Evaluative Perception Anna Bergqvist Full Chapter

Evaluative Perception Anna Bergqvist
Visit to download the full and correct content document:

https://ebookmass.com/product/evaluative-perception-anna-bergqvist/
OUP CORRECTED PROOF – FINAL, 18/4/2018, SPi
Evaluative Perception
MIND ASSOCIATION OCCASIONAL SERIES

This series consists of carefully selected volumes of significant original papers on
predefined themes, normally growing out of a conference supported by a Mind
Association Major Conference Grant. The Association nominates an editor or editors
for each collection, and may cooperate with other bodies in promoting conferences
or other scholarly activities in connection with the preparation of particular volumes.
Director, Mind Association: Julian Dodd
Publications Officer: Sarah Sawyer
RECENTLY PUBLISHED IN THE SERIES:
In the Light of Experience
Edited by Johan Gersel, Rasmus Thybo Jensen, Morten S. Thaning, and Søren
Overgaard
Perceptual Ephemera
Edited by Thomas Crowther and Clare Mac Cumhaill
Common Sense in the Scottish Enlightenment
Edited by Charles Bradford Bow
Art and Belief
Edited by Ema Sullivan-Bissett, Helen Bradley, and Paul Noordhof
The Actual and the Possible
Edited by Mark Sinclair
Thinking about the Emotions
Edited by Alix Cohen and Robert Stern
Art, Mind, and Narrative
Edited by Julian Dodd
The Social and Political Philosophy of Mary Wollstonecraft
Edited by Sandrine Bergès and Alan Coffee
The Epistemic Life of Groups
Edited by Michael S. Brady and Miranda Fricker
Reality Making
Edited by Mark Jago
Evaluative Perception
 
Anna Bergqvist
and Robert Cowan
1
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
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 in certain other countries
© the several contributors 2018
The moral rights of the authors have been asserted
First Edition published in 2018
Impression: 1
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 licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries 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
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2017961513
ISBN 978–0–19–878605–4
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Contents
List of Contributors vii
Introduction 1
Anna Bergqvist and Robert Cowan
Part I. The Existence and Nature of Evaluative Perception

1. Rich Perceptual Content and Aesthetic Properties 19
Dustin Stokes
2. Can We Visually Experience Aesthetic Properties? 42
Heather Logue
3. Moral Perception Defended 58
Robert Audi
4. Evaluative Perception as Response-Dependent Representation 80
Paul Noordhof
5. Doubts about Moral Perception 109
Pekka Väyrynen
6. Seeing Depicted Space (Or Not) 129
Mikael Pettersson
7. Perception of Absence as Value-Driven Perception 143
Anya Farennikova
Part II. The Epistemology of Evaluative Perception

8. Moral Perception and Its Rivals 161
Sarah McGrath
9. Perception and Intuition of Evaluative Properties 183
Jack C. Lyons
10. On the Epistemological Significance of Value Perception 200
Michael Milona
11. Epistemic Sentimentalism and Epistemic Reason-Responsiveness 219
Robert Cowan
Part III. Evaluative Perception and Value Theory

12. Value Perception, Properties, and the Primary Bearers of Value 239
Graham Oddie
vi 
13. Moral Perception, Thick Concepts, and Perspectivalism 258

Anna Bergqvist
14. The Primacy of the Passions 282
James Lenman
15. Sexual Objectification, Objectifying Images, and ‘Mind-Insensitive
Seeing-As’ 295
Kathleen Stock
Bibliography 311
Index 331
List of Contributors
R A is John A. O’Brien Professor of Philosophy at the University of Notre

Dame. He has research interests in wide range of areas: epistemology, ethics,
philosophy of action, philosophy of mind, philosophy of religion, and political
philosophy. He is the author of several books, including The Good in The Right
(2004), Rationality and Religious Commitment (2011), and Moral Perception (2013),
as well as a wide range of journal articles.
A B is Lecturer in Philosophy at Manchester Metropolitan University.
Her principal research interests are aesthetics and moral philosophy. She is co-editor of
Philosophy and Museums: Ethics, Aesthetics and Ontology (CUP, 2016) to which she has
contributed a thematic piece on objectivity in interpretation. She has also published on
aesthetic particularism, thick evaluative concepts, and selected issues in philosophy of
language (semantic contextualism). She also works on the intersection between
metaethics, philosophy of perception, and philosophy of psychiatry, and is currently
preparing a monograph on particularism and personalized medicine.
R C is Lecturer in Philosophy at the University of Glasgow. His
research is focused on ethics, epistemology, and the philosophy of mind. In particular
he is interested in the nature and epistemology of intuition, perception, and emotion,
as well as the connections between these and accounts of ethical knowledge. He has
recently published papers on these topics in Canadian Journal of Philosophy, Ethics,
and Philosophy and Phenomenological Research.
A F is Lecturer in Philosophy at the University of Bristol. Her
research interests lie in philosophy of mind and cognitive science. She has recently
published on the perception of absence in Philosophical Studies.
J L is Professor of Philosophy at the University of Sheffield. His
research focuses on ethics, especially metaethics. His recent work has dealt with
issues relating to moral realism, moral expressivism, moral psychology, moral epis-
temology, and moral responsibility, compatibilism, consequentialism, contractual-
ism, constructivism, and contingency.
H L is Lecturer in Philosophy at the University of Leeds. Her research
focuses on issues in metaphysics and epistemology, and particularly on issues
concerning perceptual experience. She has published and forthcoming papers on
Naïve Realism, disjunctivism, scepticism about the external world, experience of
high-level properties, and the metaphysics of colour. Recent publications include
‘Good News for the Disjunctivist about (one of) the Bad Cases’ (Philosophy and
Phenomenological Research, 2013), and ‘Experiential Content and Naïve Realism: A
Reconciliation’ (Does Perception Have Content? ed. Berit Brogaard, OUP, 2014).
J C. L is Professor of Philosophy at the University of Arkansas. His research
is mainly in epistemology, cognitive science, and philosophy of mind. He has recently
published a book entitled Perception and Basic Beliefs (OUP, 2011), and is the editor
for the journal Philosophical Topics.
viii   
S MG is Assistant Professor of Philosophy at Princeton University. Her

primary areas of interest are metaphysics and ethics; key recent publications include
‘Skepticism About Moral Expertise as a Puzzle for Moral Realism’ (Journal of
Philosophy, 2011) and ‘Moral Knowledge and Experience’ (Oxford Studies in
Metaethics, 2011).
M M completed his dissertation on the role of emotions in moral
epistemology in August 2016 at the University of Southern California. For the
2016–17 academic year, he is a postdoctoral researcher at Cornell University working
on the nature and value of hope.
P N is Anniversary Professor in Philosophy at the University of York.
His main research interests are in philosophy of mind, action theory, and metaphys-
ics. He has published extensively in these areas, and is currently writing a book which
presents a counterfactual theory of causation, and another book on the nature and
explanatory character of consciousness.
G O is Professor of Philosophy of the University of Colorado at
Boulder. His main research areas are metaphysics, value theory, metaethics, formal
epistemology, philosophical logic, and aesthetics. He is author of two monographs,
Value, Reality, and Desire (OUP, 2005) and Likeness to Truth (Reidel, 1986), and co-
editor of two other books What’s Wrong (OUP, 2004) and Justice, Ethics, and New
Zealand Society (OUP, 1992).
M P is Assistant Professor in Philosophy at the Lingnan University,
Hong Kong. His area of specialization lies within aesthetics (broadly construed), in
particular its intersection with philosophy of mind (perception and imagination) and
metaphysics (causation). He has published on the phenomenology of photography,
about the role of causation in photography, and an oft-neglected phenomenon in
pictorial representation, namely that of pictorial occlusion.
K S is Reader in Philosophy at the University of Sussex. Her main
research interests are aesthetics and philosophy of mind, especially the imagination
She has published extensively in these areas, and on fiction and definitions of art. She
is editor of Philosophers on Music (OUP, 2007) and co-editor of New Waves in
Aesthetics (Palgrave, 2008).
D S is Assistant Professor in Philosophy at the University of Utah. He
works primarily in philosophy of mind and cognitive science. His research includes
work on perception, imagination, and creative thought and behaviour. He is cur-
rently co-writing a book on Imagination and his paper ‘Cognitive Penetration and
the Perception of Art’ was the winner of the 2012 Dialectica Essay Prize.
P V̈ is Professor of Moral Philosophy at the University of Leeds. He
is the author of The Lewd, the Rude and the Nasty: A Study of Thick Concepts in Ethics
(OUP, 2013) and articles on a wide range of topics in metaethics.
Introduction
Anna Bergqvist and Robert Cowan
Evaluation is ubiquitous. It isn’t an exaggeration to say that we assess actions,

character, events, and objects as good, bad, cruel, kind, beautiful, ugly, etc., almost
every day of our lives. A paradigm of evaluation is evaluative judgement. For example,
Jen makes the judgement that the north-west Scottish Highlands are beautiful, while
Phil judges that the rise of religious intolerance is a bad thing. Indeed, if we are liberal
about how we classify the ‘evaluative’¹ such that it includes the deontic—rightness,
wrongness, obligation, blameworthiness, etc.—then the pervasiveness of evaluative
judgement is even more striking.
So far, so uncontroversial. A more contentious set of issues arise when we consider
whether evaluative judgements are expressions of cognitive or non-cognitive states,
whether they deploy distinctively evaluative concepts, whether they are ever true,
or can be epistemically justified, etc. Note, however, that recent developments in
metaethics, in particular the emergence of versions of Quasi-Realism,² somewhat
complicate debates about these questions.³ For example, if we adopt a deflationary
view of truth, then even Non-Cognitivists can agree that some evaluative judgements
are true. Given this, assuming that evaluations can be true, etc., is perhaps not as
controversial as it might first seem. In what follows we will speak as if evaluative
judgements can indeed be true and epistemically justified.
Historically, it has been thought that a distinctive kind of evaluation is perceptual or
experiential. Further, some have also believed that this sort of evaluation can be veridical,
and can play significant roles, e.g. epistemic.⁴ To illustrate: in aesthetics, many philo-
sophers have claimed that adequate aesthetic judgement must be grounded in the
appreciator’s first-hand experience of the item judged. Thus, Frank Sibley asserts that
we have to read the poem, hear the music, or see the picture (not merely have it described in
non-merit and even determinate terms if that were possible) and then judge or decide whether
an aesthetic merit term applies to it or not.⁵
¹ We won’t attempt to provide some set of necessary and sufficient conditions for what gets to count as
‘evaluative’. Suffice to say that we are assuming a broad notion.
² See e.g. Blackburn (2006). ³ See e.g. Drier (2004).
⁴ Note, that as was the case with value judgements, notions of veridicality, etc., can presumably be
understood in a deflationary sense too.
⁵ Sibley (2001), 99.
     
This claim is often treated as a truism about aesthetic discourse.⁶ Further, some
philosophers⁷ have also made a positive claim about what critical debate involves:
that its purpose is to bring one’s audience to see the object in a certain way. Some also
think that criticism does not depend for its plausibility on general aesthetic criteria, if
that means deductive reasoning from general aesthetic claims, for no such claims are
available.⁸ In any case, the point of critical discussion is not the formation of belief,
but the engendering of perception. Related to this, within the tradition from Hume
through to Sibley in analytic aesthetics, acquired sensibilities of taste (and similar
conditions such as a sense of humour) are seen as cognitively necessary in the
appreciation of aesthetic merit qualities.⁹ This is suggestive of an important role for
experience in the normative standing of aesthetic beliefs.
In ethics, Aristotle and modern-day Virtue Ethicists such as John McDowell¹⁰
characterize practical wisdom in terms of a perceptual ability. Along somewhat
similar lines, other philosophers¹¹ have emphasized the importance of ethical ‘vision’
as a matter of seeing things aright, particularly with respect to its ability to lead us to
revise our preconceptions about particular objects, persons, and events. As Iris
Murdoch puts it, goodness is
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 perfectly familiar kind of moral discipline.¹²
Elsewhere, Moral Sense Theorists like Frances Hutcheson¹³ seemed to think that an
affective experience of moral (dis)approbation constituted our primary awareness of
moral qualities.
Contemporarily, there has been continued and renewed interest in the connection
between the evaluative and the perceptual. We here note five examples.
Firstly, in aesthetics there has been growing discussion of the idea—introduced by
Richard Wollheim¹⁴—that the phenomenon of Seeing-In, which is often claimed to
be typical of pictorial experience, marks out a sui generis kind of perception.¹⁵
Second, a relatively substantial literature on the existence and nature of ethical
perception has sprung up over the last decade.¹⁶ Much of this has been informed
by recent work in the philosophy and epistemology of perception. Specifically, (and
this is our third example) it has been influenced by the emergence of High-Level
views¹⁷ about the contents of perceptual experience, i.e., roughly, views which allow
⁶ But see Livingston (2003).

