Phenomenal Socialism ^†

Abstract

Phenomenal socialism says that what we actually, directly, literally perceive is only or primarily instances of high-level phenomenal properties; this paper argues for phenomenal socialism in the weaker, primarily version. Phenomenal socialism is the philosophy of perception that goes with recognitionalism, which is the metaethics that goes with epiphanies. The first part states the recognitionalist manifesto. The second part situates this manifesto relative to some more global concerns, about naturalism, perception, the metaphysics of value, and theory vs. anti-theory in ethics. The third part rehearses two familiar views about the possible contents of perceptual experience, Phenomenal Conservativism and Phenomenal Liberalism. It notes that the usual catalogue omits two other theoretical possibilities, Phenomenal Socialism and Phenomenal Nihilism, and it defends a watered-down form of Phenomenal Socialism from four main objections. The fourth part makes some connections with the epistemology of modality and with the role of the imagination.

I believe the power of observation in numbers of very young children to be quite wonderful for its closeness and accuracy. Indeed, I think that most grown men who are remarkable in this respect may… be said not to have lost the faculty, than to have acquired it; the rather as I generally observe such men retain a certain freshness and gentleness, and capacity of being pleased, which are also an inheritance they have preserved from their childhood.

(Charles Dickens [1], David Copperfield [1850], Chapter 2)

One day during my last term at school I walked out alone in the evening and heard the birds singing in that full chorus of song, which can only be heard at that time of the year at dawn or at sunset. I remember now the shock of surprise with which the sound broke on my ears. It seemed to me that I had never heard the birds singing before and I wondered whether they sang like this all year round and I had never noticed it. As I walked I came upon some hawthorn trees in full bloom and again I thought that I had never seen such a sight or experienced such sweetness before. If I had been brought suddenly among the trees of the Garden of Paradise and heard a choir of angels singing I could not have been more surprised. I came then to where the sun was setting over the playing fields. A lark rose suddenly from the ground beside the tree where I was standing and poured out its song above my head, and then sank still singing to rest. Everything then grew still as the sunset faded and the veil of dusk began to cover the earth. I remember now the feeling of awe which came over me. I felt inclined to kneel on the ground, as though I had been standing in the presence of an angel; and I hardly dared to look on the face of the sky, because it seemed as though it was but a veil before the face of God.

(Griffiths 1979 [2], p. 9; quoted by Taylor 2007 [3], p. 5)

1. The Recognitionalist Manifesto

My metaethics is recognitionalist. Here, in a nine-point manifesto, is recognitionalism:

(1)

Values are encountered. We might almost say we bump into them. They come to us from outside; like tables [4]1 and trees and tax invoices, they are “just there”, waiting for us to notice or apprehend them. As with tables, trees, and tax invoices, we do not construct values or infer them from other more basic or immediate objects of experience. We experience the values themselves, directly.

(2)

We are, broadly, passive relative to the values that we encounter, not active. Their presence before us is, basically, none of our doing. I have to say “broadly” and “basically”, because at least since Kant, the distinction between active and passive in experience has been notoriously hard to draw. I am not denying, for example, that the way we frame and categorise and narrate our experience is shaped by our context, our history, and our psychology, that our experience is never free of conceptual loading (more about that shortly). Nor, even more obviously, am I denying that what values we encounter may depend on which ways we direct our attention, just as a table’s availability to my vision may depend on which way I turn and which way I look. Despite provisos like these, the basic relationship between us and the values that we encounter is a passive one. (And a good thing too.)

(3)

Values do not come across to us as any kind of function of our desires or preferences or overall aims, any more than trees that we encounter come across to us as such a function. The values that we encounter are not the way they are because, in any sense, we want them to be. It is nice when the values that we encounter do match well with our wishes, commitments, projects, and life plans. But it is perfectly possible and not at all uncommon for them not to. The reasons that they generate are mostly, in Williams 1981’s [5] sense, external reasons, reasons that are there no matter what is in anyone’s subjective motivational set.

(4)

Any particular encounter with value has the highest possible degree of independence in its evidential force. Insofar as it is possible for a single encounter to have evidential force irrespective of its inferential connections (or lack thereof) to anything else, including any other encounter with value and any system of theory, any single value encounter has this force.

(5)

Values are no less transculturally available for encounters than tables, trees, and tax invoices. They may also, as the case of tax invoices reminds us to add, be no more transculturally available; though, in fact, I think it is natural to be more optimistic about transcultural availability with values than it is with tax invoices. (Think of how we might come to understand a very different society: it is easy to imagine moments where insight into that society flashes upon us, because we come to grasp what they count as valuable. It is less easy to imagine such insights happening when we come to grasp what they count as a tax invoice.) Likewise, encounters with value are pervasively transculturally intelligible. As we can show by examples, we can mostly understand the responses to value of otherwise exceedingly alien cultures. (“Mostly”: certainly, we cannot always, but the cases where we cannot are perhaps overdramatised by philosophical scepticism.) [6]2.

(6)

Values are not the proper object of any single sensory modality—but then neither are tables and trees and tax invoices. One cannot be aware of values without being conditioned and disposed in the right sorts of way—but then the same is true of tables, trees, and tax invoices. And in fact, in all four cases, the conditioning and disposing required is not especially recherché and exceedingly widespread.

(7)

Encounters with values, like encounters with tables, trees, and tax invoices, are generally, though not always, such that they can give rise to reasons to act (and external ones too—cp. point (3)). They are sometimes such that they must give rise to reasons to act. But then so, when other things are equal, are encounters with tables (when they are about to fall on you), trees (when you are about to cycle into them), and tax invoices (when they say “Pay this or go to prison”).

