesthetic explanations and so no aesthetic understanding. I think this argument fails for two reasons. For a start, it is implausible that there are no valid ceteris paribus principles of taste, at least, that might support aesthetic explanation. In fact, Sibley’s own view is more complex, since he thinks that there are no lawlike principles of taste that are “bridge” principles, linking aesthetic and nonaesthetic properties. But he does think that there are important connections between thick aesthetic concepts and aesthetic value: if you attribute grace, or intensity, to a work of art without any further qualification, you imply that it is good (Sibley 2001). If this is right, there are at least valid ceteris paribus aesthetic principles that can play a role in aesthetic explanations, even if there are no principles connecting aesthetic value to nonaesthetic properties. In any case, however, I am inclined to be liberal about what kind of explanations are acceptable in aesthetics, and as a consequence, of what is required to understand why some work of art has value. I see no reason why explanations 4.   There is a large literature on explanation. Just a selection includes Leibowitz (2011), Jenkins (2008), Strevens (2008), Woodward (2003a; 2003b), Ruben (1990).



must contain strict or even ceteris paribus laws, so even if there are no genuine principles of taste, aesthetic understanding would still be possible. Explanations are asked for and given in different contexts; some require a more full explanation, or an explanation couched in particular terms, or an explanation made to a certain level. The knowledge and experience of the audience, your familiarity with them, a more or less formal setting—​all may make a difference. It may well be that claims about aesthetic value can have more than one adequate explanation: in some contexts understanding why p may require a grasp of just one such explanation; other contexts may be more demanding. It would be possible to set the conditions for understanding why an artwork is valuable so high that no one understood any such thing; but I think there is good reason to regard understanding why as coming in degrees, as I argue in the next section. So even if there are contexts in which the conditions for having full aesthetic understanding are very demanding and so out of reach for most of us, anyone in those contexts who grasps to some extent why a work of art is valuable will correspondingly have some aesthetic understanding, but not full or complete understanding.

8.4. Does Aesthetic Understanding Come in Degrees? Explanations can be more or less full and detailed. You can be better or worse at explaining yourself in your own words. You can be better or worse at making judgments about related cases. Since these constituents of understanding why come in degrees, it is tempting to think that aesthetic understanding comes in degrees too. There is some suggestive linguistic evidence supporting this idea.5 “Understands why p” is “gradable,” that is, it is similar to verbs such as “regret” or adjectives such as “tall.”6 A linguistic sign that an expression is “gradable” is that it can be modified. For instance: 5.   There are other possibilities: there might be a “cut-​off ” point, which marks the difference between having understanding and not having it all. Now there are a number of different possibilities. This crucial “cut-​off ” point could be fixed in all contexts; or it might vary, for a variety of reasons. Each of these (as well as some combination of them) has some appeal, but I favor an account according to which understanding why p comes in degrees, which I think is best supported by the linguistic evidence, as illustrated in the following section. 6.   This section applies to aesthetic understanding my discussion of the gradability of understanding why in general in Hills (2015a). See Stanley (2005, ch. 3)  for the marks of gradability, with special reference to knowledge (understanding why is not discussed). Kvanvig (2003) distinguishes between understanding and knowledge at least in part because he takes understanding but not knowledge to be gradable.

Aesthetic Understanding 1. a. b. c. d.


Adam is very tall. Ben is not very tall. Claire regrets losing her bag. Dorothy does not regret quitting her job.

It is not at all natural to modify “understands why” with “very much”: 2. a. # Adam understands very much why Middlemarch is exceptionally good. b. # Dorothy does not understand very much why Citizen Kane is an excellent film.

But there are very natural constructions with “understand why” modified by “very well”: 3. a. Ben understands very well why Mozart’s Don Giovanni is good. b. Claire does not understand very well why Rembrandt’s self-​portraits are remarkable.

And there are other possible modifiers: 4. a. Adam understands completely why Vermeer’s paintings are beautiful. b. Ben does not completely understand why Mozart’s Don Giovanni is good. c. Claire barely understands why Rembrandt’s self-​portraits are remarkable. d. Dorothy fully understands why Beethoven’s late quartets have merit.

The modified claim can be conjoined with an unmodified claim without inconsistency, as in: 5. a. Adam is tall, but not very tall. b. Ben regrets laughing, but not very much. 6. a. Adam understands why Vermeer’s paintings are beautiful, but not very well. b. Ben understands why Beethoven’s late quartets are valuable, but he does not fully understand why.



Finally, there are natural comparatives. You can compare different people who both understand why p, and you can compare the same person with regard to her understanding of two different propositions: 7. 8. 9.

Adam understands why Middlemarch is an exceptionally good novel better than he understands why Citizen Kane is an excellent film. Adam understands why Middlemarch is an exceptionally good novel better than Ben does. Adam understands why Middlemarch is an exceptionally good novel better than Ben understands why Citizen Kane is an excellent film.

This linguistic evidence suggests that there are grades of understanding. There are some grounds for taking knowledge why to be gradable too, for the following sentences are acceptable: 10. a. b. 11.

Adam knows very well why Middlemarch is exceptionally good. Ben knows full well why Citizen Kane is an excellent film. Claire does not know very well why Vermeer’s paintings are beautiful.

But there are perfectly acceptable modifiers of “understands why” that are not at all felicitous with “knows why”: 11. a. b. c. d.

# Adam knows completely why Vermeer’s paintings are beautiful. # Ben does not completely know why Mozart’s Don Giovanni is good. # Claire barely knows why Rembrandt’s self-​portraits are remarkable. # Dorothy fully knows why Beethoven’s late quartets have merit.

Knowledge why is gradable in terms of “quantity”: if there are three reasons why Middlemarch is good and Adam knows them all whereas Ben knows one of them, then Adam has greater knowledge why Middlemarch is good than Ben does. Aesthetic understanding why has the same phenomenon—​a gradability of “quantity”—​but it also has a gradability of quality (how completely, fully, barely, or partially you understand why p). This is linked to the set of abilities, cognitive control, that is an essential part of understanding why but not of knowledge why, and it is why there are some linguistic expressions of gradability that are acceptable for understanding why but not for knowledge why.

Aesthetic Understanding


8.5. Are All the Abilities Necessary: Is Aesthetic Understanding Always Articulate? The account of aesthetic understanding sketched above included what I called “cognitive control,” which consisted in six abilities to manipulate the relationship between aesthetic propositions and their explanations. Doubts might be raised over whether all six are really necessary. In particular, it might be thought that aesthetic understanding in particular, perhaps unlike other kinds of understanding in science, for instance, resists being made explicit and articulate. Surely we are all familiar with the idea of someone with excellent aesthetic judgment, sensitive to aesthetic reason, but who simply cannot articulate why a painting is beautiful or a play valuable. Of course, understanding why only requires that you normally can put an explanation into your own words and so on. It is consistent with this that you understand why p, even though you cannot explain why p is true. But even so, one might think that there are two sorts of abilities here. One is the set of abilities needed to grasp the relationship between claim and explanation in such a way that you draw the right conclusion in this and in similar cases: you correctly judge Vermeer’s paintings to be beautiful, and would do the same for similar works. The other set of abilities consist in grasping this relationship consciously, explicitly, and articulately. Is it really necessary to be able to put into words why Hamlet is a great play, or why Vermeer’s paintings are beautiful, in order to understand why they are? One might even go further, and say that it is difficult, perhaps even impossible, to articulate fully what is important about (some) valuable works of art. For one thing, interpretation of the work is a process that is open-​ended and so one cannot expect ever to reach a settled conclusion. For another, the value of (at least some kinds of ) art in part is that it gives us an insight, or a way of seeing, that is not easily expressed as a series of propositions. To understand why a work of art is valuable, is to see it (and perhaps, as a consequence, the world) in a new light. This casts doubt on two aspects of my account of aesthetic understanding. First, must aesthetic understanding essentially involve explanation? It is far from clear whether a new way of seeing is a kind of explanation. So if understanding can consist only in this, perhaps it does not involve grasping an explanation in any way. But even if the best way of interpreting this “new kind of seeing” is in terms of grasping an explanation, we might still worry that the



account of aesthetic understanding I gave is too intellectualist. For it may be that insofar as it involves grasping an explanation, it does not require you to articulate that explanation. That you have grasped why the artwork is good could be shown in other ways, for instance, in producing your own artworks that are responsive to similar aesthetic reasons. However, I think that these objections overstate the difficulties. In the first place, coming to understand why a work of art is valuable typically does involve trying to put into words what it means and what aesthetic qualities it has, even if this is not at all easy. Of course, it would be especially difficult if the explanation had to be in terms of nonaesthetic qualities, but there is no reason to think that it must. At the very gives least, one can offer an explanation: “it gives us a new way of seeing,” and, at least in some circumstances, this may be sufficient to count as an explanation and so as affording (perhaps minimal) aesthetic understanding. Of course, we might reasonably hope for more, and one of the things that we expect from those we regard as aesthetic experts, such as art critics, is that they are able to express more clearly and in more detail what is good (or not) about a work of art, in such a way that helps the rest of us gain understanding of why it is. So I do want to insist that aesthetic understanding is essentially connected to explanation, even if it can be of a rather minimal kind. On the other hand, I  do not want to deny that that explanation can be grasped and expressed (e.g., by other artists) without being articulated (perhaps without even being articulable). In response to this, I  will describe the kind of understanding I outlined previously as explicit understanding, and to distinguish a different kind. Let us say that you tacitly or implicitly understand why p, if p is true and q is why p, provided that you believe that p and under normal circumstances you can draw the conclusion that p (or that probably p) from the information that q; and you can draw the conclusion that p′ (or that probably p′) from the information that q′ (where p′ and q′ are similar to but not identical to p and q). What I mean by “draw the conclusion p from the information q” here is that you correctly believe that q and on that basis you draw the correct conclusion (p). You do not have to represent q as the basis of p or even necessarily be explicitly aware of it at all. Even the judgment that p may be implicit.7 I think that both implicit understanding and explicit understanding are important, 7.   The difference between implicit and explicit understanding is somewhat similar to Sosa’s well-​known distinction between animal and reflective knowledge (Sosa 2009, 138). See Hills (2009; 2010; 2015a) for more details. Hills (2015b) has an extensive discussion of whether explicit (rather than implicit) understanding is important in ethics.

Aesthetic Understanding


with, very crudely, explicit understanding having most significance within the practice of art criticism and appreciation; and implicit understanding playing an important role in artistic production.

8.6. The Relationship between Aesthetic Knowledge and Aesthetic Understanding It is controversial whether understanding quite generally is identical with a form of knowledge, and in particular whether understanding why is identical with knowledge why. I have written extensively about this elsewhere, and have argued that knowledge why is neither necessary nor sufficient for understanding why. I will not repeat the whole argument here, but I do want to give a very brief sketch of it.8 Knowledge and understanding both require true belief. And each has an additional requirement: the truth needs to be “tied down” or “anchored” in a certain kind of way.9 If you have knowledge, you believe that p and p is true, and you could not easily have got your beliefs wrong; your belief was not “lucky.” If you understand why p, your belief why p is true, but there is no requirement that it was not lucky. Rather, you have to be able to be sensitive to the connection between p and q in these and similar circumstances: you need cognitive control. Cognitive control is a matter of know-​how or a set of abilities. You can know why p without having these abilities—​nor are they extra pieces of propositional knowledge—​and so understanding why p is not a form of propositional knowledge.10 In fact, you can have true beliefs with cognitive control as a matter of luck, in which case you have understanding without knowledge. Because knowledge differs from understanding in not requiring cognitive control, it can play a vital social role: it can be shared quite easily though testimony. If someone who knows asserts: “q is why p” and you understand what they say and rightly trust them, you come to know why p. But that wouldn’t be true if knowledge required cognitive control, because usually this cannot be passed on through assertion alone, or at least not easily. You cannot 8.   This section, and the beginning of the following section on the value of understanding, summarize the arguments of Hills (2009; 2010), but especially Hills (2015a), which deals with these questions in much greater detail. 9.   A metaphor familiar from Plato, Meno (97e ff.). 10.   There are, I think, further reasons why understanding why p is not a form of propositional knowledge; in particular, they each have a different relationship to luck. For example, suppose that a very unreliable interlocutor correctly told you that a work of art was innovative and you yourself correctly drew the conclusion that it was aesthetically valuable, you might, I think, understand why the work was valuable without being in a position to know why it was (for you do not know that it is innovative, as the speaker whom you trust is not normally at all reliable). See Hills (2009; 2010; 2015a) for further arguments.



normally pass on the ability to draw conclusions or give explanations about similar cases simply by telling someone that q is why p (or even telling them that, plus that q′ is why p′ and so on). Cognitive control is something that typically comes with reflection and practice. Because it can be passed on easily through testimony, knowledge allows for substantial divisions of epistemic labor: you can find out why X and then tell me; I can find out why Z and then tell you. We will both end up with knowledge why X and why Z. Understanding why is different. Even if you tell me that w is why x is true and I tell you that v is why z is true, we won’t thereby have passed on the cognitive control necessary for understanding: I will understand why z but not why x, you will understand why x but not why z. Because it requires cognitive control that cannot easily be passed on by testimony, understanding why typically does not allow for a substantial division of epistemic labor. Understanding has a very different epistemic role: because it requires that you have cognitive control, it ensures that you have the abilities to make judgments about related cases. Anyone with understanding has a kind of mastery over the topic in question (that is, the proposition, its explanation, and related questions) that is important for several reasons (some of which I describe in more detail in the next section). To sum up. I  have described two epistemic states playing two different roles:  one encouraging a division of epistemic labor and allowing the state to be passed on easily through testimony, but not guaranteeing that you can make judgments about related cases; the other making it difficult if not impossible to pass on the state through testimony, but allowing you to make judgments about related cases through cognitive control. The first state is knowledge, the second is understanding why. Since no single state can play both roles, knowledge why p (plus additional propositional knowledge) is not sufficient for understanding why p.

8.7. Aesthetic Appreciation and the Value of Aesthetic Understanding Recently there has been much discussion of the value of understanding why compared to the value of knowledge.11 Here I want to consider whether aesthetic understanding, specifically, is valuable. 11.   For instance, Kvanvig argues that knowledge cannot be the supreme epistemic value because the anti-​Gettier clause will be gerrymandered and inelegant (Kvanvig 2003). Riggs argues that knowledge cannot be the supremely valuable epistemic aim because it cannot explain why certain cognitive skills are intellectual virtues (Riggs 2003). See also Pritchard (2010a; 2010b).

Aesthetic Understanding


There are very good reasons to think that it is instrumentally valuable. It is useful when you need to tackle a new question because it guarantees (as knowledge, even knowledge why, does not) that you have the know-​how to do this successfully. Moreover, if you have the set of abilities that constitute cognitive control, you can use them: I will call this exercising your understanding. And in particular you can use it to make judgments, including aesthetic judgments that are the claims that you understand. For instance, on the basis of your recognition of its maturity and humanity, you draw the conclusion that Middlemarch is a very good novel. That is, you can exercise your understanding (i.e., your cognitive control) in coming to understand why p. I suggest that forming beliefs by exercising your understanding is valuable for its own sake. It is difficult to argue for this kind of claim, and I will do so here by extending a familiar account of the value of true belief: that true beliefs are valuable as a mirror of nature. What does it mean for a belief to mirror nature? The metaphor has typically been understood in terms of the relationship between the content of the belief and the facts. A belief of the content: “Middlemarch is good” mirrors the world if Middlemarch is indeed good. But a set of beliefs might also mirror the world in virtue of their form; that is, the similarities between the relationships between those beliefs and the relationships between the facts in the world. Suppose Middlemarch is good in virtue of its being mature and humane (that is, Middlemarch’s being mature and humane explains its being good). And suppose that you draw the conclusion Middlemarch is good on the basis of your belief that Middlemarch is mature and humane (that is, your belief that it is mature and humane explains your belief that it is good), then clearly there is a similarity—​a mirroring—​ between your beliefs and the world, that cannot be explained fully in terms of content of those beliefs alone, but also must refer to the relationship between them: one of your beliefs depends on the other, just as there is a dependence between the facts in the world. Now aesthetic understanding makes available both kinds of mirroring. Like knowledge, it requires the content of your beliefs (that p, and that q is why p) to mirror the world, i.e., your beliefs must be true. In addition, by exercising your understanding, you can mirror the structure of the world within the structure of your own thoughts as well as their content. Why is this important? I want to suggest a few ways in which this might be significant in aesthetics specifically. First, when you make evaluative judgments in response to the reasons why they are true, you are responding to



aesthetic reasons. I suggest that there are appropriate or fitting cognitive as well as noncognitive responses to aesthetic reasons and aesthetic value. We are familiar with the idea that pleasure (or perhaps, specific kinds of pleasure) is the appropriate response to aesthetic beauty, for instance, and more generally that positive noncognitive attitudes fit aesthetic value, and negative ones fit works of art that lack value. But I suggest that forming beliefs that have patterns of dependence that mirror the dependence of aesthetic reasons and aesthetic value is also fitting or appropriate. Why should we think that using one’s cognitive control to form aesthetic judgments is the appropriate or fitting response? Why not just form true beliefs, of whatever provenance? Or why not try to acquire knowledge of whether a work of art is valuable, whether or not that knowledge is formed in response to aesthetic reasons or in some other way (e.g. on the basis of testimony from others whose aesthetic judgments one trusts). Responsiveness to aesthetic reasons though one’s judgments of aesthetic value is, I  suggest, one of the key components of a distinctive activity:  the appreciation of a work of art. There a couple of different ways in which you can respond to a work of art, and specifically to its aesthetic value. The first we might call classification. You classify a work of art in this sense when you judge whether or not it is any good. You might make this judgment on the basis of having seen the film and your awareness of its aesthetic qualities, but that is not essential to classification. Classifications of works of art need not be supported by aesthetic reasons, in the sense that you can classify Citizen Kane as excellent while having no idea why it is (for instance, on the basis of your knowing that it topped the Sight and Sound film poll for several years). Now suppose that you watch Citizen Kane carefully yourself. On the basis of your awareness of its aesthetic qualities, you make a judgment of its aesthetic value. This is what I will call appreciation.12 It is generally accepted that appreciation is aesthetically very important:  appreciation is, it has been said, what works of art are for (Goldman 2004, 93). There is less agreement, however, about what it is. What I mean by 12.   This is a fairly standard account of appreciation; see, for instance, Goldman (2004). However, I do not attempt to give a theory of appreciation that encompasses every kind of activity that might properly be described as appreciation, which are very varied (as Lopes points out: “People appreciate songs by listening intently for their large-​scale structural properties, by singing along, or by dancing; and they appreciate paintings by scrutinizing them visually, copying them, using them as motifs in new paintings, or writing ekphrastic verse” (Lopes 2008, 207). I  am simply giving an account of one kind of appreciation—​one that results in an judgment of aesthetic value.

Aesthetic Understanding


it, is the practice of engaging with the artwork, responding in various cognitive and noncognitive ways to it, and making a judgment of its value based on the qualities which make it valuable. What is important about aesthetic appreciation? Obviously, engaging with works of art gives us a kind of pleasure that we do not get by reading reviews of the same. There is, I  think, a distinctive pleasure in coming to understand why a work of art is valuable.13 This is not, I think, merely a way in which aesthetic understanding is instrumentally valuable, or not straightforwardly so in any case, as the pleasure is not properly separable from the understanding. Rather, there is a pleasure taken in understanding, the two are part of a valuable complex whole. It is this that leads us to read works of art and literary criticism, first, to better understand the work itself, and second, to better understand its achievements and failures. Engaging with art in this way involves a wide variety of mental faculties both cognitive and noncognitive.14 Sometimes, the different sorts of response support each other: the noncognitive response, for instance, which might be a type of pleasure or feeling, may inform the cognitive response; a belief that the artwork has merit. And a better understanding both of the work itself, the aesthetic qualities it has and how they contribute to its value, may refine your noncognitive response, changing the quality or intensity of pleasure or make available further kinds of pleasure or feeling. If this is right, understanding why, even though it is quite an intellectual response to a work of art, is crucial to appreciation. But, it might be argued, while it is possible for one’s cognitive appreciation to enhance and support one’s noncognitive response to an object, this does not always happen. The two components of appreciation can be in tension, in the sense that trying to get a better cognitive grasp of the value of a work of art undermine the immediate, unforced, and spontaneous pleasure that one takes in it. It is familiar, after all, that the close reading of a novel or poem that you love can actually decrease your enjoyment of it. And once you have (metaphorically speaking) taken a poem or novel to pieces by thinking carefully about how it “works,” you can’t put it back together again and experience it as if for the first time. 13.   This account of the experience distinctive of art owes a great deal to Budd (1995), who takes the experience of art that essentially includes understanding to be characteristic of all forms of art and to explain the value of all of them. I do not intend to make such a strong claim here. All I intend is to point out that this kind of experience is often associated with appreciating art and it is valuable for its own sake. I think that there are other reasons why appreciating art is valuable, and there are also other explanations of the value of the art itself. 14.   Some of which are described by, e.g., Goldman (2004, 102).



One might take this objection to show that aesthetic understanding is not and could not be part of aesthetic appreciation, since it can undermine the distinctive noncognitive reaction that one might take to be definitive of appreciation. But I do not think that that is the right conclusion to draw. Aesthetic appreciation is complex and has many facets. It is very pleasing when they support and enhance one another, as when a better understanding of the work and why it is valuable makes available a different, but not worse (and indeed perhaps more subtle) noncognitive response to it. But at the same time, it is not surprising when this does not happen:  appreciation in one dimension can detract from appreciation in another. But whether it does or not, there are cognitive and noncognitive dimensions to appreciation; so when one tries to use one’s understanding to make an aesthetic judgement, one is not substituting something else for genuine aesthetic appreciation, but rather substituting one kind of appreciation for another.

8.8. Conclusion and Further Work In earlier work I have written extensively about understanding why, both in general and more specifically in ethics. Here I have applied it to aesthetics to see if it can produce an account of aesthetic understanding which is both plausible in its own right and can play an important role in aesthetics. There are a number of places in which the resulting account could be questioned, notably whether aesthetic understanding requires the grasp of an explanation (where this is assumed here to be of propositional form) and whether that grasp needs to be conscious and articulate. I have defended both of these, at the least as part of the activity of critical interpretation and appreciation. But there is, no doubt, much more to be said on this issue, and I hope to do so in the future. I became interested in moral understanding in part because I  wanted to explain some puzzles about moral testimony and moral disagreement: I argued that because forming moral beliefs using moral understanding was important (morally and epistemically), this explained why it was not morally ideal to trust moral testimony, and why it is right (both morally and epistemically) not to change one’s mind on a moral matter when faced with disagreement with an epistemic peer (see Hills 2009; 2010). There are some apparently similar puzzles in aesthetics, and I would like to explore whether the account of aesthetic understanding sketched here can be used to resolve them in a similar manner.15 15.   On aesthetic testimony, see Hopkins (2000; 2001; 2011), Meskin (2004), Robson (2012). On aesthetic disagreement, McGonigal (2006), Schafer (2011).

Aesthetic Understanding


Finally, I argued that moral understanding was important because it was the essential cognitive component to morally worthy action and to moral virtue. Does aesthetic understanding have any similar role? Is there such a thing as “right action for the right reasons” in aesthetics? Is there such a thing as aesthetic virtue? And if so, what are they like, and does aesthetic understanding play a role in them? In other words, in terms of action, virtue, and cognition, are ethics and aesthetics fundamentally similar or fundamentally different? Again, these are questions I hope to pursue in further work.

References Beardsley, Monroe. 1958. Aesthetics. Indianapolis: Hackett. Budd, Malcolm. 1995. Values of Art: Picture, Poetry and Music. London: Penguin. Goldman, Alan. 2004. “Evaluating Art.” In The Blackwell Guide to Aesthetics. Ed. Peter Kivy. Oxford: Blackwell. 93–​108. Grimm, Stephen R. 2006. “Is Understanding a Species of Knowledge?” British Journal for the Philosophy of Science 57: 515–​35. —​—​—​. 2012. “The Value of Understanding.” Philosophy Compass 7: 103–​17. Hills, Alison E. 2009. “Moral Testimony and Moral Epistemology.” Ethics 120: 94–​127. —​—​—​. 2010. The Beloved Self. Oxford: Oxford University Press. —​—​—​. 2015a. “Understanding Why.” Noûs 49: 1–​28. —​—​—​. 2015b. “The Intellectuals and the Virtues.” Ethics 126: 7–​36. Hopkins, Robert. 2000. “Beauty and Testimony.” In Philosophy, the Good, the True and the Beautiful. Ed. Anthony O’ Hear. Cambridge: Cambridge University Press. 209–​36. —​—​—​. 2001. “Kant, Quasi-​Realism and the Autonomy of Aesthetic Judgement.” European Journal of Philosophy 9: 166–​89. —​—​—​. 2011. “How to Be a Pessimist about Aesthetic Testimony.” Journal of Philosophy 108: 138–​57. Jenkins, Carrie S. 2008. “IV—​Romeo, René, and the Reasons Why: What Explanation Is.” Proceedings of the Aristotelian Society 108: 61–​84. Khalifa, Kareem. 2012. “Inaugurating Understanding or Repackaging Explanation?” Philosophy of Science 79: 15–​37. Kitcher, Philip. 2002. “Scientific Knowledge.” In The Oxford Handbook of Epistemology. Ed. Paul K. Moser. Oxford: Oxford University Press. 385–​407. Kvanvig, Jonathan. 2003. The Value of Knowledge and the Pursuit of Understanding. Cambridge: Cambridge University Press. Leibowitz, Uri D. 2011. “Scientific Explanation and Moral Explanation.” Noûs 45: 472–​503. Lipton, Peter. 2004. Inference to the Best Explanation. 2nd ed. London: Routledge. Lopes, Dominic. 2008. “Virtues of Art: Good Taste.” Aristotelian Society Supplementary Volume 82: 197–​211.