⁷ See e.g. Hampshire (1967); Sibley (1959/2001), (1965/2001), and (1983/2001); Strawson (1966).
⁸ See e.g. Bergqvist (2010); Isenberg (1967); Mothersill (1984).
⁹ For further discussion of this point, see e.g. Hopkins (2006) and (2011).
¹⁰ McDowell (1998).
¹¹ See e.g. Blum (1994); DePaul (1993); McNaughton (1988); Murdoch (1970); Sherman (1989).
¹² Murdoch (1997), 330. For in-depth analysis of the implications of Murdoch’s account of moral
perception for the possibility objectivity in ethics, see Bergqvist, Ch. 13 in this volume. For its importance
to the history of moral philosophy in the twenty-first century, see Bergqvist (2015).
¹³ See e.g. Hutcheson (1728/1991). ¹⁴ See e.g. Wollheim (1980).
¹⁵ See e.g. Lopes (1996) and (2005); Hopkins (1998); Pettersson (2011).
¹⁶ See e.g. Audi (2013) and (2015); McBrayer (2010a); Väyrynen (2008a); Werner (2016).
¹⁷ See e.g. Bayne (2009); S. Siegel (2010a).
 
that we can perceive (in the canonical modalities) complex properties such as natural
kinds and other categorical properties. Interest in ethical perception has also been
informed by recent work on the view that perception in the canonical sensory
modalities is Cognitively Penetrable, i.e., roughly, perception is susceptible to non-
trivial influence from cognitive states like beliefs and desires (more on this later).¹⁸
Fourth, there has been much recent interest in the development or further refinement
of perceptual views of desires,¹⁹ emotions,²⁰ and pains²¹ according to which they are
experiences of value. Fifthly, and finally, there has been some recent work²² on the
connection between the existence of evaluative perception and views in value theory,
e.g., whether perceptual theories of the emotions are compatible with particular kinds
of Sentimentalism about value concepts and properties.
Despite this history and recent developments, there has only been limited inter-
action between philosophers working on these various topics. This volume aims to
remedy this by bringing together philosophers in aesthetics, epistemology, ethics,
philosophy of mind, and value theory, to contribute in novel ways to debates about
what we call ‘Evaluative Perception’. Specifically, they contribute to answering the
following questions:
Questions about Existence and Nature: Are there perceptual experiences of
values? If so, what is their nature? Are experiences of values sui generis? Are values
necessary for certain kinds of experience?
Questions about Epistemology: Can evaluative experiences ever justify evaluative
judgements? Are experiences of values necessary for certain kinds of justified
evaluative judgements?
Questions about Value Theory: Is the existence of evaluative experience sup-
ported or undermined by particular views in value theory? Are particular views in
value theory supported or undermined by the existence of value experience?
In the following three sections we provide an introduction to some of the main topics
of discussion, and to the volume papers.
Before doing so, the reader should note the following. As shall become clear, the
fifteen papers in this volume are all concerned with answering one or more of these
questions, and often cross-cut different areas, e.g., epistemology and value theory.
For that reason, it is somewhat artificial to divide them into discrete subgroups.
However, to aid the reader in seeing the connections between the papers, and to
contribute to the thematic unity of the volume, we have placed the papers into three
subgroups. The first is primarily concerned with the Existence and Nature of Evalu-
ative Perception. Here we have placed contributions by Dustin Stokes, Heather
Logue, Robert Audi, Paul Noordhof, Pekka Väyrynen, Mikael Pettersson, and Anya
Farennikova. The second group is mostly addressed to questions about Evaluative
Perception and Epistemology. Here we have contributions from Sarah McGrath, Jack
C. Lyons, Michael Milona, and Robert Cowan. Finally, the third subgroup is focused
¹⁸ See e.g. Firestone and Scholl (2014); Macpherson (2012); S. Siegel (2012).
¹⁹ See e.g. Oddie (2005). ²⁰ See e.g. Döring (2003). ²¹ See e.g. Bain (2013).
²² See e.g. Brady (2013); Cowan (2016); Tappolet (2011).
     
on the connections between Evaluative Perception and Value Theory. Here can be
found the contributions by Graham Oddie, Anna Bergqvist, James Lenman, and
Kathleen Stock.
1. Existence and Nature of Evaluative Perception

Many of the papers in the volume are concerned with the existence and nature of
Evaluative Perception. As we are characterizing it, this involves the experiential
representation²³ of value properties, i.e., there are some perceptual experiences with
evaluative content.
Before clarifying this and different kinds of evaluative perceptual experience, the
phenomenon should be distinguished from two others.
Firstly, there is what we call ‘De Re Perception of Values’.²⁴ That is, we perceive
objects, persons, events, states of affairs that as a matter of fact instantiate evaluative
properties, e.g., moral rightness, beauty. Everyone but the value Error Theorist can
assent to the claim that there is De Re Perception of Values. For our purposes, the
crucial difference between this and Evaluative Perception is that the De Re Percep-
tion of some F by a subject is compatible with the subject not having a perceptual
‘experience’ of F, i.e., an experience with evaluative content.
Second, there is what we call ‘Evaluative Seeing-That’. This involves making an
evaluative judgement in response to a perceptual experience (often it is assumed that
the evaluative judgement has some positive epistemic status, e.g., constitutes know-
ledge). One way in which some restrict the notion of Evaluative Seeing-That is by
insisting that it must be both psychologically and epistemically non-inferential, i.e.,
the relevant judgements mustn’t be the result of, nor be epistemically dependent
upon, inference.²⁵ Given this restriction, there is scope for philosophical debate as to
whether all putative instances of Evaluative Seeing-That involve some sort of infer-
ential epistemic dependence (for more discussion see Section 2 of this Introduction).
For our purposes, the important difference between this and Evaluative Perception is
that Evaluative Seeing-That is consistent with the perceptual experience only repre-
senting non-evaluative properties, e.g., the properties upon which the evaluative
property supervenes or is consequential.
Now to clarify Evaluative Perception. Although there is an important distinction
between non-factive perceptual experience, and perception which is factive, we will
mainly focus on the former and for ease of expression simply refer to it as ‘Evaluative
Perception’. If there is Evaluative Perception, in this central sense, then subjects can
have perceptual experiences that represent the instantiation of evaluative properties
²³ In characterizing things this way, we are using the language of an Intentionalist or Representational
theory of perception. On a Relational view of perception, by contrast, veridical experiences do not have
representational content, but instead involve the obtaining of a perceptual relation between the perceiver
and worldly objects, such that the relevant objects can be said to literally constitute one’s experience. See
e.g. M. G. F. Martin (2006). For brevity we will speak of perceptual representation (partly because most of
the papers in the volume assume it) but the Relational view ought to be kept in mind.
²⁴ Or ‘non-epistemic seeing’. See Dretske (1969), esp. ch. 2. See also Dretske (1993).
²⁵ One might go further and provide a positive characterization of the aetiology of Seeing-That,
e.g., they have their source in modular perceptual systems; see Jack Lyons (Ch. 9 in this volume).
 
of the sort described above, i.e., some perceptual experiences are accurate only if they
represent evaluative properties. Later we will introduce some other phenomena
which, although they don’t fall under this category, are worthwhile considering
alongside discussion of Evaluative Perception proper.
There are at least three different kinds of Evaluative Perception that are worth
distinguishing.
The first of these is what we call ‘Canonical Evaluative Perception’. This involves a
commitment to the representation of evaluative properties in one or more of the five
canonical sensory modalities. Some version of this view is defended or endorsed in
this volume by Robert Audi (at least on one interpretation), Paul Noordhof, and
Dustin Stokes.
The second kind of Evaluative Perception is what we refer to as ‘Affective
Evaluative Perception’. This involves the representation of evaluative properties in
an affective or conative state, such as desire, emotion, or pain. On each of these views,
the relevant mental state apparently shares important features with ordinary percep-
tual experience, e.g., they have phenomenal character, representational content, can
be recalcitrant to doxastic changes, exemplify some covariance with the subject’s
environment, and can play a non-inferential epistemic role. The point of interest for
our purposes is that if such experiences could have an evaluative content—and
proponents all seem to think that they can, e.g., Oddie²⁶ conceives of desires as
experiences of goodness, Döring²⁷ thinks that moral emotions such as guilt represent
moral properties, Bain²⁸ thinks that pain experiences represent bodily damage as bad
for the subject—then there could be perceptual experiences with evaluative content,
albeit ‘non-traditional’ ones. Versions of this view are discussed or defended by
Robert Cowan, Graham Oddie, Michael Milona, and Paul Noordhof in this volume.
The third kind of Evaluative Perception is what we call ‘Sui Generis Evaluative
Perception’. This involves the representation of evaluative properties in a sui generis
kind of experience (by ‘sui generis’ we simply mean that it doesn’t reduce to familiar
sensory or affective phenomena). One example of this, which is discussed or
defended in this volume by Robert Audi, Michael Milona, and Pekka Väyrynen, is
the view that Evaluative Perceptions are ‘integrated’ experiences, which are an
amalgam of sensory, emotional, and imaginative components. This is distinct from,
e.g., Canonical Evaluative Perception in part because the sensory component of an
integrated experience need not have an evaluative content. Another example of Sui
Generis Evaluative Perception, which is discussed by Jack Lyons in this volume, is the
view that, although evaluative properties aren’t represented in sensory experience
proper, they can be represented in perceptual seeming states. Roughly, these are
propositional non-doxastic states that are the causal upshot of sensory experience or
a sensory system.²⁹
Due to space constraints, our discussion will be focused on the prospects for
Canonical Evaluative Perception. However, this will still allow us to make extensive
²⁶ Oddie (2005) and Ch. 12 in this volume. ²⁷ Döring (2003). ²⁸ Bain (2013).
²⁹ For extensive discussion of seemings, see C. Tucker (2013).
     
reference to other kinds of Evaluative Perception in this section and throughout the
Introduction.
It is first worth noting that Canonical Evaluative Perception runs counter to the
mainstream view according to which canonical perceptual experience, e.g., vision,
only represents what are called ‘Low-Level’ properties such as colour, shape, and
motion. If there is Canonical Evaluative Perception then ‘High-Level’ properties, e.g.,
aesthetic and ethical, must also be represented in perceptual experience.
Why would anyone think that there is Canonical Evaluative Perception? One kind
of evidence—already noted in this Introduction and highlighted in chapters by Audi,
Logue, Noordhof, Stokes, Väyrynen—is that we often use perceptual language when
talking about the evaluative. For example, it would not be unusual to hear someone
say ‘the Botticelli looks incredible close up’, or, ‘I could hear her demeaning tone’.
However, this sort of evidence is quite weak. Even if we accept that such cases are
psychologically (or perhaps even epistemically) non-inferential, this doesn’t clearly
support the truth of Canonical Evaluative Perception, as opposed to some other kind
of Evaluative Perception. Indeed, it doesn’t obviously support any kind of Evaluative
Perception, if such cases can be explained as cases of Evaluative Seeing-That on the
basis of non-evaluative perceptual experience.
A more promising strategy is perhaps to appeal to what have come to be known as
‘Contrast Arguments’. This sort of argument has been presented by some philo-
sophers of perception, notably Tim Bayne³⁰ and Susanna Siegel,³¹ in support of the
High-Level (or ‘rich’ or ‘liberal’) view of perception, with a focus on natural kind
properties, e.g., being a pine tree, and causal relations. Very roughly, Contrast
Arguments involve conceiving of two experiences with very similar or identical
low-level content, but where there is plausibly a difference in the phenomenology
between them, e.g., the contrast between the experiences of looking at pine trees
before and then after acquiring a familiarity with what their characteristic look is.
The crucial move in Contrast Arguments is to say that the best explanation of the
phenomenological difference is a difference in the representational contents of
perception as opposed to, e.g., attentional differences.
Put very simply:
P1: There is a phenomenological difference between target experiences e and e*.
P2: The best explanation of the phenomenological difference between target
experiences e and e* requires positing some high-level content, c, in experience.
C: (Probably) some experiences have some high-level, c, content.
Most are willing to accept P1. Thus the central task for proponents of Contrast
Arguments is to show that positing high-level perception is indeed the best explan-
ation of the relevant cases.
Is this kind of argument more or less promising in the evaluative case? In this
volume Dustin Stokes provides a Contrast Argument in favour of the conclusion that
at least some aesthetic properties, e.g., gracefulness, are represented in canonical
experience. Interestingly he thinks that this sort of argument is more promising in
³⁰ See Bayne (2009). ³¹ See S. Siegel (2010a).