(8)

Encounters with tables, trees, and tax invoices can be evidentially decisive, or as good as decisive: while there is perhaps always room for a merely philosophical doubt, encounters of these sorts can still be such that they leave us no room for serious doubt about the reality of what we have encountered, and they typically are, though there is also a small minority of borderline cases where there is serious doubt about the reality of the things encountered. Exactly parallel remarks apply for encounters with values.

(9)

Encounters with values and with commonplace objects like tables, trees, and tax invoices are, I suggest, structurally analogous in all these eight ways. But here, to close my list of theses comparing the two kinds of encounters, is what I take to be a disanalogy between them. Encounters with value can be, and in many central and focal cases are, epiphanies. An epiphany, in a paradigm case, is an overwhelming existentially significant manifestation of value in our experience, often sudden and surprising, which feels like it “comes from outside”—it is something given, relative to which, I am a passive perceiver—which teaches us something new, which “takes us out of ourselves”, and to which there is a natural and correct response. Encounters with values can be revelations to us of something that is found or that revolutionises the whole way we see the world, the whole way we think about value in general, our entire motivational and justificatory outlook; they can be moving, awesome, inspiring; they can give us a sense of the transcendent or the infinite (whatever this may come to; I am not here committed to any particular account of what it does come to). Encounters with tables, trees, tax invoices, et cetera, cannot, in this sense, be epiphanies; except of course, when they are also encounters with value.

2. Phenomenology’s Success Conditions: Sincerity, Accuracy, and Significance

Section 1’s manifesto presents a position that I label recognitionalism. Is recognitionalism a systematic ethical or metaethical theory? No. First, the position is not in any sense reductive: it involves no attempt to say “This is all there is, and everything else boils down to this, and whatever else you may think you’ve got hold of, you’re wrong—it’s really this”. On the contrary, the world is huge and chancy. Of course, I think that epiphanies, and the class of value encounters to which they are central, are important in that world and important to ethics: I would not bother writing about them if I did not. Perhaps epiphanies are even, in some senses of an ambiguous word, “fundamental” to ethics. But I have no need to claim that epiphanies are, or recognitionalism is, the only game in town, and I do not.

S1’s recognitionalism is not offered as an argument or as the conclusion of an argument, but as phenomenological description. If the phenomenological description that S1 offers is correct, then that establishes, or helps to establish, the truth of a substantive position in philosophical ethics, one that is part of the larger view about epiphanies and experience that I develop in my book Epiphanies [7]. This position deserves the name recognitionalism, partly because it simply is a view about the ethical importance of recognition, partly in honour of other people who have taken a similar view, such as Axel Honneth: “the stance of empathetic engagement in the world, arising from the experience of the world’s significance and value [Werthaftigkeit], is prior to our acts of detached cognition. A recognitional stance therefore embodies our active and constant assessment of the value that persons or things have in themselves” (Honneth 2006 [8], p. 111, note i).

Arguments succeed when they move by valid steps from true premisses to true conclusions, whereas phenomenological descriptions succeed when they are sincere—when they are offered with a serious attempt at honesty, in good faith, and without conscious ideological bias; when they are accurate—when they capture what our experience is actually like; and when they are significant—when what they sincerely and accurately capture is existentially central.

As to the sincerity of my reporting, all I can do is hope that my readers will give me that epistemic credit.

As to its significance, judge for yourselves. Though I might as well say at this point as at any that one of my ultimate aims is a philosophy that explains what kind of reverence for nature might replace the capitalism of pillage and sacrilege that is our society’s current approach. Given the depth and urgency of the environmental crisis facing us now, I cannot see many philosophical projects or phenomenological reports more significant than that.

As to its accuracy, there are three things I can do. The first of these is the piece of hoping I have already referred to: I hope that what I say is familiar to the reader from their own experience. It had better be—given that phenomenological descriptions are meant, as I say, to be significant—to be important typical descriptions of human experience.

The second thing I can do is produce some more examples that support and illustrate my nine-point recognitionalist phenomenology of value encounters. I do that mostly elsewhere (e.g., in Epiphanies 3.3-3.5); it is not my focus here.

Thirdly, I can say something about an obvious threat to Section 1’s phenomenological sketch, and to the picture of an ethical outlook, based on value experience in general and epiphanies in particular that I want to build. The threat is that even if my nine-point sketch is accurate, it cannot be taken fully literally, because it conflicts in obvious ways with well-established theses in philosophy. Or there again, maybe what the sketch shows is that those theses are not so well-established after all? That, at any rate, will be my own conclusion here. My main focus in the rest of this paper is on a number of alternative possibilities about the metaphysics of experience, and to a possibility that I think has been neglected: this is the view that I call phenomenal socialism.

3. The Admissible Contents of Experience

My being a recognitionalist commits me to the direct availability of the evaluative in our experience. By this I mean not just awareness of the (allegedly) “thin” moral properties of rightness and wrongness, but of “thick” properties like cowardice and glory as well, and indeed of all the abundantly rich variety of evaluative properties that we find in “the War and Peace world”, the world of ordinary human life. So, it commits me to a position in a contemporary debate in the philosophical theory of perception: a view about the admissible contents of experience. I need to be able to explain how value can be part of those admissible contents; in fact, I need to explain the place in perception of what we may call conceptual loading or framing. In this paper, I achieve at least something towards that explanation.

