Aesthetics from the Visual Artists’ Viewpoint
How to characterize aesthetics has been revived with Bence Nanay’s Aesthetics as Philosophy of Perception. Reviewing criticisms made by Dustin Stokes, this paper makes the argument that Nanay’s problem is broader than what Stokes points to, as it involves the problem of property attribution and the difference between perceiving a property in a nonaesthetic situation and an aesthetic one. The latter context involves not attributing a property to an object, but rather the process of perceiving low-level features. The problem of how to characterize aesthetics is thus solved by looking at three things: recent research into gist perception, Gareth Evans’ notion of nonconceptual information, and, most importantly, the way visual artists look at the world: All of this points to the centrality of low-level perceptual features that can be doubted, revised, and rearranged. In turn, this explains the core of imaginative viewing and the pleasure of aesthetics, giving it a satisfactory characterization that distinguishes it from the nonaesthetic.
Evans; Gareth; gist perception; low-level features; Nanay, Bence
1. Problems in defining aesthetics
The question of how to characterize aesthetics has been revived in earnest in the last few years with the publication of Bence Nanay’s Aesthetics as Philosophy of Perception. The book presents Nanay’s very plausible argument that we should approach aesthetics from the perspective of attention and from within the methodology of the philosophy of perception, instead of viewing aesthetics within the more traditional and seemingly dead-end concerns of beauty or within the framework of Kantian disinterestedness: “…aesthetics as philosophy of perception.” This seems laudable and correct.
Nanay’s analysis is through the lens of attention, and it is delineated in two forms: non-distributed and distributed. Thus, our attention can be focused and nondistributed over objects or it can be distributed; likewise, our attention can be focused or distributed over properties. Aesthetics is then defined by him as: “Focused with regards to objects and distributed with regards to properties.”
But there have been some legitimate criticisms over whether or not Nanay has correctly characterized aesthetics, particularly in regard to being able to distinguish it from ordinary visual tasks, the latter of which are defined by him as “distributed with regards to objects and focused with regards to properties.” Dustin Stokes, in his review of Nanay’s book, points out that Nanay’s characterization of aesthetics in fact collapses into the category of ordinary visual tasks, and we are thus left with a definition of aesthetics that includes such (unwanted) instances as scanning for something and the exercise of expert opinions.
This problem is due, according to Stokes, in part to Nanay’s inconsistency with what counts as an individual concerning our attention to objects. For although Nanay defines ‘object’ as “sensory individual,” sometimes it seems to imply concrete individual and at other times not. For example, Nanay uses scanning a pile of socks in order to sort for the colors red and blue as an instance of a visual task, and in this instance, according to Nanay, our attention is distributed over objects and focused on properties. But Nanay also calls a landscape, which is an instance of an aesthetic experience, as focused on the object and distributed over the properties, and thus the landscape itself is the (singular) object. So, Stokes is right: why is the pile of socks instance different from a landscape? That criticism seems right, and thus in Nanay’s sock example our attention would be focused with regard to an object, just like it is in aesthetics.
Stokes also makes the argument that Nanay’s problem of aesthetics collapsing into ordinary visual tasks is additionally due to Nanay’s treatment of properties. Here, Stokes’ criticism is that the confusion is due to a conflation between types and tokens. While Nanay wants to say that in the visual search example, we are exercising focused attention in regard to properties, Stokes points out that while it may be true that we are looking for one thing in regard to the property type, for example, in Nanay’s example of socks, red or blue, we are identifying individual instances of those things, that is, individual blue socks or individual red socks. In Stokes’ words: “We attend to property-instances or features (not types).” This is true. Therefore, we must conclude that Nanay’s attempt to properly define aesthetics is less than completely satisfactory.
2. A proposed solution
I propose to approach the problem from a different point of view. Nanay, like almost all other philosophers who have written about aesthetics or the philosophy of art, looks at the problem from the point of view of the viewer. I propose looking from the point of view of the visual artist, and seeing if that can solve some of the problems that the viewer-perspective posed in Nanay, in addition to giving a more inclusive and accurate definition of aesthetics. I will show that the perspective of artists — as they have an aesthetic perceptual experience of the world — is one that relies not on the perceptions of high-level, semantic features, for example, on the linguistically defined and conceptualized object, nor on properties of that object as instances of property attribution. Instead, the perspective of artists relies on the sensed experience of low-level features of perception.