 
the aesthetic case than for other high-level properties such as natural kinds. Stokes
argues that there is no way to account for the phenomenology of aesthetic cases
without admitting perceptual representation of some organizational gestalt, e.g., an
organizational gestalt typical of impressionist works. Crucially, to experience organ-
izational aesthetic gestalts just is to experience the relevant aesthetic properties.
Interestingly, this feature is lacking in cases of natural kind or even colour properties
(a similar view is defended in Bergqvist’s discussion of value theory in this volume).
Against this kind of optimism and the general line of thought, Heather Logue in
Chapter 2 casts doubt on aesthetic Contrast Arguments. She thinks that there is
another plausible explanation (which is just as parsimonious as positing Canonical
Evaluative Perception) for the difference in phenomenology: appeal to emotional
states (Affective Evaluative Perception). Notably, this alternative explanation may be
absent in the case of other high-level properties like natural kinds. Logue also
considers and rejects arguments in favour of Canonical Evaluative Perception of
aesthetic properties which appeal to the putative ‘observationality’ of some aesthetic
properties—roughly, for an aesthetic property, F, in ideal viewing conditions, if
something visually appears to be F, it is F—or to their allegedly ‘superficial’ meta-
physical nature. Regarding the latter, Logue draws attention to the idea that aes-
thetic properties are plausibly response-dependent (a feature that might complicate
Stokes’s argument).
In Chapter 5 Pekka Väyrynen considers a Contrast Argument for moral percep-
tion,³² but finds that the phenomenological difference can be explained just as well by
a model which posits that a non-perceptual moral representation results from ‘an
implicit habitual inference or some other type of transition in thought which can be
reliably prompted by the non-moral perceptual inputs jointly with the relevant
background moral beliefs’.³³ The representations involved may be affective in nature
(hence Väyrynen may be interpreted as countenancing the existence of Affective
Evaluative Perception or Sui Generis Evaluative Perception). Further, he thinks that
this alternative model possesses theoretical virtues of simplicity and unity that give it
an advantage. Positing Canonical Moral Perception adds no explanatory power.
Thus P2 of a moral Contrast Argument should be rejected.³⁴
Another kind of argument that could be offered in favour of Canonical Evaluative
Perception is epistemological:
P1: There could be a certain kind of justified belief, J, only if there were
Canonical Evaluative Perception.
P2: There is justified belief J.
C: There is Canonical Evaluative Perception.
P1 claims that Canonical Evaluative Perception is a necessary condition for a
particular kind of epistemic justification (a similar argument could be constructed
³² Note that Werner (2016) presents a contrast argument in favour of moral perception.
³³ Pekka Väyrynen (Ch. 5 in this volume).
³⁴ Väyrynen thinks that whether a Contrast Argument is likely to succeed depends on the properties in
question, and his view is that it is a good deal less promising in the moral case as compared with other high-
level properties (as Logue argues with regard to aesthetic properties).
     
for knowledge and other positive epistemic properties). We will postpone discussion
of this idea until Section 2 of the Introduction. For now note that Logue discusses and
rejects this sort of epistemic argument in her chapter.³⁵
Even if one is somewhat sympathetic to arguments in favour of Canonical Evalu-
ative Perception, there remains, inter alia, the question of how value properties could
be represented in experience. For example, it might seem highly implausible that
value properties are represented in a similar way to that in which low-level properties
like colours and shapes are represented in vision. This point is not only noted by
opponents of Canonical Evaluative Perception—see Jack Lyons’s paper in this
volume—but also by proponents of the view, such as Audi and Noordhof (both in
this volume). Whether or not this is fatal for the view depends upon whether we
allow that there are different kinds of representation in perceptual experience. For
example, Audi distinguishes, inter alia, between what he calls the ‘perceptual’,
associated with ‘cartographic’, ‘pictorial’ and ‘basic’ representation (the sort allegedly
involved in the representation of colours and shapes), and the ‘perceptible’, which
involves non-cartographic, non-pictorial, and non-basic representation. Crucially,
he seems to think that some perception (in the canonical modalities) is of the
perceptible. Noordhof also distinguishes between ‘sensory’ and ‘non-sensory’ repre-
sentation in canonical perception.³⁶ If we admit that there can be different kinds of
perceptual representation, then Canonical Evaluative Perception may be a good deal
more plausible.
One thing that makes it difficult to determine whether there really is Canonical
Evaluative Perception is that it is hard to see how we can satisfactorily resolve
disputes between those who, like Audi and Noordhof, think that there is something
like non-sensory representation, and those, like Lyons, who think that what is being
referred to as non-sensory perceptual representation is actually something post-
perceptual, e.g., a perceptual seeming state which has phenomenal character (though
perhaps it is bland character) and conceptual content. Note, however, that Noordhof
provides arguments in favour of admitting non-sensory representation (one which
appeals to the phenomenology of chicken sexers and speech perception, the other
which appeals to epistemic considerations), as well as reasons for doubting the
existence of perceptual seemings in his chapter. Also, in this context it is worth
noting Heather Logue’s view that there may not be a fact of the matter regarding the
question of whether aesthetic properties are represented in canonical perceptual
experience.
Even if one allows that there can be non-cartographic or non-sensory perceptual
representation, e.g., perhaps this is what all high-level representation involves, some
still might think that there are specific problems for Canonical Evaluative Perception.
³⁵ Robert Audi (Ch. 3 in this volume) also suggests the following sort of argument for moral perception:
if there is perception of emotion, e.g., anger, then there is little reason to doubt the existence of moral
perception. Of course, one might doubt that the antecedent is true, and in any case might doubt the truth of
the conditional. As Pekka Väyrynen argues, admission of perceptual contents should proceed on a case-by-
case basis.
³⁶ Although some aspects of Noordhof ’s distinction appear to be similar to Audi’s, Noordhof places
emphasis on discriminability of properties in order to flesh out the difference.
 
In his chapter, Noordhof thinks that there may be some resistance to the idea of
Canonical Evaluative Perception even among those who countenance non-sensory
representation because of the common thought that value properties are in some way
response-dependent (also highlighted by Logue).³⁷ In order to address this, Noordhof
presents an account of intrinsic response-dependent representation—i.e., represen-
tation of a property that has nothing to do with the representation standing in
relation to something independently characterized in the world. He illustrates this
with reference to the perception of badness in pain and moral perception.
Suppose that a plausible argument can be made for thinking that there is Canon-
ical Evaluative Perception. To support this, proponents will need to identify psycho-
logical mechanisms by which this could take place. This is, of course, an empirical
matter that can’t plausibly be settled from the armchair.
One candidate model that has received attention in the recent literature on moral
perception is that Canonical Evaluative Perception could be brought about by a
process of Cognitive Penetration.³⁸ Roughly, Cognitive Penetration of sensory
experience is possible if and only if it’s possible for two subjects to have experiences
which differ in content and/or phenomenal character, where this difference is the
result of a causal process that traces more or less directly to states in the subjects’
cognitive system, and where we hold fixed the perceptual stimuli, the condition of the
subjects’ sensory organs, the environmental conditions, and the attentional focus of
the subjects.³⁹
Potential cognitive penetrators include moods, beliefs, desires, emotions, and
character traits. Such a model might help to explain how there can be expertise
with respect to values, e.g., in aesthetics. What the expert has, and the novice lacks, is
a set of background commitments that cognitively penetrate their sensory experience
such that it comes to have an evaluative content.
Although there is growing evidence for the Cognitive Penetrability thesis,⁴⁰ it is
still highly controversial. It is therefore worth noting that adopting the Cognitive
Penetrability model requires proponents of Canonical Evaluative Perception to
undertake substantial empirical commitments regarding the capacities of ordinary
perception and its relation to cognition. In this volume, Bergqvist, Cowan, Lyons,
Pettersson, Stokes, and Väyrynen discuss Cognitive Penetration.
Another model appeals to Perceptual Learning, which involves a repeated asso-
ciative process that takes place within the perceptual system. For example, after
repeated exposure to a particular kind of artwork (and top-down processing within
the visual system) perhaps one’s visual system may come to encode information
about aesthetic properties. Stokes discusses this possibility in Chapter 1 in this
volume. If we think that there is such a thing as aesthetic expertise, one might
think that one feature that distinguishes the expert from the novice is that they
have been repeatedly exposed to artworks. Note, however, that a similar sort of
³⁷ Noordhof also thinks that there is a lack of ‘phenomenal presence’ in the case of aesthetic and moral
representation, though not in the case of the representation of badness in pain experience.
³⁸ Cowan (2015b). ³⁹ See e.g. Vance (2014).
⁴⁰ See e.g. Bruner and Goodman (1947); Delk and Fillenbaum (1965); Hansen et al. (2006). See also
D. Stokes (2013).
     
repeated exposure may be less common in the moral case (although this might be
overstating things).
A final model for Canonical Evaluative Perception is that our perceptual systems
are hard-wired for evaluative representation (in the way that representation of
colours and shapes seems to be). One might think that a necessary condition for
perceptual hard-wiring is that possessing the relevant representational powers will be
of use to subjects who possess them in almost any environment in which they might
be placed.⁴¹ Notice that, while it is not implausible that moral properties meet the
necessary condition, it is perhaps a good deal less so for kind properties, e.g., being a
pine tree, and aesthetic properties.
In Chapter 3 in this volume, Robert Audi may be interpreted as defending a sort of
hard-wired view about Evaluative Perception; however, it is not entirely clear
whether he is willing to countenance this for Canonical Evaluative Perception or
only for a kind of Sui Generis Evaluative Perception (integrated experiences). Finally,
it is worth noting that in the case of Affective Evaluative Perception—regarding
emotion and pain in particular—philosophers seem to be more sanguine about the
claim that evaluative representation is hard-wired.
Although the majority of the papers in the volume are engaged in debates about
Evaluative Perception (as we have defined it), two of the volume papers—those by
Mikael Pettersson and Anya Farennikova—discuss distinct kinds of phenomena
which we think are usefully grouped with these other papers.
Firstly, there is the phenomenon of Seeing-In, typical of pictorial representation,
which—as noted earlier—some have thought marks out a distinctive form of per-
ception. For instance, it is often said that an important part of our experience of
pictures, such as looking at a wedding photograph in the family album, is that we see
its subject matter ‘in’ its surface, in a way that is different from watching the world
(its pictorial content) face-to-face.
In Chapter 6 in this volume, Mikael Pettersson problematizes extant accounts—
resemblance⁴² and recognitional⁴³—of Seeing-In by considering how they fare with
respect to the phenomenon of seeing empty space in pictures. Although seeing
empty space in a picture (and seeing-in more generally) isn’t a form of Evaluative
Perception (as we have defined it), the alleged distinctiveness of this form of
perception, and its apparent significance for aesthetics, e.g., the aesthetic appreci-
ation of some pictures will depend on our experience of empty space or void in the
picture, make inclusion of discussion of this topic in this collection important.
Pettersson rejects cognitive accounts of seeing empty space in pictures (what we
earlier referred to as Seeing-That) and goes on to sketch an imagination theory,
which according to one model involves the cognitive penetration of perceptual
experience by imagination.
The second phenomenon is that of perceptual experiences which, though not
themselves Evaluative Perceptions, are nevertheless allegedly causally dependent for
their occurrence upon subjects having certain sorts of evaluative commitments. For
example, perhaps possessing certain aesthetic beliefs makes one more perceptually
⁴¹ See e.g. Macpherson (2012). ⁴² See Hopkins (1998). ⁴³ See Currie (1995).
 
attentive to particular low-level features of artworks, e.g., colour and shape arrays,
even if such experiences don’t represent the instantiation of evaluative properties. In
addition, we might include affordances, as a related kind of evaluative perceptual
experience that represents not an evaluative property as such but rather features of the
perceived lived environment that bear intimately on action in reasonable perceptual
agents—such as giving up one’s seat on the bus upon seeing a visibly tired person.⁴⁴
Rather than an Evaluative Perception, these are perhaps better called ‘Value-Enabled
Perceptions’.
In Chapter 7 in this volume, Anya Farennikova argues that perceptual experiences
of absences, e.g., seeing that Pierre is not in the café, or noticing the absence of a ring
on someone’s fourth finger, are dependent upon the subject’s desires or values.
Roughly, the idea is that desiring and valuing more generally can make one percep-
tually sensitive to the existence of the absence of certain things in one’s environment.
Without those values, one wouldn’t recognize what isn’t there. Thus, Farennikova
thinks that one’s desires enable the perception of absences: they are value-driven
experiences.
2. Epistemology and Evaluative Perception

As was noted in Section 1, one kind of argument that is sometimes offered in favour
of Canonical Evaluative Perception is epistemological. What is perhaps the crucial
premise in that argument claims that Canonical Evaluative Perception is necessary
for the existence of some kind of justified evaluative belief or knowledge. One
candidate might be the existence of justified evaluative beliefs about concrete
particular cases, e.g., John’s judgement that the hoodlums were wrong to set fire
to the cat, or Lucy’s belief that the lead ballerina’s movements in the final act were
graceful.
In this volume there are several papers that argue against this sort of view. In
Chapter 2, Logue at one stage suggests that an Inferentialist picture could perhaps
accommodate justified beliefs in aesthetic cases, e.g., Lucy’s justified belief about the
gracefulness of the ballerina might be the result of her justified background belief that
features F, G, H are usually sufficient for movements to be graceful, and the belief or
perceptual registering that the ballerina possesses those features.
In Chapter 8, Sarah McGrath—focusing primarily on moral perceptual knowledge—
argues at length against versions of Inferentialism (albeit focused on moral cases):
deductive, inductive, and abductive. These models allegedly fail to adequately explain
moral knowledge in particular cases. She also considers and rejects a more sophisticated
Inferentialist model found in the work of Kieran Setiya,⁴⁵ according to which the
relevant moral ‘perceptual’ knowledge is actually inferred from non-moral evidence.
However, McGrath doesn’t think that Canonical Evaluative Perception is necessary
for justified belief or knowledge in the relevant cases. Instead, she argues that non-
inferential Evaluative Seeing-That is sufficient. Interestingly, McGrath argues that
such non-inferential ‘perceptual’ judgements about particular concrete cases—as
⁴⁴ This example is from Bengson (2016). ⁴⁵ Setiya (2012).