The two most familiar views about the possible contents of perceptual experience are often called, by Bayne 2011 [9], Phenomenal Conservativism and Phenomenal Liberalism.

Phenomenal Conservatism: What we actually, directly, literally perceive is only instances of low-level phenomenal properties.

Phenomenal Liberalism: What we actually, directly, literally perceive is not only instances of low-level phenomenal properties, but also instances of high-level phenomenal properties.

Now, as they stand, I think both these views are obviously false for quite a simple reason, namely, that it is not just things in the category of property that are perceptible to us. I see no grounds for denying that perception is syncategorematic: we perceive things as well as their properties, and we also perceive relations, positions, orientations, and indeed dispositions (see Section 4).

But even if we waive that point, and proceed in the usual way to talk only about perception of properties, Phenomenal Conservatism (PC) and Phenomenal Liberalism (PL) still look wrong to me. PC is roughly the view that Alasdair MacIntyre attacks as the picture of “facts” and “experience” that goes with, and underwrites, traditional empiricism.

‘Fact’ is in modern culture a folk concept with an aristocratic ancestry. When Lord Chancellor Bacon... enjoined his followers to abjure speculation and collect facts, he was immediately understood by such as John Aubrey to have identified facts as collectors’ items, to be gathered in with the same enthusiasm that at other times has informed the collection of Spode china or the numbers of railway engines. The other early members of the Royal Society recognised very clearly that, whatever Aubrey was doing, it was not natural science as the rest of them understood it; but they did not recognise that, on the whole, it was he rather than they who was being faithful to the letter of Bacon’s inductivism. Aubrey’s error was… not only to suppose that the natural scientist is a kind of magpie; it was also to suppose that the observer can confront a fact face-to-face without any theoretical interpretation interposing itself.

(MacIntyre 1981 [10], p. 79)

Phenomenal Conservatism is the natural home of theories of impressions and sense-datum theories, and of the appeal to “in reality all we ever really see…” that is the keynote of so many forms of empiricism. Phenomenal conservatives think that some relatively small and/or modest ranges of phenomenal properties are the real or basic ones, the “low-level” ones, and that it is these low-level properties that should be the basis of our theory of phenomenal knowledge and content. Thus, PC is characteristically (to use Bayne’s word as quoted below) an austere account of what is directly available to us in perceptual awareness—and we all know how analytical philosophers love austerity in their theories. By contrast, Phenomenal Liberalism claims that the range of phenomenal properties that are directly available to us to perceive is wider and more generous, and includes not only the “low-level” properties but the “high-level” ones too.

Everything turns, of course, on what PC and PL mean by two phrases in particular. One is (as I put it; there are plenty of similar formulations in the literature) “actually, directly, literally perceive”; the other is “low-level” (and “high-level”). The first phrase is supposed to mark a distinction between perception strictly so called on the one hand, and on the other, either figurative3 perception or inference (perhaps very quick, perhaps subconscious). Seeing that the apple is red is, as the usual story goes, literal perception, involving the specific sensory modality of vision. By contrast, seeing that the lecture is going to last a long time is (we are told) only figurative perception, involving, quite possibly, all sorts of different epistemic inputs, some of them deductive or to do with, say, having previous knowledge of the lecturer, but only some of them sensory—and literal vision might not be part of this “seeing” at all. (I might be listening to a lecture that begins with the words “I will make six points today”, and after an hour the lecturer says “And now for my second point”, and so I say to myself “I see this will take all day”.)

As for the distinction between “low-level” and “high-level”, here is Bayne 2011 [9], pp. 16–17’s exposition:

[PC holds] that the phenomenal character of visual experience is exhausted by the representation of low-level properties—colour, shape, spatial location, motion, and so on. [Likewise with hearing,] the phenomenal character of audition is exhausted by the representation of volume, pitch, timbre, and so on; the phenomenal character of gustation is exhausted by the representation of sweetness, sourness, and so on. The phenomenal world of the conservative is an austere one… [By contrast, according to PL] the phenomenal character of perception can include the representation of categorical4 (“high-level”) properties, such as being a tomato. We perceive objects and events as belonging to various high-level kinds, and this, the liberal holds, is part of perception’s phenomenal character. What it is like to see a tomato, taste a strawberry, or hear a trumpet is not limited to the representation of “low-level” sensory qualities but involves the representation of such “high-level” properties as being a tomato, a strawberry, or a trumpet.

This passage from Bayne should raise immediate questions. One of the most obvious is that Bayne seems, in fact, to have given us an incomplete list of alternatives. About whether high-level and low-level phenomenal properties are, in themselves, available to perception, there are not just the two possible positions that Bayne’s list labels as phenomenal conservatism and phenomenal liberalism. There are at least four alternatives that are obvious even before we start doing less obvious things like, e.g., positing more than two levels of phenomenal property. Here they are in

Table 1. Four alternatives.

	Low-Level Phenomenal Properties	High-Level Phenomenal Properties
Directly available to perception?	No	No	Phenomenalnihilism5
Directly available to perception?	Yes	No	Phenomenal conservatism
Directly available to perception?	Yes	Yes	Phenomenal liberalism
Directly available to perception?	No	Yes	Phenomenal socialism

:

There is the option of allowing only low-level phenomenal properties to figure into perception (phenomenal conservatism) and there is the option of allowing both low-level and high-level phenomenal properties to figure into perception (phenomenal liberalism). But there are two further options. The third is to allow neither high-level nor low-level phenomenal properties to figure into perception, which the table calls phenomenal nihilism. And the fourth is the option of allowing only high-level phenomenal properties to figure into perception: what we actually, directly, literally perceive is only instances of high-level phenomenal properties. This last is a position that, for more than one reason, I shall call phenomenal socialism, and I shall be defending it (or something close to it. In fact, I am at best a wishy-washy phenomenal socialist because I argue the position with a moderation of “only” to “primarily”; my view is that what we actually, directly, literally perceive is primarily instances of high-level phenomenal properties).