This kind of low-level perception is a perceptual process best explained through an understanding of the recent scientific research in gist perception, and thus I will briefly explain this in order to properly frame the issue of low-level perception. I will then give a philosophical account of how this process is not an instance of property attribution, as it is in ordinary visual tasks, but instead is a focus on perception as sensing, which is not of conceptual objects. This, in turn, is best explained by turning to a philosopher from the past: Gareth Evans. While not recycling the well-trod debates of the early 2000’s, instead I will use some of Evans’s views to pair with low-level perception in gist, which then will allow for the formulation of a definition of aesthetics that obviates the difficulties in Nanay’s views.
A side note is in order here. I do not mean to conflate the distinction between conceptual and nonconceptual with the distinction between low-level and high-level/semantic perceptual features. While no vision scientist categorically or precisely draws the line between low-level and high-level perception, even in terms of spatial frequency, and while the debate has not been settled on how to definitively fence off nonconceptual from conceptual, it is nevertheless uncontentious to say that the two pairs do not isomorphically map onto each other. Though this is true, high-level semantic perceptual features are defined in terms of the application of governing concepts ( it is an “apple”), and it is the latter’s function vis-à-vis general germs that gives high-level features their ability to identify objects as F of x.
With these distinctions in hand, I will argue that it is this — the artist’s perspective — seen as the perceptual experience of low-level, nonconceptual features that best provides the paradigm for aesthetic experience and thus solves the dilemma of how to separate instances of aesthetic experience from ordinary visual tasks.
‘Gist’ is generally defined as holistic perceptual experience within the first 200-300ms (milliseconds), and it is different from object recognition. Object-recognition takes much longer and requires higher spatial frequency, while we see global, holistic scenes very quickly and at much lower frequencies. Included in the definition of gist are both high-level and low-level features. Low-level features are primarily things such as color and spatial frequency. Others that have been included in the category of low-level are orientation, hue, texture, contours, motion, volume, and illumination. High-level features, on the other hand, are natural kinds (“apple”) and general semantic kinds, whether it is an indoor scene or an outdoor scene.
Researchers using whole-scalp MEG techniques tested the relationship between neural processing and visual information categorization in gist perception between 100-250ms, finding the “temporal relationship between when low-level scene-related information becomes available and when it becomes useful for behavioral categorization” are co-localized with visual information representation. In addition, many low-level features can be detected at <100ms, such as color, which can be detected at 42ms; and at the threshold of 19-67ms (reaching asymptote at 100ms) participants were able to classify naturalistic scenes (natural versus manmade) more quickly than they could make a semantic classification of the scene (for example, mountain, urban): “Remarkably, the threshold presentation duration was, on average, shorter for perception of a scene’s global properties than for perception of the scene’s basic-level category.”
While in the instances of high-level features, test subjects were able to name their perceptions after the perceptions themselves occur, this is not always the case with low-level features, most of which are without names or semantic category. For example, the particular blobby-like contours of scene will be seen at low spatial frequencies well <100ms, but that particular shape need not be identified via a name or, in fact, even have a name. Nevertheless, we are able to perceive this information as the brain exhibits “strong sensitivity to low-level summary statistics…100ms after stimulus onset confirms that this information is available early in processing.
It is not only the case that low-level features of a scene’s structure contribute to the global properties of gist perception, but these low-level features are constitutive of the mechanical, computational mechanism that absorbs the external world through sampled episodes driven by extremely fast saccadic scans. These samplings, which happen at three to four per second, are also highly edited; we do not process all that is perceptually available. While there is disagreement among researchers and philosophers regarding the exact roles of top-down and bottom-features, there is growing evidence that our (edited) perception is largely governed by so-called “saliency maps,” which, in turn, are significantly constituted by low-level features: “This model consequently represents a complete account of bottom-up saliency and does not require any top-down guidance to shift attention.” Thus, a saliency map is a selecting of the significant perceptions into usable data. Granting computational efficiency, “Saliency is independent of the nature of the particular task, operates very rapidly and is primarily driven in a bottom-up manner…” This combines with targeting provided by context and Bayesian models, giving an attentional “mechanism driven by image saliency and contextual guidance” that are “a natural consequence of the probabilistic framework,” which allows us to identify global features as well as subsequent objects.
Gist, as global whole-scene perception, then is significantly the result of processing low-level features. As one of seminal researchers in the field, Aude Oliva, states, “…the global features are originally built as combinations of local low-level filters of the type found in early visual areas.” It is important to note that recent research does not view this reliance on low-level information as a feedforward process, as was originally thought by early researchers, but that significant amounts of low-level information function in a feedback loop. This further emphasizes the role of low-level features in perception.