     
opposed to judgements about hypothetical cases which are implicitly general⁴⁶—can

play an important and hitherto unrecognized role in the process of arriving at
reflective equilibrium.
Jack Lyons, in Chapter 9, also argues—from an epistemological Reliabilist
perspective—that the sort of justification that one might have thought could only
be gained by Canonical Evaluative Perception, can also be delivered by Sui Generis
Evaluative Perception (perceptual seemings), or what he calls ‘value intuitions’, i.e.,
cognitive seemings with evaluative content.
Finally, a more extreme version of the view that Canonical Evaluative Perception
plays some epistemically essential role would claim that it is required for the
existence of any justified (substantive) evaluative belief. In Chapter 10, Michael
Milona argues against this,⁴⁷ concluding that for any substantive justification gained
from Canonical Evaluative Perception, subjects could gain similar justification by
imaginatively considering the relevant cases and having some sort of affective, e.g.,
emotional, response to it. Indeed, Milona goes further and claims that the question of
whether there is Canonical Evaluative Perception is not important at all for debates
about evaluative epistemology (but recall McGrath’s point about the epistemic sig-
nificance of perceptual evaluative judgements). Note, however, that Milona seems to
think that Affective Evaluative Perception—desires or emotions—is a necessary
condition for justified substantive evaluative beliefs. We’ll return to this view shortly.
Even if Canonical Evaluative Perception isn’t necessary for evaluative justification
(and thus epistemological arguments for its existence are likely unsound), we can ask
whether it could be sufficient for justified evaluative beliefs. Further, if it turns out
that Canonical Evaluative Perception were sufficient for basic or non-inferential
justification then it could still have epistemological significance, i.e., it could be a
source of regress-stopping justification. Is it?
On one view of non-inferential experiential justification, in order for an experience
to immediately justify a belief that p, this requires that the experience has distinct-
ively presentational content with respect to p. Being presented with p is meant to
differ in important respects from representing p (contrast seeing a red ball with
simply thinking about one). In Chapter 4, Noordhof claims that moral perception
lacks presentational character in a way that other sorts of Evaluative Perceptions, e.g.,
pain experiences, do not. He suggests that this impacts on the former’s capacity to
confer justification. On the other hand, Audi (Chapter 3) has no problem with the
idea that moral perceptions could be presentational. One explanation for the differ-
ence in views is that there are competing views of what presentational content is.⁴⁸
Which is the correct view arguably needs to be settled on independent grounds.
If one thinks that non-inferential epistemic justification or knowledge require
reliability, then one needs to give an account of how Canonical Evaluative
Perception could be reliable. Robert Audi’s considered view—in the moral
case—is that this has something to do with the possession of moral concepts.
⁴⁶ See Kagan (2001).

⁴⁷ Note that Michael Milona (Ch. 10) interprets Audi’s Integration model as a form of Canonical
Evaluative Perception. Others, e.g. Väyrynen (Ch. 5 in this volume), do not.
⁴⁸ For a particular conception, see Chudnoff (2013).
 
However, one might wonder how concept possession could make one reliable in
this way, and whether this renders the relevant knowledge/justified belief non-
empirical. An alternative view is that reliability will depend upon the background
beliefs or value commitments that one has. However, one might worry—as
Väyrynen does in Chapter 5—that this introduces an epistemic dependence that
is antithetical to non-inferential justification. In this context it is worth noting
Lyons’s argument (Chapter 9) against the view that causal cognitive influence
(penetration) on some mental item should lead us to posit an epistemic depend-
ence relation. Instead, what matters is whether the cognitive influence is such that
the mental item can be said to be based upon the influencing cognition. So even
if Canonical Evaluative Perception is brought about by cognitive penetration by
background beliefs, if this doesn’t involve basing then this would be consistent
with Canonical Evaluative Perceptions being sources of immediate justification or
knowledge (so long as they are reliable).
As was noted earlier, Michael Milona suggests that Affective Evaluative Percep-
tions could be sources of immediate justification for evaluative beliefs (indeed, he
appears to suggest that they may be necessary for such justification). However, one
might wonder whether emotions, e.g., could be sources of immediate justification, if
they are always grounded in ‘cognitive bases’,⁴⁹ e.g., beliefs, perceptions, imaginative
episodes concerning non-evaluative objects and events. For example, David’s guilt is
based upon his belief that he lied to his partner. One might think that for his guilt to
justify an evaluative belief, e.g., that he has done something wrong, his belief about
having lied must be justified. But then it looks like his emotion isn’t a source of
immediate justification. Further, this sort of basing/epistemic dependence might
make us doubt that emotions really are Affective Evaluative Perceptions (similar
points could be made about desires). In Chapter 11, Robert Cowan argues that even if
all emotions have cognitive bases, and even if this entails an epistemic dependence of
the emotion upon the cognitive base, this is compatible with emotions playing an
epistemically fundamental role with respect to evaluative propositions.
3. Value Theory and Evaluative Perception

If we define Evaluative Perception in terms of a non-factive representational state,
then it might seem as though there are little or no connections between its existence
and questions in value theory about, e.g., the metaphysics of values.
However, that’s a bit quick (see also Noordhof (Chapter 4) and Stokes (Chapter 1)).
To illustrate, consider a simple Sentimentalist view of value properties (a similar view
could be developed about concepts), according to which X possesses value property,
F, iff X elicits emotional response, E, from all who consider X. One way to understand
this view would be as a reductive analysis of value properties, i.e., the right-hand side
has explanatory priority. However, if emotional responses have evaluative content,
e.g., guilt represents wrongness, that might problematize a reductive version of this
view (since the right-hand side will make reference to value properties, albeit the
⁴⁹ See Deonna and Teroni (2012), 5.

     
representation of such properties). This might push us in the direction of some other
kind of account, e.g., a no-priority view.
Even if one has doubts about the previous line of thought, if we assume that at least
some Evaluative Perceptions are veridical, and that they can play some epistemic
role, e.g., justify evaluative beliefs, then there are arguably clearer connections
between the existence of Evaluative Perception and Value Theory. For example, it
has recently been argued⁵⁰ that a view according to which emotions are Affective
Evaluative Perceptions, which can non-inferentially justify evaluative beliefs,⁵¹ is
incompatible with Neo-Sentimentalism about value concepts—according to which
X possesses value property, F, iff X merits or justifies an emotional response from
those who contemplate X, E, e.g., X is admirable iff X merits admiration. This is
because the combination would appear to have the counter-intuitive consequence
that emotions can confer justification for themselves.⁵²
In Chapter 12 Graham Oddie defends the view that desires are Affective Evaluative
Perceptions, i.e., they are experiences of the goodness of things that can confer prima
facie justification for evaluative beliefs about goodness. However, Oddie argues that if
we think states of affairs are the primary bearers of value, then the view that desires
are value perceptions faces an isomorphism problem. On the one hand, if there is
Affective Evaluative Perception then it might seem that it ought to be reflective of the
evaluative facts. Certainly, it seems that our evaluative beliefs ought to be this way.
But on the other hand, it seems that at least some kinds of value experience—in
particular desire and emotion experiences—are legitimately perspectival, e.g., it
seems legitimate for me to prefer that my mother be saved rather than someone
else’s in a scenario where only one can be saved. After suggesting a somewhat
metaphorical way of addressing this (which appeals, inter alia, to value distance),
Oddie suggests that the isomorphism problem can be dealt with if we adopt the view
that properties/states of being, e.g., being happy, are the primary bearers of value and
objects of desire. This is because Oddie thinks properties, in particular, ‘local’
properties (a property that can be borne by one thing without everything possessing
it), have their own ‘built-in’ perspective.
In Chapter 13, Anna Bergqvist also considers the perspectival nature of evaluative
perception. On Bergqvist’s reading of Iris Murdoch, moral perception involves
not only being attuned to one’s environment thanks to cognitive penetration
through the concepts that we deploy, but also the claim that one’s conceptions of
these concepts decisively influence what we see. Bergqvist argues that we can
nonetheless make good on the robust realist claim that the salient concepts of an
individual’s life-world can be revelatory of value without appeal either to Platonism
or value constitutivism. Bergqvist distinguishes two readings on the concept of ‘non-
perspectival value’—an epistemic reading and a non-epistemic one—and argues that
commitment to the thesis that value is in some sense always value for us does not as
such rule out value’s being non-perspectival in the sense of existing independently
of any actual world views or perspectives in the non-epistemic sense. Bergqvist
⁵⁰ See Brady (2013). ⁵¹ Defended by e.g. Döring (2003); Pelser (2014).

⁵² But see Cowan (2016) for a response.
 
considers the possible objection as to how to account for the notion of structure
and unity of moral thought if we follow through on Murdoch’s suggestion and take
the central target notion of world view to be an unruly holistic admixture of evaluative
and non-evaluative concepts: are there any limits as to what might plausibly be counted
as ‘value for us’? In her estimation, what is needed is a separate argument that speaks
to the practicality of thick moral concepts as action-guiding concepts, and the notion
of action-oriented perception more generally.
By contrast with Bergqvist’s perspective-neutral view about the nature of value, in
Chapter 14 James Lenman explores the topic of Evaluative Perception within the
context of a broadly Expressivist (or Quasi-Realist) metaethical framework. Accord-
ing to this metaethical view, our reasons for action (including moral reasons) emerge
out of a relatively stable network of desire-like attitudes and commitments, arrived at
via some process of reflection and deliberation. Despite a sort of ultimate dependence
of reasons on desire, it can apparently still make sense to speak—from within our
web of commitments—of desire-independent moral reasons and moral truth. With
this view on the table, Lenman rejects the idea that Canonical Moral Perception takes
place: this sort of position would only seem to make sense if we assumed some sort of
naturalist realism. Even then, Lenman thinks that the real epistemic work would be
done by moral theorizing about the relation between natural and moral properties,
not sensory perception. Instead, he affords a limited role to Affective Evaluative
Perception (emotions and desire) in disclosing value to us, where the latter process is
understood more as self-interpretation rather than some sort of perceptual-like
engagement with an external evaluative reality. However, it is clear that Lenman is
more favourable to a picture of ethical thought as reflective and interpretative, rather
than immediate and perceptual.
Finally, in Chapter 15, Kathleen Stock appeals to a kind of perception in order to
illuminate a particular kind of practice with evaluative and normative significance:
objectifying behaviour, i.e., treating people like objects. Specifically she argues that a
mediating role between objectifying images (e.g., those found in pornography) and
objectifying behaviour is played by a distinctive kind of perception: what she calls
‘mind-insensitive seeing-as’. This amounts to a mode of perceiving people on a par
with looking at mindless inanimate objects, and involves a sort of gestalt which can
come in various types, e.g., seeing-as body, seeing-as fungible. Positing this phenom-
enon explains some important kinds of objectifying behaviours, e.g., the attentional
and cognitive habits towards members of objectified groups. Although the sorts of
experiences that Stock appeals to may not themselves have an evaluative content
(strictly speaking), they may be said to be evaluative perceptions in a broader sense,
due to the morally significant way they involve presenting their intentional objects
(persons). Even if it were objected that the phenomena that Stock points to is not a
class of Evaluative Perception, one significant feature is worth noting. The experi-
ences that Stock describes would themselves seem to be the appropriate object of
evaluation, whether or not they represent value properties, i.e., they nevertheless
constitute Morally Evaluable Perceptions.
PART I
The Existence and Nature
of Evaluative Perception
1
Rich Perceptual Content
and Aesthetic Properties
Dustin Stokes
Both common sense and dominant traditions in art criticism and philosophical
aesthetics have it that aesthetic features or properties are perceived. It is common-
place to hear someone in a gallery say something like ‘Look how balanced this piece
is’ or at a symphony performance ‘Can you hear how unified the string section is?’
or at an electronic music show ‘Can you feel the power of the bass?’ These remarks
are posed in such a way so as to direct perceptual attention to some aspect of the
object or performance. The same is true of criticism. Critical language is rife with
examples of aesthetic features being discussed in sensory and phenomenal terms.
And one finds the same in philosophical aesthetics, from Baumgarten to Hume to
Kant and to more contemporary figures like Budd and Levinson and Lopes. Here is
Frank Sibley,
[A]esthetics deals with a kind of perception. People have to see the grace or unity of a work,
hear the plaintiveness or frenzy in the music, notice the gaudiness of colour scheme . . . They
may be struck by these qualities at once, or they may come to perceive them only after repeated
viewings, hearings, readings, and with the help of critics.¹
One’s own phenomenological introspection should suggest the same: if one identifies
the sombreness of a Wyeth painting or of certain movements of a Mahler symphony,
or the frenzy of the Pollock action painting or the Bad Brains punk rock number,
does it not seem obvious that these features characterize one’s sensory experience?
One is most naturally inclined to describe the identified features as part of the
phenomenology of one’s conscious perceptual experiences. Taken alone, none of
these observations are conclusive. But together, the convergence is strongly suggest-
ive: ‘the aesthetic’ is something experienced. Here is a simple thesis that captures the
unified thought concerning aesthetic perception:
AP: Aesthetic properties are sometimes represented by perceptual experience.
¹ Sibley (1965/2001), 34. See Budd (1995); Levinson (1984) and (1994); Lopes (2005).
  