My table shows that these two further options are in-principle theoretical possibilities. Does either have any real attraction? No doubt nihilism is the least attractive of the four possibilities, but it is not entirely impossible to see ways of motivating it. Someone might be a phenomenal nihilist because, for example, they do not believe in phenomenal properties, or because they believe that no perception of properties is ever really direct, or because they believe that all perception is illusion, or because they are eliminativists about subjective experience. None of these alternatives interests me particularly (though Daniel Dennett, for example, seems at different times to flirt with all four) but that is not the point. The point is that phenomenal nihilism in all these varieties, and possibly others too, is on the menu.

So, what about phenomenal socialism, the view (in my wishy-washy version) that we have direct access primarily to high-level phenomenal properties (HLPPs) and not to low-level ones (LLPPs)? The obvious objection to that, from a liberal or conservative viewpoint, is likely to be that we can get at the high-level properties only via the low-level properties: access to HLPPs has access to LLPPs as a necessary condition, or supervenes on it, or something like that. But for reasons we will come to shortly, I do not think this objection has any real power at all. And without that objection, I see no serious argument, on the whole, against my mild form of phenomenal socialism.

But then, what argument is there in favour of it? My main argument for my wishy-washy form of phenomenal socialism—and here we see the second reason for the name—has to do with the essentially social way in which human beings learn to perceive in the first place. The low-level features that are fundamental to phenomenal perception, says Tye 1995 [11], p. 141, are things like “being an edge, being a corner, being square, being red”. How is this supposed to work as an analysis of what is basic to our experience in perception? Human beings are not born as geometricians or colour scientists, and only later learn to be babies. They begin as babies, and babies are not interested in geometrical properties or colour properties; not at any rate primarily, and not at any rate at first.

What a baby is interested in is human relationships, and in particular, their relationship with their mother or other first carer. Accordingly, the first visual object that a typical baby learns to recognise after birth is not a geometrical feature or a colour sample. It is the human face. Likewise, the first auditory object that they learns to recognise (they have already done this, in fact, before birth) is certainly their mother’s voice, and probably the voices of others who are around a lot as well—their father and siblings, perhaps. And the first smells that they learn to recognise—after birth, since smell presupposes breathing—will be their mother’s smell, the typical smell(s) of wherever they and their carers live, and of course the smell of milk.

So, the phenomenal properties that are chronologically primary for humans—the ones that humans, as such, typically learn first—are usually and centrally and naturally high-level phenomenal properties, HLPPs. (If babies learned a LLPP early, that would be a kind of accident; it could happen, but it would happen because that LLPP happened to coincide with something that interests babies.) Not only that, they are a particular kind of HLPP; I shall call them collectively social ones. Specifically, the social properties are those HLPPs that are salient in the socialisation of typical babies, and these are the phenomenal properties that teach the baby about their sources of care and food and the other human beings around them.

Phenomenal socialism is, as I defend it, the view that we have direct access primarily to HLPPs, not to LLPPs. What we can now see is that the main reason to be a phenomenal socialist comes from familiar facts about ordinary human developmental psychology (Scheler 1954 [12], pp. 260–261):

…we certainly believe ourselves to be directly acquainted with another person’s joy in his laughter, with his sorrow and pain in his tears, with his shame in his blushing, with his entreaty in his outstretched hands… and with the tenor of his thoughts in the sound of his words. If anyone tells me that this is not “perception”, for it cannot be so, in view of the fact that a perception is simply a “complex of physical sensations”… I would beg him to turn aside from such questionable theories and address himself to the phenomenological facts.

Here are three quick objections to this argument:

• The argument confuses developmental psychology with metaphysical explanation. A story about how babies learn to perceive phenomenal properties, and which phenomenal properties they learn to perceive first, is something different from a story about which phenomenal properties are metaphysically basic. Understanding how human minds develop over time is not the same thing as knowing how perception works, metaphysically speaking, at any one moment.
• The perceived properties of the face, the voice of the mother, the smell of milk, etc., are not really phenomenal properties because the baby does not really perceive them, not directly anyway; what they do is directly perceive LLPPs, and learn to infer the mother’s presence or the presence of milk from grouping those low-level properties into high-level properties.
• Even if the baby does directly perceive the HLPPs, they can only do this on the basis of and in virtue of perceiving the low-level phenomenal properties because the perception of the high-level properties supervenes on the perception of the low-level properties.

It seems to me, what is immediately striking about these three objections is their obvious weakness and question-beggingness. The first objection complains about a “confusion” of the roles of developmental psychology and of metaphysical explanations in our theory of phenomenal perception. But the whole point of phenomenal socialism, as I have just presented it, is to suggest a way that developmental psychology might be a metaphysical explanation, or at any rate be crucially relevant to it, in the case of phenomenal perception. To talk of confusion here is simply to beg the question against that proposal.

The second objection insists that the baby does not really perceive the social properties, the HLPPs associated with the mother’s face and voice and smell, the smell of milk, etc. They must, according to this objection, be inferring or constructing these HLPPs on the basis of some more direct perception of LLPPs. But why be dogmatic about this? That the baby can and does directly access the HLPPs certainly seems to be what happens. Why be so desperate to deny this appearance? (Perhaps because it conflicts with your theory of causation? But that is exactly my point, only tollensed: that a theory of causation that entails the falsehood of anything except phenomenal conservatism is ipso facto, a suspect one is just what I am saying.)