McGonigal, Andrew. 2006. “The Autonomy of Aesthetic Judgement.” British Journal of Aesthetics 46: 331–​48. Meskin, Aaron. 2004. “Aesthetic Testimony: What Can We Learn from Others about Beauty and Art?” Philosophy and Phenomenological Research 69: 65–​91. Newman, Mark. 2014. “EMU and Inference: What the Explanatory Model of Scientific Understanding Ignores.” European Journal for Philosophy of Science 4: 55–​74. Plato. 1994. Meno. In Plato’s Meno in Focus. Ed. and trans. Jane M. Day. London: Routledge. Pritchard, Duncan. 2010a. “Knowledge, Understanding and Epistemic Value.” In Epistemology. Ed. Anthony O’ Hear. Cambridge:  Cambridge University Press. 19–​43. —​ —​ —​ . 2010b. “Knowledge and Understanding.” In The Nature and Value of Knowledge: Three Investigations, by Duncan Pritchard, Adrian Haddock, and Alan Millar. Oxford: Oxford University Press. 3–​88. Riggs, Wayne D. 2003. “Understanding ‘Virtue’ and the Virtue of Understanding.” In Intellectual Virtue: Perspectives from Ethics and Epistemology. Ed. Michael DePaul and Linda Zagzebski. Oxford: Oxford University Press. 203–​26. Robson, Jon. 2012. “Aesthetic Testimony.” Philosophy Compass 7: 1–​10. Ruben, David-​Hillel. 1990. Explaining Explanation. London: Routledge. Schafer, Karl. 2011. “Faultless Disagreement and Aesthetic Realism.” Philosophy and Phenomenological Research 82: 265–​86. Sibley, Frank. 1959. “Aesthetic Concepts.” Philosophical Review 68: 421–​50. —​—​—​. 1965. “Aesthetic and Nonaesthetic.” Philosophical Review 74: 135–​59. —​ —​ —​ . 2001. “General Criteria and Reasons in Aesthetics.” In Approach to Aesthetics:  Collected Papers on Philosophical Aesthetics. Ed. John Benson, Betty Redfern, and Jeremy Roxbee Cox. Oxford: Clarendon Press. 104–​18. Sosa, Ernest. 2009. Reflective Knowledge. Vol. 2 of Apt Belief and Reflective Knowledge. Oxford: Oxford University Press. Stanley, Jason. 2005. Knowledge and Practical Interests. Oxford: Oxford University Press. Strevens, Michael. 2008. Depth:  An Account of Scientific Explanation. Cambridge, MA: Harvard University Press. Woodward, James. 2003a. Making Things Happen:  A  Theory of Causal Explanation. Oxford: Oxford University Press. —​—​—​. 2003b. “Scientific Explanation.” In Stanford Encyclopedia of Philosophy. Summer 2003 ed. Ed. Edward N. Zalta. http://​​archives/​sum2003/​ entries/​scientific-​explanation/​. Zagzebski, Linda. 2001. “Recovering Understanding.” In Knowledge, Truth, and Duty:  Essays on Epistemic Justification, Responsibility and Virtue. Ed. Matthias Steup. Oxford: Oxford University Press. 235–​51. Zangwill, Nick. 2014. “Aesthetic Judgment.” In The Stanford Encyclopedia of Philosophy. Fall 2014 ed. Ed. Edward N. Zalta. http://​​archives/​fall2014/​ entries/​aesthetic-​judgment/​.


Why (Study) the Humanities? The View from Science Jenann Ismael

9.1. Introduction Questions about the value of the humanities and the relationship between the sciences and humanities have been very much in the news recently.1 Just a brief review in the public press shows scientists and humanists weighing in and responding to one another. Public opinion is shifting in favor of scientific and technological education. There are two related challenges that have been leveled about the value of the humanities. 1. There is a threat to the perceived value of the humanities in the culture at large. This manifests itself in two ways: reduced public support for humanities research, and students being steered away from studying the humanities in university.2 One often hears complaints of the form, “Why should we as a culture invest resources in humanities education? Why should we spend good money for our children to study French literature, or why should the state subsidize degrees in philosophy? Science and engineering degrees are effective ways of getting jobs and we (as a country) need

1.   Stanley Fish, Steven Pinker, Philip Kitcher, Daniel Dennett, and Martha Nussbaum have all weighed in on the public discussion. A new report commissioned by a bipartisan quartet of lawmakers looks at the role of the humanities and social sciences in public education: http://​www.humanitiescommission. org/​. 2.   This is the threat addressed here: http://​​2013/​06/​humanitiesreport.



more scientists and engineers, but what is a degree in cultural anthropology worth?” 2. And then there is the threat to the humanities mounted by those who claim that in a scientific worldview the human sciences will ultimately be absorbed into (or replaced by) the hard sciences, and there will be no place for the humanities as a source of knowledge.3 Those who defend this position base it on the claim that we are bits of matter, alongside other bits of matter, governed by material laws, and so understanding ourselves is not different in principle from understanding celery or cells. The great complexity of the human has fostered the illusion that human behavior is different in kind, but that position is indefensible from the point of view of science. On the face of it, these are two quite different challenges. In response to the second, the position that some have retreated to is to relinquish the claim that the humanities provide a source of knowledge about the human being and hold that it “makes our lives better” in other ways. In response to the first, some have argued that contrary to appearances, a humanities education will make you a better lawyer, businessman, get you a better job, or make you better at public relations (see Rand 1999). Martha Nussbaum has argued that it makes you a better citizen and so is crucial to the success of the polis.4 A more dismissive answer is provided by Stanley Fish, who rejects corporate or economic values as the sole or ultimate arbiter of value. He demands to know why the humanities should have to justify themselves by those standards any more than corporate values have to justify themselves by the standards of the humanities. One might make more money as an engineer, but one would be culturally illiterate, historically ignorant, and uninsightful. Rhetorically, Fish is right. There is no reason that the humanities should have to justify themselves by the standards of the sciences. His response, however, makes it sound as though the humanities are a pleasurable diversion to be enjoyed by those with the privilege of leisure, and that is not the kind of defense that 3.   Alex Rosenberg is the most outspoken advocate of the position, though one finds glimmers of it in the opinions of many scientists and philosophers, and it is a challenge that any self-​styled naturalist has to address in their own thinking. 4.   Nussbaum (2010). Fareed Zakaria, in an interestingly related recent piece in the Atlantic, has argued that education in the humanities, as opposed to STEM education, makes one more creative and entrepreneurial and that is what has allowed America to flourish economically despite lagging behind in STEM education.

Why (Study) the Humanities?


will persuade a parent to support a child’s decision to study art history rather than (say) physics. A deeper defense would say what kind of knowledge the humanities provide, how it differs from that of the sciences, and why it is indispensable in a well-​lived life. It would address the relationship between the scientific vision of the human being and the humanistic one explicitly, and say whether the humanistic vision is undermined by what science is teaching us about ourselves. This way of putting it brings it close to the second problem, so I want to begin by addressing that one.

9.2. The Physics of Open Systems Every physical system falls under the scope of general microscopic laws that govern the universe as a whole. Those laws are exact and exceptionless. At the level of human behavior those laws are also local and deterministic.5 Consider any subsystem of the universe and an enclosing sphere of any diameter around that system.6 We call the variables that characterize the world on the boundary and outside the sphere ‘exogenous,’ and those that characterize the interior of the sphere ‘endogenous.’ Locality entails that the values of variables on the surface of the sphere screen off (render irrelevant) the values of variables outside the sphere. This, together with determinism, entails that if we know the initial state of the matter inside the sphere and all of the forces that impinge on the surface of the sphere over any interval of time, we can predict the behavior of anything inside the sphere with certainty. This holds regardless of the size of the sphere and regardless of what kind of matter it encloses. Let’s call this the DL principle (for deterministic-​local). The DL principle holds for the collection of dust particles in this little bit of space. It holds for planets. It holds for toasters, and tree frogs. And it holds also for the human being. So, here’s a human being. Here are the exogenous influences impinging on a sphere enclosing it. What physics tells us is that if we have the initial state of the body and we know the forces impinging on the

5.   The classical setting simply gives us a precise physical framework in which the challenge takes its sharpest and most pressing form. Until the interpretation of quantum mechanics is settled, it is impossible to say definitively whether it makes a difference to human action, though there are no definite physical reasons right now for thinking it will make any significant difference. At the level of brain function relevant to human action, classical physics seems to be the effective theory. 6.   We make it a sphere just for convenience. We can try to get it as close to the boundaries of the body as possible, though the boundaries of the body are a little vague (is the hair on your skin part of your body? What about the skin cells just flaking off . . . etc.).



surface of the sphere through some interval, we can calculate how the body will move over that interval. That sounds like a very strong result. But it is actually quite weak for the following reason. The number of degrees of freedom in the microscopic state of the enclosed system = (6 × the number of particles that compose the system). The number of degrees of freedom in the exogenous variables =roughly (6 × the number of particles in the rest universe).7 The DL principle says if we know the precise values of all of those variables, we can calculate with microscopic precision how any physical system enclosed in that sphere will behave. It does not entail that knowing anything less than that will let us calculate anything about that system. It entails, that is to say, that if we know everything we can calculate everything. It doesn’t entail that knowing less than everything will let us calculate anything. Here’s what I mean. Let’s consider a coin, and let’s suppose that the only thing we really care about concerning its behavior is whether, if it is tossed right now, it will land heads. The DL principle tells us that if we know its initial microstate and the values of all exogenous variables impinging on it through the course of the toss, we can calculate with certainty how it will land. But it also tells us—​and here is the crucial part—​that nothing less will do. The reason is that the dynamical laws entail that whether a coin lands heads or tails on a given toss is so sensitive to the microscopic values of so many exogenous variables—​e.g., the exact angle at which it is released, the Brownian motion of dust particles in the air, indiscernible fluctuations in speed and direction of wind—​that unless we know all of this with perfect precision, or can control their effects, the outcome of the toss cannot be predicted.8 What is happening here is that coin tosses amplify ignorance. They transfer any ignorance we have about the microscopic state of any of the particles that might make some impact on the boundary of the sphere into ignorance of the result of a given toss.9

7.   One qualification is necessary to make this strictly correct, but it makes no practical difference here. If the universe is big enough and the sphere is small enough, and we live in a universe in which information cannot be packed densely in any region of space, there may be fewer degrees of freedom on the boundary of the sphere. 8.   In the case of coin tosses, we can attach probabilities to outcomes (fifty-​fifty for fair coins, different probabilities for weighted coins), because there are stable relative frequencies over the dynamically relevant exogenous variables, but that itself is not something that is generally available. If we can control the effects of exogenous variables, the outcome can be reliably predicted. It is because of the lack of knowledge and absence of control in everyday circumstances that makes coin tosses effectively unpredictable. 9.   Note here that this is a different phenomenon from the sort of unpredictability that arises with chaotic systems. In the case of chaotic systems, the unpredictability has its source in the nonlinearity of the

Why (Study) the Humanities?


9.3. Science Is about Generality; Humans Are All Specificity When you see how weak the DL principle is, you might be surprised that we can have a predictive science of any open subsystem of the world. And, indeed, the vast majority of open subsystems (if by “open subsystem” we mean the matter contained in any region of space-​time around which we can draw a closed boundary) do not exhibit the kind of regularity that makes it possible to formulate simple, predictive laws that express their behavior as a function of the state of the environment.10 But some do. And we can look at the kind of dynamics that a bit of matter has to have to make a predictive science of its behavior possible. Let’s start with a toaster, and let’s suppose that we are interested primarily in its gross, discernible macroscopic behavior. The toaster sits inactive when the lever is up. Depressing the lever lowers a chassis in which bread is placed and initiates a process in which electricity heats internal grills to a certain temperature for a fixed amount of time. When the process is done, the chassis returns to its normal position. There are simple laws for this behavior of toasters because a toaster has a (relatively) fixed internal structure that (moreover) makes only a small number of variables relevant to that behavior. The position of the lever and the knob to determine grill time matter, but not the presence of wind or the absence of noise. Small differences in input don’t produce grossly different responses. It doesn’t matter exactly how hard or fast you push the lever.11 The internal wiring is hard, and designed to produce the

equations and arises for closed systems as well as open ones. When dynamical equations are nonlinear, tiny differences in initial state can lead to radically different outcomes. That means that anything less than perfect precision in knowledge of the initial state can leave us with very great uncertainty (uncertainty spread all over phase space) about the final state. The sort of failure of unpredictability under discussion here has nothing to do with nonlinearity of dynamical equations. It has to do, rather, with the openness of the systems and their sensitivity to exogenous variables. It arises even if the dynamical equations are linear. 10.   Philosophical usage often counts as a law only perfectly universal, fundamental generalizations. I am using the term loosely, to include defeasible, counterfactual supporting regularities of all kinds. So what we are looking for is a description of the gross behavior of a system (typically its movements) as a function of environmental impact. In the case of living organisms, these kinds of laws are often thought of in terms of its responses to stimuli. 11.   Of course, this is all true only if we describe the toaster at a very coarse-​grained level, and restrict the predicted behaviors. If we include temperature in “discernible behavior” and allow very fine discriminations, none of this would be true. The behavior would exhibit a high sensitivity to microscopic changes in its environment.



same response for the life of the toaster.12 We don’t need to know very much either about the environment, or what goes on inside a toaster to know how to expect it to behave, so long as it is operating normally. Using this as a model, we can say that there are simple laws that allow us to express the behavior of an open system as a function of its environment wherever there exists a reduced variable subspace of the physics of the universe that makes only certain variables relevant to behaviors of interest, and the system’s internal structure is fixed so that the impact of those variables is constant (over short time scales), and linear (or approximately so).13 Toasters break down and wear out.14 But breakdowns are infrequent, so we can ignore them at a small cost in the exactness of the laws, and wearing out is a gradual, predictable change that can be incorporated into predictive laws. Many systems, often of tremendous complexity, conform to this model. For example, in self-​organizing systems like termite colonies or slime molds, even though these systems have an enormous number of microscopic parts and the interactions among those parts contain feedback loops that would make the physical equations for their conjoined behavior effectively unsolvable even if we knew the initial state of each of them, there is emergent behavior regular enough to permit a reduced set of variables in terms of which we can find simple, predictive laws for the configuration. These simple, predictive laws, moreover, can be discerned without a good understanding of the underlying physics.15 There is no simple, univocal account of the dynamical underpinnings of systems that exhibit this structure. Scientifically, we’ve only begun to understand them. Although the open systems that exhibit this structure are the ones that tend to attract our scientific attention, it should be noted

12.   What I mean by “hard” here is fixed, relative to a range of contexts. One way of making things hard, in this sense, is by making them rigid. But we can also make a connection between A and B hard relative to contexts C by having a lawlike regularity connecting A and B scaffolded by structure present in C. So, for example, the connection between the button on my garage opener and the garage is hard when the two are in spatial proximity and all of the background things that need to be in place for it to function properly are there. 13.   There are three things to note: (1) the existence of such laws is relative to behaviors of interest (see fn.13), (2) in saying that the internal structure is fixed, we don’t mean that it is static, but that it changes in ways that support a fixed relationship between input and output variables, and (3) simplicity has been left unanalyzed here. More can be said, but an intuitive conception of simplicity will do well enough for our purposes. 14.   We do need to know how they work to fix them, or to explain their behavior when they start behaving anomalously. 15.   This is not an exhaustive list.

Why (Study) the Humanities?


that among all the open systems in the universe, they constitute a very small minority. Living organisms often exhibit this structure. Beginning with the macromolecules of DNA and RNA, we can trace step by step, as animal life moved from the simple ability to respond to frequently recurring environmental conditions to much more powerful mechanisms for producing behavior finely attuned to circumstance. Living systems have parts that are bound together in a relatively fixed configuration designed over ecological time to produce advantageous responses to stimuli.16 The frog brain, for example, is a remarkably well-​designed instrument for (among other things) getting frog bodies to respond in reliably predictable and adaptive ways to stimuli. The frog responds to the image of a passing fly with a flick of the tongue because that is what it has been designed to do.17 There is a lot going on in the frog brain, but the activity is designed (in part) to produce regular macroscopic responses to a particular class of stimuli: to filter out the noise, ignore the differences between one flyspeck and another, and get the tongue where it needs to be. Because this is what the frog brain was designed to do, one frog will do the same as another, and the behavior is more or less constant over time. The same goes for mongooses, mole rats, and three-​toed sloths.18 Viewed as part of this progression, what is special about the human being is that in the human mind we see the development of a cognitive platform for the emergence of a new behavior management strategy involving deliberation and choice. Instead of passing through a set of internal filters designed to keep behavior covarying reliably with features of the local environment, the effect of the stimulus on behavior is mediated by a process that seems almost perversely geared to undermine any possibility of general laws of human behavior. Consider a mundane example of choice. A mother walks into a shop to grab a coffee on her way home. She sees her young daughter, who is supposed to be at the library, talking to a boy that the mother doesn’t trust. She retreats before being seen, and walks home ruminating about what to do. The incident makes her realize that her daughter is growing up, a time she has long been preparing herself for psychologically,

16.   I borrow the phrase “wiring and connections” from Peter Godfrey-​Smith (2004). 17.   Designed by natural selection or whatever hidden hand shaped the universe. 18.   Responses to environmental stimuli can adapt, but adaptation in an individual frog happens slowly, and only with sustained pressure from the outside. Like the wearing out of the wiring of the toaster, it can typically be anticipated and incorporated into the laws.



but it seems too soon. She begins to think about her own life and her plans to have other children. A reconfiguration in her view of her daughter begins to occur. She wonders how much she doesn’t know about her daughter, and feels instinctively that this is a delicate time in her daughter’s life. By the time she is home, she has resolved to spend more time with her daughter and reaches for the phone to make dinner reservations at an old, favorite place. The deliberative process is bringing into the causal chain between stimulus and response, here, a whole lot of stored information collected over a lifetime of personal experience. This includes beliefs not only about the world, but also about herself, her daughter, their place in the world, personal plans, memories, intentions, and commitments. If a fly’s tongue snaps out unreflectively at a passing fly, the mature human adult runs his experience through a much more complex transformation that can—​in the most reflective decisions—​call up everything he is and has become. We are not Hamlets at every moment of our lives, but we have Hamlet moments. The result is that the choice-​governed aspects of human behavior do not just depend on the immediate stimulus, but are open to influence from an in-​principle unlimited number of sources, all stored in memory and encoded in the soft structure of the brain. As if that weren’t bad enough, the bearing of this information on behavior is filtered through a quite complex set of higher-​order principles for choice (goals, values, priorities, beliefs about who we are and who we want to be) that themselves vary from one person to the next and are constantly evolving. Where do these higher-​order principles come from? They are the products of experience, in some sense, but they are forged under the hot fire of personal reflection. And reflection is one of those processes that has the hallmarks of unpredictability. It is holistic, self-​feeding, and ongoing. Even if there were a deterministic equation that someone could write down that would describe it, the feedback would make the equation unsolvable within a few steps. So not only does choice make all of one’s personal history potentially relevant to one’s present behavior, it makes the bearing of personal history on behavior subject to second-​order principles that are themselves highly variable both across the population and over the history of a single subject. And even if we held all of that fixed, very small differences in stimulus can produce huge differences in response. We are highly attuned to tones of voice, subtle cues we are scarcely conscious of (e.g., the something in the air that tips the balance between accepting and declining an invitation to speak in New Zealand). Small difference makers can lead to very large differences. The effect of all of this is something that common sense knows, viz., that human behavior is highly individual and deeply unpredictable.

Why (Study) the Humanities?


Responses to stimuli vary from one person to the next and over time in the history of a single person.19 In a toaster, the structure that mediates input and output is “hard,”: ordinarily changing slowly and in predictable ways. The structure that mediates stimulus and response in humans is “soft” and changes at lightning speed in ways that are directly relevant to its discernible behavior. The very processes that are designed to stabilize regular behaviors both across the population and over time in other animals are geared to produce variability and differentiation in the human being. All frogs flick their tongues at passing flies, birds flock, and bees dance. The choice-​governed behavior of the individual human being, by contrast, is so highly sensitive to the specificities of personal history and belief, all encoded in the soft structure in the brain that, from an external perspective, choice effectively randomizes the effect of stimulus on response. There is no simple, general relationship between environmental stimulus and behavior. Whereas explanations of frog and fish behavior typically refer to the environmental stimulus and general laws of frog and fish behavior, explanations of human behavior typically appeal to accidents of individual history of which there are no traces in the stimulus and that can’t be collected under general laws.20 Laws seek to capture generalities, and human beings are all specificity.21

19.   We are still physical systems, and so in principle if we knew the microphysical state of the world at some cross section of a person’s back light cone and we knew the laws that govern the universe as a whole, we could predict their behavior. The same is true for any physical system. The point is that in the case of the human being (or any system whose macroscopic behavior is determined a mechanism that draws on a fund of information evolving as quickly and idiosyncratically as personal belief ), nothing less will do. 20.   Philip Kitcher’s strategy is to break down the difference between science and humanities, to make the difference, as he puts it, one of degree rather than kind. That would make the humanities cousins of the sciences aiming for the same kind of understanding, employing the same kinds of methods, but with a less impressive history of success (though he tries to argue that this too is overstated). I think that he is right in much of what he says, and that the differences that I emphasize might be differences of degree rather than kind. 21.   The line between humans and animal cognition is more complex than the contrast between frog tongue-​flicking and human choice suggests. On one end of the spectrum, we see simple organisms exhibiting very regular responses to the environment. On the other, we see humans exhibiting highly irregular responses to the environment on the other. As we move from one end of this spectrum to the other, we see increasing complexity both in the character of the stimulus and in the subtlety of the response. Experts disagree on whether there is a hard line to be drawn anywhere along this spectrum. It remains true that choice has the effect of bringing into the chain between stimulus and response information encoded in memory, and so the more information is encoded in memory, and the more that information varies across the population and over time, and the more behavior is governed by choice, the less regular responses to the environment will be.



This is not to say that there aren’t very useful algorithms for predicting how people will act. We rely on these algorithms every day of our lives. Their success depends on the fact that although we weren’t designed simply to respond in predictable ways to impact from the environment, we were designed to be deliberators. If we want to understand systems whose behavior is governed by choice—​i.e., to know what makes them tick, to guess how they will respond, to influence their behavior, and to interact with them effectively—​we have to get good at psychological interpretation. We have to become skilled at understanding what other people believe and desire and feel. We need to understand one another, not in the way that we understand toasters and electrons, but as persons, i.e., as subject of experience and emotion, as believers and agents. Anybody that can operate in the social world has rudimentary skill at psychological interpretation, but really learning to see things through other people’s eyes, to understand what they are thinking and feeling, what they hope and fear and value, having a rich sense of the complex inner world of another human being in all of its emotional and psychological complexity, is a kind of understanding that can be fostered and developed by an education in the humanities. Someone who has grown up reading novels will learn to understand the intricate hidden rumination that goes on inside another human being. Someone who has studied history will have an appreciation for the complex currents of culture that govern the unfolding of civilizations. Each of the fields traditionally classified among the humanities makes a distinctive contribution to this kind of understanding. Reasons and explanations aren’t just about prediction. They will also make a contribution to a richer kind of understanding, a kind of understanding that is not just a matter of being able to predict how other persons will behave, but being able to see things through their eyes, being fair and generous and empathetic. It enriches our understanding of the world by helping us understand people in the terms in which they understand themselves. That is something that we need to be able to do if we are relate to them not as material systems, but as subjects of experience and sources of agency. For we are social animals: partners, friends, coworkers, mothers, and teachers. The better we are at understanding one another in the way that we each understand ourselves, the better we will be in these capacities. These types of human understanding are not something that one can get from knowledge of physical law. Explanations that invoke reasons teach us how to interact with people as rational agents and seekers of value, to affect their behavior by persuasion rather than by trying to control them causally. They teach us to address the rational standpoint

Why (Study) the Humanities?


and offer reasons for them to act as we want them to act, so that our effect on their behavior is mediated by their own deliberative processes.22 They guide in seeing things through other people’s eyes and being able to construct a narrative that is fair to all viewpoints. Insight in human affairs requires that kind of understanding. Science is not, and never will be, a substitute for that kind of understanding.