The ‘sometimes’ indicates that aesthetic properties need not be categorically

perceptually represented, while it is plausible that they are very often perceptually
represented. For all that (AP) has going for it, it turns out there is quite a cast of
reasons to be sceptical of the thesis. Here are three.
First, the traditional representationalist model of perception supposes that per-
ceptual experience represents only low-level properties. In the context of art appre-
ciation, one visually perceives the colours and shapes of the painting on the gallery
wall, but aesthetic properties like ‘being graceful’ or ‘being vivid’, like ‘being a
mountain’ or ‘being a dancer’ (as represented by the painting), are post-perceptual,
represented by a judgement or a belief. So this is a general view about perception and
cognitive architecture:
Sc: Perception represents only low-level properties.
Sc implies that AP is false: aesthetic properties are high-level properties, and so are
cognized at the level of judgement or evaluation or belief.
The second reason for scepticism regarding AP is ontological. On almost all
theories of perception, perceptual experience is descriptive: perception functions to
accurately describe or report features of one’s environment. Accordingly, if one
wants to claim that one sees some aesthetic property F, then one thereby commits
to the objective, mind-independent reality of Fness. Barring instances of illusion and
hallucination, if one is seeing F in some object o, then o must really possess feature F.
And there is a long tradition of worrying about objective aesthetic and moral
properties both. So we have:
So: There are no mind-independent aesthetic properties.
The revision is to say that although one may claim that one sees aesthetic property
F in o, the ‘sees’ here should be interpreted cognitively or epistemically: one only
judges that o is F.
The third reason is often coupled with this last one. But it centres not on the
ontology but the normativity of aesthetic attributions. Aesthetic responses may vary
dramatically from subject to subject. Although not the most illuminating character-
ization, responses to artworks are, as we say, highly subjective. Thus,
Sh: Aesthetic responses are purely subjective.
This claim is often couched in hedonistic terms: different subjects take pleasure (or
not) in different things and experiences. Accordingly, an aesthetic response is
something more like a report of how one is struck, pleasurably or not, by an artwork.
This implies that matters of taste are not matters of fact. Now, if perception is largely
descriptive and veridical, then perceiving aesthetic properties would require repre-
senting objective features of the world. But this looks incompatible with the subject-
ive variability of aesthetic response. So, AP must be false and aesthetic assessment a
matter of something non-perceptual.
Taken together, these three propositions provide a substantial hurdle for the
proponent of AP, and from a variety of theoretical angles: mental architecture,
ontology, normativity. They are, however, separable and so AP can largely be
defended against each one independently. A substantial amount of work has been
   
done on the second and third reasons for doubt.² Comparatively less has been written
on the first challenge, Sc. For this reason, and in the context of this particular volume
of research, the following discussion offers an attempted defence of AP against Sc. If
successful, this contributes to vindicating an intuition common to appreciator and
theorist alike, namely, that the aesthetic is a perceptual phenomenon.
1. For and Against the Sceptical Challenge from Sc

1.1. Sparse versus rich perceptual content
From a lay perspective, we do not often distinguish the features that ‘just appear’ to
us from those that require some interpretation or judgement. But cognitive science
has made a tradition of attempting to do just that, and one dominant view is
conservative with respect to those features that simply appear (to all normal per-
ceivers), while a lot more is admitted into the contents of interpretation, judgement,
and belief. So as you look out your window, vision carries information about the
colours and shapes of the afternoon sky. You see blues, yellows, whites, and greys,
and the outlines of the clouds and distant mountain range. But representation that an
object is of a natural or artefactual kind is done at the level of cognition. So you judge
the presence of the mountain range or of the clouds. These properties—being a
mountain, being a cloud—while mentally represented by you, are not represented
by vision.
It is worth taking a brief moment to clarify some motivation for the traditional,
sparse view. Although it takes some time to develop beyond the Jamesian blooming,
buzzing confusion of sensory stimuli, the vast majority of human beings develop
normal perceptual capacities, essential for making sense of (literally) and acting upon
objects and events in the immediate environment. And for each sense modality, there
is a dominant set of norms, a convergence on colour discrimination, tone discrim-
ination, and so on. Add to this the apparent fact that once developed, perceptual
systems work extremely quickly, with no person-level effort (once proprietary input
has been received by the relevant receptors), and in a way that most typically
accurately represents the immediate environment. This has encouraged many the-
orists to think that perceptual systems must be biologically hard-wired, and must
process limited, modality-specific classes of information. Put one way, input from
other non-sensory parts of the cognitive system (what one knows, one’s goals, and so on)
² There is too much literature to list, so here is just one defence of AP for each of the other two sceptical
worries. On the second, ontological, reason for doubt, Frank Sibley (in Sibley and Tanner 1968) argues by
analogy that aesthetic attribution meets (enough of) the same standards as colour property attribution.
And insofar as meeting these standards suffices for some kind of objectivity (not necessarily ‘purely’ mind-
independent objectivity), then aesthetic qualities may be properly understood as perceivable, genuine
properties. Regarding the third, hedonic, reason for doubt, David Hume’s (1757/1874–5) solution to the
problem of taste remains one of the best. Aesthetic experience is, for Hume, grounded in perceptual
experience. And although different subjects respond differently to the same work, there are a number of
ways that perception can be deficient. Objectivity in this context is thus revealed by identification of an
ideal judge who enjoys unhindered perceptual experience of works. So although subjective, one’s aesthetic
responses are more or less appropriate in a way that depends upon how well one’s perceptual experience of
works approximates that of the ideal judge.
  
would undermine the observed cross-perceiver convergence on discrimination,

speed, and objectivity and so, by inference to the best explanation, perceptual systems
must function largely independently of those cognitive systems, processing only
basic, context-neutral features of the environment.
Generalizing, we have the claim that grounds the relevant sceptical challenge
to AP:
Sc: Perception represents only low-level properties.
The incompatibility approaches logical contrariness: If perception represents only
low-level properties, and all theorists agree that aesthetic properties are not low-level
properties (if ‘real’ properties at all), then AP is false: aesthetic properties are never
represented by perceptual experience.³
1.2. Two aesthetic cases and two categories of explanation
Being dynamic and being serene are standard examples of aesthetic properties.⁴ They
also typically oppose one another: it is rare that one finds an object to be both
dynamic and serene. Now consider the following thought experiment, modelled on a
famous example given by Kendall Walton. First, suppose you are in Madrid and you
visit the Museo Nacional Centro de Arte Reina Sofía, which houses Picasso’s Guer-
nica. As most readers would agree, experience of Guernica, in this context, is well
described as dynamic, violent, disturbing, vital. Now imagine that, on a different
occasion, you visit the Gallery of Guernicas. This is the gallery of Walton’s hypo-
thetical society of artmakers who produce works of art in the category of guernicas.
Works in this category are
like versions of Picasso’s Guernica done in various bas-relief dimensions. All of them surfaces
with colors and shapes of Picasso’s Guernica, but the surfaces molded to protrude from the
wall like relief maps of different kinds of terrain. Some guernicas have rolling surfaces, others
are sharp and jagged, still others contain several relatively flat planes at various angles to each
other, and so forth.⁵
Now there is no actual category of guernicas for reference, so the reader must simply
imagine the perceptible contrast between guernica-makers’ guernicas and Picasso’s
Guernica. And then imagine that in the Gallery of Guernicas, after many viewings of
the guernica-makers’ works, you turn the corner and encounter Picasso’s Guernica. It
is very likely that in this rather different viewing situation you would describe your
³ There is no standard definition of ‘low-level’ or ‘high-level’ as pertains to perceived properties, but

instead a standard contrast. Susanna Siegel characterizes high-level properties negatively, where high-level
properties for vision are those ‘other than color, shape, illumination, motion, and their co-instantiation
in objects’ (S. Siegel (2006), 481). The discussion here will just follow this convention. Likewise ‘sparse
content’ will characterize experiences, as theorized, that involve representation of only those low-level
properties; ‘rich content’ will characterize experiences, as theorized, that involve representation of some-
thing more than those basic, or low-level properties.
⁴ ‘Standard’ at least according to one common picture given in analytic aesthetics. See Sibley (1959/
2001). There is of course substantial debate regarding what makes a property (or feature or concept or
term) aesthetic vs. non-aesthetic. See Sibley (1965); Cohen (1973); Kivy (1975). Here it is just assumed that
there is some, perhaps loosely delineated, set of paradigm aesthetic properties.
⁵ Walton (1970), 347.
   
experience of Picasso’s work as cold, or lifeless, or serene, or ‘perhaps bland, dull,

boring—but in any case not violent, dynamic, and vital’.⁶ To select a property each
from Walton’s lists, in the first case one attributes being dynamic to Guernica; in the
second case, one attributes being serene to the same work, Guernica. Take this as our
first case of a contrast in aesthetic reaction.
Consider a second case. Most of us are familiar with impressionist paintings. But
of course each of us had to learn about this particular movement (and learn by
viewing, either in person or through reproductions). Once learned, it is plausible that
one thereby learns to identify, by sight, impressionist paintings as such (with varying
levels of precision and reliability of course). So at one point, perhaps as a child, one
did not have the capacity to attribute ‘impressionist’ to paintings with any reliability
and, at a later point, one acquired this capacity and then readily and accurately
identifies paintings in just this way. A natural description of the latter says that some
paintings just strike one as impressionist. Furthermore, it is natural to describe
impressionism in terms of a gestalt, even if that gestalt cannot be defined in rigorous
terms. Impressionist paintings are typified by a number of perceptible features:
highlighting of natural light and reflection; a regular (but not categorical) use of
lighter colours; identifiable quick, short strokes of paint; an emphasis on a scene
rather than any one figure or group of figures; use of angles and composition creating
a candid rather than posed depiction of people and events. To know impressionism,
is to know and respond to some cluster of these features. And this is an important
difference in the aesthetic reaction of the naïve vs. experienced viewer.⁷
Summarizing, these are both cases of contrast: one where one subject responds to
an artwork (or category of artwork) in one way, and a distinct subject (or the same
subject in a different context) responds to the same artwork in a clearly different way.
This difference in aesthetic reaction entails a difference in the overall mental experi-
ence of the two types of subjects. Plausibly, this difference will involve a difference in
the overall feel or phenomenology of the contextually distinct viewings (in the first
case) and the naïve viewing vs. the informed viewing (in the second case).⁸ And by
hypothesis, this difference depends upon some difference in learning, experience, or
knowledge. The question is how this difference should be further explained and
whether the most plausible explanation favours Sc or AP?
Two categories of explanation are relevant. The first explanation says that the
aesthetic difference is explained post-perceptually: the informed viewer makes dif-
ferent judgements or evaluations, premised on the knowledge she has and the naïve
lack. But this is not, and does not depend upon, a perceptual difference.⁹ Put in its
⁶ Walton (1970), 347.

⁷ One might maintain that ‘being impressionist’ is an artistic property, but not an aesthetic property,
perhaps because it seems to be a property instantiated only by artworks, by contrast to more broadly-
instantiated properties like ‘being dynamic’ or ‘being balanced’. Granting this distinction makes no
difference to the argument that follows.
⁸ This is not meant to be question-begging in the current dialectical context: the overall phenomenology
of one’s mental experience may include more than just sensory phenomenology. See S. Siegel (2006).
⁹ A great deal more could be said about candidate cognitive states or processes here. For example, is
there a difference between a judgement and an evaluation? What is the role of (occurrent) belief? These
questions should be analysed, but not here. All that matters here is that there are a number of cognitive
  
strongest form, the perceptual experiences of the distinct subjects are the same
(in phenomenal character and representational content). If pressed, this theorist
could concede a sensory phenomenal difference, but this difference would be
explained in terms of sparse content only.
The motivation is this: if aesthetic properties like ‘being dynamic’ and ‘being
impressionist’ are high-level properties, then they are not represented by experience.
Again, perception is fast, hard-wired, and objective across perceivers and cultures, and
so plausibly only represents basic properties of the environment. An explanation of this
kind is further motivated by acknowledgement of the (hypothesized) fact that the naïve
versus informed subjects differ in what they have learned about artworks and aesthetic
features. So, the reasoning would go, the resultant (aesthetic) differences will be at the
level of post-perceptual cognition. One learns some typifying features of impressionism
and is then able to judge and report which paintings are impressionist. And note finally
how these two points work in tandem: if one is already committed to the claim that
perception represents only low-level properties, and the difference (between naïve and
informed subjects) involves high-level properties the identification of which depends
on background learning, then it must follow that those properties will (post-learning)
be represented in later cognitive (non-perceptual) processes.
Now for the second category of explanation. Consider once more our two
examples. In each case, there is a difference in the aesthetic reaction of the two
subjects that depends upon some background learning or cognition. And plausibly,
the overall experience between naïve and informed subject will differ in broad
phenomenology. How can the proponent of AP explain these differences?
The general type of explanation says that the phenomenal contrast is best
explained as a difference in rich perceptual representation. There is a line of argu-
ment now standard in the literature on admissible contents that appeals to phenom-
enology. Susanna Siegel argues that the overall phenomenal difference before and
after one learns to recognize pine trees is a perceptual difference. For the pine-tree
spotter, vision represents pine trees; and for the pine-tree naïve, it does not.¹⁰ This
position has been buttressed with some empirical support. For example, Tim Bayne
argues that the difference between the visual associative agnosic and a normal human
perceiver is that while the former possesses intact low-level perceptual capacities, she
lacks intact high-level perceptual capacities.¹¹ So the agnosic sees the telephone (its
spatial and colour properties) but does not visually recognize the telephone. Bayne
thus draws an inference on the basis of what’s missing for the agnosic: normal human
perception represents high-level properties like ‘being a telephone’. One can see how
this line of argument provides a template for thinking about rich perceptual repre-
sentation of aesthetic properties.¹²
states typically theorized and, their possible differences notwithstanding, they are all supposed to be non-
sensory, post-perceptual states. Accordingly, this is a category of explanation unified by its distinction from
the second category involving perceptual explanation.
¹⁰ See S. Siegel (2006) and (2010a).
¹¹ See Bayne (2009). One can find similar arguments, or at least suggestions, in Van Gulick (1994) and
Siewert (1998).
¹² For a discussion of this sort, see D. Stokes (2014).
   