There is not even a clear argument for denying that the baby has direct access to HLPPs unless the third objection is viable. But actually, it very obviously is not viable. It is a well-known datum of developmental psychology that the baby’s perception of the HLPPs does not supervene on their perception of the LLPPs. Consider more closely the best known example of all, an example I have already been using. As any parent can tell you, babies track faces; faces as such, not the subvening properties of faces. The gestalt of a visible face can and does undergo all sorts of changes in the properties that constitute it, and still remains the same gestalt: it is seen from different angles; it gets red with exertion or tanned with the sun; it looks different in ordinary daylight, inside a neon-lit supermarket, and in the low light of the bedtime nursery; it acquires new wrinkles or loses old pimples; bits of it disappear behind sunglasses or makeup or new hairstyles or facial hair; and so on. These are the subvening properties of the face and they change all the time. The baby’s perception of the face may possibly supervene on some smaller set of the most salient properties of the face, but not even that much is clear, and even if it does, the identity of this smaller set has very fuzzy edges. Babies, even newborn ones, will “lock on” to very simple representations of faces, like this one 😊, but they are strikingly unfussy about how such representations are composed, i.e., about the precise constituent lower-level properties of such representations.

So, it is very clear that in the case of babies’ facial recognition capacity, there is at the very least massive “multiple realisation” of the lower-order properties that a baby will take to be constitutive of representations of faces. But to say this, and to admit also, as we rather have to, that there is no way of closing the disjunctive range of possible alternative lower-order properties that the baby will see in this way, is to wave goodbye to any clear notion at all of the supervenience of the higher-order properties on the lower ones. And if this argument generalises—as it surely does—from facial recognition to the recognition of other social HLPPs, then the third objection against phenomenal socialism fails, and at the same time, the case for phenomenal socialism becomes a lot stronger. So quite generally, it seems obvious that humans, in their development from birth (and indeed before), very often perceive HLPPs as the primary objects of their perception, and without any prior dependence on LLPPs; in fact, the very idea of LLPPs cannot necessarily come first in the order of psychological development. And, to state this conclusion is just to state (my moderate version of) the thesis of phenomenal socialism.

With these points in mind, perhaps we can now see how to handle a fourth—and more considerable—objection to any position that says that we can and do directly perceive high-level phenomenal properties, such as phenomenal liberalism and phenomenal socialism. Suppose someone first sees only the lower-level components of some high-level phenomenal property or object, such as a face, or a tune, or an inscription of the name SOPHIE GRACE. Then, by whatever kind of aspect or gestalt shift, they come to see the high-level phenomenal property too. That is, they come to see the relevant shapes as a face, or to hear the component sounds as a tune, or to see some squiggles on a page as the written words SOPHIE GRACE. If you like, they acquire a capacity to parse or “read” what is in front of them: in the case of the writing, literally a capacity to read it. But surely, whatever kind of change in the spectator there may be in such cases, it cannot be a change in what they perceive, only in how they interpret it since, according to the hypothesis, the low-level phenomenal properties remain unchanged throughout.

That last inference, of course, presupposes something that I have already rejected, viz. the assumption that the HLPPs must supervene on the LLPPs. If there is (as I have argued) no such supervenience, then it is not true that there can be no change or continuity in the HLPPs without some underlying change or continuity in the LLPPs. So, as pointed out above, we can perceive the same HLPPs and wildly different LLPPs, e.g., the face that the baby tracks as the same face can undergo all sorts of changes in different lighting conditions and details of its appearance. Or we can hear the very same tune “Flower of Scotland” played on different instruments in different keys and with different tones, and yet hear it as the same tune. Or the letters in the inscription of SOPHIE GRACE can change colour or font or brightness while we are learning to read them, without that necessarily impeding the learning process. Conversely, I now add, we can perceive the very same LLPPs and yet come to perceive quite different HLPPs. And that is exactly what happens when we learn to “read” a face, or to read an inscription, or to see a duck or a rabbit as a rabbit or a duck (or as a duck–rabbit).

An obstinately resistant sense remains that this cannot be a matter of a change in our perception. After all, it will be said, in cases like these, at any given moment, the exact same low-level phenomenal properties are presented to the reader and the non-reader. So surely, their perceptual experiences are identical at that time. If there is a difference between them, and evidently there is, it cannot be in their perceptual experience; it must be in how they interpret that experience.

This obstinate resistance is, of course, still begging the question against the thesis of phenomenal liberalism or socialism. The claim at issue is precisely that there can be changes in HLPPs without changes in LLPPs, and it is no argument against that claim simply to say “surely not”. Nor is there any force to the suggestion that we cannot imagine what it would be like for us to experience a change in HLPPs without a change in LLPPs. Certainly, we can imagine this—it is what we experience when we see a gestalt shift.

That said, it is perhaps a more constructive response to this “obstinate resistance” to offer two further moves. The first picks up the theme of conceptual framing or loading that has been bubbling under throughout this paper: it is to question the hard line that is being drawn between perception and interpretation. If Sellars [13] was right in his famous attack on the “myth of the given”, then perceiving is—essentially and always—interpreting the world, and there is no such thing as perception without interpretation. So, to change our interpretation of what we perceive is also to change what we perceive (more about this below).