9.4. Science Is about Description; the Humanities Are (Partly) about Guidance Of course that is only half of the story. What was said above was looking at the processes that mediate stimulus and response in a human being from a third-​person standpoint. When we turn from a third-​person to a first-​person perspective, something else emerges that makes the indispensability of the humanities even more manifest. Again, it has to do with choice, but this time it concerns the special status that choices have for the person making them. To other people’s lives, we are observers. We watch them make choices, try to gauge what they are thinking, guess how they feel, what they care about, and why they act the way they do. But we are not mere observers of our own lives. We live our own lives and we make up our own minds about what to think, and how to act. We bear a special relationship to our own thoughts and experiences that makes it impossible to take a purely detached attitude toward them. There is no way of abdicating our active participation in the making of our choices. From the mundane to the momentous (i.e., whether deciding which socks to wear or whom to marry), the universe will not make those choices on our behalves.23 The inescapability of choice is our situation in nature. And to be a chooser is to have a special kind of creative role in the production of our lives. It is worth noticing how little of our own lives, for many of us, is dictated by the practicalities of survival. People need to eat and sleep to survive. In the state of nature, that meant that our daily lives were organized around the necessities of obtaining food, caring for our young, and maintaining shelter, in terms largely dictated by our situation. One made a shelter with available

22.   The contrast is with manipulation or coercion. Manipulation tries to control the output of the decision process by controlling the input. Coercion tries to bypass the decision process entirely by using physical force to move the body to move directly. 23.   We can choose to be passive, but choosing to be passive is a way of choosing.



materials, and ate what one could get one’s hands on. The structure and content of daily life were not, to a very great extent, a matter of choice. Things are vastly different nowadays. The landscape of opportunities that the world presents is radically expanded. There are countless ways to make a living, and countless ways of maintaining house and home. You get to choose how to make a living, where to live, and whether to have children. And that is to say nothing of what to eat, what to wear, what music to listen to, what newspaper to read, and what to watch on TV. The choices we make are ultimately choices about who we are and who we want to be. Our identities as persons and agents is constituted by them. They give shape and definition to our lives in the way that the hammering of the sculptor gives shape and definition to the unformed stone, transforming an indefinite multitude of potential shapes into a single actuality. It is in equipping us with tools to address the very personal questions—​ “What should I do? How should I live?”—​that I see the humanities as making an indispensable contribution. A  humanities education can, among other things, open up the imagination to the rich array of possibilities of what to be. The bookish child of a farmer who reads Jude the Obscure for the first time can see a world open up that he hadn’t known. Books—​by showing us examples of successful and failed lives—​help us decide which of them are worth wanting. What I learned from Plato, Aeschylus, Dante, Goethe, Tolstoy, Joyce, Elliot, and Mann played a very personal role in making me who I am. Once you see all of that—​i.e., once you see your role in creating history, rather than being a passive observer—​you see that the humanities are not a recherché pursuit undertaken in leisure from which people with highbrow tastes draw enjoyment. They help us decide what to make of our lives. They are the tools of our becoming. It is sometimes said that if our scientific knowledge were good enough, at least in a deterministic world, we would be able to predict what we will do and we could just sit back and let it happen. One expression of the threat is embodied in the scientific challenge to free will.24 The thought seems to be that a completed scientific understanding of the world will push forward the boundaries of prediction, leaving no room for choice. But that there is something wrong with this line of thought should be clear from the observation that without our activity, there would be nothing to predict. Our choices don’t get made unless we make them (see Ismael 2016). Learning physics is not

24.   This is explicit in Rosenberg (2011), for example.

Why (Study) the Humanities?


going to relieve you (in practice or in theory) of the burden of making choices. And it is not going to relieve you (in practice or in theory) of the burden of running your life, or making yourself into what you will be. Understanding ourselves will always be an art as much as a science because it involves a form of creation.

9.5. Deciding What to Think and How to Feel When R.  W. Hepburn remarked that “One may look upon the “material” of one’s life . . . rather as an artist regards his canvas and paint or a sculptor his stone” (Hepburn and Murdoch 1956, 14–​58), he meant to be both calling attention to the creative role that we play in making our lives what they are, and observing that within the bounds of the given facts of our lives, there is a great deal of discretionary leeway in how we understand our lives. We each have to sift through the raw materials of our own experience and cull from them an understanding of what has happened to us and who we are. And this observation extends beyond our narrow understanding of our own lives to our understanding of the people around us and the world quite generally. We are always deciding what to think and how to feel about things. This is not a passive matter of simply opening our eyes, but a complex interpretive task that requires imagination and discernment. It requires imagination because it requires to us to be able to see the interpretive possibilities. The exercise of the interpretive imagination is something at which art, history, literature, and autobiography all excel. The right portrait can make the seedy seem romantic, the innocent seem menacing, and the sublime seem ridiculous. It can exalt the debased and deflate the exalted. Tom Waits’s descriptions of waitresses in cheap diners, Knut Hamsun’s scathing portraits of intellectuals, Toulouse Lautrec’s romantic visions of Paris brothels, and Degas’s gorgeous portraits of the ballet have a revelatory character, making us see their objects differently. Interpretation isn’t a matter of getting a complete or detached view of things. It doesn’t strive for the kind of forensic accuracy that is prized in scientific representation. It is a matter of selecting and suppressing, foregrounding and enhancing.25 It is designed to bring out one particular pattern, suggest an evaluation, and often to encourage us to feel a certain way. New ways of seeing can

25.   Science is also very interested capturing patterns, but it tends to be interested in patterns that reveal regularity. It is interested in laws, and in induction, so it tries to isolate the shared and generalizable elements in nature.



transform how we experience the world, making us see familiar objects in a novel way and discover value where we hadn’t seen it before.26 It demands discernment, because entertaining interpretive possibilities is only part of the task. We also have to decide what to think. When we have a fight with a spouse or feel affronted by an interaction with a colleague, we need to sort through what actually happened in our minds and arrive at some interpretation. In doing that, we are not trying to settle the simple narration of events (what he said, what I said, and in what order); we are trying to understand what happened in an evaluatively rich sense. We are trying to understand whether we have been wronged or are in the wrong, whether we should be hurt or apologetic, or whether we should be insulted or indifferent. On one interpretation of the fight with the spouse, I came home after a very difficult day, he attacked me for no reason, I responded defensively, and instead of staying and sorting it out, he left the house, leaving me alone hurting and bewildered. On another interpretation, I walked in the door that night already wounded and needy. When he made a careless remark, I lashed out violently, and he left the house only to keep from angering me further. Getting this right matters in obvious ways, and it is not a simple task. It demands a disciplined and willful effort to see things from other people’s point of view, an unwillingness to acquiesce in self-​serving interpretations, and the resolve to be fair and honest in our assessment. We portray others in a manner that is flattering to ourselves, and often misdiagnose the sources of our emotions. Iris Murdoch has done more, perhaps, than anyone to describe the moral rigors of what I’ve called “getting it right.” Her most famous example concerns a mother-​in-​law who undergoes a transformation in her view of her daughter-​in-​law, and it is worth quoting in full. A mother, whom I shall call M, feels hostility to her daughter-​in-​law, whom I shall call D. M finds D quite a good-​hearted girl, but while not exactly common yet certainly unpolished and lacking in dignity and refinement. D is inclined to be pert and familiar, insufficiently ceremonious, brusque, sometimes positively rude, always tiresomely juvenile. M does not like D’s accent or the way D dresses. M feels that her son

26.   It is tempting to say that interpretation is about evaluation rather than fact, but this is to suppose a separation between fact and evaluation that is not psychologically faithful. The most basic descriptive terms that we use to represent a situation—​‘seedy’ vs. ‘romantic’, ‘innocent’ vs. ‘menacing’, ‘sublime’ vs. ‘ridiculous’—​don’t have a clearly identifiable, shared factual core.

Why (Study) the Humanities?


has married beneath him. Let us assume for purposes of the example that the mother, who is a very “correct” person, behaves beautifully to the girl throughout, not allowing her real opinion to appear in any way.  .  .  .  [T]‌ime passes, and it could be that M settles down with a hardened sense of grievance and a fixed picture of D, imprisoned . . . by the cliché: my poor son has married a silly vulgar girl. However, the M of the example is an intelligent and well-​intentioned person, capable of self-​criticism, capable of giving careful and just attention to an object which confronts her. M tells herself: “I am old-​fashioned and conventional. I may be prejudiced and narrow-​minded. I may be snobbish. I  am certainly jealous. Let me look again.” Here I  assume that M observes D or at least reflects deliberately about D, until gradually her vision of D alters . . . . D is discovered to be not vulgar but refreshingly simple, not undignified but spontaneous, not noisy but gay, not tiresomely juvenile but delightfully youthful, and so on. (Murdoch 1970, 16–​17) We are made to understand from Murdoch’s discussion that the altered interpretation of D comes closer to getting it right, and that getting it right is an epistemic matter, though not one that is simply a matter of looking. I think that we all understand what she is pointing to here. We have to interpret people and events all the time, and suffer the interpretations of others. Sometimes (as M’s case) getting things right is a matter of being kinder and gentler in our vision of others, but sometimes it is a matter of seeing them under the cold, harsh light of sober assessment. It is not easy to recognize that your lover is an overconfident bore, or that your child is a cruel bully. Getting things right doesn’t come easily. It brings with it a kind of truth and rigor that are quite different from the sort that one finds in the sciences. Murdoch said that it consists of “a refined and honest perception of what is really the case, a patient and just discernment and exploration of what confronts one, which is the result not simply of opening one’s eyes but of a certain and perfectly familiar kind of moral discipline” (Murdoch 1999, 330). Humanistic disciplines—​literature, history, and art, most obviously, but also philosophy, anthropology, and languages—​engage the interpretive muscles. In so doing, they help us see more deeply into ourselves and others, and cultivate the kind of understanding that helps us get it right in our own lives. There is no monolithic account of what the humanities are and what they do. These are just a couple of examples of the myriad complex ways in which the humanities help us make sense of ourselves and the world in which we live.



One of the complaints or frustrations that people who gravitate to science or math sometimes make about the humanities is the squishiness of the subject matter. They complain that everything is qualitative, impressionistic, and a matter of discretion or judgment. They say that there is no proof, no certainty, and no truth. It is correct that there is no proof and no certainty, but that does not mean that there is no truth. The kind of truth that the humanities strive for is softer than the kind of truth we have in math or in the sciences. One has to get comfortable with ambiguity and squishiness and the lack of full resolution—​i.e., with interpretation rather than calculation—​to operate in that environment. But it resembles in that respect the messy world of human affairs. Not everybody needs to be a writer, artist, or historian to make some knowledge of literature, art, and history valuable. Reading Dostoyevsky, Thomas Mann, Milton, Joyce, Herodotus, Omar Khayyam, Plato, and Charlemagne teaches things that will enhance our perception and enrich our experience of the world. It will deepen our understanding of ourselves and other people, tutor us in how to live, how to love, and how to feel. That is enough to make it a valued part of education. The upshot of all of this is that sciences and humanities don’t compete or conflict. They are entirely complementary, answering to different needs. The humanities provide a type of insight that is both essential to human living and not readily attainable from science. The everyday notion of understanding makes room for both. Of course, that leaves open the vexed question about what universities are for. That is a question that we need to address as a culture. We do need to prepare a workforce and produce researchers who will help us cure cancer and save the earth. But education does not have to be just about that. It can also be about helping us make better choices, getting us to care about the right things, and opening up the imagination to new ways of thinking and being. The academy should be a place where all of these things can happen. If we ask ourselves what we would like to pass on to our children, I think that many of us would say that, alongside hoping that they learn a trade or a job that will support them, we hope that they learn how to be imaginative and resourceful and fair, that they become the kind of human beings that know how to love and that approach other people with subtlety, perceptiveness, and understanding. These are valuable qualities within the workplace and without.

9.6. Conclusion It has taken time for science to mature so that we can see the importance of the humanities as emerging from within the scientific conception of the human

Why (Study) the Humanities?


being. If there was a time when the sciences and the humanities seemed to offer competing visions of the human being, that time is past. It is now possible to say on scientific grounds what is wrong with the idea that the sciences will ever replace (or displace) the humanities. The structure of human knowledge is complex, but it forms a single fabric, in which the humanities have their place alongside the sciences, and which every part makes a contribution to understanding of ourselves and our place in the universe.

Acknowledgments This work was supported by the generous support of the Templeton Foundation. I  owe a large debt of gratitude to Stephen Grimm. The chapter would not have been written without him, and his insightful comments made it much better than it would have been otherwise. I  would also like to thank Elisabeth Camp and Barbara Tversky for very helpful comments and the audience at the “Varieties of Understanding” conference, Fordham University, June 2015.

References Godfrey-​Smith, Peter. 2004. “On Folk Psychology and Mental Representation.” In Representation in Mind:  New Approaches to Mental Representation. Ed. Hugh Clapin, Phillip Staines, and Peter Slezak. Amsterdam: Elsevier. 147–​62. Hepburn, Ronald W., and Iris Murdoch. 1956. “Vision and Choice in Morality.” Proceedings of the Aristotelian Society, Supplementary Volumes 30: 14–​58. Ismael, Jenann. 2016. How Physics Makes Us Free. New York: Oxford University Press. Kitcher, Philip. 2012. “The Trouble with Scientism.” New Republic, 24 May. http:// ​ w ​ a rticle/ ​ b ooks-​ a nd-​ a rts/​ m agazine/​ 103086/​ scientism-​humanities-​knowledge-​theory-​everything-​arts-​science. Murdoch, Iris. 1970. The Sovereignty of Good. New York: Routledge. —​—​—​. 1999. Existentialists and Mystics:  Writings on Philosophy and Literature. New York: Penguin. Nussbaum, Martha. 2010. Not for Profit:  Why Democracy Needs the Humanities. Princeton, NJ: Princeton University Press. Rand, Ayn. 1999. Why Businessmen Need Philosophy. New York: Ayn Rand Institute. Rosenberg, Alex. 2011. Atheist’s Guide to Reality. New York: Norton. Zakaria, Fareed. 2015. “Why America’s Obsession with STEM Education Is Dangerous.” Washington Post, 26 March. https://​​opinions/​why-​stem-​wont-​make-​us-​successful/​2015/​03/​26/​5f4604f2-​d2a5-​11e4-​ab77-​ 9646eea6a4c7_​story.html.


Understanding and Coming to Understand Michael Patrick Lynch

10.1. Introduction Suppose I ask you how to get to Larissa, and you give me the right answer. Suppose further your answer is not a guess; you have some grounds for it. There are lots of different ways that could happen. For example, you might: • Look it up on your phone • Remember how you got there last year • Do both of these things but also explain why certain routes that look good on the map are actually impossible or difficult because of the geography and road conditions

As I  see it, all three of these actions might result in conditions that could ground your knowledge of how to get to Larissa. Such conditions represent three different ways our opinions can be grounded, by being based on: • Reliable sources • Experience or reasons that we possess • A grasp of the causal relations between local conditions and the feasibility of local travel routes

If, like me, you are tolerant of a sensible pluralism about knowledge, you’ll be comfortable saying that these different kinds of grounding give rise to different

Understanding and Coming to Understand


kinds of knowing.1 The first sort of knowing is the sort we engage in when we absorb information from expert textbooks or good Internet resources. The second is the sort of knowing that occurs whenever possessing reasons or experience matters. And the third is different still—​it is the sort of knowing we expect from experts—​even if those experts are more intuitive than discursive in their abilities. This is what I’ll call understanding. Understanding in this sense is what we have when we know not only the “what” but the “how” or the “why” (see Kvanvig 2003; Grimm 2006). Understanding is what the scientist is after when trying to find out the root causes of Ebola outbreaks (not just predict how the disease spreads). It is what you are after when you want to know why your friend is so often depressed (as opposed to knowing that she is).2 Many philosophers take understanding to be a distinctive kind of knowledge that is particularly valuable. The kind of knowledge in question concerns grasping dependency relations. This chapter aims to investigate and address two well-​known puzzles that arise from this conception. The first concerns the nature of understanding itself—​in particular, the nature of the “grasping” relationship that understanding is thought to involve. The second concerns the source of understanding’s distinctive value. In what follows, I’ll argue that we can shed light on both puzzles by recognizing, first, the importance of the distinction between the act of coming to understand and the state of understanding; and, second, that coming to understand is a creative act.

10.2. Understanding: The Functional Role “Understanding,” like “perceiving,” displays a typical state/​act ambiguity. Taken in the first sense, it refers to a particular kind of epistemically valuable intentional cognitive state.3 While the details differ, most views of understanding agree on certain common features they take the state to have. These

1.   Ernest Sosa is the leading pluralist about knowledge in this sense. See Sosa (2010). 2.   Philosophical accounts of the state of understanding often differ over whether to take understanding as a form of knowledge or not. This is an important difference, although how important may depend on one’s account of knowledge; pluralist accounts, like the one I favor, are willing to take “knowledge” as multiply realizable. While I will continue to take understanding as a form of knowledge in what follows, the substantive contributions of this view are consistent with holding that the concepts are more distinct. For discussion, see Zagzebski (2001), Kvanvig (2003), Grimm (2006). 3.   The state I go on to describe is degree-​theoretic. One can understand more or less.



commonly cited properties of understanding can be used to help fix the reference of the “state” use of the term, and help us get clear on what we are talking about. First, the state obviously is meant to be capable of having positive epistemic status—​moreover, it can convey such status on other states. The state of understanding has probative force. Second, it is directed at how or why something is the case. Thus, you understand more about the civil rights movement if you understand why and how it came about; you understand string theory if you understand why it predicts certain events; you understand a person to the extent you don’t just know that she is unhappy, but what makes her unhappy. We can take it that what is common between understanding how and understanding why is that we know something about the structure of the whole (see Grimm 2006; 2011). This sounds grand, and it can be, as when we understand how a proof works or why a great historical event occurred. But it can also happen on a smaller scale. Consider, for example, the lucky person who understands how her car works. She has this understanding in part, as we’ll discuss more fully in a moment, because she has certain skills, skills that give her the ability to see how various parts of a machine depend on one another: you can’t get the car to move without the battery and the battery won’t be charged without the alternator. You understand when you see not just the isolated bits, but how those bits hang together. Similarly with understanding why something is the case. When we understand why something is the case—such as why a certain disease spreads, or why your friend is unhappy, or why a given apple tree produces good apples—we do so because we grasp various relationships. These relationships are what allow us to see the difference between possibilities, between one hypothesis and another. If this is so, we might say that a third common thought about understanding is that understanding why or how is the result of grasping actual dependency relations, not just correlations. An instructive example is Plato’s Euthyphronic contrast: x is holy when, and only when, x is loved by the gods. Instances of this schema will be universally true. They might be true in all possible worlds. But simply grasping the instances doesn’t add up to understanding why what is holy is loved by the gods, or how holiness and the will of the gods are metaphysically situated in terms of one another. Therefore, it doesn’t

Understanding and Coming to Understand


add up to understanding the nature of holiness. To truly understand, you also need to know what depends on what. The dependency relations we grasp when we understand can come in different forms. Some relations might be about cause and effect. Think of a game of chess: if I move my bishop to a certain square, I cause it to change its position. But they might also be logical: if I move my bishop to this square, it will be vulnerable to your pawn. Or semantic: the bishop can move to that square because the rules define it as being able to move diagonally across the board. In other words, the first important element of understanding a game like chess is grasping dependency relations: having systematic knowledge of how things both fit together and depend on one another, causally, logically, and otherwise.4 Of course, knowledge of certain kinds of dependency relations might be particularly relevant to certain kinds of understanding. It seems plausible that scientific understanding, for example, gives pride of place to knowledge of causal relations. Understanding, seen as stemming from the grasp of dependency relations, is consistent with, if it does not entail, holding that understanding can be directed at both theories and persons. Understanding a theory, on this view, would involve understanding the dependency relationships between the principles and theorems that constitute the theory. Likewise, understanding a person would amount to understanding the relationship between their cognitive and emotional states and their behavior in certain contexts. In this sense, understanding a person amounts to grasping what “makes them tick,” as we say.5 The state of understanding is also seen as being related, directly or indirectly, to other cognitive states and an agent’s behavior. Thus the idea that understanding is the grasping of dependency relations supports the idea that understanding, as numerous commentators have noted, is tied to explanation. On some views, understanding of certain kinds involves having the grasp of a correct explanation, or at least having the potential for such a grasp (see Strevens 2013). But even if one does not take the (potential) to supply a

4.   This is a broadly Aristotelian account of understanding. See Greco (2014). See also Grimm (2006). (Not everyone sees understanding as involving knowledge. See Zagzebski 2001.) 5.   One complicating factor is whether, as Stephen Grimm has noted, one needs to also take a person’s beliefs and desires to be intelligible in order to understand her. Here, I think it is relevant to remember that understanding is a matter of degree. I understand you to some degree if I understand why you do what you do. I understand you more if I understand why you do what you do and find your beliefs and desires intelligible. See Grimm (2016).



correct explanation as a necessary condition for being in the state of understanding, it is plausible, at the very least, that understanding is conducive of good explanations.6 A related point is that the agent who understands thereby has certain abilities. The Oracle of Delphi supposedly announced that no one was wiser than Socrates. He famously replied that he only knew that he knew very little, or what he didn’t know. So what sort of knowledge did he have? Well, consider what he was truly good at. One thing, surely, was asking questions. This came from a combination of knowing facts and the ability to draw connections between them. As a result he had know-​which, as it were. He knew which questions to ask. This is suggestive. The person who understands is, to some degree, discerning not only the actual situation, but also why various hypotheses and explanations won’t work as well as how to ask what would (again, see Strevens 2013). They know that kicking the refrigerator here and not there will help get it working. This is something that experts in general can do. Indeed, experts—​ those who understand a given subject best—​are often able to increase their understanding even further because they have the ability to know which question they should ask in the face of new information. By so doing, they can, for example, reveal that Euthyphro knows nothing of piety. Arguably, however, the skill of being able to ask good questions itself hinges, at least in part, on a simpler (and less overtly verbally orientated) cognitive capacity: the ability to make inferences and draw out a position’s consequences—​and not just the actual consequences of, say, a given position on what causes apples to be tasty, but also the consequences of that position in certain counterfactual situations. This is precisely the skill that a good doctor employs when considering whether to administer a drug, or a lawyer uses when considering an argument. It is also, arguably, the skill a good mechanic employs when considering whether to disassemble a head gasket, or an apple farmer uses when deciding whether another farmer’s advice is reasonable. And those who have the capacity to cognitively engage, should they have the requisite verbal and linguistic abilities, will know which questions they should ask in order to carry their inquiries even further. This list of commonly cited characteristics of understanding is hardly exhaustive. But it can be used to give a partial functional characterization of 6.   Thus understanding need not be factive, although the deeper it becomes, the more it will approach factivity. To understand perfectly, perhaps, is factive. (For further discussion see Elgin 2009; see also Zagzebski 2001.)

Understanding and Coming to Understand


the state of understanding. Even a partial functionalist characterization can be used to fix the reference of the term. One suggestion might be this: U: A state of some agent plays the understanding-​role with regard to some subject when its content concerns dependency relations between propositions or states of affairs relevant to the subject; it is conducive of the agent’s ability to offer justified explanations of the relevant subject;7 and it disposes the agent to make further justified inferences both factual and counterfactual about the subject. Such a state would presumably have positive epistemic status and probative force. Like any functional description, this one still leaves much to be said. In particular, it leaves open the underlying psychological nature of the state or states that can play the role, and the value those states may or may not have when playing that role. Moreover, as we shall see, there is more to be said about the etiology of understanding—​the distinctive causal antecedents of states playing the understanding-​role.