It is important to note two things: 1) gist is not the same as object-recognition, and 2) gist is not the same as ensemble perception. I will address the former first. Research has shown that gist perception is not primarily accomplished in the same region of the brain as is object-recognition. Known as the para hippocampal region, this part of the medial occipitotemporal cortex responds to scenes but not to objects. While early visual research focused on objects at the atomic units of perception, that has shifted to an emphasis on global, whole scene perception. As Oliva, states, “There are many interesting properties of a real-world scene that can be defined independently of the objects.” The latter, for example, ensemble perception, is an analysis of how the brain represents a group of similarly constituted objects, with research showing that the statistical information in terms of the mean is more accurate than perception of individual membership. This distinguishes itself from whole-scene gist perception of low-level features, in that ensemble perception is a kind of summary perception, where the visual system represents sets of similar items, such as blades of grass, whereas gist applies to perceptual experiences regardless of whether or not it is a similarly constituted set. Nevertheless, there is an important relationship between ensemble perception and gist. In their discussion of ensemble perception, Oliva and Torralba note, “This suggests that a reliable scene representation can be built, in a feed-forward manner, from the same low-level features used for local neural representations of an image.”
Since low-level features play a significant role in gist perception and thus perception itself, and as they are experienced independently of object recognition, it therefore allows us to think of perception as something other than of entirely linguistically defined objects. As an empirical analysis of the role of low-level features, gist research allows us to step back from viewing percepts as entirely constitutive of objects. In other words, it allows us to focus on the entities that are the percepts of those low-level features of sensory perception, despite them being unnamed. And this, in turn, brings us to the consideration of perception conceived of as something other than linguistically identified objects: a consideration of Gareth Evans.
Writing in the 1980s, Evans’ posthumous The Varieties of Reference addressed concerns of representation and semantics and significantly emphasized the role of sensing, most pertinently in an oft-quoted footnote: “…the senses yield non-conceptual information, whereas language embodies conceptual information…” Evans’ view is that objects are not only perceived in terms of being an instance of a named category — “apple,” but instead proposed that part of the information that we receive from perceptual experience comes from sensing and that that is non-semantic.
His emphasis on sensing was not new, of course. One need think not only of the British Empiricists or of such recent philosophers as Christopher Peacocke, but even more pertinently of Bertrand Russell’s distinction between knowledge by description and knowledge by acquaintance. Evans, in discussing this issue in Russell, wrote, “All the items with which a subject may be acquainted have what may be termed the Cartesian property…But Russell, like any Cartesian, would wish to decompose such predicates into strictly mental and non-mental components, the claim about infallibility extending only to the former.” Evans’ work is, in many ways, an attempt to explain perception such that it doesn’t run afoul of what he calls Russell’s Principle, that a subject cannot make a judgment about something unless he knows which object his judgment is about, while yet explaining, contrary to Russell’s Cartesian requirement, that we make errors in perceptual judgments. As Evans explains, “however, there does not seem to me to be anything incoherent in the idea that it may be, for a subject, exactly as though he were thinking about a physical object (say) which he can see, and yet that, precisely because there is no physical object he is seeing, he may fail to have a thought of the kind he supposes himself to have.”
What Evans is pointing to is that while something might initially seem to be a certain way to us, we can yet be mistaken about that initial sensing information, that is, our later sensing can diverge. Thus, while his emphasis on sensing was not new, what was new was Evans’ claim that it is of a distinct kind, a kind that gives us nonconceptual information. This distinction between the kind of potentially fallible information given in a sensing, nonconceptual experience and the kind of information given in a subsequent judgment was thus explained by him as follows:
The informational states which a subject acquires through perception are non-conceptual, or non-conceptualized. Judgements based upon such states necessarily involve conceptualization: in moving from a perceptual experience to a judgement about the world (usually expressible in some verbal form), one will be exercising basic conceptual skills. But this formulation (in terms of moving from an experience to a judgement) must not be allowed to obscure the general picture. Although the subject’s judgements are based upon his experience (i.e. upon the unconceptualized information available to him), his judgements are not about the information state. The process of conceptualization of judgement takes the subject from his being in one kind of informational state (with a content of a certain kind, namely, non-conceptual content) to his being in another kind of cognitive state (with a content of a different kind, namely, conceptual content).
