What S in a Look

M.G.F. Martin

What’s in a Look?

This paper is concerned with the superficial. Its main concern is with the looks of things, and with how we manage to speak of these through talk of how things look. But in addition, the picture I want to sketch both of our talk and of what we talk about aims to be more minimalist and parsimonious than many discussions of these matters.[1]

One of the most detailed, and influential, discussions of the semantics of looks statements is to be found in the first few chapters of Frank Jackson’s Perception: A Representative Theory. Jackson hopes to convince us that we should endorse a sense-datum theory of vision through proper attention to the logical form of looks statements. So he hopes to find in our lexical knowledge confirmation of a substantive and controversial view of the nature of visual experience. Some more recent discussions reflect similar ambitions: it is becoming common to note that we mark a difference between how things look to us and how they are; and this contrast has been taken to indicate our commitment to representational nature of sense experience. That is, some have supposed that the correct account of our looks talk reveals a commitment to viewing sensory states as representational states or even propositional attitudes.[2]

In contrast to both of these approaches, I will propose a minimalist approach to the semantics of looks-statements. On the whole, we convey information to each other both about the looks of objects, and about sensory episodes in which things look some way or other to us through the use of implicitly comparative claims: that the object, or the sensory state, is relevantly similar to some other, paradigm case. That we succeed in conveying information to each other through such talk presupposes that we have more knowledge of looks and visual experience than is explicitly required to grasp the form of the statements we make. On such a minimalist approach, it would be mistaken to look to our semantic competence to argue in favour of one substantive theory of sense experience over any other.

Such minimalism in the semantics of looks talk, then, puts more pressure on our knowledge of the nature of looks, and of states of things looking some way to one. In turn, here, too, I propose we adopt a strategy of parsimony. Various recent discussions of appearances in general, and looks in particular, have been moved by concerns with conflicting appearances to posit a range of properties in the world as the looks of things over and above the other properties that we are committed to supposing objects have, such as their shapes and colours. According to these accounts we need to recognize such things as perspectival shapes (‘oval from this angle’, for example) and apparent colours to play the role of the looks of things through which we come to experience their colours and shapes. In contrast, the parsimonious view of looks that I will sketch proposes that we identify the looks of objects with their basic visible properties, including their colours and shapes.[3] Accommodating within our ontology the looks of objects does not require that we posit additional features of these objects over and above those properties we are otherwise committed to supposing them to have through what we can know of them through perception.

This leads to a final introductory comment about the approach to be sketched here. A common strategy in discussing the looks of things is immediately to offer an account of corresponding psychological states: the visual experiences that we enjoy in looking at things. In offering semantic theories, this has led writers to propose accounts of looks-statements which take as fundamental ascriptions used to talk about psychological states. Both Jackson and Tye focus on giving accounts of sentences of the form ‘O looks F to S’, rather than the seemingly simpler form ‘O looks F’. Connected with this are at least two lines of thought. The first is that our talk of appearances in general and looks in particular exhibit a subjectivity of use which parallels avowals of psychological states, so one might take this as an indicator that this is best explained by supposing that at root such talk avows the presence of relevant psychological states. Secondly, one may suspect that with such elements as looks, philosophers are inclined to accept a phenomenalism which they are more likely to reject for what they take to be the intrinsic properties of physical objects: namely, that facts about the looks of objects are constituted by facts about actual or possible visual experiences one may have of them. Although the assumption of phenomenalism seems widespread, and with it the assumed priority of statements about how things look to x over statements about how things look, the account here reverses the order. We will not assume that phenomenalism is true, and we will seek a sketch of the meaning and use of statements of the form ‘O looks F’ in advance of what will be taken to be semantically more complex, ‘O looks F to S’ or ‘O looks F from l’. I won’t here directly argue against the more common assumption, but rather indirectly launch a challenge through indicating how the reverse order is perfectly coherent and gives us a clear picture of what the looks of objects are, and how our talk of them may be subjective without having to assume phenomenalism.[4]

The first part of this sketch focuses principally on questions about the semantics of looks talk. I take off from Chisholm’s and Jackson’s proposals that there are three uses and indeed three senses of looks statement: the epistemic, comparative and non-comparative or phenomenal. In the discussion which follows I’ll broadly concur with this idea, at least with respect to how we can talk in English about the looks of things. But, I shall argue, Jackson is mistaken to suppose that a phenomenal sense of looks statements reveals much about the looks of the ordinary objects that we perceive; instead I will conclude that our talk about how things look red or look oval, examples that Jackson takes to be paradigms of phenomenal talk, instead to be best understood as comparative uses.