10.3. Grasping and Coming to Understand So what is the state that plays the u-​role? A natural suggestion, given what we’ve said so far, might be a distinctive cognitive attitude we’ve called “grasping.” While I think this suggestion is intuitive, and has something to be said for it, I don’t think it is productive, for two reasons. First, and as Stephen Grimm has noted, the psychology of such relations is difficult to parse—​especially when we take it as constitutive of a stable state (Grimm 2011). One reason for that is that the root metaphor at work in “grasping” is obviously active. Grasping is something we do, and insofar as we think of it has having a distinct phenomenological character, it is a cognitive act, available to conscious attention. In contrast, understanding, taken as a state of mind, seems (like belief ) dispositional. Just as one might be in a state of decision (or indecision, as the case may be) without doing anything in particular,

7.   This clause is intended to be neutral with regard whether agents have such an ability in the first place. If they do have such ability, the state playing the understanding-​role will be conducive of that agent manifesting it. Children, for example, might lack that ability in certain contexts while still possessing some understanding. Thanks to Stephen Grimm for discussion.



or even being conscious of being in that state, we can understand why something is the case without being consciously aware of that fact, and without the understanding being available for conscious attention. Your understanding in such a case is tacit or implicit. Arguably, much of what we understand we understand in this way, and we can forget that we understand something that we do understand. As noted above, understanding is also thought to have distinctive value. A second problem with the suggestion that grasping is what plays the u-​role, is that it doesn’t particularly help answer this question. One might think at first that the reverse is the case. The thought would be that understanding is valuable because grasping dependency relations is valuable. That seems true, but it doesn’t say what is distinctively valuable about understanding. If cognitive contact with dependency relations is valuable, then it seems possible that we might have epistemically positive cognitive contact with such relations in other ways. That is, we might know about the relations without understanding them. If so, what constitutes the distinctive value of the grasping of those relations? One tactic is to argue that the value of grasping rests in the fact that understanding is “active” and not passive: as a result understanding, unlike other epistemic states, is a cognitive achievement (see Pritchard 2008). That is plausible. But it raises some questions all on its own.8 First, we might wonder what type of achievement understanding is. Second, if being in the state of understanding is an achievement, then being in that state must itself be the result of an act—​the act of achieving understanding. And generally, when we talk about achievements, we think that part of what makes the achievement admirable is that the agent voluntarily did something to achieve the goal. Achievements are partly valuable because, well, they were . . . achieved. In sum, I  think that it is implausible that grasping is what plays the u-​ role. On the one hand, it seems to overcomplicate matters: it posits an active process to play the role of what is plausibly a dispositional or implicit state. Second, it underexplains: it doesn’t by itself, at least, explain why understanding is held to have distinctive value over and above other kinds of knowledge. That said, I  think it clear that grasping should figure in any account of what understanding is. The question is where.

8.   Similar problems arise with the possibility that understanding is valuable because knowledge is valuable, and understanding either is a type of knowledge or leads to knowledge. That just seems less than illuminating, since it doesn’t actually tell us what is distinctively valuable about understanding itself and/​or makes that value instrumental to another.

Understanding and Coming to Understand


I suggest we can make a promising start on this project by paying attention to the other half of the “understanding” state/​act ambiguity. More precisely, I think it is worth thinking about the process of coming to understand. This should shed light on understanding (the state) itself. That’s because, like other kinds of mental states, the state’s etiology, or typical causal antecedents, help to pick out its functional role. Think of the state of believing that certain things are in your visual field. Being in that state is the causal result of having visual experience. Likewise, to understand, one must first come to understand. My hypothesis is that an analysis of the cognitive act of coming to understand can help shed light on why we are tempted to say that understanding (the state) is a cognitive achievement, and why it involves an activity like grasping. In the view we’ll entertain here, grasping is constitutive not so much of the state of understanding, but of the causally prior act of coming to understand; moreover, this prior act is partly definitive of the state it produces. If this is right, then the state of understanding is distinctive in part because of its etiology; one comes to be in that state only in virtue of having been caused to do so by first engaging in an active psychological process which is available to conscious attention.9 But what is it to come to understand? Coming to understand is a mental act in the same way that reflecting or deciding are mental acts. They are activities that your mind engages in. They take effort and increase the total cognitive load. A full description of the act of grasping is of course the job of empirical psychology and cognitive science; but prior theoretical reflection sharpens, here as elsewhere, our empirical inquiries. The sharpening I suggest is this: coming to understand, and therefore the grasping that helps to constitute it, is a creative act. In order to begin to see why this is plausible, and how it sheds light on understanding itself and its value, consider a (probably apocryphal) story about Descartes. Descartes was a late riser. His habit, when possible, was to stay in bed till around noon—​musing. One day, according to legend, he was watching a fly zoom around above his head when, suddenly, he realized that he could track its position by measuring its distance from the walls and the ceiling. He understood how to plot its flight path in space . . . and voilà! We get Cartesian coordinates, or so the story goes. The story of Descartes’s fly—​and others like it, such as those about Newton’s apple or Einstein’s clock—​are instructive because they emphasize that the

9.   Something can be available to conscious attention without, at that moment, being attended to.



moment of coming to understand can involve sudden insight. Such moments are often called “aha moments” and, in the psychological literature, are collectively taken to signify the “Eureka effect” (so named after Archimedes, who after a moment of great insight shouted “Eureka!”). Of course, most acts of understanding do not require the sudden novel inspiration that Descartes had. But all of them do involve some level of insight. Having such an insight is part of why understanding is fundamentally a creative act. Creativity, or creative acts, are complex. They are marked by having a combination of characteristics, characteristics that other acts can have but which taken together help to distinguish creative acts from other things humans do. One subset of creative acts—​a particularly important subset—​are creative mental or psychological acts. It is their characteristics I discuss here. First, creative mental acts are new or novel. As Margaret Boden has famously emphasized, creative ideas needn’t be historically novel—​like Descartes’s new geometrical ideas—​but they are psychologically novel or novel to the creator (see Boden 2004). Thus, being creative isn’t the same as being original. People can have ideas that are creative for them. As Boden says, “Suppose a twelve-​year-​old girl, who’d never read Macbeth, compared the healing power of sleep with someone knitting up a raveled sleeve. Would you refuse to say she was creative, just because the Bard said it first?” (2004, 2). I don’t think so, and neither does Boden. Creativity is relative to a person. Second, creative mental acts are generative of valuable, not just psychologically novel, ideas. Creative ideas are valuable to the person’s cognitive workspace. They move things forward on the conceptual field on which they are currently playing. They are useful and fecund. They have progeny, and they contribute to the problems at hand. Third, creative mental acts are typically the result of the cognitive effort distinctive of synthetic imagination. The psychological act involved in composing a song requires the ability to put together a complex string of different ideas about harmony, melody, and rhythm; the process of creating a coherent and believable fictional character likewise involves psychologically combining ideas of personality and physical description. That is, we think that creative mental acts often put things together in new ways. So far, coming to understand seems to fit this model of creative psychological acts: it involves generating in a synthetically imaginative way new and valuable ideas. Which ideas? Those that concern dependency relationships—​ how things fit together. The “grasping” of those relationships, which lies at the heart of understanding, is what makes understanding creative. This is most obvious in paradigmatic, historic cases of new understanding, like Descartes’s

Understanding and Coming to Understand


insight into how point location in Euclidean space can be plotted algebraically, or Einstein’s flash of understanding relativity upon seeing a clock. But what about less historically original acts of understanding? Consider again a child who comes to understand, for the first time, why 0.150 is smaller than 0.5. At that moment, the child is also having an insight—​a realization of how things are related. Or consider again our student above, coming to understand for the first time why Lady Macbeth sees blood on her hands, or why sailing is more pleasant and efficient when the wind is not behind you. Each of these acts of coming to understand are creative insights for the person in question, even though they are in no way novel. The three characteristics of creative mental acts we’ve canvassed so far are not exhaustive; there is a fourth—​one emphasized by Boden and more recently by Nanay (2014). Creative mental acts have a distinctive phenomenological feel. The phenomenology in question might be described as something akin to surprise. Boden calls this their “impossible” aspect—​that is, an idea is creative for a person when she has a felt sense that the idea could not have been had prior to the moment of creation. Conditions were right, and the person suddenly “sees.” It seems clear that the act of grasping/​coming to understand shares this characteristic as well. Indeed, it is particularly striking in this case. Coming to understand has a particular phenomenological appearance. In cases of sudden insight, this phenomenological aspect of creativity either constitutes, or leads to, the “eureka” feeling. But creative acts can be surprising even if they do not necessarily provoke that “aha” moment. Consider coming to slowly understand, for example, why a particular theorem followed from a particular set of premises—​you understood, as we say, the proof. Even when coming to understand happens gradually over time it still feels “new”—​as if you couldn’t have understood it prior to that moment. It feels as if you’ve made forward progress. That’s why it makes sense to say that the act of coming to understand is also surprising—​again, not necessarily in the “eureka” sense—​because the person who comes to understand feels as if they could not, relative to their past evidence and cognitive context, have understood it before that moment. It might be thought that not all acts of coming to understand can be creative in this sense. Surely, one might think, coming to understand the simple logical entailments of what else I understand cannot be a creative. In many cases, those entailments, and seeing why they follow, is too trivial to count as truly creative. In some cases, coming to know what follows from a proposition you understand, and why it follows, is not at all trivial. It requires great effort



and insight. In those cases you do come to understand the entailment. But in other cases such knowledge may well be trivial, and nothing in the account rules that out. That’s because coming to know a given entailment may not be an act of coming to understand. Understanding is a matter of degree, and our account of the state implies that the greater one understands, the more one is able to draw the relevant inferences. Part of drawing the relevant inferences is knowing why certain entailments hold from what you do understand. Thus, if one understands

to a sufficient degree one will, just by virtue of being in that state, know why certain entailments hold. That’s consistent with saying that while coming to understand

might be creative, one doesn’t separately come to understand (in the target sense) . That’s because if one understands

sufficiently, one is by virtue of that fact already disposed to know both and .10 Some might protest that this account of creative mental acts is too permissive. According to this line of thought, originating a new proof is creative. You are the first person that comes up with it. But simply coming to understand why the proof works isn’t creative. This objection confuses the ways in which something can be creative. A novel discovery or origination of a new proof is undoubtedly creative. Call this special type of creativity, which is very rare, o-​creativity (for “original”). But as I pointed out above, not all creative acts are o-​creative. (Consider, for example, the fact that someone might originate or discover a proof, the proof be forgotten for a thousand years, and then someone else might originate or discover it again.) Thus your act of first originating a proof might be o-​creative, but my act of coming to understand it for myself might be creative for me. And that’s creativity enough. Finally, it is worth emphasizing that the surprising or “impossible” aspect of creativity makes creating seem at once something we do (which it is) and at the same time something happening to us. The muse suddenly strikes. Realization comes in a flash. Coming to understand is like this as well. It involves insight, and insight, as the very word suggests, is like the voluntary opening of a door—a “disclosing,” as Heidegger said. One acts by opening the door, and then one is acted upon by seeing what lies beyond.

10.   In addition, it bears noting that even if one understands

, knows , and on that basis, comes to know , that fact alone doesn’t mean you come to understand since you may not understand .

Understanding and Coming to Understand


10.4. The Value of Understanding We can summarize the hypotheses floated in the last section as follows: To understand requires first coming to understand. And coming to understand involves actively grasping certain dependency relations, where grasping is a conative state of mind (both directed and active) with features associated with a creative mental act. In particular, the act of coming to understand/​ grasping is creative for a person to the extent that it generates ideas that are, for that person: • Novel • Valuable • Resultant from synthetic imagination • Distinctive in their phenomenology: their “surprising” or “impossible” aspect

Obviously, this description of the mental act of coming to understand/​grasping doesn’t tell us everything about its nature. But it does help us see what is distinctive and special about the act of grasping, and as a result, can help to guide further investigation. When we “look for” grasping in our psychological theorizing and experimentation, the present suggestion is that we look for a mental act with these characteristics. The above suggestion also allows us to adopt a more straightforward answer to the question of what occupies the u-​role—​that is, what kind of mental state is at play when we are in the state of understanding. The straightforward answer is that we are in a state of belief whose properties (both epistemic and psychological) allow it to play the understanding role. To understand is to believe in a certain way. But part of what it is to believe in that way is to arrive at the belief in a distinctive manner, to come to understand by grasping the relevant dependency relations. If so, then we can revise our functional analysis as follows: U*:  A  belief (of some agent) plays the understanding-​role with regard to some subject when the agent has been caused to be in that state by grasping dependency relations between propositions or states of affairs relevant to the subject; the ensuing belief is about those relations; having the belief is conducive of the agent’s ability to offer justified explanations of the relevant subject; and it disposes the agent to make further justified inferences both factual and counterfactual about the subject.



The above analysis is consistent, of course, with some kind of psychological state other than belief playing the understanding role in some agents. But it seems likely that it is a kind of belief that most often realizes the role in human beings. I began by suggesting that a focus on the act of coming to understand can shed light on two puzzles about understanding. The first puzzle was metaphysical, and concerned its nature. The above analysis provides an overall lesson. Like many other targets of psychological and epistemic analysis, understanding is distinguished by (1)  how we come be in the state; (2)  its properties and content; (3)  its effects on our behavior and dispositions to behave. Nonetheless, understanding is still a deeply interesting and important state of mind, one that our analysis predicts as requiring cognitive effort to achieve, the result of the act of grasping how things hang together. The second puzzle about understanding our suggestion may help to solve concerns its value. But before dealing with this question directly, it is illuminating to look at another long-​standing issue about understanding: its relation to testimony. Understanding is often said to be different from other forms of knowledge precisely because it is not directly conveyed by testimony—​and thus not directly teachable (see Zagzebski 1999; 2001). The thought is that you can give someone the basis for understanding via testimony, including the knowledge that they must have in order to achieve that understanding. But in the usual cases, you can’t directly convey the understanding itself. The question is why this should be. What is it about understanding that makes it difficult or impossible to convey via testimony? The present suggestion supplies an answer: in order to first understand, one must come to understand. And coming to understand is a creative act. As such, it requires a cognitive, generative psychological action on the part of the agent over and above whatever knowledge might be conveyed by another. An art teacher, for example, can give me the basis for creative thought by teaching me the rudiments of painting. She can give me ideas of what to paint and how to paint it. But I did not create these ideas; I create when I move beyond imitating to interpret these ideas in my own way. Likewise, you can give me a theorem without my understanding why it is true. And if I  do come to understand why it is true, I do so because I’ve expended some effort—​I’ve drawn the right logical connections. Coming to understand is something you must do for yourself. Let’s contrast this with other kinds of knowledge. I can download ordinary factual knowledge directly from you. You tell me that whales are mammals;

Understanding and Coming to Understand


I believe it, and if you are a reliable source and the proposition in question is true, I know in the receptive way. No effort needed. Or consider responsible belief:  you give me some evidence for whales being mammals. You tell me that leading scientists believe it. If the evidence is good, then if I believe it, I’m doing so responsibly. But in neither case do I thereby directly understand why whales are or aren’t mammals. You can, of course, give me the explanation (assuming you have it). But to understand it, I must first grasp it myself. Understanding can’t be outsourced. Earlier we noted that while it is intuitive that understanding has distinctive, perhaps intrinsic, value, it is unclear what the basis of that value happens to be. Our reflections on why understanding isn’t conveyed by testimony rely on the fact that understanding is partly defined by its etiology; to be in the state, one must first come to understand. This same fact helps to explain the distinctive value of understanding as well. Earlier we noted that we typically take understanding to be a cognitive achievement, and that fact is part of the explanation for why we think it is valuable. But seeing understanding as an achievement, we noted, means that the state must be something we do out of an act of will. It has to be, as it were, achieved. The present account dovetails with, and explains, this fact. We achieve understanding because we first come to understand—​an act that requires effort. Moreover, coming to understand is a creative act. And the creativity of that act helps to explain our intuitive sense that understanding is a cognitive state of supreme value and importance, not just for where it gets us but in itself. Creativity matters to human beings. That’s partly because the creative problem-​solver is more apt to survive, or at least to get what she wants. But we also value it as an end. It is something we care about for its own sake. And that goes for coming to understand as well. It is an expression of one of the deepest parts of our humanity.

Acknowledgments Thanks to John Greco, Duncan Pritchard, and Nathan Sheff for comments and discussion of this chapter, and to audiences at the University of Edinburgh, St. Louis University, Yale University, and Henderson State University.

References Boden, Margaret A. 2004. The Creative Mind:  Myths and Mechanisms. New  York: Routledge.



Elgin, Catherine. 2009. “Is Understanding Factive?” In Epistemic Value. Ed. Adrian Haddock, Alan Millar, and Duncan Pritchard. Oxford: Oxford University Press. 322–​30. Greco, John. 2014. “Episteme:  Knowledge and Understanding.” In Virtues and Their Vices. Ed. Kevin Timpe and Craig A. Boyd. Oxford. Oxford:  Oxford University Press. 285–​301. Grimm, Stephen R. 2006. “Is Understanding a Species of Knowledge?” British Journal for the Philosophy of Science 57: 515–​35. —​—​—​. 2011. “Understanding.” In The Routledge Companion to Epistemology. Ed. Sven Berneker and Duncan Pritchard. New York: Routledge. 84–​94. —​—​—​. 2016. “How Understanding People Differs from Understanding the Natural World.” Philosophical Issues 26.1: 209–​25. Kvanvig, Jonathan L. 2003. The Value of Knowledge and the Pursuit of Understanding. Cambridge: Cambridge University Press. Nanay, Bence. 2014. “An Experiential Account of Creativity.” In The Philosophy of Creativity:  New Essays. Ed. Elliot Samuel Paul and Scott Barry Kaufman. Oxford: Oxford University Press. 17–​35. Pritchard, Duncan. 2008. “Knowing the Answer, Understanding and Epistemic Value.” Grazer Philosophische Studien 77: 325–​39. Sosa, Ernest. 2010. Knowing Full Well. Princeton, NJ: Princeton University Press. Strevens, Michael. 2013. “No Understanding without Explanation.” Studies in History and Philosophy of Science Part A 44.3: 510–​15. Zagzebski, Linda. 1999. “What Is Knowledge?” In The Blackwell Guide to Epistemology. Ed. John Greco and Ernest Sosa. Oxford: Blackwell. 92–​116. —​—​—​. 2001. “Recovering Understanding.” Knowledge, Truth, and Duty:  Essays on Epistemic Justification, Responsibility, and Virtue. Ed. Matthias Steup. Oxford: Oxford University Press. 235–​51.


Against Understanding (as a Condition on Explanation) Bradford Skow

11.1. Understanding as a Condition on Explanation? It is often said that there is an important relationship between explanation and understanding. Like so many ideas about explanation, this one makes an appearance in Carl Hempel’s “Aspects of Scientific Explanation” (1965), the founding document of contemporary thinking about the topic. In the second section of that essay Hempel introduces the Deductive-​Nomological Model of explanation, according to which (I’ll risk the reminder) an explanation of a fact X is a sound argument for X that essentially contains a law-​stating premise. Hempel then writes: [A]‌D-​N explanation answers the question “Why did the explanandum-​ phenomenon occur?” by showing that the phenomenon resulted from certain particular circumstances, specified in C1, C2, . . . , Ck, in accordance with the laws L1, L2, . . . , Lr. By pointing this out, the argument shows that, given the particular circumstances and the laws in question, the occurrence of the phenomenon was to be expected; and it is in this sense that the explanation enables us to understand why the phenomenon occurred. (337) What is going on in this passage? In particular, why did Hempel think it important to say the things he says in that last sentence? One natural



interpretation has Hempel assuming, in the background, a necessary condition on explanation: The Explanation-​Understanding Condition: Something E is an explanation of some fact F only if someone who possesses E understands F—​ at least in normal circumstances. Then, this interpretation continues, what Hempel does explicitly in this passage is assert (it would be too much to call it an argument) that the DN model satisfies the necessary condition the Explanation-​Understanding Condition states. The Explanation-​Understanding Condition, if true, looks like it could be a really useful tool for evaluating theories of explanation. One basic way to evaluate a theory of explanation is to “directly check it against your intuitions”: find a body of fact that the theory counts as an explanation, and ask yourself, is it an explanation? Or, find an explanation, and then ask yourself, does the theory count it as an explanation? If the answer to either question is no, that’s a mark against the theory.1 But it would be nice to have other ways to evaluate a theory, besides directly checking it against intuitions. The Explanation-​Understanding Condition provides another way. Find a body of fact E that the theory counts as an explanation of F, and check if someone could possess E without understanding F.  If that is possible, then the theory must be false.2 In this way, it looks like the phenomenon of understanding can be used as an independent check on theories of explanation. The Explanation-​ Understanding Condition has been, and continues to be, widely accepted. Michael Friedman, for example, in his 1974 paper “Explanation and Scientific Understanding,” asks, “What is it about  .  .  . scientific explanations  .  .  . that [give] us understanding of the world?” (5). A page later he asks again, “what is the relation between phenomena in virtue of which one phenomenon can constitute an explanation of another, and what is it about this relation that gives understanding of the explained 1.   I take it that this is the method philosophers should call “checking your intuitions,” even though my statement of the method never uses the word “intuition” or any similar word. I realize this is controversial, but that controversy is not relevant here. (I am largely in agreement with Williamson [2008] and Cappelen [2012] about the role of so-​called intuitions in philosophy.) 2.   Of course, this method involves using your “intuitions” about whether someone in some hypothetical circumstances understands F.

Against Understanding (as a Condition on Explanation)


phenomenon?” (6). Friedman’s questions presuppose the Explanation-​ Understanding Condition. I could go on with examples like this for a long time. Here are two more: in 1984 Elliott Sober asserted that “Explanations afford understanding” (78). In 2003 James Woodward asserted that “It is a plausible constraint on what an explanation is that it must be something that provides understanding” (179). (I will have more to say about these authors’ use of the Explanation-​Understanding Condition later.) I am “against” the Explanation-​Understanding Condition. On one interpretation, the condition is false. On another interpretation, it can’t, after all, be any use for evaluating theories of explanation. A certain idea has shadowed philosophical debates about explanation like an assassin. That idea is that theories of explanation really are, or really should be, theories of answers to why-​questions. I believe in this idea. I believe that all that philosophers of science should want out of a theory of explanation is a theory of answers to why-​questions. And I hold that thinking about the phenomenon of understanding is no help at all in evaluating theories of answers to why-​questions.

11.2. Explanations as Answers No theory of explanation a philosopher has proposed has ever really been a theory of all kinds of explanation. Instead they have always been, at best, theories of explaining why. Suppose Smith asks Jones why the moon is waning, and in response Jones explains why the moon is waning by asserting some propositions (presumably about, among other things, the earth’s and moon’s orbits). The DN model, and every theory that has followed, can be read as aiming to say something about what it took for Jones to have succeeded in explaining (to Smith) why the moon is waning. But now suppose that Smith asks Jones what a gene is, and in response Jones explains what a gene is by asserting some propositions (presumably about inheritance, maybe also about chromosomes and DNA). There is no way to read either the DN model or any theory of explanation that has followed as aiming to say anything about what it took for Jones to have succeeded in explaining (to Smith) what a gene is. I said in the last paragraph that every theory of explanation since the DN model can be read as a theory of explaining why. That’s not entirely true. Bas Van Fraassen’s theory of explanation, in The Scientific Image (1980), can’t. His theory takes the form of a theory of why-​questions and their answers. That’s not the same thing as a theory of explaining why. Assuming that the



“canonical form” of an answer to the question why Q is something of the form “Q because R,”3 a theory of answers to why-​questions will fill in the schema (S1)  Q because R iff . . . A theory of explaining why, by contrast, will fill in the schema (S2)  Person P1 explained (to person P2) why Q iff . . . Since explaining is a speech act, to complete (S2) is to lay down the conditions required for someone to have successfully carried out this speech act. One does not have to lay down such conditions to complete (S1). What I just said might suggest that a theory of explaining why must do “more” than a theory of answers to why-​questions, in the sense that a theory of explaining why must contain a theory of answers to why-​questions as a part. This suggestion is not true. A theory of explaining why will certainly make use of the notion of an answer to a why-​question. But it can do this without containing, or entailing, any particular theory of answers to why-​ questions. In support of this, here is how I think the true theory of explaining why, the true completion of (S2), will start: (S3)  P1 explained to P2 why Q iff P1 told P2 the answer to the question why Q in the following way/​manner: . . . Right there after “told P2” is the reference to the answer to the question why Q.  But (S3), while it mentions answers to why-​questions, doesn’t contain a theory of such answers. Of course (S3) is not itself a theory; it trails off into ellipsis. But I don’t think that the true completion of (S3) will contain a theory of such answers either. The stuff that comes after “way/​manner” will (obviously) specify a way of telling someone a proposition, but I can’t think of any way of specifying a way of telling that involves stating a theory of answers to why-​questions. When Van Fraassen proposed his theory of explanation, which, again, was a theory of why-​questions, he did not say that the criteria for judging his

3.   Certainly not all answers have this form; the answer to the question why that plant is closing its stomata may be that it is doing this in order to conserve water. I happen to believe that every “in order to” answer—​sometimes called teleological answers—​is equivalent to a because-​answer, but will not defend this claim here. (I defend it in ­chapter 6 of Skow 2016.)

Against Understanding (as a Condition on Explanation)


theory were different from those for judging earlier theories of explanation, like the DN model. But if his is a theory of why-​questions and their answers, while earlier theories were theories of explaining why, shouldn’t the criteria be different? The answer to this question would be yes—​if the question’s presupposition, that earlier theories were theories of explaining why, were true. The fact that Van Fraassen did not say that his theory should be judged on different criteria is some evidence that he, at least, thought that this presupposition was false. Anyway, it is my view that the presupposition is false. Although earlier theories can be read as theories of explaining why, I don’t think they should be read that way. I don’t think they’re best interpreted that way. It’s not only Van Fraassen’s theory of explanation that is really a theory of answers to why-​questions; that’s what all of them really are.4 I don’t take these observations to be terribly novel or even all that revolutionary. They’ve been around in some form or another from the beginning: Hempel himself says that explanations are answers to why-​questions.5 But even if the idea that philosophers of science should really be after answers to why-​questions is in fact widely accepted, it is not really taken to heart. For philosophers rarely phrase their theories as theories of answers to why-​ questions. They phrase them as theories of explanation, phrasing them as completions of one of these: • Fact X explains fact Y iff . . . • Fact X is an explanation of fact Y iff . . .