In the discussion that follows, a new argument for rich perceptual representation
is offered. This argument employs Siegel’s contrast method, as well as some of her
original analysis, but focuses on how to best understand mental occurrences
described as ‘seeing-as’, with an emphasis on the phenomenology of visual percep-
tion of ambiguous figures. It takes inspiration from related discussions in Wittgen-
stein and N. R. Hanson.¹³ The results of this argument are then applied to the cases of
aesthetic contrast, where the aesthetic cases are specially illuminating. This general
explanation can then be extended by appeal to three different kinds of mental
mechanism. This abundance of explanatory options for the proponent of AP shifts
the burden of proof to the critic of AP.
2. Ambiguous Figures and the Argument

from Seeing-As
2.1. The argument
Figure 1.1 includes two famous examples of ambiguous figures: the duck/rabbit and
the Rubin goblet.¹⁴ Consider the duck/rabbit. First see it as a duck. Then see it as a
rabbit. The term ‘seeing-as’ comes naturally here.¹⁵ The same goes for the Rubin
goblet. First see it as a goblet. Then see it as a pair of faces directed at one another.
Here again, you see as a goblet or see as a pair of faces (and cannot do both
simultaneously). Plausibly, the overall experiences of seeing the image as a duck
versus seeing the image as a rabbit differ in phenomenology: what it’s like to see
something as a duck is broadly different from what it’s like to see it as a rabbit. The
most natural explanation of these observations is that the switch from seeing-as a
duck to seeing-as a rabbit involves a change in visual representation. The difference is
Figure 1.1
¹³ See Wittgenstein (1953/2009); Hanson (1958) and (1969).

¹⁴ The reader should note the particular duck/rabbit image used. The choice of this image over the also
common, mere two-dimensional outline shape image is deliberate, since the latter may not work for the
line of reasoning that follows, and for reasons that hopefully become clear.
¹⁵ Indeed, for whatever it is worth, it is incredibly difficult to write these instructions in some non-
’seeing’ or ‘perceiving’-involving way. ‘Judge it to be a duck’, ‘Interpret it as a duck’, ‘Cognize it as a duck’,
‘Believe it is a duck’ . . . are all forced at best.
  
a genuine perceptual difference. Same goes for the Rubin goblet and many other
ambiguous figures.
How might a proponent of Sc explain these switches while denying any rich
perceptual difference? Here the low-level theorist will encounter a challenge in the
form of a trilemma. First, a low-level theorist could claim that there is no phenom-
enal difference between seeing-as a duck and seeing-as a rabbit. Second, one could
grant that difference but claim that it is an entirely cognitive phenomenal difference.
Third, one could claim that there is a difference in judgement accompanied by a
change in low-level perception only. Each of these explanations is either less plausible
than the opposing high-level perceptual explanation, or collapses into that very
explanation.
The first horn has the low-level theorist rejecting the claim that there is a difference
in the phenomenology of seeing-as a duck versus seeing-as a rabbit. There is no
difference in what it’s like to have these two experiences. This position is implausible.
A theory that denies that it feels different to see an image as a duck versus see it as a
rabbit is almost entirely lacking in intuitive force, looking instead like an unfortunate
theoretical consequence. Intuition and introspection are fallible, but both strongly
favour phenomenal differences between the two experiences.¹⁶
Second horn: a low-level theorist may grant the difference in phenomenology, but
explain it as deriving from a difference in the non-perceptual cognizing of the two
images. This claim is also problematic. First, cognitive phenomenology (by contrast
to sensory phenomenology) is controversial. Many have argued that there is nothing
it is like to have a belief or make a judgement.¹⁷ More substantively and less theory-
committal, examples where judgement and phenomenology come apart are easy to
come by. First, there are cases where a change in judgement causes no change in
phenomenology; visual illusions illustrate this point. Most are familiar with the
Muller-Lyer illusion (Figure 1.2), but of course each of us learned at some point
that it is an illusion. It is commonly accepted that as one’s beliefs (or judgements)
about the lengths of the lines change (one learns that the two lines are in fact of the
same length), one’s phenomenology stays the same. The illusion persists and there is
no obvious difference in what it is like across the pairs of experience.¹⁸ Second, there
are many cases where phenomenology changes while judgement remains the same.
Consider colour constancy. Imagine a pink book on your desk, illuminated by the
setting sun, moving from bright sunlight at t₁ to the last sunlight of the day at t₂.
Across this entire window of time, your belief that the book is pink is stable. And, in
one sense, you still perceive the book as pink (per constancy mechanisms) and would
¹⁶ This does not seem to be the sort of claim for which further argument is appropriate, since it flatly
appeals to intuition. That said, and for what it’s worth, a number of disparate theorists share the intuition,
as discussed in Macpherson (2006). See Goldstone (1988); Millar (1991); Tye (1995).
¹⁷ One classic critic of cognitive phenomenology is Dennett (1988). See also Kim (2005). For arguments
that cognitive states enjoy distinctive phenomenology, see Siewert (1998); Flanagan (1991); Horgan and
Tienson (2002); Loar (2003). For a recent volume of papers on this topic, see Bayne and Montague (2011b).
¹⁸ Macpherson (2006) makes this point, partly in service of a challenge to non-conceptual intentionalist
theories of perception which, she argues, struggle to explain experiences of ambiguous figures, once it is
granted that the gestalt shifts involve shifts in phenomenal character. The discussion of ambiguous figures
in the present paper is indebted to Macpherson’s analysis.
   
Figure 1.2
judge it accordingly. However, it is undeniable that there is a difference in what it’s

like to see the book at t₁ versus what it’s like to see the book at t₂.¹⁹ So, phenomen-
ology comes apart from cognition (judgement) in both directions. The theorist under
consideration would have to claim, implausibly, that all cases of seeing-as are
somehow different, where changes in relevant judgement (that is a duck) always
come with changes in non-sensory phenomenology.
The intermediate conclusion to this point is that seeing-as is most plausibly
understood as a sensory perceptual phenomenon. For some object o, seeing o as an
F is a distinct perceptual experience from seeing o as a G.²⁰ The final step is to extend
the argument to defend a rich perceptual content thesis. This third horn of the
trilemma requires more lengthy discussion.
First consider the ambiguous figures just discussed. It is eminently plausible that
(assuming a non-naïve viewer), when one sees the duck/rabbit figure, say, as a duck,
some of a cluster of properties—‘being a duck’ or ‘being a duck appearance’ or ‘being
an image of a duck’ or ‘being like a duck’—are thereby mentally represented as
instantiated.²¹ And note that each of the properties in this disjunction is a clear
example of a high-level property. If this is correct, and the first two horns of the
trilemma are successfully challenging, then there remains one way for the low-level
theorist to maintain her position. What this theorist must say is that the phenomenal
difference (from switch to switch) is at the level of low-level property representation,
¹⁹ Some will note that constancy phenomena may, for reasons intimated here, be taken as a challenge to
intentionalist theories of perception, since it appears that we have an instance where representational
content is the same across phenomenal changes. An early example of this challenge can be found in
Peacocke (1983). More recently, see B. Millar (2013). For a representationalist defence, see Dretske (2003).
And for alleged explanations of constancy phenomena in terms of Fregean content, see Chalmers (2006);
Thompson (2009). For general discussion, see S. Siegel (2010b).
²⁰ This should be qualified with ‘sometimes’ or ‘for many Fs and many Gs’, since this is not intended as,
nor need it be, a categorical claim about all instances of seeing-as. For example, in cases of indiscernibles—
say a pair of identical twins—there may be no perceptual, phenomenal difference between seeing the same
object as being one or the other of the indiscernible pair (for instance, seeing an individual as being one or
the other twin). The argument here only needs to be plausible for many, but not all, uses of ‘seeing-as’. And
furthermore, there may be deviation in, or different uses of, ‘seeing-as’.
²¹ This qualification is made for a pair of reasons. First, there are extra complications introduced since,
strictly speaking, the duck/rabbit figure is a mere depiction and so one does not take there to be (in any
sense of mental representation) a real, living, breathing duck in one’s visual field (when viewing the duck/
rabbit figure). Second, there is an open question, not sufficiently discussed in the literature, about whether
the rich content theorist needs to claim that kind properties (e.g. the natural kind property ‘being a duck’)
are themselves part of the content of perceptual experience. It is plausible, as discussed momentarily, that a
much weaker claim is sufficient for rich perceptual content.
  
while ‘being a duck’ and the other relevant high-level properties are cognitive
contents (e.g. one interprets the image as a duck, or as a duck representation, or as
being a duck appearance, and so on). The question is what the former difference in
perceptual content amounts to. It is both introspectively and psychologically
implausible that each time the switch occurs, perception represents distinct colours
or distinct edges (with the same duck/rabbit figure in view throughout). Moreover, as
again introspection should reveal, it isn’t just (or even) the outline shape that
changes. Instead, such a change would have to be in something like an overall gestalt,
some organization of low-level properties. A familiar way to put this is that the duck
(or the rabbit) ‘pops out’. So, for the duck/rabbit image, when one sees it as a duck,
call the candidate, pop-out feature of content ‘being organized like a duck’. Is it
correct to insist that this kind of perceptual representation of an organization of low-
level features involves only low-level content?
This is a fine line to walk. Note that the switches in question are reliable, in the
sense that (with these and most ambiguous figures) one switches between only two
things: now it’s a duck, now it’s a rabbit, now it’s a duck, etc., and one will tend to
have the organized-like-a-duck experience only when appropriate stimuli are avail-
able (one will not have that kind of experience when viewing, say, the Rubin goblet or
a sunset). Furthermore, the phenomenology is, all else being equal, relatively stable
across these switches: it isn’t as if the figure strikes the perceiver, across a range of
switches, with a wide variety of duck-appearances.²² All of this is supposed to be
explained by the low-level theorist under consideration as a switch in interpretation
(between ‘that’s a duck’ and ‘that’s a rabbit’) and an accordant shift between visual
experience representing the world as containing the relevant organizational gestalt
(as containing an overall organization of edge, colour, and shape typical of a duck,
then of a rabbit, etc.). These considerations in hand, what reason is left to deny that
this content is rich content of visual experience?
One way to put things is in terms of discriminatory capacities. The low-level
theorist in question says that one’s visual experience characterizes some cluster of
low-level properties (colours, edge, shapes) as being ‘organized like a duck’, where
this is the (or at least a) appearance—a coherent organization of looks—typical of
a duck. And she must grant further that tokens of this visual experience type
reliably co-vary with ducks or duck images or duck appearing objects. This implies
a visual capacity for discriminating a kind of thing: things whose basic features are
organized so as to appear like a duck. And this is just another way to say that vision
²² In this way, at least when applied to ambiguous figures, this proposal is distinct from the one
criticized in S. Siegel (2006), in her defence of perceptual representation of kinds. One rejoinder from
the low-level theorist she considers is that, in the case of pine-tree spotting, one perceptually represents a
general pine-tree gestalt that is invariant across differences between individual pine trees. Siegel argues,
plausibly, that this is phenomenologically implausible, considering at least a series of viewing distinct pine
trees. But the suggestion here is that two distinct overall organizational gestalts, and fairly stable ones, are
likely to correspond, phenomenologically, to the switch between seeing-as a duck and seeing-as a rabbit.
And in this case, unlike the pine-tree case, there is not a variety of distinct stimuli (since one is just viewing
the same duck/rabbit figure). Generalizing, the suggestion here is not that, when seeing-as, we token some
stored, individual-invariant gestalt representation. This is what Siegel seems to have in mind, but surely
this is not the only way to think about perceptual organization vis-à-vis ‘gestalts’.
   