It is important to stop for a moment and think about how he is defining the more general notion of content and its relationship to concepts. The content of a thought is conceptual — in this way, it is a type — and this fact is used to constitute Evans’ emphasized principle that governs the flexible combinatorics of subjects and predicates. Therefore, content is a type when it is constitutive of judgments, which are implemented through the use of concepts: “Thus, if a subject can be credited with the thought that a is F, then he must have the conceptual resources for entertaining the thought that a is G, for every property of being G of which he has a conception. This is the condition that I call ‘The Generality Constraint.’” But content would not be a type in the information-states that are constituted by the nonconceptual, for while Evans emphasizes our ability to uncouple and recouple predicates and subjects, he is adamant in his focus on the atomistic, nonconceptual constituents of perception:
Philosophers have been prepared to attach considerable theoretical importance to recognitional capacities in their accounts of what it is for a subject to have an idea of a property, or kind, of particulars, while relegating recognitional capacities for particulars to a theoretically insignificant position in their account of what it is for a subject to have an Idea of a particular; it is the aim of the present chapter to redress the balance.
If we reflect for even a brief moment about the recent research into gist perception, we immediately realize that perception is not entirely of objects construed as linguistically constructed entities. Instead, as was shown, perception includes low-level information that is non-semantic and more fungible and thus less concept-driven. The elemental low-level features do not come to us as instances of a larger, conceptual category, but as distinct elemental percepts in their own right. They have yet to be assigned to linguistically named categories. We might see the contours of a shape but not have a name for either those contours or that shape.
If we take seriously recent scientific research and agree that we experience many of these kinds of low-level percepts in not only surprisingly few milliseconds but separately from semantic identification, we must furthermore agree that this makes perception something more than of objects defined as linguistic entities. Perception includes sensing. And, when we ask the question what is it that we are sensing, we now see that sensing gives us a lower ontological level of elemental percepts, a level that is unnamed. As an ontological level that is non-semantic, it is, thus, nonconceptual.
Evans’ broader notion of the nonconceptual thus refers not only to the epistemological process of sensing but also to those entities that are the source of the sensing: The low-level percepts that are perceived non-semantically.
Positing low-level percepts as the entities that provide the perceiver with nonconceptual information through sensing also allows for a framework with which to understand Evans’ arguments concerning the contents of perception, in particular to what he referred to as “seemings.” Let us remember, there is, the act of sensing, the entity that is sensed, and the contents of that perception. For Evans, neither objects nor their constituent parts enter “without objective content.” Everything enters with some degree of content, for Evans. We perceive things within a schema. This is a core mechanism within his notion of perception.
Robert Stalnaker explains Evans’ view with the following: “Evans regards it as important to identify information-bearing states of perceptual systems with states of seeming since he is anxious to avoid the traditional epistemologist’s picture according to which the subject receives, through the perceptual systems, sensory data that is ‘intrinsically without objective content,’ but which forms the basis for inferences about the world that causes them.”
Evans is arguing against a view of the world that sees information coming to us neutrally, shorn of meaning: “without objective content.” That view sees the objects of perception as arriving at the doorstep of our perceptual faculties in a state that guarantees their objectivity; they do not carry content with them. Thus, they are purely objective entities on which we can now use our operations of logical inferences and thus be guaranteed satisfactory truth conditions for knowledge-acquisition.
But this, Evans was arguing, is not the case. Rather than operating on data that is without objective content, what we actually do is receive information that is already imbued with meaning: Information that is already part of a schema; hence, “seemings.” We might resist it, we might doubt it, as it is at least potentially “resistible,” but it nevertheless already comes imbued with meaning. As a named object, it is not neutral. Rather, it is the entity that is the consequence of applying the inferences to which Evans was referring and the entity that has come to us already within a system of assigned meanings and associations. This is because we, as a social unit, have already in the past named these objects prior to any of us, as individuals, perceiving it in the present moment. This is an important point. This is why it is “impossibly the wrong way round:” When we refer to objects, we are referring to those linguistic entities that have already been named. They have already been brought in within a schema. They are already imbued with meaning and thus with content. Thus, all objects come to us as “seemings:” They seem to be a certain way.
4.2 Seeing is not believing
But objects do not have to be the way they seem. In the paragraph next after the one referred to by Stalnaker, Evans makes the distinction between being in a state of perception and in a state of belief, and he wants the two to remain as separate parts of the informational structure. He states, “…the subject’s being in an information state is independent of whether or not he believes that the state is veridical.” The next paragraph amplifies this and connects it to art: “The belief-independence of the state of the informational system is not merely a curiosity, not even merely a suggestive curiosity. Upon it depends the whole of representational art.”