At one level this is just a matter of terminology:  there is, I  think, a well-​ established practice of using the noun “explanation,” in some contexts at least, as a general term for answers to why-​questions (only). In some contexts, “I was confused about the photoelectric effect; then Professor Smith gave me the explanation” means “I was confused about why the photoelectric effect 4.   There are some exceptions. For example, Achinstein is explicit that the theory he presents (1983) is a theory of the speech act of explaining. I think this makes his theory a kind of theory that philosophy of science should not be interested in. I will defend this claim below. 5.   However, Hempel did not say that every answer to a why-​question was an explanation; he isolated (or tried to isolate) a special class of why-​questions that he called “explanation-​seeking” (see Hempel 1965, 335). I argue against the existence of such a special class, and defend more generally the idea that theories of explanation really are, or at least should be read as, theories of answers to why-​questions, in ­chapter 2 of Skow (2016). Some of the distinctions I’ve been drawing were first drawn by Bromberger, in the 1960s. He observed that one can explain things other than answers to why-​questions, and emphasized that explaining an answer is different from telling someone the answer; see the papers collected in Bromberger (1992).



happens; then professor Smith told me (the answer to the question) why the photoelectric effect happens.” Still, overuse of the word “explanation” can seduce us into false claims and mislead us into accepting bad arguments. The Explanation-​Understanding Condition is a good example of this.

11.3. Against the Explanation-​Understanding Condition Here is the Condition again: Something E is an explanation of some fact F only if someone who possesses E understands F—​at least in normal circumstances. Is this right? That depends on how it is interpreted. The first thing we need to do, to make it more precise, is to reword it so that it does not speak of facts as the things that are explained, or understood. We do talk of explaining or understanding facts, but it is misleading to talk this way in this context. An explanation of the fact that the moon is waning is, in this context, just an explanation of why the moon is waning. Similarly, to understand the fact that the moon is waning is, in this context, to understand why the moon is waning. So let’s make the “whys” that belong in the statement of the Explanation-​ Understanding Condition explicit (from now on I’m going to leave off the qualification about normal circumstances): (EU) Something E is an explanation of why Q only if someone who possesses E understands why Q. Now I want to distinguish two readings of (EU). One reading takes “explanation of why Q,” as it appears in (EU), to mean nothing more than “answer to the question why Q.” Another reading takes talk of an “explanation of why Q” to be talk of acts of performing the speech act of explaining. Here is the first reading: (EU1) A proposition P is the answer to the question why Q only if anyone who knows P understands why Q. I should note that to get (EU1) from (EU) I did more than replace reference to an explanation with reference to an answer. There’s another difficult bit of

Against Understanding (as a Condition on Explanation)


wording in (EU): it speaks of “possessing” an explanation. Once we eliminate “explanation,” we are left with talk of someone “possessing” an answer to a why-​question. But what is it to “possess” an answer to a why-​question? The most natural thought is that possessing an answer is just knowing that answer. Now we have one interpretation of the Explanation-​ Understanding Condition. Do we have in it an interesting constraint on theories of answers to why-​questions? I don’t think so, because I don’t think (EU1) is even true. Since knowing why Q is the same thing as knowing the proposition that is the answer to the question why Q, (EU1) amounts to saying that knowing why Q is sufficient for understanding why Q. But this claim is false. Understanding is a greater achievement than knowledge, not a lesser one. Consider, for example, Lester. Lester has never taken a chemistry class, or studied the subject on his own. But like all of us he has heard people use the words “acid” and “base,” and words related to them, like “acidic.” He has heard people say things like “Lemon juice is acidic,” and “Baking soda is basic.” He’s had enough exposure to these words for them to be part of his vocabulary. For example, when he called his doctor recently with a stomach ache, and she asked him if he’s had anything acidic to eat or drink recently, he understood her question, and replied that he had had some orange juice earlier in the evening. Now the other day Lester’s niece was showing him the lab experiment she’d done in her chemistry class. She dipped a piece of litmus paper into a liquid, and it turned red. Lester was very curious about this result. “Why did the litmus paper turn red?” he asked her niece. “Because this stuff I dipped it into is an acid” she replied. Here Lester’s niece told him the correct answer to the why-​question he asked. Moreover, she knew that was the correct answer, having studied this experiment in her chemistry class. Furthermore, Lester has no reason to doubt her testimony. So after hearing his niece’s answer Lester knew that the paper turned red because it was dipped it in acid. In other words, Lester knew why the litmus paper turned red. But, I claim, he did not understand why the litmus paper turned red. This is a counterexample to (EU1). What more would it take for Lester to understand why the litmus paper turned red? I am not really sure. It doesn’t really matter for what is to follow. But for what it is worth, here is some speculation. One possibility is that Lester needs more knowledge. He needs to know more than just the answer to the question why the litmus paper turned red in order to understand why the litmus paper turned red. Maybe what he needs to know, in addition, is some chemistry. More specifically, maybe he needs to know something about why dipping litmus paper into acids turns it red. If he knew that, he would



know something about the “connection” between the event the why-​question concerns (the color-​change), and the fact offered in the answer to that why-​ question (that the paper was dipped in acid). And this is not just any old connection; instead, it appears to be the connection in virtue of which that answer counts as the answer. That is, the fact that acid causes litmus paper to turn red via such-​and-​such chemical process appears to be the (or at least an) answer to the question “Why is it true that the paper turned red because it was dipped in acid?” If these speculations are on the right track, then they suggest a generalization:  maybe, in general, understanding why Q requires not just that one know the proposition that is the answer to the question why Q, but also know something about why that proposition is the answer. In terminology I introduce in (Skow 2016), the claim is that understanding why Q requires not just knowing some or all of the reasons why Q, but also knowing why those reasons are reasons.6 So the first interpretation of (EU), namely (EU1), is not a principle that can be used to evaluate theories of answers to why-​questions, because it is false. I  turn now to a second interpretation of (EU). This interpretation focuses, not on the answer to the question why Q, but on the act of explaining why Q. It interprets (EU) as stating a necessary condition on having performed this speech act. Here is one way to state this condition: (EU2) Person P1 explained to P2 why Q only if, as a result of what P1 did, P2 understood why Q. What (EU2) is saying is that to count as having explained to P2 why Q, as opposed to, say, telling P2 why Q, one must do more than get him to believe the answer to the question why Q. What more that is (EU2) does not say—​ except that it is whatever it takes for P2 to end up understanding why Q. So is (EU2) right? I honestly have no idea. I can see reasons to doubt it. There are plenty of principles that compete with it. Here are a few: 6.   Stephen Grimm holds that understanding why E happened requires knowing, not just that C (a cause of E) happened, but also knowing something about the “modal relationship” between C and E (Grimm 2014). It may be that the connection between C and E in virtue of which “because C happened” is the answer to the question why E happened just is the modal connection Grimm focuses on. But I will not pause here for a detailed comparison of the two views. Michael Strevens holds that understanding why E happened requires directly mentally apprehending the explanation of E, where knowing a proposition is not sufficient for directly mentally apprehending it (Strevens 2013). If we assume that “the explanation of E” here denotes the answer to the question why E happened, then I disagree; one need not directly mentally apprehend the answer, but one does need to know other propositions in addition to the answer.

Against Understanding (as a Condition on Explanation)


(EK) P1 explained to P2 why Q only if, as a result of what P1 did, P2 knew why Q. (EP) P1 explained to P2 why Q only if, as a result of what P1 did, P2 was in a position to know why Q. When I say that these principles compete with (EU2) I don’t mean that they are incompatible with (EU2); in fact, assuming that understanding why Q entails knowing why Q, (EU2) entails each of them. They compete with (EU2) not for truth, but for acceptance. Maybe having one’s audience (merely) know why Q is the strongest (relevant) necessary condition on successfully explaining why Q; maybe one’s audience doesn’t also need to have achieved so much as having come to understand why Q. When I compare (EU2) to (EK) and (EP), I just don’t know how I would go about choosing between them. But I also think it doesn’t matter. What reason does a philosopher of science have to care whether (EU2), rather than one of the others, is true? The people who have reason to care are philosophers who aim to have a theory of the speech act of explaining. But, as I argued above, that’s not what philosophers have been looking for under the heading of a theory of explanation. Of course, it’s possible that while philosophers of science haven’t been looking for a theory of the speech act of explaining, they should have been. Now if philosophers of science should be looking for a theory of explaining, that must be because a complete philosophy of science requires such a theory. But in fact I think that a complete philosophy of science shouldn’t include a theory of explaining. A complete philosophy of science would have to include a theory of explaining if, but only if, explaining were in some way part of the nature of science. Now we do sometimes say things that suggest that it is. Some philosophers say that one of the aims of science is to explain why things happen. They say this in opposition to those who say that science aims to provide (just) a comprehensive description of what happens. Whether they are right is contentious; this debate is part of the debate over “scientific realism.” But just what claim is it that the debate is about? Does “science aims to explain why things happen” mean just “science aims to answer why-​questions,” or does it mean “science aims to inform the public (or whomever), by means of performing the speech act of explaining, why things happen”? Once this distinction is made, I think it’s obvious that only the first interpretation has a chance of being true. Or I should say, it is only the first interpretation the scientific realists mean to affirm.



What exactly is scientific realism? It doesn’t seem to be the same thing to everyone. But certain doctrines appear to be central:  the doctrine that our current scientific theories are close to being true; the doctrine that it is a constitutive aim of science, as a practice, to produce true theories; and—​I think—​the doctrine that it is a constitutive aim of science to produce theories that are not just true, but also contain answers to why-​questions. But I just can’t see why anyone would think it was a constitutive aim of science that its practitioners perform any particular speech acts. Couldn’t a mute Robinson Crusoe have been an excellent scientist? Maybe he finds himself alone in Geneva in a postapocalyptic world. He fixes up the Large Hadron Collider and does a bunch of experiments. He improves on the current formulation of quantum field theory, devising a theory that is closer to the truth. He discovers the answer to the question why electrons, or Higgs bosons, do this or that. But he never says anything (not to anyone else, there being no one else there, and not to himself either). He never performs any speech acts. In particular, he never explains anything to anyone. Would any scientific realist really maintain that Crusoe was in some way defective as a scientist—​that he was failing to pursue all of the aims of science? Let’s take stock. I  distinguished two readings of the Explanation-​ Understanding Condition. One reading, (EU1), says that some fact about understanding is a necessary condition on something’s being an answer to a why-​question. But (EU1) is false. The other reading, (EU2), says that some fact about understanding is a necessary condition on someone’s having (successfully) performed the speech act of explaining. Maybe (EU2) is true, but whether it is is of no interest to the philosophy of science.

11.4. An Example: Sober on Explanation and Causation What I want to do next is look at a few examples of philosophers invoking the Explanation-​Understanding Condition. We are now in a position to see the mistakes these invocations have led to. I selected the examples to look at from important works in the philosophy of science, but the selection was otherwise random. There are a great deal more I could have looked at. In his 1984 book The Nature of Selection Elliott Sober took up the question of whether an organism’s fitness causes, or explains, its survival. This led him to a discussion of whether a disposition, say a sugar cube’s being soluble,

Against Understanding (as a Condition on Explanation)


causes, or explains, its manifestation (the sugar cube’s dissolving). Sober accepted that the solubility causes the dissolving, but claimed that “we feel somewhat disappointed when told that the sugar lump dissolved in water because it was water-​soluble” (77). Being told that the sugar lump is water-​ soluble is satisfying when we have asked what caused it to dissolve, but is (in his words) “disappointing” when we have asked why it dissolved. To reconcile the claim that the solubility is a cause with the claim that citing is does not make for a “satisfying” explanation, Sober set out to “disentangle issues of causation from issues of explanation”: Consider the question “What caused Y?” One correct answer would be “the cause of Y.” However, this answer might rightly be classified as unexplanatory, since it does not provide us with any better understanding of why Y occurred than we started with. Explanations afford understanding; therefore, claims asserting that X explains Y are true or false partly because of the relation of those two terms to a third—​ namely, us. Causality is not similarly “consciousness-​ dependent.” Whether or not X caused Y is not in general influenced by whether we might find it interesting to be told this. (1984, 78) Let’s look closely at the argument Sober is offering. Here it is as he stated it: 1. “Explanations afford understanding.” 2. “Therefore, claims asserting that X explains Y are true or false partly because of the relation of those two terms to a third—​namely, us.” 3. “Causality is not similarly ‘consciousness-​dependent.’ ” 4. [Implicit conclusion:] It is false that, or at least not always true that, if X is a cause of Y, then X explains Y. I don’t think that this is a very good argument. Its problems start with line 1.  Line 1 is a version of the Explanation-​Understanding Condition. I  have proposed two readings of this condition. But the first reading, (EU1), is false, and the second, (EU2), is not an interesting thesis (from the perspective of the philosophy of science anyway). So the first premise of Sober’s argument is either false or uninteresting. The charge of uninterestingness might seem itself uninteresting. Surely what matters is how interesting Sober’s conclusion is, not how interesting his premises are. So before dismissing the reading of Sober’s argument that takes



line 1 to be (EU2), we should at least see what the argument’s conclusion looks like on that reading. Since (EU2) comes from (EU) by interpreting talk of explanations as talk of the speech act of explaining, we should interpret line 4 to also make a claim about the speech act of explaining. Here is one stab at an interpretation of line 4 like that: 4*. The following is false: for any description C of a cause of E, someone can explain to someone else why E happened by saying ⌜E happened because C happened.⌝ I think that 4* may well be true, but that whatever interest it has does not essentially depend on its focus on the speech act of explaining. Here is how I would put the main thought Sober wants to get at in the quoted passage above, that 4* seeks to capture: if I ask you why some event E happened, and you reply that E happened because its cause happened (let’s pretend that E has only one relevant cause), you will have failed in some way to respond to my question appropriately. Now I can see a route one might try to take from this claim to an interesting conclusion, a conclusion about what can and cannot be an answer to a why-​question. Here is how it goes, starting from the claim I ended the last paragraph with: 5. If, in response to the question why E happened, someone X responds by asserting that E happened because its cause happened, then X will have given a bad response to the question. 6. If, by asserting P, someone gives a bad response to a question, then P is not an answer to that question. 7. So the proposition that E happened because its cause happened is not an answer to the question why E happened. 8. So even though the cause of E is a cause of E, it is false that E happened because its cause happened; it is false that for every description C of a cause of E, ⌜E happened because C happened⌝ is true. I want to make two points about this argument. The first is about who this argument is supposed to target. Elsewhere in his book Sober takes issue with the thesis that to answer the question why E happened one must cite, or describe, causes of E. The argument (5)—​(8), if sound, would show this thesis to be false, under one interpretation. That interpretation says that ⌜E happened

Against Understanding (as a Condition on Explanation)


because R⌝ is true if and only if the sentence R in some sense describes a cause of E—​where any way of describing that cause is permitted. But this does not seem to me a strong point against those who think that answers must cite or describe causes. They will just opt for a weaker interpretation of their thesis; they will find a principled way to restrict which descriptions of the E’s causes are allowed as values of R. My second point is that the argument’s premise 6 is false. Asserting P can be a bad response even if P is an answer. How? Here is one way. It could be that P, while an answer, is not the answer the asker is looking for. I  might watch a window break after being hit by a rock,7 but not know whether the window was fragile, or whether instead it was made out of superglass that had just one point of weakness, the point where the rock was lucky enough to hit. I ask why the window broke. If you tell me that it broke because it was hit by a rock, that’s a bad response. But the problem with it is not that what you said was false. “Because it was hit by a rock” is true, and is (therefore) an answer to the question why the window broke. The problem with your response is, instead, that it was not the answer I was looking for. The point can be made without using a why-​question as an example. For any question, there are many true answers to that question that it would be wrong to assert when asked that question. Suppose you and I  are in some Boston greenspace or other, but I, being from out of town, don’t know which. Are we on the Common, or in the Public Garden, or what? I asked you where we are. If you reply “We are in Boston,” you give a true answer to the question “Where are we?” But it was wrong of you to assert that answer. It’s not the answer I was looking for. Similarly, suppose I ask you who came to the party, and you answer “The people who came to the party came to the party.” You haven’t said something false. You’ve given me an answer to my question. But it’s an answer I already knew. It’s not the answer I wanted. This could be what is going on with the response Sober is interested in, the response to the question why E happened that one gives by saying “E happened because its cause happened.” That’s certainly a bad response; but it could be bad not because it’s not an answer, but only because it’s not the answer the asker wanted. If this is right, then Sober’s observation that it would be bad to say “Because its cause happened” in response to the question of why E happened does not show that there is any restriction on what descriptions of E’s causes can appear in an answer to the question why E happened.

7.   After the window is hit by a rock, that is.



I have said that “We are in Boston” and “The people who came to the party came to the party” are correct answers to the relevant questions. This might be challenged. (I believe there are theories of questions and answers on which they are not answers.) But I don’t think Sober is in a position to challenge them. He accepts, right at the beginning of the long quotation above, that “The cause of E” is a correct answer to “What caused E?” To sum up: Sober, in the quoted passage, tried to use a connection between explanation and understanding to reach a conclusion about when descriptions of the causes of E can, and cannot, constitute explanations of the occurrence of E. I have argued that his argument was based on a false claim about the connection between explanation and understanding, and also sketched a way that someone who thinks that any description of the causes of E can appear in an answer to the question why E happened can resist his conclusion.

11.5. Another Example: Woodward against the DN Model We saw Hempel claim that the DN model satisfies the Explanation-​ Understanding Condition. In Making Things Happen, James Woodward turns this around: he argues that the DN model is false on the ground that it fails to satisfy the Explanation-​Understanding Condition. Woodward in his argument focuses on explanations that do not appear to cite any laws, such as Michael Scriven’s canonical example, Ink bottle:  The carpet is stained because Jones knocked over the ink bottle.8 Hempel’s view about Ink bottle, and other similar examples, was roughly that the sentence in Ink bottle is true, even though no law-​statements follow the word “because,” because that sentence, or someone who uses that sentence, in some way conveys information about the DN argument that is the “ideal” explanation of the carpet stain. Woodard calls the strategy behind this response “the hidden structure strategy.” The “hidden structure” is the ideal explanation; it is hidden because someone could truly assert that the carpet is stained because Jones knocked over the ink bottle without knowing what laws appear in the ideal DN argument. The hidden structure strategy is to say 8.   It became canonical by being quoted at length and discussed by Hempel (1965, 360).

Against Understanding (as a Condition on Explanation)


that a “because” statement can be true by “conveying information” about a structure that is hidden from (not fully known to) the person asserting that statement, or his audience. Woodward does not like the hidden structure strategy, and thinks that any theory of explanation that makes use of this strategy will fall afoul of the Explanation-​Understanding Condition. Here is how Woodward puts his argument: There is yet another reason for rejecting the hidden structure strategy. This derives from a general point about the epistemology of explanation and the connection between explanation and understanding.  .  .  .  It is a plausible constraint on what an explanation is that it must be something that provides understanding. To say that certain information is “part” of an explanation or contributes to its explanatory import is to say that this information contributes to the understanding provided by the explanation. This in turn imposes an epistemic constraint on what information can be part of an explanation and can contribute to its explanatory import: such information must be epistemically accessible to those who use the explanation. Put slightly differently, the idea is that the features of the explanation that endow it with explanatory import—​that make it an explanation—​ must be features that can be known or grasped or recognized by those who use the explanation; if not, it isn’t in virtue of possessing those features that the explanation produces understanding. On this way of looking at matters, there is something deeply puzzling about the suggestion that claims like [Ink bottle] explain or convey understanding in virtue of providing information about the existence of some underlying epistemically hidden structure, whether a DN argument or an ideal explanatory text. The mere obtaining of this structure, independently of anyone’s awareness of its existence, cannot be what accounts for people’s judgment that, for example, the impact of the knee on the desk is explanatorily relevant to the tipping over of the inkwell. If this line of thought is correct, it seems to follow that to the extent that information about the laws or structures that underlie singular-​causal (or other sorts of causal) claims is epistemically hidden from those who use such explanations, it cannot be that this information contributes to the explanatory import of these explanations. (179–​80)



While I agree with Woodward that the DN model is false, I don’t think there is a good argument against it here in this passage. Let’s walk through Woodward’s argument slowly. Woodward starts with a version of the Explanation-​Understanding Condition: 1. “It is a plausible constraint on what an explanation is that it must be something that provides understanding.” He also asserts a version of this condition about “parts” of explanations: 2.  “To say that certain information is ‘part’ of an explanation or contributes to its explanatory import is to say that this information contributes to the understanding provided by the explanation.” Woodward then says that claim 2 “imposes an epistemic constraint on what information can be part of an explanation and can contribute to its explanatory import.” The epistemic constraint that he infers from claim 2 is this: 3.  Information can be part of an explanation and can contribute to its explanatory import only if it is “epistemically accessible to those who use the explanation.” Presumably the argument proceeds from here by appealing to the implicit premise: 4.  Plenty of people use, or have used, the explanation in Ink bottle, to whom no law of nature is epistemically accessible. The argument’s conclusion then follows from lines 3 and 4: 5.  Therefore, no law of nature is part of the explanation in Ink bottle, and no law of nature contributes to its explanatory import. How does all this look if we eliminate “explanation” in favor of “answer to a why-​question”? It is not entirely obvious how to translate each of the premises. Here is one attempt. Line 1 becomes (EU1): 1*. A proposition A is an answer to the question why Q only if someone who knows A understands why Q.

Against Understanding (as a Condition on Explanation)


Line 2 takes line 1 and turns it into a claim about parts of an explanation, so line 2* should take line 1* and turn it into a claim about parts of an answer: 2*. If P is part of an answer A to the question why Q, then anyone who knows A-​P (“A minus P,” the information left over when P is “subtracted” from A) understands why Q to a lesser degree than someone who knows (all of ) A. Line 3 required parts of explanations to be epistemically accessible, and was inferred from line 2, so from line 2* we should infer a claim about parts of answers being epistemically accessible: 3*. If P is part of an answer A to the question why Q, anyone who answers the question why Q by asserting A knows P. Let me pause briefly: premise 3 spoke both of being part of an explanation, and also of contributing to that explanation’s explanatory import. But it’s hard for me to see why adding the stuff about explanatory import is needed; how could something be part of an explanation without contributing to that explanation’s explanatory import? In terms of answers, how could something be part of an answer without contributing to that answer? As far as I can tell, it couldn’t. So I’m leaving out the stuff about explanatory import in the *-​ed interpretation of Woodward’s argument. Let’s continue. There’s one premise left, and then the conclusion. I’ll put the whole argument here, with the last premise and the conclusion at the end: 1*. A proposition A is an answer to the question why Q only if someone who knows A understands why Q. 2*. If P is part of an answer A to the question why Q, then anyone who knows A-​P understands why Q to a lesser degree than someone who knows (all of ) A. 3*. (So,) if P is part of an answer A to the question why Q, anyone who answers the question why Q by asserting A knows P. 4*. Plenty of people who have answered the question why the carpet is stained by asserting the proposition Ink bottle did not know any law of nature. 5*. Therefore, no law of nature is part of the answer to the question why the carpet is stained in Ink bottle.



Is Woodward’s argument, interpreted this way, any good? Premise 1*, again, is (EU1), which I have rejected. That’s enough for the argument to fail. But let’s look at the rest of it. Maybe there’s a way to fix the argument, so that it doesn’t rely on 1*? I think there is. The argument’s intermediate conclusion, line 3*, looks plausible to me. Surely if you assert A you know A—​or at least this is usually the case. And if you know A you know (or at least are in a position to know) its parts. Of course, the idea of a “part” of a proposition is not an ordinary one. It is some kind of technical notion.9 But the paradigm case must be conjunctions: the proposition that X is part of the proposition that X and Y. In this case a proposition entails each of its parts (though not vice versa). Presumably this holds in general: if P is part of a proposition A, then A entails P. And it is not too wild to assume that if you know A, and A entails P, you are at least in a position to know P. (Maybe this looks better if we require you to know that A entails P; but such niceties don’t matter here.) So, interestingly, the *-​ed version of the argument arrives at a plausible intermediate conclusion, line 3*, an intermediate conclusion that, as I have stated it, says nothing about understanding; this despite the fact that it is reached from suspect claims about the connection between understanding why Q and being an answer to the question why Q. In fact, the rest of the *-​ed interpretation of Woodward’s argument goes fine as well. I think that 4* is true: plenty of people have used Ink bottle without knowing any laws (or, at least, we can easily consider a hypothetical scenario in which this is so). And the conclusion 5* follows from 3* and 4*. Despite all this, there is still a big problem with the *-​ed argument. The problem is not that it is unsound; the problem is that its conclusion is not the conclusion Woodward wants. Sophisticated defenders of the DN model, or of theories that descend from the DN model, Peter Railton, for example (whose views Woodward discusses at length in the chapter from which I am quoting), do not say that some law of nature is part of the answer in Ink bottle. They say instead that Ink bottle is itself only part of a larger and more complete answer, and that that larger and more complete answer contains laws of nature (see, e.g., Railton 1981). 9.   It receives a detailed development in Yablo (2014).