discriminates—that is, represents—more than just colours, shapes, edges, motion,

and location. It represents objects in the world as organized, where those organiza-
tional gestalts correspond in some systematic way to the kinds of objects that
populate one’s environment. This perceptual discrimination manifests in instances
where we recognize an object (or image) as being of this kind or that kind, or at least,
recognize an object as appearing as being of this kind or that kind. We see the image
as a duck and then as a rabbit.²³
Now recall the two aesthetic cases. In the first case, a perceiver, in two importantly
distinct contexts of appreciation, ascribes first ‘dynamic’ and then ‘serene’ to Picas-
so’s Guernica. The second case contrasts a subject with the capacity for attributing
‘impressionist’ to works with a subject who lacks this capacity. In both cases, there is
a contrast in aesthetic reaction. Here again, it is natural to describe the contrast as a
perceptual one: one sees the Guernica as dynamic or sees it as serene. One sees a work
as impressionist or not. An opponent of AP who wants to maintain that this is either
a purely cognitive difference or a difference only in low-level perception encounters
the same trilemma.
The first and second horns of the trilemma are applied here as they were to
ambiguous figures. First, a claim that there is no phenomenological difference
between aesthetic reactions is implausible. What it is like to react to an artwork as
being dynamic is phenomenally distinct from what it is like to react to the work as
serene. Second, the claim that all phenomenal differences are purely cognitive is
implausible, saddling the objector with rich cognitive phenomenology, and contrary
to common instances where judgement (or other cognitive states) and phenomenally
characterized experience come apart. These two horns are sharpened in the aesthetics
cases. Both ordinary opinion and ordinary language suggest that aesthetic reactions
to artworks are sensory in character. Anyone unversed in sophisticated philosophical
debate about these topics would default to a sensory-phenomenal characterization of
recognizing the dynamism or impressionism in a work. Same goes for both artists
and art critics. Visual artists, for example, create works with intentions of establishing
certain aesthetic ‘looks’ or ‘appearances’, and assume that their uptake is visual. And
critical discourse is full of aesthetic description couched in sensory terms.²⁴
The final horn challenges the theorist who grants a phenomenal change, but
maintains that it is a feature only of low-level perception. So, the difference between
the impressionist spotter and the impressionist naïve is that the first, and not the
second, judges of a work ‘that is impressionist’, and this is accompanied by some
difference in low-level perceptual representation. At least in this type of case, it is
implausible that there is a difference in basic colour or shape or edge perception.
More plausibly, there is some difference in the way these basic features are percep-
tually organized. The impressionist spotter enjoys perceptual experience as of an
²³ Bear in mind that the point is not one about linguistic representation. Figuring in the content of
perception does not require some tokening of a linguistic symbol. This is equally true for the property
‘being red’ as it is for the property ‘being a duck’; it is not a condition on perceptual content that ‘red’ or
‘duck’ is mentally represented.
²⁴ See also D. Stokes (2014) for discussion of these matters, and in relation to the cognitive penetrability
of perception.
  
impressionist gestalt. And here, as with ambiguous figures, the most plausible way
to understand this option is in a way that concedes rich perceptual representation.
And with the aesthetic case, once again, things can be sharpened. Unlike natural
and artefactual kind properties, which on many accounts have a non-perceptible
essence or kind-determining underlying structure, many aesthetic properties are
exhausted by appearance features.²⁵ There is nothing to the property of ‘being
dynamic’ or ‘being impressionist’ (at least when in the context of aesthetics and
artworks) beyond an organization or gestalt of basic features.²⁶ Accordingly, if a work
is recognized as ‘being impressionist’, and perceptually recognized as an overall
organization, then ‘being impressionist’ is perceptually represented. Therefore, in
the aesthetic case, if the low-level theorist (or other sceptic of AP) grants a sensory
difference in terms of the perceptual organization of features, this is equivalent to
granting that aesthetic properties are perceived. And to grant this is to grant that
some high-level properties figure in the content of experience.
A clarification of the argument at this stage will be useful. The argument is not,
simply, that because (many) aesthetic properties are appearance properties, they
must be perceptually represented (if mentally represented at all). Instead, considering
pairs of contrasting aesthetic reactions, attempts to explain the contrast in strictly
non-phenomenal ways (Horn 1) or strictly cognitive ways (Horn 2) are implausible.
The opponent of AP is then left with a hybrid explanation, where a contrast in
judgement is accompanied by distinctive sensory phenomenology (Horn 3). The
latter perceptual differences are not sufficiently explained in terms of only colour or
shape or edge perception, but instead require some appeal to an overall organization
of those basic features. But, to perceptually experience that kind of organization—an
impressionist gestalt—just is to perceptually experience the relevant aesthetic prop-
erty. Therefore, aesthetic reaction sometimes involves perceptual representation of
high-level properties. It involves rich perceptual content.²⁷
2.2. A rejoinder and some general lessons

The low-level theorist might reply to the above argument by taking on Horn 3, as
follows. Grant that these instances of seeing-as involve some rich content, while
maintaining that genuine kind properties are not themselves perceptual contents.
One way to articulate this rejoinder is in familiar terms of accuracy conditions. If
perceptual experience is representational then, in a rather intuitive sense, it says to
²⁵ The ‘many’ qualification here is deliberate. As already noted, just what is and is not an aesthetic
property (or quality, term, or concept) is controversial. And one point of debate is whether aesthetic
properties are, or are exhausted by, appearance properties. For example, it is plausible that literary works
have aesthetic qualities, but at least some of those qualities have nothing to do with the visual or auditory
appearance of the words as they appear on the page or as they are spoken. This general debate can be
reasonably avoided, since the case being made here for rich perceptual content of aesthetic properties only
requires that some aesthetic properties are exhausted by appearance. The discussion will assume this
qualification, even if it is not always made explicit with ‘some’ or ‘many’.
²⁶ Note that the question about the relation between aesthetic properties and basic properties is
important, but orthogonal to the present line of reasoning. Note also that this is not to suggest that
there is no objectivity to aesthetic properties or that there are no correctness conditions for aesthetic
responses and attributions.
²⁷ This discussion owes much to Siewert’s (1998) discussion of perceptual recognition and organization.
   
the perceiver how the world is here and now.²⁸ Representations can be more or less
accurate with respect to how they represent; perceptual experiences can be more or
less accurate in what they say about how the world is here and now. Accordingly, for
any experience token e, one can identify the way(s) w the world would need to be
(here and now) for e to be accurate (setting to one side difficulties about degree of
accuracy). If the world is way w, then e is accurate (and if not, then not). To specify
these accuracy conditions is to specify what any experience e is about; it is to specify
the content of e. The debate about admissible contents concerns the types of
properties that are needed in order to specify accuracy conditions and, thereby, to
adequately capture the content of perceptual experience. Here two questions should
be distinguished: (1) Are basic properties sufficient (so, for vision: colour, edges,
shape, motion, location)? (2) Are natural or artefactual kind properties needed? It is
often assumed that if one answers ‘no’ to question (1) then one answers ‘yes’ (perhaps
must answer ‘yes’) to question (2). Working through this final rejoinder reveals that
this may be an unnecessarily coarse way to think about the richness of perceptual
representation.
Consider once more the ambiguous figures. The option being considered (Horn 3)
is that seeing-as a duck in the presence of the duck/rabbit image involves a judgement
that ‘that’s a duck’ plus a perceptual experience that is best characterized as ‘being
organized like a duck’. The low-level theorist might attempt to maintain that the
content of this perceptual experience is low-level by reasoning as follows. Plausibly,
one will have this experience in a variety of circumstances, only some of which
involve the instantiation of the property ‘being a duck’. Thus, one will token this
(broadly individuated) experience type in the presence of living, breathing ducks, or
in the presence of plastic duck decoys, or certain landscape paintings or, as in the case
in question, in the presence of the duck/rabbit image in Figure 1.1. Furthermore, it
seems plausible that the experience is accurate if one is in the presence of any of these
stimuli. And notice that the natural kind property ‘being a duck’ is instantiated only
in the first circumstance listed here. So it cannot be the case that the kind property
‘being a duck’ is part of the content of the relevant experience, since it only figures
(and need only figure) as one factor in a disjunctive set of accuracy conditions. In
other words, there is an array of circumstances, any one of which would render the
experience accurate but only one of which involves instantiation of the relevant kind
property. By contrast, the various low-level properties are part of the content of the
experience. So, if part of the character of the experience involves a vertical line or
edge (say part of the duck-outline), then this experience (or sub-experience, if one
prefers) is accurate only if there is a vertical line or edge in one’s presence. The basic,
low-level property ‘being a line (with such-and-such orientation)’ must be instanti-
ated here and now. Accordingly, this low-level property is part of the content of the
experience. Notice further how this will be part of the content no matter which
of the listed circumstances obtain (real duck, duck decoy, duck/rabbit image, etc.).
So the low-level theorist argues by disanalogy, and concludes that while the relevant
²⁸ S. Siegel (2010) argues that a rich content thesis could be suitably adjusted even if a content thesis
fails, for example, if one favours some direct realist theory of perception.
  
experience involves the ‘being organized like a duck’ character, that character is
sufficiently captured by appeal to basic properties only. In particular, the property
‘being a duck’ is unnecessary in specifying the representational content of the
experience. And the disanalogy generalizes, low-level properties are necessary to
specify perceptual content (as accuracy conditions) while kind properties are not
necessary to specify content.
There is an important insight in this rejoinder, but it will not service the low-level
theorist’s defence. The argument was that kind properties are unnecessary for
specification of content (as accuracy conditions), and by an alleged disanalogy with
perceptual representation of basic features like colours and edges. The theorist is
right about kind properties, but wrong about the disanalogy: the same is plausibly
true for low-level properties as well. Consider colour. Suppose one has a properly
functioning visual system, and has an experience as of a red globe. One might employ
the same technique as above and claim that this experience is accurate just in case
there is some object, here and now, that instantiates the property ‘being red’. There
must really be, as we sometimes say, some red globe in one’s visual field. So, the
property REDNESS is part of the content of one’s experience. But brief reflection
reveals that this is too strong. First, notice how it depends upon some form of colour
realism being true. If all forms of colour realism should turn out false, then colour
perception is systematically inaccurate: there isn’t really a red thing in one’s visual
field, ever. Perhaps this is unsurprising to some philosophers of colour (perhaps it is
even the makings of a transcendental argument that some realists might employ). But
it suggests that something has gone wrong with the theory if accuracy of basic forms
of perception is beholden to accuracy of ontological theorizing. Second, ontological
theories to one side, this specification of content still renders much of experience
inaccurate. Suppose in having the experience as of a red globe, with the content
specified as per above, one in fact is not in the presence of a red globe. Instead, one is
in the presence of a white globe, perfectly illuminated by hidden red lighting so as to
have the appearance as of a red globe. In this circumstance, the object of perception
(the globe) does not instantiate the property ‘being red’ but only appears red (to any
normal perceiver). Indeed, as described, no single object here and now instantiates
‘being red’. But as specified, the experience represents the object (the globe) as
instantiating the property ‘being red’. Accordingly, the experience is inaccurate.
This is an odd result. Situations like these are entirely ordinary and one has, by
hypothesis, a normally functioning visual system. Extended, this would render a great
deal of perfectly normal, appearance-tracking perception, inaccurate.
The conclusion should be that perceptual representation of high-level or kind
properties and perceptual representation of low-level properties are analogous
(instead of disanalogous) with respect to how properties figure into content. In
both cases, there is a risk concerning the fineness of grain in identifying accuracy
conditions (and thus content). Whether an experience has a reddish character or a
duck-ish character, if for accuracy of the experience it is required that the properties
(‘being red’ or ‘being a duck’) be, strictly speaking, instantiated here and now, then
inaccuracy among normal perceivers is rampant. In other words, by parity of
reasoning, if the low-level theorist maintains that the property ‘being a duck’ is
unnecessary to adequately capture the content of the experience type with the
   