I would say the whole of art, not merely representational art. But that is the only part with which I disagree. Evans has it right. There is seeming and there is believing. And they are very different things. It is possible, as Evans argues, to see something and not believe; it is possible to doubt, to realize one might be wrong — to look again.
This is easily understood once we remember that, for Evans, data comes in already in a context and not neutrally. Thus, we have a very large role in how we edit the data, what system in which to place it, and how to construe it; phenomenal experience comes to us adulterated by prior beliefs and its membership in an already established mental schema. There is no virgin eye.
Because it is not the case that phenomenal perception is of neutral/untainted percepts that lead us directly to truth, we are able to uncouple these two functions of perception and belief. This allows us to not only separate the content of perception from the content of belief but, in addition and upon reflection, allows us to fully characterize the nonconceptual as an important part of the constituents of perception; for it is with the experience of the low-level, nonconceptual percept that doubt is primarily possible. It is in the experience of the low-level features of perception that we are best able to doubt, revise, and edit anew.
I will try to illustrate this by giving an example in words. My eyes light on the object that I don’t yet recognize as a dog. I sense the look of it, and I am aware of the surface as a thing that would be pleasant to touch. If I were to consciously articulate it and name it, I would call it “fuzzy,” although clearly this quality designation and naming is not necessary for the sensing experience. But I decide to double-check this and look again and now sense the visual phenomenon as coarse-haired. It does not now seem particularly touchable to me, at least in the former inviting way. My sensing has altered. In these few milliseconds, my sensing was doubted, edited, and replaced. And the point here is that we do this uncountably many times with many, if not most, of our perceptions. Sensing is a rapid-fire testing of the waters: accepting, believing, discarding, doubting, revising. It is the realm of flexibility of belief.
And each object that we see is constituted by those minute sensings. As an object, the dog uncountably has many sensings that are constitutive of it. The fur itself is an instance of nameable qualities, such coarseness, thickness, fuzziness, color-qualities, cleanliness, single-coated or double-coated, and so on. And those are just the ones we can name. Other percepts that would be included are subtle variations of color, blobbiness of various shapes within the fur, the orientation of individual hairs, the differing degrees of illumination and light, and on and on. The summation of all of these low-level features that have been not disbelieved or discarded by us, and instead are conjoined together in the voting process that is part of our consenting to the veridicality of our perceptions, are then matched with a memory of similar objects experienced in the past; thus, the designation of a linguistic object. The object-seeing — the distal accomplishment — is in coordination with this multitude of nonconceptual information that is visual processing.
5. How artists see the world
Now that we have an overview of the science of low-level features in gist perception and a philosophical account of sensing, we can turn to a description of the artist’s low-level perceptual experience of the world. As each artist’s looking at the world is a first-person experience and thus, almost by definition, impossible to encapsulate in a descriptive third-person account, what we can do instead, as a way into this viewpoint, is to examine the emblematic teaching of art. How artists are taught to look at the world is reflective of how they do, in fact, look at the world. Thus, this examination can be accomplished by drawing on the lessons of the mid-twentieth century artist and teacher, Josef Albers, a well-known German abstract painter who moved to the U.S. after the infamous 1933 Nazi Entartete exhibit and joined the John Dewey-influenced Black Mountain College in North Carolina. He wrote, in his 1963 Interaction of Color:
In visual perception, a color is almost never seen as it really is
-as it physically is.
This fact makes color the most relative medium in art…
…What counts here – first and last – is not so-called knowledge
of so-called facts, but vision – seeing.
Seeing here implies Schauen (as in Weltanschauung) and is coupled
with fantasy, with imagination (stanza form in original)
In an interview in 1970, he states,
I made true the first English sentence that I uttered (better stuttered) on our arrival at Black Mountain College in November 1933 (right after the closing of the Bauhaus art school in Dessau Germany, by the Nazi’s) when a student asked me what I was going to teach I said: ‘to open eyes’…I have taught – until 10 years ago – for nearly 40 years, that is almost half of my life. And when I think that over – now afterwards, I come to the conclusion, namely that I did not teach arts as such, but philosophy and psychology of art.