Against Understanding (as a Condition on Explanation)


Is there a better interpretation of Woodward’s argument, an interpretation that gets him to the right conclusion? Well there is at least a different interpretation; let us see if it is better. If you look back at the long quotation above, you will see that after Woodward drew the intermediate conclusion 3, he wrote: Put slightly differently, the idea is that the features of the explanation that endow it with explanatory import—​that make it an explanation—​must be features that can be known or grasped or recognized by those who use the explanation; if not, it isn’t in virtue of possessing those features that the explanation produces understanding. I actually do not think that this is conclusion 3  “put slightly differently.” I think that this is a very different idea. To ask what features of an explanation endow it with explanatory import, or to ask what features make it an explanation, sounds to me like asking why it is an explanation. If we focus, as we should, on answers to why-​questions, not explanation, the question we are attending to becomes the question of why a given answer to a why-​question is an answer. What do Woodward’s remarks suggest about the answer to this question? Here’s what I think they suggest: 6. If P is the answer to the question of why is it that R is the answer to the question why Q, then anyone who asserts R as an answer to the question why Q must be in a position to know P. Here P, the answer to the question of why R is the answer to the question why Q, describes the “features of the explanation”—​that is, the answer—​“that make it an explanation,” or answer. Unlike claim 3*, claim 6 can get us to the conclusion Woodward wants—​if it is true. Let’s first see how it gets Woodward where he wants to go, and then turn to the question of whether it is true. The argument from 6 against the DN model is, I think, straightforward. We need just two more premises: 7. If the DN model is correct, then some law of nature is (at least part of ) the answer to the question of why Ink bottle is the answer to the question why the carpet is stained.



8. Plenty of people who have answered the question why the carpet is stained by asserting the proposition Ink bottle did not know (and were not in a position to know) any law of nature. (This is premise 4* from the *-​ed argument, slightly revised.) From 6, 7, and 8, the desired conclusion follows: 9. The DN model is not correct. What about premises 7 and 8? Line 8 is certainly plausible, for the same reason that premise 4* from the *-​ed argument was. I myself don’t know of any laws of nature that could be added to Ink bottle to turn it into a DN argument. The plausibility of line 7 is harder to assess. Rather than dwelling on line 7, though, I will skip to the end: ultimately this argument does not succeed, because its crucial premise, 6, is false. At least that’s what I think. I think that someone could know the answer to the question why Q, and could tell someone who wants to know why Q that answer, without having any idea why that answer is the answer. Suppose Jones is on a tour of the Old Mill. The tour guide points to some charred wood, and says that these are the parts of the mill that burned in the fire of ’03. Jones asks why the mill caught on fire in ’03. The tour guide replies that it caught on fire because it was struck by lightning. Jones then persists in his questioning. He says, “That’s great, but I have another question. Why is that the answer? Why is it that ‘because it was struck by lightning’ is the answer to the question of why the mill caught on fire?” The tour guide is baffled, and a little annoyed. “I don’t know” he says. He gives Jones a look to get him to shut up. Still, despite not knowing the answer to Jones’s second question, the tour guide knew why the mill caught on fire. He also successfully told Jones why the mill caught on fire. It is important to distinguish the baffling question Jones asked from another one. Jones did not ask why the answer is true. If, after the tour guide had said “Because it was struck by lightning,” Jones had said, as my two-​year-​ old automatically says in response to pretty much everything I say, “Why?,” we would most naturally take him to be asking why the mill was struck by lightning. To ask that is to ask why the tour guide’s answer is true. That’s not the same as asking why it is an answer. I extracted premise 6 from some things Woodward said. But he didn’t just say some things that suggested 6; he also offered a little argument for those things. Maybe 6 will seem more plausible if we look at that argument? Here, again, is what Woodward wrote; I’ve italicized the relevant bit:

Against Understanding (as a Condition on Explanation)


the idea is that the features of the explanation that endow it with explanatory import—​that make it an explanation—​must be features that can be known or grasped or recognized by those who use the explanation; if not, it isn’t in virtue of possessing those features that the explanation produces understanding. This argument is worthy of our attention not just because Woodward offers it, but also because it appeals to a connection between explanation and understanding, and such connections are the main concern of this chapter. So what is the argument? I think it is a good idea to start by making the argument more explicit using Woodward’s own terminology, before trying to see how to put it as an argument for 6. Here is what the argument seems to me to be: 10. The features of an explanation that make it an explanation are (also) features in virtue of which that explanation produces understanding. 11. The features of an explanation in virtue of which it produces understanding are features that those who use that explanation can know it to have. 12. Therefore, the features of an explanation that make it an explanation are features that those who use that explanation can know it to have. The conclusion 12, of course, corresponds to premise 6 of the earlier argument: 6. If P is the answer to the question of why is it that R is the answer to the question why Q, then anyone who asserts R as an answer to the question why Q must be in a position to know P. What we need to do is “translate” lines 10 and 11 so that they can serve as premises in an argument for 6. I don’t think this is hard to do, in light of my earlier discussions of what could be meant by “explanation produces understanding,” and my earlier claim that “X in virtue of W” can be rendered as “W is the answer to the question why X.” I’ll label the translations of lines 10 and 11 with *s: 10*. If P is the answer to the question of why is it that R is the answer to the question why Q, then P is the answer to the question of why anyone who knows R understands why Q. 11*. If P is the answer to the question of why anyone who knows R understands why Q, then anyone who offers R as an answer to the question why Q is in a position to know P.



6. (Therefore,) if P is the answer to the question of why is it that R is the answer to the question why Q, then anyone who asserts R as an answer to the question why Q must be in a position to know P. This argument is valid, but the premises 10* and 11* don’t seem to me to have anything going for them. Start with 10*: the problem with it is that it presupposes that it is true that anyone who knows the answer to the question why Q understands why Q. I’ve said why I think this is false. I also don’t see much reason to accept premise 11*. I can see no reason to think that you can only answer a why-​question if you know why the answer you’re giving produces understanding in those who hear it. I have been examining Woodard’s argument that the DN model fails to respect some connection between explanation and understanding. I  found two different arguments that Woodward might have been giving; I  have argued that both of them fail.

11.6. Conclusion The idea that there is an important connection between explanation and understanding is a common one. The alleged connection, vaguely stated, is that explanations produce understanding. If it existed, this connection would open up a new avenue for searching for the true theory of explanation: just investigate what things produce understanding. But I think that the promise that comes with this alleged connection is not fulfilled. On one precisification of the vague idea, the connection does not exist—​knowing the answer to the question of why some event E happened is not sufficient for understanding why it happened. On another precisification the connection is this: someone has performed the speech act of explaining with respect to the question of why E happened only if their audience came to understand why it happened. Maybe so, but a theory of this speech act is not what philosophers of science have been after under the heading “theory of explanation.” After all this negativity I want to draw attention to the fact that my title is not “Against Understanding.” While I have opposed a certain use of the notion of understanding in theorizing about answers to why-​questions, I do not think that it is a useless notion for other purposes, or that the question of what it takes to understand why some event E happened is not itself independently philosophically interesting. In fact at least one place where the notion of understanding may be important is not hard to find. I said earlier that one of the aims of science is to answer why-​questions (at least, according to some

Against Understanding (as a Condition on Explanation)


scientific realists). But it is plausible that science aims at more. Certainly science aims not just to “have” the answer to the question of, for example, why the dinosaurs went extinct, in some thin sense of “have,” where believing it on very weak evidence is enough. Science aims to know why the dinosaurs went extinct. But even this might undersell how lofty the aims of science are. I find it plausible that science aims to understand why the dinosaurs went extinct; more generally, it aims, for each fact in its domain, to understand why that fact obtains. Even if we do not need to understand understanding in order to figure out what it takes to be an answer to a why-​question, we do need to if we want to understand science.

Acknowledgment I’d like to thank Fred Feldman for discussing this material with me.

References Achinstein, Peter. 1983. The Nature of Explanation. New York: Oxford University Press. Bromberger, Sylvain. 1992. On What We Know We Don’t Know. Chicago: University of Chicago Press; Stanford, CA: Center for the Study of Language and Information. Cappelen, Herman. 2012. Philosophy without Intuitions. New  York:  Oxford University Press. Friedman, Michael. 1974. “Explanation and Scientific Understanding.” Journal of Philosophy 71: 5–​19. Grimm, Stephen R. 2014. “Understanding as Knowledge of Causes.” In Virtue Epistemology Naturalized:  Bridges between Virtue Epistemology and Philosophy of Science. Ed. Abrol Fairweather. Heidelberg: Springer. 329–​45. Hempel, Carl. 1965. “Aspects of Scientific Explanation.” In Aspects of Scientific Explanation and Other Essays in the Philosophy of Science. New York:  Free Press, 331–​496. Railton, Peter. 1981. “Probability, Explanation, and Information.” Synthese 48: 233–​56. Skow, Bradford. 2016. Reasons Why. New York: Oxford University Press. Sober, Elliott. 1984. The Nature of Selection. Chicago: University of Chicago Press. Strevens, Michael. 2013. “No Understanding without Explanation.” Studies in History and Philosophy of Science 44: 510–​15. Van Fraassen, Bas C. 1980. The Scientific Image. New York: Oxford University Press. Williamson, Timothy. 2008. The Philosophy of Philosophy. Oxford: Wiley-​Blackwell. Woodward, James. 2003. Making Things Happen. New York: Oxford University Press. Yablo, Stephen. 2014. Aboutness. Princeton, NJ: Princeton University Press.


Understanding and Fluency J. D. Trout

12.1. Introduction In philosophy and psychology, there is now a widely used notion of the sense of understanding, typically a feeling conveyed by an explanation that may be true or false, and invoked to explain why people make the choices they do. This sense is most prominent when we experience wonder, or moments of insight. The notion of understanding itself, however, may not include this feeling of fluency, or possess a distinctive phenomenology. As a description of its use in both technical and nontechnical contexts, ‘Understanding’ seems loosely associated with properties like transparency (things we understand we can also introspect), or voluntary (cognitive) control (things we understand we can turn over in our mind). I will discuss recent empirical literature on attention and memory in which there are many kinds of candidate cases of understanding that lack these properties of transparency and voluntary (conscious) control. In fact, ‘understanding’ may be a term that denotes an unprincipled stew of states, processes, capacities, and goals that are sometimes present and sometimes not when we apply the term or deploy the concept. In light of the evidence I consider, understanding may be a metacognitive motley of unobservable states of memory and attention, a hidden assemblage of mechanisms that, together, yield cognitive capacities that can contribute to intellectual control and success. There is nothing in this description that supports the bold pronouncements of some philosophers that understanding must possess a specific property, usually one available to introspection. There may be value in seeking a unified

Understanding and Fluency


account of understanding, but it is an open question whether it consists in a set of necessary and sufficient conditions. We may experience a sense of understanding, a feeling of cognitive control or fluency that inclines us to believe we have mastered a skill or concept. Yet, in ordinary life, that cue is notoriously unreliable. Any unity we find in this sense comes not from the involvement of common mechanisms across diverse cases, but rather the phenomenological residue of these messy cognitive activities in the common goal of pursuing the truth. Accordingly, philosophical declarations about the requirements of understanding are not premature but misplaced. In my view, the very fact that people address such proposals at length shows that philosophical fluency does not reliably identify and weed out the products of unpromising methods and frivolous constructs in the discipline.

12.2. The Conceptual Analysis of ‘Understanding’ Imagine that you are given a rich assortment of measurement tools and theoretical models to characterize some prototheoretical process or event in our psychology, like understanding: reaction times and error rates to measure processing speeds and accuracy, and MRIs and PET scans to chart areas and levels of neural activation. Instead of using those tools and models, and remaining open to the possibility that the nature of the underlying processes may not be traceable by our introspection and brute associations, we reject those tools and models and rely on methods of conceptual analysis that haven’t changed in centuries. Sadly, this is no parody. In fact, this would seem an accurate description of the philosophical study of understanding. The concept of understanding, like the concept of love or anger or joy or sorrow, is many things to many people. ‘Understanding’ is a notoriously difficult concept to define. But then again, so are lots of words that have a life at once in the lay and the scientific worlds that serve many purposes—​words like ‘belief ’, ‘cloud’, ‘desire’, etc. But that is not news. If the model of philosophical analysis is to propose candidate definitions until one meets a certain conception of philosophical precision, we shouldn’t be surprised to find that most of these concepts just can’t carry the freight. Understanding is a kind of lurking presence in contemporary work in epistemology and on explanation. It is sufficiently related to these



important topics to bear mentioning, but not concrete enough to get any traction. This has left work on the nature of understanding half-​baked, tentative, impressionistic, and just plain programmatic. It is also in a terrible state. Unlike epistemology and explanation, which enjoy broad agreement about many of the basic issues (truth matters to knowledge, justification is an important feature of knowledge), participants can’t agree on even the most basic features of understanding. Is ‘understanding’ a success word (does it require “downstream successes”?) Is understanding a representation? Is it propositional? Is it a recognitional capacity, and if so, why doesn’t its cue sensitivity make the process of understanding just like the epistemic process of truth tracking, and so a species of knowledge? Existing analyses have, accordingly, expanded the senses of ‘understanding’ to understanding how, understanding why, understanding of, and understanding that. As with other concepts that populate the inventory of lay psychology, like belief or desire or intention or intelligence, its boundaries are fuzzy and its uses are driven by many forces of different strength and opposing or orthogonal direction. Such forces have little to do with truth or satisfaction conditions, like the affective need for cognitive closure, the pragmatic demands of workaday communication, and personal idiosyncrasy. Not surprisingly, then, understanding is a mongrel concept, used ambiguously, vaguely, and unevenly across languages, cultures, and ages. Accordingly, the research on understanding oriented to capture this character and variability is carried out by intellectual hybrids who have at once a deep interest in foundational issues and a deep respect for empirical approaches (see, e.g., Wilkenfeld, Plunkett, and Lombrozo 2016). Still, contemporary analytic philosophers tend to toil in their narrow and traditional specializations, and they have some surprisingly specific and often conflicting ideas about the extension and proper application of the concept of understanding. All of them are held with great conviction. On the basis of their own phenomenological and narrative evidence, some philosophers seem to have their own primitive theories of how psychological capacities are characterized and interact. For example, understanding is “a state that is constituted by a state of conscious transparency” (Zagzebski 2001, 246). You are aware that you understand whenever you do. Understanding is also said to have a special relation to curiosity; it supplies the “aha” moments that act as “legitimate closure of inquiry” (Kvanvig 2011, 89). Only understanding, not knowledge, has the power to slake our curiosity. (Never mind that some of the best-​known “aha” moments in history are fantastic blunders, marking

Understanding and Fluency


closures of inquiry that are stupendously illegitimate.)1 On yet another view, apparently, understanding has nothing special to do with transparency or curiosity, but is characterized primarily by reflective equilibrium: “The individual commitments that comprise such a system must be reasonable in light of one another, and the system as a whole must be at least as reasonable as any available alternative in light of relevant antecedent commitments.”2 It is hard to know what status these claims have. None of them strike me as obviously true, but even if they did, we should recognize that many false, even stupid, assertions strike humans as obviously true. Nor are these assertions the consequence of scientific theorizing. What, then, are we to make of their seemingly authoritative delivery and with it, the idea that this approach properly addresses deep and durable question? I will identify the origin of this conceit in a sense of understanding, a feeling of fluency, conveyed by entrenched methods and concepts in contemporary philosophy. In the meantime, let me suggest an alternative vision of many protoscientific concepts, including understanding. ‘Understanding’ is like a chimera. It is not a unified or monolithic ‘thing’, but a mongrel, an awkward and unprincipled mixture of traits. In its objective use, ‘understanding’ covers a psychologically loose assemblage of cognitive achievements. I will show how a systematic empirical perspective shows by decomposition how the term ‘understanding’ has no role to play in a working science, but it is worth noting that the term also has few of the hallmarks of unification even in ordinary usage. What tendencies it has in that direction are largely superficial artifacts of some prominent and compelling phenomenological experiences associated with general cognitive achievements. Philosophers engaged in this conceptual analysis might cast their efforts not as the final act, but a “place to start” in the analysis of the concept of understanding. But that is the standard claim of conceptual analysis, and it now has only the ring of ceremony. The specific conceptual analysis of understanding we’ve considered is not oriented toward projects confirmable by

1.   Trout (2002); also see Trout (2016). This tendency reaches into many domains, such as finance and investment. Increases in the sense of understanding, for instance, reduce perceived risk (Long, Fernbach, and De Langhe 2016). 2.   Elgin (2007, 2). Not surprisingly, philosophers more likely to take a philosophy-​of-​science approach to the study of understanding are less prone to this basic, distinctly philosophical, conceptual analysis. For example, marking his own view off from Zagzebski’s, Grimm says: “[U]‌nderstanding requires truth, is not transparent, and can be Gettiered” (2006, 516). Also see Strevens (2013) for a somewhat more naturalistic approach. For more work by philosophers who canvas empirical approaches to the notions of understanding and explanation, see Wilkenfeld (2017) and Khalifa (2017).



scientific (as opposed to anecdotal) support. It is not couched in a vocabulary deployed in research on memory, attention, or computation generally. It does not cite those sources. It does not entertain the possibility that its characteristics are moderated by individual variables, by demographic facts, by culture, or by epoch. In short, there is no evidence in these conceptual analyses that they are conceived as a place for psychologists and other interested parties to begin their investigation. On the contrary, the firewall of neglect they have built around the concept indicates that such empirical investigations are at best unnecessary, and at worst, unwelcome. In our common uses of ‘understanding’, lots of consequences are not just unarticulated, but undetermined. It is not as though there is some settled extension of the term, but we have not fully captured it; the extension is no more unfinished than the extension for ‘thought’ or ‘intention’ or ‘desire’. Focused efforts advertise the extension as under construction, and there isn’t much evidence of progress. Rather than the philosopher unearthing the implicit consequences of the notion of understanding, philosophers are often trying to make up a coherent notion of understanding that nothing in our use or theoretical practice dictates. Humans have inbuilt cognitive limitations that place some explanations permanently beyond their grasp. This is especially so for accounts of explanation that demand much from understanding. Our faculties of attention and memory arise from a rigid neurological architecture. The ordinary standards of understanding—​normally requiring that the causes responsible for the effect be introspectable and separately trackable—​cannot be met by an architecture that squeezes efficiency out of processing by rendering so much processing automatic, shallow, and opaque. Our perceptual and cognitive systems are capable of maintaining only very small numbers of memory and attention contents from just a few information sources. If we had undistorted and deliberate access to the content of attentional windows of arbitrary size, could track those events with less constrained powers of identification and discrimination, and could integrate the many temporal levels and modal qualities of information, we might be able to meet the standards of grasping or understanding normally imposed when we attempt to ‘understand’. In other words, if we had capacities that we don’t, we might have been able to do things we can’t.3 I don’t think you want an account of understanding that requires that you appreciate things like separable

3.   A conceptual treatment of this issue is found in McGinn (1991).

Understanding and Fluency


dependence relations, because almost nothing in the world is understandable in that sense. I will focus on two features of existing philosophical accounts of understanding. The first is that they include concepts like transparency, awareness, and control that enjoy a certain fluency in the philosophical tradition but belong to no scientific category. Second, and relatedly, the existing philosophical accounts of this fundamentally psychological notion—​understanding—​ are not formulated in way that could be confirmed by scientific evidence. Instead, they characterize understanding by performing conceptual analysis on the experience of understanding, dissecting and reassembling it using only those capacities that our best psychological science already demonstrates are unfit for routine identification and estimation. Although conceptual analysis aspires to yield a clear and specific product, some philosophers begin with very specific, even tailored, ideas about what understanding requires. Understanding, Zagzebski says, “is a state that is constituted by a type of conscious transparency” (2001, 246). But some lay and technical uses conflict with this conviction. We often say that a person understands a function even when they can’t articulate it or be aware of its features. Zagzebski continues: It may be possible to know without knowing that one knows, but it is impossible to understand without understanding that one understands.  .  .  .  [U]‌nderstanding is a state in which I  am directly aware of the object of my understanding, and conscious transparency is a criterion for understanding. Those beleaguered by skeptical doubts therefore can be more confident of the trustworthiness of putative understanding states than virtually any other state. (2001, 246–​47) When philosophers announce a condition for the application of the term ‘understanding’, it can be hard to tell whether they are playing a normative role like the prim grammarian, or a merely descriptive one of the anthropological lexicographer. Is this a claim about how we use the word? Is it an attempt to summarize a supposed consensus about the nature of understanding? With an evidence base so uncertain, we might think we should treat every assertion of the meaning of ‘understanding’ as speculative or risky, until the necessary evidence is specified. Nevertheless, Zagzebski claims that understanding is “The state of comprehension of nonpropositional structures of reality” (Zagzebski 2001, 242). Sooner or later, advocates admit to a strategy behind their specific treatment.



Some hold that an account of understanding is best treated as a virtue. Some seek an account continuous with our philosophical heritage in ancient Greece. Some believe that an account must properly navigate a particular strain of usage. “In the case of understanding, the objects would be something along the lines of ‘structures’ (Linda Zagzebski), or ‘systems’ ( Julius Moravcsik), or ‘information chunks’ ( Jonathan Kvanvig), or ‘dependency relations’ ( Jaegwon Kim and [Stephen Grimm])” (Grimm 2012, 105). Other accounts are very broad, leaving open the reasonable possibility that there are no features of understanding common to the cases we might be inclined to apply the term: “As a very crude first approximation, I suggest that understanding is a grasp of a comprehensive body of information that is grounded in fact, is duly responsive to evidence, and enables non-​trivial inference, argument, and perhaps action regarding the subject the information pertains to” (Elgin 2007, 9). Although it is good to be open about one’s philosophical strategies, admission is not absolution, still less a substitute for a reason that one might think understanding deserves a unified treatment. Other approaches fare no better. One might try to gain insight into understanding generally by looking at scientific understanding specifically, suggesting that scientific understanding is the state produced, and only produced, by grasping a true explanation.4 Much depends on how “grasping” gets worked out. But taken at face value, “grasping” does not belong to any scientific category. And, if it is circular to define ‘understanding’ in terms of the synonym ‘grasp’, you also don’t want it defined in terms of ‘comprehension’ either. “While these descriptions differ in various ways, if there is a common idea here it seems to be that understanding is directed at a complex of some kind—​in particular, at a complex with parts or elements that depend upon, and relate to, one another, and that the mind grasps or apprehends when it understands” (Grimm 2012, 105). If we had a psychologically informed and empirically rigorous account of ‘grasping’, that would go a long way toward characterizing understanding. But we don’t. In keeping with these dependency approaches, I  have a special place in my account of understanding for causation. After all, with the exception of a few special domains, only causal taxonomy promotes understanding. But my account gives priority to science. When it comes to uncovering causal taxonomies, science has no peers. The goal of this search for the causal details of

4.   An account friendly to this one can be found in Strevens (2013).

Understanding and Fluency


cognitive achievement is to provide the best theoretical systematization of the empirical facts delivered by cognitive and comparative psychologists.5

12.3. What the Concept of Understanding Is Made of The philosophical study of understanding involves a circle of interdefined concepts. Appreciating the structure of the situation, grasping, comprehending, manipulating representations—​all with their conscious and unconscious variants—​these are the central intellectual notions designed to elucidate understanding. Although the sense of understanding, when present, is palpable, it is really hard to know when genuine understanding is present. We can’t elucidate understanding until we look at the processes of memory and understanding beneath it. This is not a plea for reductionism; it is a call for explanatory independence,6 a description of notions that seem related to, if not constitutive of, understanding, that are implicated in a wide range of independently specifiable notions. The bottom line is, there are simply too many ways of representing cognitive achievements and failures equally as understanding. Memory and attention also have dual lives in the lab and the public square, and both make understanding possible. When we consider the application of those words by the many, we see the assumption that in order to understand a statement, concept, theory, rule, etc. we must use our memory capacity. In order to understand, we must recall the characteristics of the object of understanding. And in order to characterize those properties, we must be able to focus our attention on them. At the same time, memory and attention are not endless resources (Cowan 2005). There are limits on the number of items in memory that can be constantly activated ( Just and Carpenter 1992), the number of relations among

5.   De Regt and Dieks (2005) suggest two principles can explicate the notion of understanding, albeit of a scientific variety: 1.  A phenomenon P can be understood if a theory T of P exists that is intelligible. 2. A  scientific theory T is intelligible for scientists (in a context C) if they can recognize qualitatively characteristic consequences of T without performing exact calculations. Principle 2 is too weak, of course, because there is a potentially infinite number of “qualitatively characteristic consequences of T” that are deductive consequences but utterly uninformative about the causes producing those consequences. In those cases, I doubt we would want to say that we understand phenomenon P. 6.   For one account that emphasizes unification, see Kitcher (1989).