‘being organized like a duck’ character (since it would render inaccurate many
relevant experiences, including the experience of the duck/rabbit), then one should
maintain the same for standard low-level experiences, including colour experiences.
Comparing the results of this line of reasoning to the case of aesthetic properties
suggests some more surprising conclusions.
Suppose for the moment that some objectivism about all three ‘types’ of property is
true—basic properties like colour, kind properties like ‘being a duck’, aesthetic
properties like ‘being dynamic’ are all real in at least the minimal sense that ascription
of each type is truth-evaluable. Consideration of aesthetic properties is instructive in
this context. Supposing an expert perceiver (perhaps a visual artist), perception
provides full access to aesthetic properties. Aesthetic properties (or again, at least
many of them) are something approaching pure appearance properties.²⁹ And so a
suitably trained perceiver can ascribe and identify aesthetic properties with high
frequency just on the basis of her experiences. This is because there is no hidden
nature or essence to a property like ‘being graceful’ or ‘being impressionist’, and this
contrasts with both kind properties and basic colour properties. Glossing over a vast
amount of theoretical detail and controversy, what makes something an instance of
the natural kind DUCK is some underlying structure, not detectable by perception
alone. And so there are many circumstances where uncertainty or mistakes about
duck-identification will not be correctable by perception alone.³⁰ Put in terms of
expertise, there is no relevant expert perceiver (say an ornithologist) who, knowing
all there is to know about ducks, will not have the organized-like-a-duck experience
in the presence of, say, some hereto unencountered duck-looking non-duck creature
from a faraway land. Correcting misascription in this case requires more than what
even expert perception can provide. And note, crucially, that this correction will
correct judgement but will not correct perception: the expert still has an experience
(of the non-duck) characterized as ‘being organized like a duck’. A similar story
would be told for artefactual kinds, where here the essence might involve some
conventional function or role. And finally, the same is true for colour properties.
Assuming again some expert perceiver (perhaps an expert colourist), there are many
²⁹ Note that this is compatible with variation in capacity for recognizing these appearances. So, we can
train ourselves, and sometimes may need to. We can better recognize appearances by repeated experience
or learning of some kind. As discussed in Section 3, this could be achieved by a variety of cognitive
mechanisms.
³⁰ For simplicity, this discussion (like all of the discussion of perception of kind properties) proceeds on
the assumption that some kind of realism about kinds is true and, moreover, that we are liberal in our kind
taxonomy. Thus, ducks and pine trees are natural kinds. However, the discussion could be adjusted in
various ways without undermining the argument. For example, if one is a conventionalist about kinds, then
the conventions (rather than some underlying microstructure) are what cannot be perceived. Or, perhaps
one does not think that things like ducks or pines are kinds, but instead are species (understood as
individuals or classes or sets). Here again, there will be criteria for being a member of a species that exhaust
the perceptible, macrostructural features of any perceived biological thing, for example, the capacity to
successfully interbreed or the sharing of a common ancestry. Indeed, this was a point anticipated even by
Darwin in The Origin of Species, ‘adaptive characters, although of the utmost importance to the welfare of
the being, are almost valueless to the systematist. For animals belonging to two most distinct lines of
descent, may readily become adapted to similar conditions, and thus assume a close external resemblance;
but such resemblances will not reveal—will rather tend to conceal their blood-relationship to their proper
lines of descent.’
  
perceptual circumstances where vision will present the world as containing an object
with some coloured appearance, but where that object in fact lacks whatever under-
lying (non-perceptible) features are required for an object to really be that colour.
One can, at the cognitive level, correct any such misidentification by further inves-
tigation (perhaps manipulating the environment by checking the light, or asking
about the light, etc.), but this does nothing to correct, in any relevant sense, the
experience. The white globe in red light will still appear red to the expert, even after
she knows it is not really red.
The important difference between aesthetic properties vs. kind and colour prop-
erties is that sensory perception can (often) provide full access to the first but to
neither of the second. Again, put somewhat crudely, aesthetic properties are some-
thing approaching pure appearance properties (exhausted by organizational gestalts);
kind and colour properties are not (for reasons given just above). One might then
reason that it follows that perception of the second will just more often be in error,
while perception of the first is rarely in error. But this locates the difference in the
wrong place, and saddles both low-level and putative high-level perception with too
much inaccuracy. Instead, one should locate the difference in the world: in the way
that these various properties are, or at least are theorized by our best methods of
taxonomy. Accordingly, there will be differences in cognition of or judgement about
or empirical investigation of these broad types of properties (i.e. how they are
ascribed and identified), but there need be no general difference in how accurately
they are perceptually experienced.
Here then is a radical lesson, cautiously drawn from the attempted disanalogy in
the rejoinder above. Perhaps properties the natures of which are not exhausted by
appearance should never figure in the contents of experience. This is clearest with
kind properties. Natural kinds have essential structures hidden to perception. Sen-
sory perception alone cannot provide access to kind properties. And so kinds are not
contents of perception, because they do not figure in the accuracy conditions for
experience. But what is true for kind representation is true for colour perception: if
instantiation of the property ‘being red’ is a condition for the accuracy of perception,
then many normal visual perceivers in ordinary circumstances, will suffer inaccurate
experiences. Here again there are aspects to colour properties that go beyond
appearances, that go beyond what perception is able to access. We are left with the
awkward result that because my experience of the red-illuminated white globe says
‘that is red’, it is impugned as inaccurate. This holds a well-functioning visual system
accountable to something that it cannot access because of the nature (or theorized
nature) of the relevant properties.³¹
Instead, we should include in the content of perceptual experience only those
properties or aspects of properties that are accessible to perception. Both kind and
colour properties have characteristic appearances. That is, for most kinds and most
³¹ As an anonymous referee notes, this line of reasoning would seem to commit to a claim (of some
strength) that phenomenal content (or character) determines perceptual content, and claims of this kind
are controversial. This is another way in which the radical lesson drawn is a cautious one. To fully support
the lesson (which is simply not the central aim or intention of this paper), then, one would need to provide
a thorough analysis of the relation between sensory phenomenology and perceptual content.
   
basic properties, there is a typical feature or aspect (or cluster of aspects) that, by
normal subjects, is detectable by perception alone. The aspect ‘being organized like a
duck’ is such an example. These aspects of the relevant properties are more appro-
priate candidates for contents qua accuracy conditions. Note what follows. In one
respect, perception of aesthetic properties, kind properties, and colour properties are
on a par, since in all cases those experiences are about the typical appearance aspects
of said properties. In another related respect, only perception of an aesthetic property
is perception of the property, since there is nothing to an aesthetic property beyond
some complex organizational gestalt. But one never, strictly speaking, perceptually
represents kind or colour properties.
All of this is to think of the accuracy of perception as appearance-aptness but not as
truth-aptness. The latter is often assumed, but this may smuggle in considerations
better reserved for doxastic cognitive states. To count as knowledge, beliefs must be
true and justified. And so, because truth is the aim of belief (getting all of the facts we
might say), it is appropriate that accuracy of belief qua representation is understood
accordingly. A belief of the white globe illuminated in red light that it is red fails at its
aim, and is accordingly inaccurate. What is different is that belief can, in this
circumstance, still achieve its aim. Investigation may be required, but belief can be
corrected according to its aim. So when it fails, it is not impugned for something it
cannot do. This contrasts with (visual) perception. Vision cannot do any better than
it does in this circumstance: even upon receipt of information about the lighting in
the room, one will still experience the white globe as red. And so it shouldn’t, in this
case, be impugned as inaccurate. Including the property ‘being red’ in the accuracy
conditions, and therefore content, of the perceptual experience delivers this very
verdict of inaccuracy. So, properties of this kind, as well as kind properties, by
contrast with aesthetic properties, should not be included in the specification of
perceptual contents.³²
2.3. Concluding the argument: rich perceptual content

of aesthetic properties
There are less revisionary (or at least different) ways to conclude from this analysis. So
one might grant that kind properties are not (or are rarely) contents of experience,
while colours and more basic properties are. One possible motivation here is the
defensible view that colours just are appearance properties, by contrast to most kind
properties.³³ Or, one might maintain that inclusion of a property in the accuracy
conditions of experience does not entail that the properties actually be instantiated in
the perceiver’s environment. It is only required that some member of a cluster of
relevant properties (some of them perhaps gerrymandered) is instantiated, where the
cluster involves the range of stimuli that typically cause the relevant broadly typed
³² Siewert (2012) suggests worries about inclusion of kind properties as such in perceptual contents.
Another worry looming in the vicinity here, in response to the hypothetical rejoinder, is to conclude that
there is a problem not with inclusion of properties in contents but, more fundamentally, with specifying
perceptual contents by specifying accuracy conditions.
³³ At least with respect to perceptual representation of basic properties like colour, a related proposal is
found in Shoemaker (1994).
  
experience in normal perceivers. Thus the experience with the character ‘being
organized like a duck’ is accurate just in case the perceiver is in the presence of a
living, breathing duck or a duck decoy or a duck depiction or . . . This kind of proposal
might be further defended if coupled with a liberal, but surely revisionary, theory of
kind properties, where any instance of one of the disjuncts from the list just given is an
instantiation of ‘being a duck’. The important point for the proponent of AP is that, no
matter one’s choice about these more general matters, the work has already been done
to defend the claim that aesthetic properties are admissible contents of experience.
The argument from seeing-as, briefly applied to the aesthetic contrast cases once
more, ran as follows. Explanations of the contrast in strictly non-phenomenal ways
or in strictly cognitive-phenomenal ways are both (comparatively) implausible. The
best move for the critic of AP then is to grant the phenomenal difference, but explain
it in terms of differing judgements accompanied by a sensory-phenomenal differ-
ence. The sensory-phenomenal difference is not plausibly one in perceived colours or
edges or other basic features; it is most plausibly a difference in the overall perceptual
organization of those basic features. The impressionist spotter enjoys an experience
of an organizational gestalt typical of impressionist works. And finally, to come this
far is to grant rich perceptual content. An experience of this type represents more
than basic colours and edges and shapes; it represents a kind of organization of
features, distinctive of impressionism.
A return to the questions posed in Section 2.2 helps to further show that this all
implies rich perceptual content, and motivation for AP. Those questions concerned
what is needed to specify accuracy conditions (and so content) for perceptual
experience. They were, (1) Are basic properties sufficient (so, for vision: colour,
edges, shape, motion, location)? (2) Are natural or artefactual kind properties
needed? The argument from seeing-as suggests a clear ‘no’ to question (1). The
relevant part of the argument implies that specification of the contents of the
contrasting aesthetic experiences (or the contrasting seeing-as a duck and seeing-as
a rabbit experiences) just in terms of colours, shapes, and edges, would indicate no
contrast, and so clearly not explain the apparent contrast. What is needed is speci-
fication of relevant organizational gestalts. But note also that, when the emphasis is
on the aesthetic cases, the answer to (2) doesn’t matter. Put most strongly, suppose
one accepts one version of the radical lesson provided in Section 2.2, that kind
properties are not contents of perceptual experience. Then, trivially, one would
have to answer ‘no’ to question (2). In effect, one can have rich perceptual content
without perceptual representation of kind properties. Perception of aesthetic prop-
erties involves rich perceptual content.
3. Further Motivating AP: Mechanisms

for Explanation
Additional support for AP can be adduced by noting the various cognitive mechan-
isms or processes that might underlie, and therefore explain, apparent instances of
rich perceptual experience of aesthetic properties. Here, the proponent of AP enjoys a
great deal of explanatory freedom.
Another random document with
no related content on Scribd:
distributed Project Gutenberg™ eBooks with only a loose
network of volunteer support.
Project Gutenberg™ eBooks are often created from several

printed editions, all of which are confirmed as not protected by
copyright in the U.S. unless a copyright notice is included. Thus,
we do not necessarily keep eBooks in compliance with any
particular paper edition.
Most people start at our website which has the main PG search
facility: www.gutenberg.org.
This website includes information about Project Gutenberg™,

including how to make donations to the Project Gutenberg
Literary Archive Foundation, how to help produce our new
eBooks, and how to subscribe to our email newsletter to hear
about new eBooks.

Evaluative Perception Anna Bergqvist Full Chapter

Uploaded by

Copyright:

Available Formats

Evaluative Perception Anna Bergqvist Full Chapter

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Evaluative Perception Anna Bergqvist Full Chapter

Uploaded by

Copyright:

Available Formats

Evaluative Perception Anna Bergqvist

Visit to download the full and correct content document:

MIND ASSOCIATION OCCASIONAL SERIES

List of Contributors vii

Part I. The Existence and Nature of Evaluative Perception

Part II. The Epistemology of Evaluative Perception

Part III. Evaluative Perception and Value Theory

13. Moral Perception, Thick Concepts, and Perspectivalism 258

R A is John A. O’Brien Professor of Philosophy at the University of Notre

viii   

S MG is Assistant Professor of Philosophy at Princeton University. Her

Evaluation is ubiquitous. It isn’t an exaggeration to say that we assess actions,

     

⁶ But see Livingston (2003).

     

1. Existence and Nature of Evaluative Perception

     

³⁰ See Bayne (2009). ³¹ See S. Siegel (2010a).

     

     

2. Epistemology and Evaluative Perception

⁴⁴ This example is from Bengson (2016). ⁴⁵ Setiya (2012).

     

opposed to judgements about hypothetical cases which are implicitly general⁴⁶—can

⁴⁶ See Kagan (2001).

3. Value Theory and Evaluative Perception

⁴⁹ See Deonna and Teroni (2012), 5.

     

⁵⁰ See Brady (2013). ⁵¹ Defended by e.g. Döring (2003); Pelser (2014).

The ‘sometimes’ indicates that aesthetic properties need not be categorically

   

1. For and Against the Sceptical Challenge from Sc

would undermine the observed cross-perceiver convergence on discrimination,

³ There is no standard deﬁnition of ‘low-level’ or ‘high-level’ as pertains to perceived properties, but

   

experience of Picasso’s work as cold, or lifeless, or serene, or ‘perhaps bland, dull,

⁶ Walton (1970), 347.

   

2. Ambiguous Figures and the Argument

¹³ See Wittgenstein (1953/2009); Hanson (1958) and (1969).

   

judge it accordingly. However, it is undeniable that there is a difference in what it’s

   

discriminates—that is, represents—more than just colours, shapes, edges, motion,

2.2. A rejoinder and some general lessons

   

   

   

2.3. Concluding the argument: rich perceptual content

3. Further Motivating AP: Mechanisms

Project Gutenberg™ eBooks are often created from several

This website includes information about Project Gutenberg™,

You might also like