Secondly, we might also say that a characteristically Humean error seems to be motivating the resistance. This Humean error—we might even call it a Zenonian error—is the fallacy of the time slice: the fallacy of assuming that the basic unit for metaphysical analysis is always and only the instant. Pick any instant and at that instant, we will be unable to see any difference in perception between the reader, the person who sees the words SOPHIE GRACE, and the non-reader, who only sees the shapes of the words (if even those). But all time is composed out of instants. So, to say that there is no difference in perception between the reader and the non-reader at any instant is to say that there is no difference between them at any time. And, to say there is no difference between them at any time is to say that there is no difference between them at all.

Now it is not just in this case that the time-slice approach leads us to sceptical conclusions. As Hume himself gleefully pointed out, the time-slice approach also wreaks sceptical havoc on the commonsense intuition that we use to perceive causal connections. Hume could equally well have pointed out that it wreaks sceptical havoc, too, on the commonsense intuition that things move (actually move, as opposed to showing up in one place at one instant, and another place at another instant), which is why I call the time-slice approach Zenonian. Most fundamentally, Hume saw that his time-slice approach implied scepticism about the existence of substance—about the existence of objects that have properties and endure through time.

Hume apparently did not mind sceptical conclusions—or if he did mind, he took the natural indolence of the human understanding to be a sufficient remedy to them. He never seems to have taken seriously the possibility that these direly sceptical consequences of his time-slice approach give us reason to reject the whole approach. But, we should take it seriously. For at least some purposes of metaphysical analysis, including the present ones about perception, causation, and substance, we should say that the basic unit is not the instant but the (suitable-length) time span, the specious present [14].6

For perception, to stick with that case, is not (just) a state; it is (also) a process: “[C]ognition is not the representation of a pre-given world by a pre-given mind but is rather the enactment of a world and a mind on the basis of a history of the variety of actions that a being in the world performs” (Varela, Thompson, Rosch 1992 [15], p. 9).

Accordingly, the main difference between a high-level perception, such as reading SOPHIE GRACE, and a low-level perception, such as seeing only the shapes of those words (or perhaps even only the visual arrays in which those shapes are present), will be a difference not between two states, captured in snapshots at some instant, but between two processes, “filmed” (as it were) as they unfold in real time. Then, the differences between the reader and the non-reader will be differences about what is salient to the two observers, and hence differences about what properties these two observers track over time. Once they have noticed them, the reader latches onto the inscription of the two words, and looks in the context for the endurance-through-change of that inscription, perhaps for the appearance of further messages—that kind of thing. The non-reader, by contrast, simply takes in an array of shapes (or just an array, in which they do not even notice shapes). If some part of that array is particularly salient to them, it will not (ex hypothesi) be the words. They as a non-reader will not track the words’ persistence, or their failure to persist, over time; their attention will be elsewhere.

So, what we have seen is that phenomenal conservatives insist that only low-level phenomenal properties can be directly perceived. But this is not true; high-level phenomenal properties can be directly perceived too, and in fact are the first objects of our direct perception. If, as I want to claim, evaluative properties are high-level phenomenal properties too, then they too can be directly perceived (synaesthetically, not via a special sense).

4. A Lightning Raid on Modal Epistemology: What It Takes to Build a Wall

These differences about what the reader and the non-reader actually track in the visual array in front of them are paralleled, too, by differences in what they would track, if the conditions were otherwise. So, it is not just that the reader’s perception is an activity of tracking some particular salient features of the array in front of them, rather than a bare registering of that array. Besides being such an activity, their perception is also dispositionally loaded. If the array were to change in some given respect, then their attention would shift in this or that way in response to that change, and in line with the aspect of the array that interests them—namely, the words that they sees in it. As it will seem from their point of view, the pattern that interests them in the array in front of the, is a pattern that has actual properties, and it has modal properties too. It is part of what they perceives as the pattern changing, and it is also part of what they perceives could change, and that if it does, then they are dispositionally ready to follow its changes—and remains ready to do that, even if it does not actually happen [16].7

Our perceptions of ordinary objects in the world are, I am wanting to suggest, modally loaded in just this sort of way. A rock balanced on a clifftop can look precariously poised to us—and can look that way to us as a matter of immediate perception. A cheetah on the savannah can look lithe and fast to us—and can look that way to us even when this lithe fast creature is sleeping straddled across an old tree branch. (Such looks can of course be deceptive, but normally—at least in cases where we are suitably skilled observers—they are not) [17]8.

Something like this seems to happen too with competent readers and speakers of inflected languages. When a francophone hears nous avons, they hear the possibility of vous avez and j’ai “in the background” along with nous avons; when a well-trained classicist reads the classical Greek word μέλῐ (meli, “honey”), they stack “inside” that one word μέλῐ, as possibilities not yet realised, but immediately present in it, all the other cases and numbers of the declension of μέλῐ—μέλῐτος (melitos) (genitive), μέλῐτῐ (meliti) (dative), and so on. Their perceptual experience of this simple disyllable will then be quite different from someone who can read Greek letters, but knows no classical Greek grammar, and different again will be the auditory case where we contrast a competent classicist hearing the word μέλῐ, with someone hearing it who does not know that it is a Greek word (or even a word). Modal epistemology, just like the rest of epistemology, begins with the senses; for the competent perceiver, the modal possibilities all come together as part and parcel of the perception of the word. (If “metaphysics is the shadow of grammar”, perhaps the inflections of classical Greek and Latin are the psychological background that make the theory of innate dispositions come so naturally to Aristotelian or Thomist philosophers.)