items in memory that can be integrated (Halford, Wilson, and Phillips 1998), and the limit enforced to avoid interference between items in working memory (Oberauer and Kliegl 2001). Feelings of understanding aside, cognitive success depends on the delicate alignment of multiple psychological mechanisms, far more intricate than crude schema like “activated portion of long-​ term memory,” “focus of attention,” and a set of “time-​locked sensory buffers.” For example, behind long-​term memory is a process called “long-​term potentiation” in which a memory is encoded by a biological process made stable by a chemical state. Potentiation does not have phenomenological correlates, and so we cannot characterize its theoretical properties by the crude gaze of unaided introspection. Memory is not the only complex process whose components are difficult to track by introspection or casual inspection. To cite another case, the faculty of attention proceeds by activating three broad component stages: Alerting, Orienting, and Executive. Both imaging and lesioning studies indicate that each of these processes is served by distinctive neural pathways, pathways whose functions are unlikely to be captured by analytic constructs arrived at through introspection of phenomenological contents (Geva et  al. 2013). When we acquire understanding through observation, for example, an observation of even simple objects requires attentional capture of separable dimensions of shape, surface contour, color, and motion, to mention just a few factors. That capture process is measured in tens of milliseconds. Any less time, it cannot be captured in the window, and it is lost to later processes of storage and report. Any more time, and it interferes with processes that come after or crowds out other analyses that co-​occur, placing so much load on working memory that we are unable to retain the event. Even the sensory analysis of a solitary dimension of color places measurable demands on one of many distinctive attentional processes (see, e.g., Belopolsky and Theeuwes 2010). Recent studies have confirmed and extended these observations to other dopamine genes and to the orienting network. In two different studies employing other conflict tasks, the catecholamine-​O-​methyltransferase gene was linked to the mental operations related to resolving conflict (Blasi and Mattay 2005; Diamond and Briand 2004). Different alleles of cholinergic genes were also related to performance on orienting tasks such as visual search (Parasuraman and Greenwood 2005), thus confirming the link between orienting and the neuromodulator acetylcholine. It may be that there is no understanding without explanation (Strevens 2013). But it is also true that there is also no understanding without memory, without attention, and without awareness. This causal understanding portrays

Understanding and Fluency


the object’s actual or potential changes over time, and it typically takes time to represent or imagine temporal sequences. We must be able to focus on some part of the problem and relate it to other parts. Sustained focus requires that we remember the details of those components for as long as we need to in order to represent the problem. All existing theories of attention and of memory treat these processes as incremental, and composed of identifiable ingredients. Many common explanations for complex phenomena are true but people find them hard to understand. The idea is this. Humans have severe limits on memory and attention, and this places limits on what they can (explain and) understand. But many phenomena that we try to explain involve many variables of different causal direction and potency. I have in mind domains like population ecology, or models of watershed pollution, which include dozens of variables. Understanding these is very different from understanding, say, why a billiard ball moves the way it does after impact, in which only a few factors are involved. Models of these domains can potentially contain hundreds of candidate predictor variables (Bishop and Trout 2005)—​demographic factors (age, sex, race, education level, region of the country), genetic profile factors (sometimes dozens of them), behavioral factors (nutrition, exposure to certain chemicals, etc.). With many diseases, no single variable is a good predictor on its own; most of the variables are weakly predictive. So weak, in fact, that it is pretty routine to use some kind of statistical variable selection process to prune down the number of variables in the model in order to arrive at something that’s “clinically interpretable”—​in other words, to work around the limits of our processing ability! Consider a simple model for predicting coronary artery disease (CAD), in which the model identifies patients at high risk for a coronary event. By allowing us to track changes in the risk of coronary events with the addition of each variable, the model helps us to understand the causes that bring about the coronary effects. But with twenty-​two variables in the model, there is not a chance in the world that we could have understood this risk anecdotally, nor could we have followed the causal influences without the aid of a computational model. The number of items we would have to recall is more than 300% of the number we can attend to tracking them unassisted. To make our infirmity more emphatic, we cannot track the potency of the causes and their direction. Some of the causes are orthogonal and there are interaction effects.7

7.   The twenty-​two predictors in the model were age, sex, CAD diagnosis, deprivation, smoking, hypertension, diabetes, lipids, heart failure, peripheral arterial disease, atrial fibrillation, stroke, chronic kidney disease, chronic pulmonary disease, liver disease, cancer, depression, anxiety, heart rate, creatinine, white cell count, and hemoglobin. See Rapsomaniki et al. (2014).



This model edges us toward scientific understanding of CAD, but also refines and corrects our understanding of CAD in lay cases. Our lay experience with the effect of coronary artery disease may be a massive coronary event, open to casual inspection: A man of about sixty years of age clutches his chest, struggles to breathe, and falls to the ground, unconscious. We try to understand. Why did this happen? Nonexperts are told “Heart attack.” That may seem like a useful category, but if we are looking for causes, it is a term of convenience used to pick out causes, conditions, and events, too numerous and tedious to list, and too diverse to be useful. We could continue to build an ontology on it, or employ a taxonomy of heart ailments based on the casually observable actions. That is the philosopher’s discretion, and nothing about the organization of the profession prevents it. But no effort at understanding could rationalize doing so. Like calling a manic depressive “crazy” or “possessed,” this amounts to the willful imposition of archaic categories of proven inaccuracy. The same goes for theories of understanding. The limitations on memory and attention are best seen if you have theoretical knowledge of the causes. Philosophers propose accounts of understanding and are silent about the constitutive role of memory and attention. But they could see that they were mistaken about prospective outcomes had they simply considered cognitive limitations under ordinary stresses. Consider a common example of cognitive limitation showcased by Paul Meehl: Surely we all know that the human brain is poor at weighting and computing. When you check out at a supermarket, you don’t eyeball the heap of purchases and say to the clerk, “Well it looks to me as if it’s about $17.00 worth; what do you think?” The clerk adds it up. There are no strong arguments from the armchair or from empirical studies . . . for believing that human beings can assign optimal weights in equations subjectively or that they apply their own weights consistently. (Meehl 1986, 372) Notice that in Meehl’s grocery example, we know that a simple addition is the right calculation to apply and the variable values (i.e., the prices) are usually stamped right on the products. But suppose that the computation required was much more complex. This of course would make matters even worse. Suppose instead that the supermarket pricing rule were, “Whenever both beef and fresh vegetables are involved, multiply the logarithm of

Understanding and Fluency


0.78 of the meat price by the square root of twice the vegetable price”; would the clerk and customer eyeball any better? Worse, almost certainly. When human judges perform poorly at estimating and applying the parameters of a simple or component mathematical function, they should not be expected to do better when required to weigh a complex composite of those variables. (Dawes, Faust, and Meehl 1989, 1672) These processes require attention to the kinds and quantities being measured. In order to attend, people must hold in working memory the items like beef and fresh vegetables. And while doing so, we must have also made a decision about what kind of meat to attend to, or whether to classify candidate plants as “vegetables.” If these processes are constitutive of understanding grocery price, then the fact that we can’t fluently or transparently estimate a value shows us that the opacity of understanding is routine. To dramatize the human memory/​attention limits involved in understanding any complex topic, consider routine efforts to understand genetic risks. It is quite common now for people to explain, and so purport to understand, a medical condition in terms of “having the gene for it.” Perhaps the local character of causation creates the impression that its forces are easier to cognitively process. However, most conditions that have clear genetic influences (say, breast cancer) aren’t just a result of a single point mutation but instead are related to hundreds of single-​nucleotide polymorphisms (SNPs). Each mutation alone is not going to give you breast cancer, but as a whole, the aggregation of many relevant mutations can strongly affect risk. I think there really might be some human memory/​attention limits involved with failing to understand this, because the news media and personal genetic profiling companies always seem obsessed with finding “the” [insert disease] gene, even though it is very rarely that simple. It may be that philosophers, too, suffer from an errant fluency in their construction of theories of understanding. Even a few stops on a tour of philosophical theories of understanding reveals a troubled landscape: definitions of understanding that appeal to grasping, pronouncements that understanding is nonpropositional, transparent. Compared to a psychological theory that traces the temporal and spatial contours of attention, the powers of memory, and the chemistry of encoding—​all features of a complete scientific theory of understanding—​distinctly philosophical approaches to theorizing about understanding seem proudly medieval. Consider a more distinctively empirical way that an approach to understanding might take. Begin with situations that have similar structure. You



build a kind of model of some phenomenon. This model or schema incorporates the forces thought to be most important to the effects considered. If it is a weather pattern, the structure will include such factors as wind turbulence, dew points, and temperature. If the system or model concerns the stress on a steel beam, the structure will include the strength of the material, the length of the run, and the weight it bears. If it is crime in an urban neighborhood, you will focus on the kind of infraction, the demographics of the perpetrators, and the temptations in the environment that transforms onlookers into criminals. The motion of a projectile may be depicted by a model that includes forces that propel the projectile and produce recoil, factors like gravity that operate on its weight and air that creates drag. If understanding seems a difficult achievement, at least in some domains, it may be because some phenomena in the world are complex in ways that challenge the mind’s delicate balance of attentional gaze, perceptual buffer, and memory capacity. So we simplify. We conceive of an atom as a solar system, or electricity as a liquid. This abbreviation is, of course, a mnemonic, not an accuracy-​inducing device. But without it, understanding even something like electricity would be difficult for most of us. Now think of understanding the operation of complex systems—​population ecology, evolution, the psychology of morality, disease, poverty, and consciousness. What would be required in order to say that we “understand” any of these systems?

12.4. Philosophical Fluency When a concept is accessible to our thoughts, like a triangle, a dog, or for the physicist, an electron, we can use it fluently. We can talk about the concept, hold it before our attention, relate it to others, and gauge its many powers. This experience of fluency, gives rise to confidence in a judgment or decision. Psychologists call this experience by a variety of names: Feeling of Rightness, Feeling of Knowing, and Judgments of Learning, and this experience is a memory-​based judgment, and is based on experiences that are cued automatically. But the causes of these experiences—​the processes of attention, and the mechanisms of memory workspace—​are not introspectable (Brewer and Sampaio 2006; also see Koriat et al. 2004; also see Matvey et al. 2001). Thus, the experiential component of these Feelings—​of Rightness, Judgment, and Learning—​leaves one with a feeling of confidence without knowledge about the basis of that confidence. As an indicator of truth, how reliable is the Feeling of Understanding? It is a good thing that difficult problem-​solving settings provoke our analytic,

Understanding and Fluency


deliberative skills, because if left unchecked humans will succumb to overconfidence and hindsight biases to embrace the subjective cue that a given task is easier than it actually is. The sense of understanding feels good. We rush to get the feeling, and we are hooked on it. But it is a tragic addiction. Stephen J. Gould viewed it instead as comical farce, as though we have a homunculus in our head, jumping up and down, and shouting out the wrong answers (Gould 1991, 469). Whatever image we choose, these errors stemming from not actually understanding are large, and they are often about weighty matters. Surely the feeling of understanding is at least sometimes right, even right much of the time. But in the common cases of knowing simple facts, we get no jolt of satisfying understanding. When I prepare for a left turn in my car, I have the belief that oncoming traffic will adjust, and that I have correctly judged the distance. I believe that eggs will cook if I heat them, that the local bodega around the corner is open, that I can barely ski, but that a stick frame wall built to code has plenty of strength to carry a heavy snow load in a temperate zone. We could continue this list indefinitely. You might think that monitoring these biases is pointless, because their behavior is so difficult to control. But even mediocre performance can be an improvement. And some people have more room for improvement than others. People experience the Feeling of Rightness with different regularity, because they monitor and control their feelings differently.8 Some are better at spotting and corrrecting errors. Did you catch the typo in the last sentence? If so, that is an example of successful monitoring. People who monitor their psychological processes better are typically also better on selected reading tasks (Lin, Moore, and Zabrucky 2001), better in science classes when given optical principles to learn (Prins, Veenman, and Elshout 2006), and better at mathematical problem solving (Desoete and Roeyers 2006; also see Lucangeli, Tressoldi, and Cendron 1998). The extent to which we monitor our feelings determines whether we accept an existing outcome or seek another (Mazzoni and Cornoldi 1993; also see Son and Metcalfe 2000; also see Nelson 1993; also see Son 2004). If you feel that your swing is good, you may stop taking batting practice; if you don’t, you may change stance. If you are confident that you have understood the above passage, you won’t reread it. If you are confident that you have correctly remembered the name of a person you have just run into, you will address that person by name; if not, you may choose a more generic greeting. If you doubt

8.   See Hertzog and Robinson (2005) for a review.



that you will remember the bread on the way home, you might arrange a cue to trigger your memory. The success of your judgment depends not only on the accuracy of your memory, but also on your ability to monitor your mental processes and take appropriate action (Koriat and Levy-​Sadot 1999). This is the function of metacognition. Philosophers have their own conceptual repertoire, and they traffic in it with great fluency. The courting of intuitions is one example, leading to judgments about “what we want to say” about a particular case. In response to the claim that understanding must be transparent, we might generate intuitions by imagining a mathematician working on a proof that the square root of 2 is irrational. She works at it so persistently that by the time she solves the problem she is too tired to appreciate the solution. She falls asleep and when she awakens she looks at the solution she had written. The proof is perfect, but the last series of steps in reasoning are unrecognizable to her. We might want to say that she solved the problem but doesn’t understand the solution. She doesn’t understand it because she doesn’t have an awareness of the factors responsible for the solution. Appeal to awareness has the satisfying sound of transparency, but it fails to move the ball. We might understand how to ride a bike but not have an awareness of the factors responsible for this skill. Proliferating senses of ‘awareness’, however, is no way to provide a unified account of understanding. The option seems to be a concept of understanding that captures only a small class of intellectual and practical achievements—​the ones that exhibit a kind of transparent awareness of the principles in virtue of which the achievement is accomplished—​or treat recognitional capacities or discriminative sensitivities too, as expressions of understanding. Understanding feels like it goes with awareness, because we expect people who understand a subject to be able to instruct others about it, and instruction typically requires that we use that awareness to articulate the factors that contribute to the subject of instruction. Control or manipulability feels like it goes with understanding, because understanding an object seems to require knowing its causal powers. If an object can surprise us, acting unpredictably or beyond our abilities to regulate its conduct, we would not say that we understand the object, however we might fear or mistrust the object. But aren’t these all just expressions of conceptual fluency, a fluency that requires not correspondence but mere coherence? I develop this psychological view at greater length in Wondrous Truths:  The Improbable Triumph of Modern Science. If we get the same good sense of understanding whether our belief is true or false, or our theory is good or bad, how does that feeling advance the theoretical cause? After all, the feeling supplies conviction

Understanding and Fluency


rather than truth, the kind of conviction that people have that going outside in cold weather with wet hair will make you sick (it won’t), or that shaving will make your hair grow in thicker (it won’t). But after thousands of hours of implementing familiar concepts, laypeople, like scientists, have a conviction that is fully automatized, even when their relevant beliefs are false. In “The Will to Believe,” William James dramatizes the power that ritualized behavior has not just to sustain belief but to create it. As he says: “Go, then, and take holy water, and have masses said: belief will come and stupefy your scruples” ( James 1896/​1907, 6). The processing benefits of fluent concepts is clear. Familiar concepts, like familiar methods, are used most easily when most familiar. When identifying objects from an array, people display fluency (through faster reaction times) for living categories like human faces (Rhodes and Tremewan 1996; also see Rhodes et al. 2001) as well as fish, dogs, and birds (Halberstadt and Rhodes 2000; 2003). But the drive is so strong to find the center that humans do it even with categories of nonliving objects, like color patches (Martindale and Moore 1988), and even artifactual objects like furniture (Whitfield and Slatter 1979), wristwatches, and automobiles (Halberstadt and Rhodes 2000; 2003). So robust is this drive, in fact, that objects that clearly do not admit of gradations receive this treatment from humans. Whole numbers are one such example: certain odd numbers are reported by people to be “more odd” than others (Armstrong, Gleitman, and Gleitman 1983). Who would have guessed that 7 and 13 are the oddiest of odd numbers, and 15 and 23 the least odd of them? Or that 8 and 22 are the eveniest even numbers, 30 and 18 are the least eveny of the even numbers? And in science, just like in everyday life, explanatory prototypes that free up processing space in the brain are deemed more attractive, and more accurate, whether or not they actually are. Fluency feels good, and disfluency feels bad. Something like understanding has a healthy epistemic role to play in theory testing and development. But this naturalistic theory of scientific explanation should also account for explanation’s failures. I have described a number of the psychological sources of failure that lie behind our tendency to be seduced by the sense of understanding (also see Trout 2002). Other attempts to specify effective conditions for the acceptance of an explanation underestimate the theory-​dependence of these judgments. It may be that people are more likely to accept an explanation if it seems applicable to other settings of prediction and manipulation (Lombrozo and Carey 2006). But this is a deeply theory-​dependent judgment, and if your theories are poor you are likely to identify poor explanations as useful. Surely this was part of



the unifying promise of supernaturalism, of alchemy, or of the humoral theory of health. There seems no merely instrumental or pragmatic measure, such as usefulness, that tells us when an explanation should be accepted, even if there are more robust theoretical standards that might recommend acceptance (for example, that a theory is mature and has unified a diverse range of phenomena). Moreover, the understanding conveyed by a good explanation may be a community achievement. Except for the simplest of events, explanatory understanding is not essentially an achievement of an individual. And any alternative account of explanation that requires the transmission of a sense of understanding must address this criticism. My positive account of scientific explanation asserts that, as a contingent matter of fact, the only feature of an explanation that can render explanation epistemic is its systematic tendency to produce increasingly accurate theories. In effect, only explanations capable of sustaining theoretical progress are good explanations. This pronouncement may not help us to decide now, rather than in retrospect, which explanation to take seriously. But this is not the job of an account of explanation. A theory of scientific explanation should not attempt to predict the future history of science, but set out what scientific explanation is, and what standards should be met by a proper explanation. In order to accord explanation the epistemic role it seems to play in successful theory selection in contemporary science, we must abandon our attachment to the comforting idea that the “sense of understanding” is a cue to at least a working version of the truth. But this will not be easy. Explanation is a backward-​looking affair and thirty years of research on judgment shows both that people are not good at tracking how they are affected by knowledge of outcomes and that they are not good at admitting this limitation. A distinctly philosophical analysis of explanatory understanding may include a role for the sense of understanding. But it is unlikely that this role will be a justificatory one. In the series of cognitive steps that lead to understanding, phenomenology is a latecomer. In the absence of independent evidence of its reliable role—​that its presence covaries importantly with progressive findings—​and in light of the psychological and historical evidence that it is an unreliable cue, the sense of understanding is not a promising route to genuine understanding. In fact, no one has the vaguest idea how this phenomenology is related to getting things right, so it is a field ripe for exploration. Scientific realists can assign a robust role to objective factors in explanation—​such as statistical and causal relevance—​and value the contribution of explanation to scientific progress. If your focus is not balanced by a positive account of the sense of

Understanding and Fluency


understanding in a scientific theory of explanation, then it is easy to portray any criticism of the sense of understanding as the first step toward explanatory nihilism. But explanatory nihilism is surely premature. On the one hand, the sense of wonder when explaining fills us with joy. But on the other, the illusion of explanatory depth shows that our joy may be misplaced. Why would evolution have designed us so that the sense of understanding feels good, and yet so that we regularly stop before we fully understand? Alison Gopnik has the first half of the answer (Gopnik 1998). Evolution has designed us so that we have a drive to learn. And what better way to design that drive than to have a body and mind that responds to the world with wonder, so that we want to find out how these wonderful things work? But why the paradoxical termination of inquiry prior to actually fully understanding? Why would evolution make us feel so giddy when we seem to be making cognitive progress, yet at the same time hinder our learning through phenomena like the Illusion of Explanatory Depth? Of course, few people can accurately boast a complete understanding of the workings of a toilet, tumbler lock, or tectonic stress. Even the best scientists, if they are honest, will cry uncle at some point when probed about their understanding of their favorite natural phenomenon. The Illusion of Explanatory Depth, while sobering, does not resolve our question. How do we develop such beautiful and accurate theories when our understanding is so routinely incomplete and our attitude so robustly overconfident? And how can we develop a scientific account of understanding with the feeble instruments of conceptual analysis, no matter how confidently wielded?

12.5. Conclusion If understanding begins in flights of wonder, the limits of human understanding bring it back to earth. The need for speedy processing imposes forbidding constraints on memory and attention, two of the chief faculties underlying understanding (Cowan 2005). Understanding requires a complicated choreography of moves whose execution limits the number of items in memory that can be constantly activated ( Just and Carpenter 1992), the number of relations among items in memory that can be integrated (Halford, Wilson, and Phillips 1998), and the ability to mentally maintain the distinction between items in working memory (Oberauer and Kliegl 2001). Just like in juggling, the error profile is what psychologists call catastrophic; you drop one ball, you drop them all. The process that manages the juggling of attention and memory is metacognition, and metacognition has its limits too.



The evidence for the nature of understanding is empirical (even when the empirical evidence is anecdotal). You can certainly choose to ignore the scientific evidence, and generate categories tied to bygone traditions like introspectionism, respects of similarity associated with magical thinking, and theorizing based on ancient categories. You could use concepts like transparency, nonpropositionality, and grasping, notions unrecognizable to scientific cognitive and comparative psychologists. Philosophers are, of course, at leisure to use whatever methods and concepts they choose. But that is only to observe that philosophy has no certification procedure that prohibits the use of fruitless methods, and no barrier to entry whose standard of success is based on truth. By now it should be clear that the sense of understanding, by itself, is an unreliable cue to truth. True beliefs may carry more or less of the sense of understanding, a sense of fluency or metacognitive control. But this does not add up to unifying diverse cognitive achievements. It is convenient to have a name for the event or state when our struggle with a particular cognitive challenge strikes the proper combination of attentional capture, available working memory, accurate shortcuts, or heuristics to reduce the conflict. But we shouldn’t become so smitten that we attribute to the nominal orthography ‘understanding’ a unifying power, or suppose we can squeeze a new pedigree out of an old mongrel.

Acknowledgments I would like to thank Maya Mathur for providing, and discussing with me, some of the key examples of opaque explanation provided in this chapter. Kathleen Parker provided several apt examples as well. Thanks to Clinton Neptune for detailed and useful comments, Paul Abela for discussion of several examples, and Michael Strevens for many conversations about the relations between explanation and understanding. Support from the Templeton Foundation is gratefully acknowledged.

References Armstrong, Sharon Lee, Lila R. Gleitman, and Henry Gleitman. 1983. “What Some Concepts Might Not Be.” Cognition 13: 263–​308. Belopolsky, Artem V., and Jan Theeuwes. 2010. “No Capture outside the Attentional Window.” Vision Research 50: 2543–​50. Bishop, Michael A., and J. D. Trout. 2005. Epistemology and the Psychology of Human Judgment. New York: Oxford University Press.

Understanding and Fluency


Blasi, Guiseppe, and Venkata S. Mattay. 2005. “Effect of Catechol-​O-​methyltransferase val158met Genotype on Attentional Control.” Journal of Neuroscience 25: 5038–​45. Brewer, William F., and Cristina Sampaio. 2006. “Processes Leading to Confidence and Accuracy in Sentence Recognition:  A  Metamemory Approach.” Memory 14: 540–​52. Cowan, Nelson. 2005. Working Memory Capacity. Hove, East Sussex, UK: Psychology Press. Dawes, Robyn M., David Faust, and Paul E. Meehl. 1989. “Clinical versus Actuarial Judgment.” Science 243: 1668–​74. De Regt, Henk W., and Dennis Dieks. 2005. “A Contextual Approach to Scientific Understanding.” Synthese 144: 137–​70. Desoete, Annemie, and Herbert Roeyers. 2006. “Metacognitive Macroevaluations in Mathematical Problem Solving.” Learning and Instruction 16: 12–​25. Diamond, Adele, Lisa Briand, John Fossella, and Lorrie Gehlbach. 2004. “Genetic and Neurochemical Modulation of Prefrontal Cognitive Functions in Children.” American Journal of Psychiatry 161: 125–​32. Elgin, Catherine. 2007. “Understanding and the Facts.” Philosophical Studies 132: 33–​42. Fonseca, Brenda A., and Micheline T.  H. Chi. 2011. “Instruction Based on Self-​ Explanation.” In The Handbook of Research on Learning and Instruction. Ed. Richard E. Mayer and Patricia A. Alexander. New  York:  Routledge, Taylor and Francis. 296–​321. Geva, Ronny, Michal Zivan, Aviv Warsha, and Dov Olchik. 2013. “Alerting, Orienting or Executive Attention Networks: Differential Patters of Pupil Dilations.” Frontiers in Behavioral Neuroscience 7 (October 14): 145. doi:10.3389/​fnbeh.2013.00145. Gopnik, Alison. 1998. “Explanation as Orgasm.” Minds and Machines 8: 101–​18. Gould, Stephen J. 1991. Bully for Brontosaurus. New York: Norton. Grimm, Stephen R. 2006. “Is Understanding a Species of Knowledge?” British Journal for the Philosophy of Science 57: 515–​35. —​—​—​. 2012. “The Value of Understanding.” Philosophy Compass 7: 103–​17. Halberstadt, Jamin, and Gillian Rhodes. 2000. “The Attractiveness of Non-​face Averages:  Implications for an Evolutionary Explanation of the Attractiveness of Average Faces.” Psychological Science 11: 285–​89. —​—​—​. 2003. “It’s Not Just Average Faces That Are Attractive: Computer-​Manipulated Averageness Makes Birds, Fish, and Automobiles Attractive.” Psychonomic Bulletin and Review 10: 149–​56. Halford, Graeme S., William H. Wilson, and Steven Phillips. 1998. “Processing Capacity Defined by Relational Complexity: Implications for Comparative, Developmental and Cognitive Psychology.” Behavioral Brain Sciences 21: 803–​31. Hertzog, Christopher, and A. Emanuel Robinson. 2005. “Metacognition and Intelligence.” In Understanding and Measuring Intelligence. Ed. Oliver Wilhelm and Randall W. Engle. London: Sage. 101–​23. James, William. 1896/​1907. “The Will to Believe.” In The Will to Believe: And Other Essays in Popular Philosophy. New York: Longmans, Greens and Co. 1–​31.