As I will explain, this indicates that our ability to discuss informatively the looks of objects around us presupposes common knowledge of how things look directly (i.e. without reference simply to how similar or different they are in look from other things). In turn, then, an account of our talk of looks needs to give some account of what the nature of looks might be such that ordinary speakers who have knowledge of the ways things are through ordinary perceptual capacities of sight thereby are also in a position to know about those things’ looks as well. In the second half of the paper I will sketch an account which shows why we should favour a parsimonious view of looks that identifies them with (logical constructions out of) manifest visible properties such as shape and colour. In this discussion I will explore how such manifest visible properties can succeed in playing the epistemological role of looks and make true statements about how things look. Finally I will explain how a challenge from the ancient problem of conflicting appearances together with the pragmatic marker of the subjectivity of much looks talk is consistent with supposing that ways of looking that objects have are among their intrinsic properties and not relations to aspects of their environment or to perceivers, their psychological states, or perceptual capacities.

PART ONE

1. On the surface there is a wide variety of ways of characterizing how something looks: one can say that something looked red, or small, or nearly square; but equally one can say of someone that they look ready for the Queen’s Bench, or in need of a good beating. Looking at an unkempt office one may say that it looks as though a hurricane has hit it; staring at the edge of a design one may say that it looks straight to me. One can say of someone that they look like a Russian shot-putter, or like their mother. Given the diversity in talk of looks and similarity of looks that are permitted linguistically, one might suppose that there is just too much disunity here for anything in general to be said about visual appearances.

It is no surprise, therefore, to find philosophers seeking to impose order by suggesting different uses or senses that we can put looks talk to, some and not other such uses cleaving to the nature of visual appearances as such. The two most extensive discussions of the semantics of appearance talk in English-speaking philosophy occur in Roderick Chisholm’s Perception, from 1959, and nearly twenty years later Frank Jackson in Perception: A Representative Theory. According to both authors, we should distinguish three principal uses, and indeed senses, of looks statements: epistemic, comparative, and non-comparative or phenomenal uses. In the case of Chisholm and Jackson, the intention is particularly to focus on our concerns on the non-comparative or phenomenal sense of looks statements in order better to read off from this our knowledge of the nature of the looks of things. Some critics have expressed scepticism of whether there is any variation in sense here, and in particular whether there is any phenomenal sense of appearance talk.[5]

In what follows, I shall argue that there is something broadly right about Chisholm’s and Jackson’s proposals. But this will only give them half of what they want, though. Although I think we can pretty much demonstrate that there is structural semantic ambiguity between phenomenal and comparative looks statements, there are no true phenomenal looks statements which concern things looking red or looking square, the kinds of examples which Jackson takes to be a paradigm of the phenomenal.

2. Although there is an application for Chisholm’s and Jackson’s three-fold distinction of uses (and, in fact, senses), we need first to embed it in a two-fold division noted by the linguist Nikolas Gisborne, and echoed in Charles Travis’s writings on looks.[6] For one thing we can do with looks talk is to put into play a proposition, indicating that there is visual evidence for it. One may say, ‘It looks like it is going to rain’, avoiding thereby committing oneself to an outright assertion of the proposition that it is going to rain, but still putting the proposition into play in the conversational context. One can do the same with simpler sentences, such as, commenting on a neighbour, saying, ‘She looks pregnant’, thereby offering the proposition that she is pregnant for consideration by one’s interlocutors.

In this evidential use with a sentence of the form ‘o is F’ the focus of conversational concern is on the sentence ‘o is F’. That is, we are to consider how the adjective or adjectival phrase F applies to the subject of the superordinate verb. Likewise with a form that has an explicit propositional complement, as in forms such as ‘It looks as if p’, or ‘It looks like p’, focus is on that complement proposition. Some writers, noting the lack of all-out assertion are inclined to talk of a tentative, or guarded assertion.[7] However while it may often be a consequence of advancing this speech act that one is taken by an audience likely to be committed to the proposition in question (on failure of being properly conversationally co-operative) it is not clear that assertion itself should come in degrees. So perhaps we just think that the minimum that is here required is that the proposition in question be epistemically possible, or rather that it be presented as epistemically possible (since the speaker may be dissimulating about their attitudes towards it): that is, that its truth not be inconsistent with what is common knowledge or evidence or is presupposed in the conversational context.