The case of our perception of modal properties helps us to see how little clarity or stability there really is to the distinctions that are supposed to exclude it as a possibility—and how closely and philosophically linked the exclusion of modal properties is to the exclusion of moral properties. Both exclusions have their roots in a Humean insistence that at some basic level “all we ever really see” is the time slice of bare sensational inputs currently in front of us. The basic reality is, in honour of Christopher Isherwood, an I-am-a-camera world, in which my perception’s role is simply to take snapshots of what is in front of me. The cure to both kinds of scepticism is to insist that what we really see is things in process and things moving (or looking apt to move) in our environment—where we ourselves are agents in motion, with agendas and practical concerns of our own, and where our notion of “things” is tied to our social concerns, and where the “things” as we typically conceptualise them are sufficiently like Aristotelian substances to have natures, in the sense that some possibilities are natural for them and others, though still possibilities, are not. (An acorn can turn into an oak tree, and an acorn can turn into a pig. But these are very different kinds of “can”. The idea that all possibilities are more or less equally possible is a deep mistake, but it follows fairly readily from the thought that “all things in themselves are entirely loose and disconnected”.)

Developing my parallel between evaluative and modal epistemology a little further will also bring to light the pervasive role of the imagination in human thought. I am not much of a fan of just-so genealogical stories about how humanly important things first came to be, especially if they are supposed to be, or to suggest complete and unique explanations. But within those limits, just-so stories have their uses sometimes; this seems to be one of those times.

Consider then, a basic human situation, the rather Wittgensteinian situation of wall building (or house building, if you are averse to the building of walls). The builder wants to know if this slab here will fit in that gap there. The slab is heavy, and there are many similar slabs, so it is costly to try and fit it in by actual manipulation. So instead, the builder does some “spatial reasoning”. They visualise lifting the slab up, manipulating it, turning it, and adjusting it; They takes a mind’s eye tour around its three-dimensional shape, and also around the shape of the gap it is meant to go in. In short, they imagine fitting the slab into the gap.

This imaginative exercise is low-cost compared with what it simulates. It is constrained in at least three ways. First, imagining for the builder is only good if it gets the right answer to the simple question “Will the slab fit in there or will not it?” Secondly, the imagining needs to obtain this right answer in an as fruitful and illuminating a way as can be, shedding light on more general questions about the shape and fit and building as wider concepts. Thirdly, the imagining needs to appropriately deal with disruptive objections—that is, with objections to the whole procedure as envisaged.

(Examples of disruptive objections: “But the wall is made of plywood and this slab is made of granite—so it’ll fit all right, but it’ll bring the whole wall down”; or “But you’re not really moving the slab, you’re just pretending to, so you can’t really tell”; or “Building is all very well, but you need to go and hunt musk-oxen now”; or “Why not put the block in that gap over there instead? It would fit there too, and you wouldn’t have to risk doing your back in.” Such objections typically involve refusing to take the question as given; that is why I call them disruptive. They can be anything from utterly obtuse to total eureka moments (indeed epiphanies), revelatory exercises of lateral thinking. Disruptive objections are key to practical reasoning in general, and in particular, to what is often called moral reasoning. But more about them another time (see, e.g., Chappell 2015 [18] Chapter 2, on “closed” versus “open” models of deliberation).

Disruptive objections aside, what the builder imagines is short sequences of movements of the slab, and of the slab as positioned here or there and then rotated through this arc or the other, the smoothness of the surface onto which the slab is supposed to be placed, etc. This is diachronic imaging. It is spatial imagining: its bodiliness is essential to it even though that bodiliness is virtual. And it is modal imagining: it is imagining what can be done with the slab and how it would be if we tried to put it here or there. In another vocabulary, it is becoming aware of, and indeed perceiving, affordances.

Perhaps there lies in such thoughts, the germ of an account of imagination in general (as opposed to specifically artistic imagination), and a just-so story about where it might have come from, evolutionarily speaking [19].9

As I said, I do not in general believe in just-so stories in philosophy as a way of explaining a concept we have by giving it a genealogy, and not at any rate if they are supposed to be uniquely true. Life is complicated and few things of any social, practical, or ethical importance have simple origins. But the above account perhaps captures a truth, or two, about imagination, even if it is unlikely to be the truth. Imagination may well be other things besides modal exploring; but, modal exploring is certainly one important thing that imagination is.

As well as being an embryonic account of what imagination is, this thought experiment of the builders is also a sketch for an approach to modal epistemology, and indeed, to general epistemology. My argument in this paper supports a lesson for general epistemology that I was already gesturing towards when, at the beginning of my discussion of Phenomenal Conservatism and Liberalism, I objected to the shared assumption of both positions, that perception is exclusively of properties. In my discussion so far, I have allowed that assumption to go unchallenged for the sake of argument. But it is time to come back to it, and say a bit more about why I reject it.

My slogan in the philosophy of perception is objects first. I have argued elsewhere (in Chappell 2015 [18] Chapter 11) for a parallel claim about knowledge: that, despite a huge and powerful consensus in contemporary philosophy, the first object of our knowledge is not propositions. Nor is it ways of doing things. Nor even what-it-is-like-nesses. The first object of epistemology is objects: things as known to us in our interactions with them (things as items of Zuhandenheit first and of Vorhandenheit, if at all, only later) and the transformations and potentialities of things: what they are, what powers or dispositions they have, and how what they are dictates what we can do with them. (There are of course links here with enactivist and embodied cognition approaches to mind here like those of Andy Clark [16] and Dan Hutto [20] and Dan Zahavi [21] and Alva Noe [22] and Michael Wheeler [23].) And so, likewise, I want to suggest, with perception. Perception is a syncategorematic capacity—we can perceive things in all the different Aristotelian categories (or whatever other table of categories we accept)—but the first thing that perception is directed towards is, indeed, things—objects—in the world. In this sense, perception and knowledge stand, in an obvious way, in close parallel.