Just, Marcel Adam, and Patricia A. Carpenter. 1992. “A Capacity Theory of Comprehension: Individual Differences in Working Memory.” Psychological Review 99: 122–​49. Khalifa, K. 2017. Understanding, Explanation, and Scientific Knowledge. Cambridge: Cambridge University Press. Kitcher, Philip. 1989. “Explanatory Unification and the Causal Structure of the World.” In Scientific Explanation. Ed. Philip Kitcher and Wesley C. Salmon. Minneapolis: University of Minnesota Press. 410–​505. Koriat, Asher, Robert A. Bjork, Limor Sheffer, and Sarah K. Bar. 2004. “Predicting One’s Own Forgetting: The Role of Experience-​Based and Theory-​Based Processes.” Journal of Experimental Psychology: General 133: 643–​56. Koriat, Asher, and Ravit Levy-​Sadot. 1999. “Processes Underlying Metacognitive Judgments:  Information-​Based and Experience-​Based Monitoring of One’s Own Knowledge.” In Dual Process Theories in Social Psychology. Ed. Shelly Chaiken and Yaacov Trope. New York: Guilford Press. 483–​502. Kvanvig, Jonathan. 2011. “Millar on the Value of Knowledge.” Aristotelian Society Supplementary Volume 85: 83–​99. Lin, Lin-​Maio, Dewayne Moore, and Karen M. Zabrucky. 2001. “An Assessment of Students’ Calibration of Comprehension and Calibration of Performance Using Multiple Measures.” Journal of Reading Psychology 22: 111–​28. Lombrozo, Tania, and Susan Carey. 2006. “Functional Explanation and the Function of Explanation.” Cognition 99: 167–​204. Long, A. R., P. M. Fernbach, and B. De Langhe. 2016. “The Influence of Subjective Understanding on Risk Perception.” Working paper. Lucangeli, Daniela, Patrizio E. Tressoldi, and Michela Cendron. 1998. “Cognitive and Metacognitive Abilities Involved in the Solution of Mathematical Word Problems:  Validation of a Comprehensive Model.” Contemporary Educational Psychology 23: 257–​75. Martindale, Colin, and Kathleen Moore. 1988. “Priming, Prototypicality, and Preference.” Journal of Experimental Psychology: Human Perception and Performance 14: 661–​70. Matvey, Greg, John Dunlosky, and Robert Guttentag. 2001. “Fluency of Retrieval at Study Affects Judgements of Learning ( JOLS): An Analytic or Nonanalytic Basis of JOLS?” Memory and Cognition 29: 222–​33. Mazzoni, Guiliana, and Cesare Cornoldi. 1993. “Strategies in Study Time Allocation: Why Is Study Time Sometimes Not Effective?” Journal of Experimental Psychology: General 122: 47–​60. McGinn, Colin. 1991. The Problem of Consciousness:  Essays toward a Resolution. Cambridge, MA: Blackwell. Meehl, Paul. 1986. “Causes and Effects of My Disturbing Little Book.” Journal of Personality Assessment 50: 370–​75.

Understanding and Fluency


Nelson, Thomas O. 1993. “Judgments of Learning and the Allocation of Study Time.” Journal of Experimental Psychology: General 122: 269–​73. Oberauer, Klaus, and Reinhold Kliegl. 2001. “A Formal Model of Capacity Limits in Working Memory.” Journal of Memory and Language 55: 601–​26. Parasuraman, Raja, and Pamela M. Greenwood. 2005. “Beyond Heritability:  Neurotransmitter Genes Differentially Modulate Visuospatial Attention and Working Memory.” Psychological Science 16: 200–​207. Prins, Frans J., Marcel V. J. Veenman, and Jan J. Elshout. 2006. “The Impact of Intellectual Ability and Metacognition on Learning:  New Support for the Threshold of Problematicity Theory.” Learning and Instruction 16: 374–​87. Rapsomaniki, Eleni, Anoop D. Shah, Pablo Perel, Spiros Denaxas, Julie George, Owen Nicholas, Ruzan Udumyan, et al. 2014. “Prognostic Models for Stable Coronary Artery Disease Based on Electronic Health Record Cohort of 102,023 Patients.” European Heart Journal 35: 844–​52. Rhodes, Gillian, Sakiko Yoshikawa, Alison Clark, Kang Lee, Ryan McKay, and Shigeru Akamatsu. 2001. “Attractiveness of Facial Averageness and Symmetry in Nonwestern Cultures:  In Search of Biologically Based Standards of Beauty.” Perception 30: 611–​25. Rhodes, Gillian, and Tanya Tremewan. 1996. “Averageness, Exaggeration and Facial Attractiveness.” Psychological Science 7: 105–​10. Son, Lisa K. 2004. “Spacing One’s Study:  Evidence for a Metacognitive Control Strategy.” Journal of Experimental Psychology:  Learning, Memory, and Cognition 30: 601–​4. Son, Lisa K., and Janet Metcalfe. 2000. “Metacognitive and Control Strategies in Study-​Time Allocation.” Journal of Experimental Psychology 26: 204–​21. Strevens, Michael. 2013. “No Understanding without Explanation.” Studies in History and Philosophy of Science 44: 510–​15. Trout, J. D. 2002. “Scientific Explanation and the Sense of Understanding.” Philosophy of Science 69.2: 212–​33. —​—​—​. 2016. Wondrous Truths:  The Improbable Triumph of Modern Science. New York: Oxford University Press. Whitfield, T. W.  A., and P. E. Slatter. 1979. “The Effects of Categorization and Prototypicality on Aesthetic Choice in a Furniture Selection Task.” British Journal of Psychology 70: 65–​75. Wilkenfeld, Daniel. “MUDdy Understanding.” Synthese 194 (4): 1273–​93. Wilkenfeld, Daniel, David Plunkett, and Tania Lombrozo. 2016. “Depth and Deference:  When and Why We Attribute Understanding.” Philosophical Studies 173: 373–​93. Zagzebski, Linda. 2001. “Recovering Understanding.” In Knowledge, Truth, and Duty:  Essays on Epistemic Justification, Responsibility, and Virtue. Ed. Matthias Steup. New York: Oxford University Press. 235–​51.


abilitism, 31–​32, 34, 31n35. See also practical understanding; understanding Achinstein, Peter, 213n4 action, 43. See also activity action theory, 16 activity, 31n35 practical, 15 skillful, 20–​21, 44 See also action; method; skill aesthetic appreciation, 170–​74, 172n12. See also aesthetics; aesthetic understanding aesthetic proposition aesthetic value and the, 162 explanation of an, 162–​64 See also aesthetics aesthetics, 10, 105, 159–​75 and beliefs, 171–​72 judgments in, 160, 167, 171–​72 knowledge in, 159–​60 realism about, 160 truth in, 160 value in, 160, 163, 172 See also aesthetic appreciation; aesthetic understanding; art aesthetic understanding, 159–​75 aesthetic appreciation and, 170–​74 aesthetic knowledge and, 169–​71 articulateness of, 167–​69

and explanation, 167–​68 gradability of, 164–​66, 164n6 See also aesthetic proposition; aesthetics; understanding ambiguity, 192, 201, 140n9 anger, 132–​33 anthropology, 191 Aquinas, St. Thomas, 123 Aristotle, 102, 15n2, 25n23, 197n4 Armstrong, David, 79 art, 62, 109, 160, 162–​64, 167–​68, 172–​73, 189, 191, 173n13. See also aesthetics; art criticism; music art criticism, 168. See also aesthetics; art; literary criticism atheism, 81, 88–​89 attention, 232, 236, 239–​43, 250 awareness, 237, 240, 246 Bacon, Francis, 102 Bain, Alexander, 104, 106 Baker, Lynne Rudder, 66n13 Beardsley, Monroe, 163 belief, 66, 120–​21, 139, 147–​51, 171, 245, 247, 28n28, 36n44, 139n6, 148n14 false, 247, 150n16 and indexicals, 66n12 jointy, 147–​50 moral, 174 particular, 66



belief (cont.) political, 65 state of, 205–​6 true, 110, 169, 171, 250 See also knowledge; truth; understanding Bentham, Jeremy, 105 blame, 121–​32, 131n11. See also criticism; forgiveness; resentment Boden, Margaret, 202–​3 Bonhoeffer, Dietrich, 128 Brandom, Robert, 30 Brewer, Talbot, 90, 92, 90n18, 91n19 Buckley, Michael J., 88, 90–​91 Burrell, David, 78, 84, 90–​91, 94 Butterfield, Herbert, 98n1 Campbell, John, 21, 27, 30, 21n17 causation, 2, 197, 218–​22, 238. See also explanation Christianity, 131 Coakley, Sarah, 91 coalescence, 37, 38n47 external, 38, 40–​41, 43–​44, 38n49 internal, 38, 40–​41, 43–​44 cognitive control, 162, 166–​67, 169–​71. See also belief; knowledge; understanding communication, interpersonal, 69 compassion, 117–​33. See also ethics; understanding conception, 35–​39, 44n54 coalescence of a, 41 completeness of a, 36, 40, 43 the content of a, 47, 47n59 correctness of a, 36, 40, 43 guiding, 48 illuminating, 40–​42, 41n52 impoverished, 42 mastery of a, 38–​41, 43, 38n49, 39n50 noetic, 35–​41, 43–​45, 47 as a standing psychological, 40

conscience, 132 consistency, 37 copresence, 24–​25, 25n23. See also unity Cottingham, John, 84–​94. See also religious understanding creative acts, 202, 204–​7 criticism, 125–​26, 130. See also blame curiosity, 235 Davidson, Donald, 63n10 Deductive-​Nomological Model, 209–​11, 213, 222–​30 de Graffigny, Françoise, 57 dependence relations, 6. See also explanation; understanding Descartes, René, 142, 201–​3 descriptions, 67 desire, 91–​95, 120, 197n5 determinism, 179 Dilthey, Wilhelm, 57 Dretske, Fred, 30 Dreyfus, Hubert, 30, 30n34 Driver, Julia, 131n11 dualism, 34, 46 DuBois, W. E. B., 57, 62, 65, 67–​68 Dupré, John, 79–​80 economics, 105 education, 177–​78, 186, 188, 192. See also humanities; science Einstein, Albert, 203 Elgin, Catherine, 71n17 emotion, 48 sources of our, 190 vindictive, 131 empathy, 48 empiricism, 4 enrichment, 25–​26, 31, 46, 49 epistemic oppression, 54–​61, 72–​75. See also oppression; power epistemology, 1–​4, 16, 86, 135–​36, 141–​43, 233–​34, 22n19, 71n17

Index old, 69 understanding in, 2–​4, 72 See also knowledge; understanding ethics, 72, 82–​83, 123 and aesthetics, 175 and understanding, 132, 174 See also moral philosophy evidentials, 18n8 experience, 66–​67, 69 explanation, 1, 28, 137, 152, 163–​64, 186, 222–​30, 233–​34 acceptance of an, 247 aesthetic, 163 as answers, 211–​14 “ideal agent” text of an, 5 laws and, 163 mechanisms revealed in, 6–​7 models of, 63n10 representation view of, 136–​45, 152, 155–​56 scientific, 248–​49 as a speech act, 212, 217, 213n4 theories of, 210–​11, 213, 230 understanding and, 4–​7, 112, 152, 197–​98, 222, 229–​30, 236, 240, 248 See also causation; explanationism; explanatory structure; knowledge; understanding explanationism, 28–​32, 34, 41n52. See also explanation; understanding Explanation-​Understanding Condition, 210–​11, 214–​19, 222–​24 explanatory structure, 135–​56 and “natural joints”, 145–​47 the representation of, 144, 152 socially constructed, 155 See also explanation; structure; understanding Fanon, Frantz, 57, 60n7 Fish, Stanley, 178


fluency, 232–​50. See also philosophy; understanding forgiveness, 10, 117–​33, 123n8, 131n11. See also blame; ethics; understanding Fricker, Miranda, 59, 73 Friedman, Michael, 210–​11 Geertz, Clifford, 57 generality, 21, 44, 44n54. See also laws God, 80, 82–​89, 91–​94. See also theism Goldman, Alvin, 120 Gopnik, Alison, 249 Gordon, Robert, 120 Gould, Stephen J., 245 Greco, John, 28, 33, 33n40 Griffin, James, 80, 82 Grimm, Stephen, 199, 197n5, 216n6, 235n2 ground, 26, 34–​35, 49 denial of, 27 gaining, 39–​44 metaphysics of, 26n25 opposition to, 27 See also understanding; unity; U-​profile Hamilton, Sir William, 110, 112 Hanh, Thich Nhat, 118–​19 Heal, Jane, 120 Heidegger, Martin, 8, 17, 27, 204 Hempel, Carl, 209–​10, 213, 222, 63n10, 213n5 Hepburn, R. W., 189 hermeneutic injustice, 59–​61, 75, 59n4. See also epistemic oppression; testimonial injustice history, 108–​9, 111, 186, 188–​89, 191, 98n1. See also history of science; humanities history of science, 98n1



humanities, 177–​93, 185n20 education and the, 186, 188 guidance and the, 187–​89 See also history; history of science; literature Hume, David, 99–​100, 102–​3, 106–​9 imperialism, 34 incidentalism, 34–​35, 45, 47, 49. See also understanding inequality, 57–​58. See also power information, 3, 6, 236 inquiry, 114, 149–​51, 235 intellectualism, 45, 16n3 intellectual know-​how, 161. See also knowledge interpretation, 58–​59, 189–​91, 220–​21, 190n26 of experience, 66–​68, 67n14 of works of art, 71 intuition, 246 James, William, 247 Johnston, Mark, 91n20 Jones, Henry, 99–​100, 102, 109–​10 Kant, Immanuel, 23, 102, 105, 117 Kelly, Sean, 30 Kim, Jaegwon, 28 Kitcher, Philip, 185n20 knowledge, 1, 15, 65, 72, 111–​12, 169–​71, 28n28, 170n11 aesthetic, 159–​61 as factive, 138–​41 and the humanities, 178–​79 illusory, 3 knowledge-​how, 15–​16, 21, 21n14, 16n3, 44n54 knowledge-​that, 15–​16, 42, 16n3 luck and, 2 metaphysical conditions for the possibility of, 155

oppressed people and, 55–​61 perceptual, 4 “phrasebook”, 21 pluralism about, 194–​95 social relations and, 58 understanding and, 137–​38, 170, 195, 206, 161n2, 200n8 See also belief; epistemology; moral knowledge; propositional knowledge; social epistemology; understanding Kuhn, Thomas, 101 Kvanvig, Jonathan, 17, 27–​28, 28n27 laws, 185–​86, 228. See also generality; physics Lessing, Gotthold Ephraim, 110 Levinas, Emmanuel, 9, 79, 82–​83, 87, 92–​95, 83n8, 92n22 Lewis, David, 146 Lipton, Peter, 17, 27–​28 literary criticism, 173. See also art criticism; literature literature, 186, 189, 191. See also humanities Locke, John, 102, 106–​8, 143 logic, 104–​5, 108 of dependency relations, 197 fallacies of, 105 individuals and operations of, 146n12, 147n13 semantics, psychology, and,  36n44 understanding and entailments of, 203 See also mathematics; philosophy luck, 2, 39n51, 169n10 Madoff, Bernie, 120–​21 Marx, Karl, 57, 60n7 mathematics, 192, 246, 155n19. See also logic; science

Index McCabe, Herbert, 82n7 McDowell, John, 80–​81, 83, 81n4 McGeer, Victoria, 124, 129–​31 Medina, José, 60–​61, 61n9 Meehl, Paul, 242 memory, 232, 236, 239–​43, 246, 250 Merleau-​Ponty, Maurice, 17, 30 metaphysics, 10 method, 44, 29n32. See also explanation; skill Milgram, Stanley, 131 Mill, John Stuart, 48, 103, 105, 110, 105n2 moral knowledge, 121–​24, 129, 122n7. See also knowledge moral philosophy, 103, 106. See also ethics; philosophy Murdoch, Iris, 190–​91 music, 62, 109–​10, 114. See also art; poetry Nagel, Thomas, 79 Nanay, Bence, 203 naturalism, 78–83, 94 expansive, 81–83, 89–90, 94–95 reductive, 79–80, 95 scientific, 81–82 See also science; supernaturalism natural theology, 85, 87–91. See also theism; theology nature, 80–81, 83, 87–91, 94, 171 Newton, Isaac, 102, 107 Noë, Alva, 30 noetic conception, 35–41, 43–47. See also conception; understanding; U-profile Nussbaum, Martha, 119, 178 Oakeshott, Michael, 113–​14 ontology, 81, 83 oppression, 54–​58, 60–​61, 64, 74. See also epistemic oppression; inequality; power


pantheism, 82. See also theism perspective, 64–​67, 70 internal, 2 of the oppressed, 73 peculiar, 65–​68, 75 personal, 63–​64, 72, 74 privileged, 2 phenomenology, 16, 205, 232, 248 philosophy, 98–​100, 102–​14, 191, 232–​36, 250, 235n2 as the handmaid of science, 106–​9 history of, 100, 102 inconclusiveness of, 101–​2, 110 the normativity of, 109–​12 and personal quest, 112–​14 as the pursuit of understanding, 112 See also logic; moral philosophy; philosophy of mind; philosophy of science philosophy of mind, 16, 101–​2 philosophy of science, 1–​2, 16, 211, 213, 217–​18, 22n19 structure in the, 155 theory of explanation in the, 230 understanding in the, 4–​7, 17 See also philosophy; science physics, 179–​80, 182, 179n5. See also laws; science Plato, 19, 100, 145, 196, 25n23, 169n9 poetry, 109–​10. See also art; music power, 9, 57–​58 problem of, 54–​75 relations of, 58–​59 See also inequality; oppression; social epistemology practical understanding, 9, 14–​20, 22–​31, 35, 45, 16n4, 26n24 action-​g uiding nature of, 23–​24, 43 dualism about, 32 guiding conceptions of, 42–​44 See also abilitism; skill; understanding



problem of mindreading, 54. See also social cognition problem of other minds, 54. See also skepticism properties, 145–​47, 155, 162, 146nn11–​12 aesthetic, 162 jointy, 146–​47, 152–​53, 155 natural, 146 race, 153–​54, 153n17 propositional attitude, 36n44 propositional knowledge, 137–​38, 142, 144, 152, 155, 139n6, 169n10. See also knowledge psychology, 237–​39, 243, 239n5 Rahner, Karl, 82n6 Railton, Peter, 81, 226, 81n5 rationalism, 4 Reid, Thomas, 107–​8 religious understanding, 8–​9, 78–​95. See also understanding representation, 136–​37, 140, 144, 139n6, 144n10 resentment, 124–​30, 132. See also blame Rorty, Richard, 139n6 Rosenberg, Alex, 178n3 Ryle, Gilbert, 16n3, 43n53 Schopenhauer, Arthur, 129 science, 9–​10, 16, 33, 80–​81, 84, 88, 98, 100, 111, 146, 192–​93, 230–​31, 189n25 causal taxonomies in, 238–​39 description and, 187–​89, 217 free will and, 188 generality and, 181–​87, 181n10 and the humanities, 177–​93, 185n20 philosophy and, 98–​109 progress in, 101, 248 theory selection in, 248 understanding in, 167, 197, 231, 235 See also explanation; philosophy of science; scientific realism; understanding

scientific realism, 217–​18, 231. See also science Scriven, Michael, 222 sense-​making, 22, 40. See also understanding Sibley, Frank, 163 Sider, Ted, 145, 147, 149–​50, 150n16 skepticism, 16, 90, 101, 141–​42. See also understanding skill, 21–​22, 32–​33, 43–​44, 22n18, 30n33. See also knowledge; method Smith, Holly, 120 Smith, John E., 88 Sober, Elliott, 211, 218–​22 social cognition, 54, 69 social epistemology, 9, 54, 57–​58. See also epistemology social practices, 153–​54 social science, 98–​99. See also social practices Socrates, 135 Solomon, Robert, 80, 86 Sosa, Ernest, 195n1 speech acts, 11, 212, 217–​18, 220, 230, 213n4 Stampe, Dennis, 18n9 St. Paul, 122 Strevens, Michael, 28, 216n6 structure, 135–​56, 196, 237. See also explanatory structure; understanding supernaturalism, 93, 95. See also naturalism systematicity, 21, 44, 44n54 testimonial injustice, 58, 60, 75. See also hermeneutic injustice theism, 80, 82, 88 atheism and, 88 explanatory, 85, 87–​91 See also God; natural theology; theology theology, 87, 108. See also natural theology; theism

Index theoretical understanding, 9, 14–​19, 22–​29, 31, 35, 42, 26n24, 39n51, 41n52 dualism about, 32 illuminating nature of, 24, 40–​42 See also understanding transparency, 232, 234–​35, 237, 246 trope, 73, 75, 73n19 descriptive, 74 social, 74 truth, 112–​14, 148–​51, 160, 192, 247–​4 8, 250. See also belief; knowledge understanding, 1, 14, 17, 27, 69, 151–​55, 170, 206–​7, 249–​50, 154n18 characteristics of, 198–​99 coherence of, 19 comprehensive, 49 concept of, 239–​44 conceptual analysis of, 233–​239 as a condition on explanation,  209–​31 deficiency of, 39, 39n50 enrichment of, 46 the ethics of, 116–​33 explanation and, 4–​7, 135–​45, 155–​56, 222, 229–​30, 236, 240, 248 as factive, 138–​41 feeling of, 244–​45 and fluency, 232–​50 full, 71, 73 functional role of, 195–​99 genuine, 18–​22, 24, 41–​42, 46, 239 gradability of, 19, 204 grasp of dependence relations in, 7–​8, 11, 196–​97, 200, 202, 205, 236 heteronomous nature, 17 holistic, 63–​66, 70–​72, 74–​75, 63n11 of human frailty and fault, 124, 127–​28 humanistic, 7–​8


implicit and explicit, 168–​69,  168n7 intellectualism and, 45–​46 intelligent character of, 19 and knowledge, 170, 195, 206, 161n2, 164n6, 195n2, 200n8 luck and, 39n51, 169n10 as making sense, 19 moral, 174–​75, 161n1 mutual, 116 objectivity of, 4, 19, 19n11 order of, 19 perfect, 72, 75 of persons, 62–​65, 68–​75 of a perspective, 64, 68–​70 in the philosophy of science, 4–​7 as a praiseworthy good, 19 psychological component of, 19 robust character of, 19 scientific, 101, 230–​31, 238, 249 sense of, 232 and skepticism, 141–​42 social, 73 as a standing state, 19, 23 and structure, 135–​56 subjective, 3, 19 and testimony, 143–​45, 206–​7 theory of, 34, 243 the value of, 205–​7 See also aesthetic understanding; explanation; knowledge; making sense; practical understanding; religious understanding; theoretical understanding unity, 24–​27 obstacles to, 27–​35 of understanding, 14–​49 of the virtues, 24 See also copresence; ground; understanding unpredictability, 184 U-​profile, 20–​22, 26, 30–​34, 39–​43, 22n18, 26n24, 29n30. See also ground; understanding



values, 150, 200 Van Fraassen, Bas, 211–​13 virtue moral, 30n33 understanding as a, 132, 238 voluntary control, 232 Wallace, R. Jay, 124 Weil, Simone, 57, 68, 94n24 why-​question, 11, 211–​18, 220–​21, 224, 227, 230–​31. See also science

wisdom, 8 Wittgenstein, Ludwig, 111 Wolf, Susan, 118, 124, 126 Wollstonecraft, Mary, 57 Woodward, James, 211, 222–​30 Zagzebski, Linda, 32, 141, 237, 235n2 Zakaria, Fareed, 178n3