Learning how to see anew — in Albers’ words “to open eyes” — is the most fundamental aspect of learning how to paint and draw. Anyone who has learned to paint a painting knows the lesson that one doesn’t paint one object and then move to the next. To focus on the couch and then the floor and the figure dooms a painting. Instead, one must focus on the more amorphous spaces in between the named objects and see them as continuous with the named objects: what is commonly called the “negative spaces.” These are such things as the qualities of trailing light seen not on an object or in the “light” as a whole but in each discrete bit of illumination – the almost infinitesimal changes from one tiny spot to the next. Thus, seeing in painting and drawing is not recognizing the named objects and replicating them; it is exactly the opposite of that. Seeing is both seeing in between those objects, and through those objects, so to speak, so that one can get the “gist” of the whole scene; it is going back over each discrete moment in the scene and treating it as a thing in itself, not as merely an appendage to whatever it is that it might be named as part of. This is emphasizing the experience of the low-level features of perception, as each are sensed in their own particularity. It is what makes artists’ perceptual experience unique and marks off the domain of the creative.
The artist’s experience of seeing is thus much more than seeing objects, defined in the usual linguistic way, and this description just given of learning how to paint and draw, a kind of seeing which is consistent with the mature artist’s perceptual experience, is a description of learning how to isolate and elevate the perception of low-level features. This sensed particularity of artists’ perceptual experience is the phenomenon that also more fully explains Evans’ notion of “seemings.” As we have seen, for Evans all objects enter into our phenomenal experience as “seemings,” entities that already come imbued with meaning, which thus “seem” to be a certain thing. This notion of “seemings” is made clearer if we reflect on artists’ sensed perceptual experience of low-level features and, in turn, their role in how objects are identified.
As I have tried to make clear, seeing as artists see is not a perceptual activity that is limited to objects as linguistically construed, conceptual entities. Instead, artists’ way of seeing is the dissolution of linguistically defined objects through such practices such as: 1) seeing negative space, and 2) the perceptual practice of sensing the individual percept independent of its usual association with a particular concept, for example, the fuzziness of the dog’s fur is sensed independently of the concept “dog” or “fuzziness.” Thus, a percept is not experienced as an instance of something or an instantiation of a general term, but in its own sensed particularity.
This is very distinct from seeing properties. To see properties, even in the version presented by Nanay of seeing properties in a distributed manner, is to see those things that belong to objects. This is what properties are. But to perceive properties as those things that belong to objects is to see them, of course, as attributable to objects. This process of property attribution is exactly what makes Nanay’s visual tasks indistinguishable from aesthetic experience. If we are looking at the distributed instances of red in a painting, we want that process to be distinguishable from sorting red socks. But, if the experience of the instance of red in the painting is also an instance of seeing the red blob as a property of the object “the artwork,” then that process is, in fact, the same as sorting for red socks.
5.2 Seeing is not believing
Instead, we must see the instances of red in the painting as low-level information, separable, at least potentially, from its ownership by an object.
And that is to see it as artists see the world.
This artist-driven view gives us the elemental percept at a level that is in the realm of flexibility of belief, and thus, in the realm of creativity. Looking at the way artists see is a way to conceive of percepts that allow for the existence and role of the nonconceptual, low-level features of perceptual experience and thus to see the world in a way that can be something other than the way it “seemed” initially; it is to open up the possibility of creativity.
6. Conclusion: a different definition of aesthetics
Nanay’s difficulty with defining aesthetics was, we can now see, the result of the more general problems of characterizing perception. It is a problem of bifurcating perception into linguistically-identified objects and properties, the latter of which are recognized in virtue of being attributable to the former. To return to his sorting socks example: We are finding the instances of red and attributing that property to those socks. But aesthetic experience had the same paradigm: It was attributing the property red to the object artwork. The difficulty is not what Stokes thought it was: the distributed/non-distributed distinction in attention. Instead, that ends up being irrelevant. The irrelevance of that is made evident when we realize that when we experience red aesthetically, we are not doing property attribution: The relationship is not of that paradigm of property to object.
Instead, in an aesthetic experience, we experience the color as a low-level feature in its particularity. This is what the artists’ view illustrates. To perceive the world as the artist does is not to establish a relationship to the linguistically defined object; in fact, it is the opposite of that. The point to the artists’ view is that low-level features are experienced independently of their association with linguistically identified objects. Herein lies the freedom in imagination that artists experience: The emphasis on the role of low-level features potentially disassociated from named objects. This is the view that allows for the creation of new objects, newly constituted ontological entities in their own right. The rapid-fire doubting, revising, looking again, and ultimately forming a new belief is the core of the enervating process of aesthetic experience. It is the core of imaginative viewing and the pleasure of aesthetics.