As for the lesson for modal epistemology, it is this: the first things we learn about are things, and to learn about things is to learn what things can do and what can be done with things. This knowledge is necessarily and inherently diachronic, and it is necessarily and inherently modal.

This is why a lot of inquiry in modal epistemology gives me the same feeling as a lot of inquiry in “moral epistemology” (as it is usually called, though I would prefer to call it by the more inclusive and less misleading name “evaluative epistemology”). In both inquiries, the usual initial question is systematically misleading. “How in the world is it possible for us to acquire knowledge of modal truths? Mustn’t they be deeply spooky entities?” Such a metaphysical othering of the modal only makes sense if you presuppose that something like the Humean I-am-a-camera world makes sense first, and then, we have to understand modal truths on top of that world. But that presupposition is wrong; the Humean world does not make sense as the first truth about reality but the objects-first story does. And as I said, it brings modality along with it from the start: what the builder does in visualising what can be done with reorientations of a slab that they are trying to fit into a wall is a key application of objectual knowledge. It rests crucially upon spatial awareness; but, spatial awareness is already modal knowledge, because shapes—even shapes!—are already dispositional properties, at least, as soon as you think of shapes practically and diachronically. As Husserl and Bergson both emphasised (in their different ways), thought is practical and diachronic or it is not thought at all. But if thought is practical and diachronic, then it is ipso facto modal too, and value-involving as well.

I am a (mild) phenomenal socialist, as well as some other kinds of socialist, because, for human beings, the social comes first. For us, perception begins with the social properties (and entities): the properties and entities that are salient for a neonate, an infant, or an otherwise young human being—above all, in the most usual case, mother. Moreover, I have suggested, saying that these properties and entities are developmentally primary need not and should not be sharply separated from saying that they are explanatorily primary. These social properties and entities are a subset of the properties and entities that the literature usually picks out as the “high-level” phenomenal properties.

(Perhaps this is a terminology that we should reconsider, incidentally. As a matter of actual human development, there is nothing particularly “low-level” about the phenomenal properties that usually go by that name. It actually takes a particular kind of sophistication even to notice most of them: “painterly” sophistication in the case of simple colour samples and geometrician’s sophistication in the case of, e.g., “being an edge”. Perspective drawing’s notion of how things “actually appear to us” is a foreign one to most human minds: millennia and centuries passed before any society got hold of the idea of drawing, e.g., a square object seen from below as, “literally”, a rhombus on the canvas, in line with that notion. According to phenomenal socialism, it is not with such allegedly basic phenomenal properties that our perception of the world begins, and it is not in terms of them that we should seek an understanding of perception or of its metaphysics—it is with the social phenomenal properties.)

Phenomenal socialism, as I have now developed it, gets us all the way to direct perceptual realism about both the modal and the evaluative—provided that, along the way, we accept two moves in particular that need to be emphasised.

The first of these moves is the suggestion that we are directly aware, by “figurative seeing”, of ethically relevant properties like arrogance, kindness, humiliation, grace, and so on; these sorts of properties are everywhere in the world of ordinary life, and it is possible for us to perceive them accurately as features of situations that we can find ourselves in. This of course is (as I call it) figurative or synaesthetic perception—it is not the sort of perception that is associate with any single sensory modality, but rather the kind of perception that is involved in “I see your determination to get this job”.

The second move to note here is the denial that there is any clear boundary between sensory and figurative perception (or indeed between them and one of the main topics of this section: imagination). For, as we might put it, the whole point of sensory perception is to achieve figurative perception; that is to say, the whole point of using the senses to become aware of our situation is to become aware of our situation. The ultimate aim of sensory perception is knowledge of and understanding the world around us. How directly this happens is not merely given, but also a matter of our capacities: we can learn to grasp truths about our environment directly and non-inferentially, straight from the deliverances of our senses. We can also learn to grasp truths about how our environment can be, and what ways we might change it, by applying our imaginations to those deliverances. What the senses tell us, then, is a series of contributions to overall understanding: sensory awareness is, in the end, just one kind of route to awareness.

Perceptual consciousness is a special style of access to the world. But access is not something bare, brute, or found. The ground of access is our possession of knowledge, understanding, and skills. Without understanding, there is no access and so no perception.

(Noe 2009 [22], p. 105)

There is no deep categorial gap between the kind of perception that makes us aware of a piece of writing in our visual field, and the kind that makes us aware that someone is sad or angry or behaving embarrassingly. The former typically involves just one sensory modality, the latter typically involves synaesthesia—often of a quite general kind—that brings into play a whole range of sensory modalities and other ways of accessing particular kinds of knowledge. But both are perceptual routes to knowledge. If we can see an inscription of the words SOPHIE GRACE by accessing high-level phenomenal properties, then we can also see an instance of bullying by accessing high-level phenomenal properties. Now, as I have argued in this paper, the world of ordinary life is full of high-level phenomenal properties of this latter kind; in ethics, they (or some of them) are often called the thick properties. Such properties—and our direct experience of them—are the working materials of ordinary ethical consciousness, and they are part of what makes it as involved with reality as any other kind of consciousness.10