One can put the usage in sharper focus by making a contrast with a very different usage. For there are uses where the status of the derived proposition is clearly not the focus of discussion. Consider for example looking at the image which first made Saatchi & Saatchi’s name:

One might, commenting on the effectiveness of the advertisement, remark on how well the agency succeeded and say:

(1)That model looks pregnant.

In this situation one might not suppose that there is any chance at all that the male model is pregnant or could be; it may be a presupposition among the conversant that male pregnancy is nomologically impossible.[8]The purpose of one’s statement, then, is not to discuss the evidential status of the proposition that the model is pregnant,nor is it to allude to some visual evidential ground for that. Instead it seems that in this case the predicate complement is qualifying not the argument of the main verb but rather the way of looking introduced by that verb: it is helping to specify for us which way of looking the model in question has. It marks out that they haven’t just made him look as he would were he holding a pillow under his jumper, or if he had simply enjoyed the average number of pints of beer of an English middle-aged male.

As we will see shortly, such uses Jackson and Chisholm assimilate to what they call ‘the comparative use’. Here what matters is the contrast in conversational focus (at least for the form ‘o is F’) between a case where one is concerned with a proposition predicating the complement adjectival phrase of the superordinate subject, i.e. where one is concerned with whether oisF, and one where the complement in some way seems to qualify the main verb ‘looks’ and the question is simply whether o does look that way.

Now the contrast between the evidential and non-evidential (what Gisborne calls ‘evaluative’ use; it also corresponds to what Travis talks of as ‘demonstrable looks’) is marked initially as a pragmatic one. And it requires us to focus briefly on the difficult topic of evidentiality. In perhaps a quarter of the world’s languages there are syntactic markers (in most cases obligatory) which indicate the evidential standing of the proposition put forward: for example whether the speaker has direct evidence for it, whether the source is testimony, or is inference, or whether what is at issue is a psychological avowal.[9] English has no such syntactic markers, but we can provide our interlocutors with evidential information through use of sense and appearance verbs. For example, where a proposition is evident to us, we may say ‘I see that...’, where we know of something through testimony, we say, ‘I heard that...’. There are certainly some instances which can be understood purely pragmatically, but there are other evidential in some languages which seem to involve a semantic component.

Certainly the focus just on the proposition that o is F cannot be affected when the looks sentence is embedded, for example as the antecedent of a conditional. Moreover, as we will note when working our way through the Chisholm and Jackson categories, sentences which otherwise have very different uses can be employed evidentially. So the simplest hypothesis may yet be that this is a pragmatic effect. On the other hand, Gisborne argues that historically the evidential use appears earlier than the qualitative, and so one may suppose that this is some evidence that there is a distinctive sense associated with it. Since our main concern is with varieties of non-evidential uses, I shall from here leave aside the question whether the evidential can be handled purely pragmatically.

The most important thing to note is that the evidential use is not to be confused with what we may call, following Chisholm and Jackson, the epistemic sense. Jackson introduces it so:

The epistemic use is propositional in that statements containing it are in (or can naturally be cast into) the form ‘It looks as if p’, where ‘p’ is a sentence expressing a proposition: examples are ‘It looks as if the sun is sinking into the sea’, ‘It looks as if these tomatoes are ripe’, and ‘It looks as if it is about to rain.’ (Jackson, 1977, p. 30.)

The first of these is clearly not something we would employ in an evidential use, for we don’t suppose it possible that the sun should sink into the sea. So this suggests that there is a contrast between being epistemic and having the evidential role. And, I suggest, we can find other examples in English of constructions which seem epistemic in this way, without having to be taken to be being used evidentially. We have in English the peculiar infinitival form of complement which recommends solely an epistemic reading without thereby having the evidential use indicated above:

(2)John looks to be on the verge of tears

(3)Mary looks to have once been a nun

(4)Veronica looks soon to be a vice-president.

All of these are naturally read as having an epistemic sense: that there is a visually grounded evidence for the proposition that John is on the verge of tears, that Mary has once been a nun, that Veronica will be a vice-president soon. In these cases as in the first of Jackson’s examples, they do not have to be used evidentially, in the sense we proposed above, of introducing into discourse a proposition and the existence of a visual ground for it. With Jackson’s example, one might take the presence of the comparative particle ‘as’ as indicative of some kind of comparison: the way the situation looks is one which is relevantly similar to that in which it would be reasonable to infer that the sun is sinking (i.e. if one lacked sufficient knowledge of astronomy). Likewise, one might find assertible(3) even in a context of common knowledge that Mary has only ever been a bond dealer, indicating that her appearance is nonetheless one on the basis of which it would be reasonable to infer the presence of a sacred vocation.