Dena Shottenkirk is an Assistant Professor in the Philosophy Department at Brooklyn College. Her books include Perception, Cognition & Aesthetics (ed. Dena Shottenkirk, Manuel Curado, and Steven Gouveia, NY: Routledge, 2019); Cover Up the Dirty Parts! Arts Funding, Fighting, and the First Amendment (UK: Cambridge Scholars Press, 2010); Nominalism and Its Aftermath: The Philosophy of Nelson Goodman (The Netherlands: Springer Verlag, 2009); Art as Cognition: Framing the Aesthetic Experience as Conversation (forthcoming, Springer). She is also an artist and the founder of the public philosophy/art nonprofit talkPOPc, found at: talkpopc.org.
Published March 4, 2021.
Cite this article: Dena Shottenkirk, “Aesthetics from the Visual Artists’ Viewpoint,” Contemporary Aesthetics, Volume 19 (2021), accessed date.
 Bence Nanay, Aesthetics as Philosophy of Perception (Oxford: OUP, 2016), p. 2.
 See also, Dena Shottenkirk, review of Aesthetics as Philosophy of Perception, by Bence Nanay, Grazer Philosophische Studien, 97, 2 (June, 2020), 343-349.
 Bence Nanay, p. 24.
 Dustin Stokes, review of Aesthetics as Philosophy of Perception, by Bence Nanay, Notre Dame Philosophical Reviews (August 10, 2016), 1-2; ref. on 2.
 Bence Nanay, Aesthetics as Philosophy of Perception (Oxford: OUP, 2016), p. 24.
 See also Bence Nanay, “Aesthetic Attention,” Journal of Consciousness Studies, 22, 5 (2015), 96-118; ref. on 106.
 Dustin Stokes, 1-2, ref. on 2.
 The type/token distinction should also not be thought of as strictly isomorphic to either of the two distinctions above. Although again some caveat must be added: Traditionally defined as an ontological distinction between general and particular terms, the connection has been made to universals, as per Nelson Goodman: “The distinction between a word ‘type’ and its ‘tokens’ was stressed by Pierce…The type is the universal or class of which marks are instances or members.” (Nelson Goodman, Languages of Art (Indianapolis, IN: Hackett, 1976), fn p. 131).
 Aude Oliva and Antonio Torralba, “Building the gist of a scene: the role of global image features in recognition,” in Progress in Brain Research, 155 (2006), 23-36; ref. on 24.
 Aude Oliva, “Gist of the Scene,” Neurobiology of Attention, ed. Laurent Itti (NY: Elsevier, 2005), pp. 251-256, ref. on 251.
 Pavan Ramkumar, Bruce C. Hansen, Sebastian Pannasch, & Lester C. Loschky, “Visual information representation and rapid-scene categorization are simultaneous across cortex: An MEG study,” Neuroimage, 134 (2016), 295-304; ref. on 298.
 Monica S. Castelhano & John M. Henderson, “The influence of color on the perception of scene gist,” Journal of Experimental Psychology: Human perception and performance, 34, 3 (2008), 660-675; ref. on 660.
 Michelle R. Greene and Aude Oliva, “The briefest of glances: The time course of natural scene understanding,” Psychological Science, 20, 4 (2009), 464-472; ref. on 471.
 Iris I.A. Groen, Sennay Ghebreab, Hielke Prins, Victor AF Lamme, & H. Steven Scholte, “From image statistics to scene gist: evoked neural activity reveals transition from low-level natural image structure to scene category,” Journal of Neuroscience, 33, 48 (2013), 18814-18824; ref. on 18821.
 Michelle R. Greene and Aude Oliva, 464-472; ref. on 465.
 Benjamin W. Tatler, Roland J. Baddeley, and Iain D. Gilchrist. “Visual correlates of fixation selection: Effects of scale and time,” Vision research, 45, 5 (2005), 643-659; ref. on 643.
 Laurent Itti, Christof Koch, & Ernst Niebur, “A model of saliency-based visual attention for rapid scene analysis,” IEEE Transactions on pattern analysis and machine intelligence, 20, 11 (1998), 1254-1259; ref. on 1254.
 Laurent Itti & Christof Koch, “Computational modelling of visual attention,” Nature reviews neuroscience, 2, 3 (2001), 194-203; ref. on 194.
 Antonio Torralba et al., “Contextual Guidance of Eye Movements and Attention in Real-World Scenes: The Role of Global Features in Object Search,” Psychological Review, 113, 4 (2005), 776-786; ref. on p. 783.
 Aude Oliva & Antonio Torralba, 23-36; ref. on 34.
 Michael H. Herzog, Evelina Thunell & Haluk Ögmen, “Putting low-level vision into global context: Why vision cannot be reduced to basic circuits,” Vision Research, 17 (October 2015), 9-18, ref. on 11. See also: Andy Clark, Surfing Uncertainty (Oxford: Oxford University Press, 2015).
 Russell Epstein & Nancy Kanwisher, “A cortical representation of the local visual environment,” Nature, 392, 6676 (1998), 598-601; ref. on 598.
 Russell Epstein, “The cortical basis of visual scene processing,” Visual Cognition, 12, 6 (2005), 954-978; ref. on 954.
 Michelle R. Greene & Aude Oliva, “Recognition of natural scenes from global properties: Seeing the forest without representing the trees,” Cognitive Psychology, 58 (2009), 137-176; ref. on 139.
 Aude Oliva & Antonio Torralba, 23-36; ref. on p. 31.
 Dan Ariely, “Seeing sets: Representation by statistical properties,” Psychological science, 12, 2 (2001), 157-162; ref. on 160.
 Aude Oliva & Antonio Torralba, 23-36; ref. on p. 28.
 Gareth Evans, The Varieties of Reference (Oxford: OUP, 1982), p. 123 fn 5.
 At issue is whether or not one is speaking here of nonconceptual experience or of the nonconceptual content of subpersonal information-processing states. Many conceptualists concede that subpersonal processing engages in perceptual processes that are nonconceptual. Of course. Therefore, my focus here – and I believe it was Evans’ as well – is on nonconceptual experience that is not merely subpersonal, but is instead the experience of perceptual phenomena that attention can augment and amplify. And that perceptual phenomenon is experienced in the conscious act of sensing. It is an experience that one can be aware of despite the fact that it is non-semantic. And thus one can also be aware of what the experience is of. His claim is a positive one: we can gain nonconceptual information from sensing.
See also Jacob Beck, “The Generality Constraint and the Structure of Thought,” in Mind, 121 no. 483 (2012): 563-600, for a related discussion of sensed perception and the nonconceptual in Evans.
 Christopher Peacocke, “Sensational properties: Theses to accept and theses to reject,” Revue internationale de philosophie 1 (2008), 7-24.
 Gareth Evans, The Varieties of Reference (Oxford: OUP, 1982), p. 44-45.
 Ibid., p. 89.
 Ibid., p. 45.
 Ibid., p. 227.
 Ibid., p. 104.
 Ibid., p. 270.
 Robert Stalnaker, “What Might Nonconceptual Content Be?,” Philosophical Issues, 9 (1998), 339-352; ref. on 352. Note: The “traditional” epistemologist here would include the (Humean) empiricist as well as both the radical empiricist who views objects as the neutral given and the realist who views our reception of external reality as direct, completely mind-independent, and thus “objective.”
 Gareth Evans, p. 124.
 Nelson Goodman, Languages of Art (Indianapolis, IN: Hackett, 1976), p. 7.
 This could occur for many reasons all of which are irrelevant for this discussion: for example., I remember that I’ve been wrong in the past, I am worried about having an unpleasant petting experience, I note that my vision is not always so great, and so on. The point is that I may doubt my sensory perception.
 Josef Albers, Interaction of Color (New Haven: Yale University Press, 1975), pp. 1-2.
 John H. Holloway & John A. Weil, “A Conversation with Josef Albers,” Leonardo, 3, 4 (1970), 459-464; ref. on 459.
 For example, think of the way the famous Cezanne portrait of his wife (“Portrait of Madame Cezanne”) shows the shadow down the nose looking green. He painted what he saw – the way it seemed to him at the moment – and not because he would have claimed that the green shadow had objective content, a fact agreed upon by others and claimed as veridical. Seeing as an artist therefore is being receptive to the way things look apart from the way one expects them to look, for example, apart from how objects are defined as members of their respective linguistic category. Seeing as an artist is seeing discreet percepts that are non-object, non-linguistically identified things – that is largely what makes reality feel like and thus be reality.
 Nanay qualifies this emphasis on attention in a few later publications. He states that he may have made “a marketing mistake,” as it was his “third, not at all central, aim to argue that a special way of exercising our attention, in a manner that focuses on an object but distributed among its properties, plays an important role in some instance of aesthetic experiences. The marketing mistake was to start the book with this…” (Nanay, Bence. “Responses to Irvin and Schellekens” in Estetika LVI/XII, 1 (2019), p. 118.) While his intention cannot be determined by his readers, suffice it to say that the book was organized around attention and that is the way readers and critics have approached it; I thus do the same in this paper. (See also Bence Nanay, “Aesthetics as Philosophy of Perception: A Précis” in Estetika LVI/XII, 1 (2019), pp. 91–94).