Version classiqueVersion mobile

Resemblance and Representation

 | 
Ben Blumson

2. Defining Depiction

Texte intégral

1“The aim of philosophy …”, according to Wilfrid Sellars, “… is to understand how things in the broadest possible sense of the term hang together in the broadest possible sense of the term” (Sellars, 1962, 369). One way of pursuing this aim is to pursue the reductive analysis of everything in terms of the more or most fundamental things. The aim may, for example, be a reductive analysis of everything in terms of the physical, or the phenomenological, or a combination of the physical and the phenomenological. As I conceive of it, the purpose of the philosophy of pictures is to complete one small step in this ambitious reductive program.

2The reductive program may take various paths. In philosophy of mind and language, it’s common to distinguish the Sellarsian program, which pursues a reductive analysis of mind in terms of language, from the Gricean program, which pursues an analysis of language in terms of mind (see McKinsey (1983, 1-2) and Bennett (1976, 24-27)). Whereas the Sellarsian program analyses the belief that it’s raining, for example, in terms of the assertion that it’s raining, the Gricean program analyses the assertion that it’s raining in terms of the belief that it’s raining. The next four chapters pursue an analysis of depiction as part of this Gricean program.

3The Gricean program in the philosophy of language employs a strategy of divide and conquer by distinguishing speaker meaning, to be analysed in terms of intention, from sentence meaning, to be analysed in terms of convention. Whereas this chapter and the following extend the Gricean analysis of speaker meaning in terms of intention to encompass depiction, chapters four and five modify the Gricean analysis of sentence meaning in terms of convention by jettisoning arbitrariness in favour of resemblance to form an analysis of depictive symbol systems. The analysis of pictures, I will argue in these chapters, slides seamlessly into the Gricean program as a whole.

4So this chapter argues for defining depiction as a kind of intentional representation. In particular, it argues for defining depiction by combining resemblance with an analysis of speaker meaning in terms of intentions (following Abell, 2005a; 2009; as well as Blumson, 2009). The Mona Lisa depicts Lisa, for example, because Leo intended his audience to infer from the Mona Lisa’s resemblance to Lisa his intention to induce an effect in them, by means of recognition of his intentions. This analysis, I will argue, supports both a close analogy between depiction and description and the platitude that whereas description is mediated by convention, depiction is mediated by resemblance.

5The first section introduces the analysis of speaker meaning in terms of intention and argues for its application to depiction. The second section takes up and argues against the natural suggestion of defining depiction by substituting experienced resemblance for the intended effect in the analysis of speaker meaning. The third section argues for defining depiction by substituting resemblance for the intended feature of utterances in the analysis of speaker meaning from which the speaker’s intentions are intended to be inferred. The fourth section compares the analysis with Catharine Abell’s (2009), which also defines depiction by combining resemblance with speaker meaning.

2.1 Grice’s analysis of speaker meaning

6There’s an important distinction between the meaning of sentences, called “sentence meaning”, and what speakers use those sentences to mean, called “speaker meaning” (Grice, 1957, 217). Suppose, for example, I say, after Smith’s just insulted me, ‘Smith is a fine friend’. The sentence means that Smith is a fine friend, but this is not what I mean by the sentence: rather, my utterance of the sentence on this occasion means that Smith is a foul friend, for having insulted me. It’s plausible that speaker meaning should be analysed directly in terms of intentions, whereas sentence meaning should be analysed jointly in terms of convention and speaker meaning.

7This order of analysis is justified by the dependence of meaning on use: the purpose of linguistic utterances is to achieve various effects in audiences. Indicative sentences, for example, are usually uttered to produce beliefs. Imperative sentences, in contrast, are usually uttered to produce actions rather than beliefs. If, for example, I mean you to stop by uttering ‘stop!’, then my purpose is to induce you to stop. This suggests that a person means something by an utterance if and only if the person intends to induce an effect in an audience (Grice, 1957, 217). I might mean that it’s raining by uttering ‘it’s raining’, for example, because I intend you to believe it’s raining.

8But the following is a counterexample. Suppose that I want to frame you for a murder I’ve committed. In order to do so I leave your handkerchief at the scene of the crime, stained in the blood of the victim. I intend to induce the police to believe, on finding the bloodstained handkerchief, that you are the murderer, so the conditions of the analysis are met. Nevertheless, it is not the case that by leaving your handkerchief at the scene I mean that you are the murderer. I intend my action to produce an effect in my audience, but I nevertheless do not mean anything by my action, so intending to induce an effect in an audience is insufficient for speaker meaning (Grice, 1957, 217).

9The problem the example raises is that it is not sufficient for an utterance to have meaning that it be used for a certain purpose, even if that purpose is characteristically linguistic, because it’s always possible to achieve that purpose in some other way, without communicating one’s intentions. The example shows an adequate analysis of speaker meaning has to characterise not only the effects, such as inducing beliefs and actions, that meaningful utterances are intended to accomplish, but also the distinctive way in which meaningful utterances bring about those effects. Merely causing beliefs and actions is not the same as communicating them.

10In meaning something by an utterance, one’s intentions are explicit in a way one’s intentions in other actions aren’t: telling someone something, for example, is explicit in a way that tricking someone is not (Grice, 1957, 218). This suggests:

(5) A person means something by an utterance if and only if the person intends the utterance to produce an effect in an audience by means of recognition of this intention (Grice, 1957, 220).

11So I mean that it’s raining by uttering ‘it’s raining’, for example, because I intend to induce you to believe that it’s raining by means of recognition of my intention.

12Requiring the intended effect of an utterance to produce an effect in an audience by means of recognition of intention captures the distinctive way in which meaningful utterances produce their effects, because it captures the fact that one’s intentions in meaning something by an utterance are explicit in a way one’s intentions in other actions aren’t. In leaving your bloodstained handkerchief at the scene of the crime, for example, I don’t mean that you’re the murderer, because although I intend the police to believe that you are the murderer, I don’t intend this effect to be produced by means of recognition of my intention.

13Depictions, like sentences, are aimed at the production of various effects. Maps, for example, are depictions aimed at producing beliefs in audiences about terrain. Lego instructions are depictions which are intended to induce audiences to arrange Lego in a certain way, rather than to induce them to believe that Lego is arranged in that way. Caricatures, like jokes, are aimed at producing amusement. This suggests analysing depiction, like speaker meaning, in terms of the effects pictures are intended to induce in their audiences, by means of recognition of these intentions: depiction, in other words, should be analysed as a special case of speaker meaning.

14It would be a problematic disanalogy with meaning if there were any depiction which did not have the purpose of producing an effect in its audience. The fact that the purposes of some depictions are exclusively aesthetic, rather than directed towards a practical end such as belief or action, might be thought to show this. But, firstly, this is not a disanalogy since there are examples of descriptions, such as in poetry or song, which are also exclusively aesthetic and, secondly, even a depiction with an exclusively aesthetic purpose is at least intended to produce some aesthetic effect, such as pleasure or amusement, concerning what it represents in its audience.

15An example analogous to the example of the handkerchief shows that depictive, like descriptive, representation requires the effect in the audience to be induced explicitly and openly. Suppose that I have a secret tunnel in my office. One panel of the wall is actually the door to the secret tunnel, which I have disguised to look exactly like an ordinary part of my bookshelf. Hence, the tunnel door is intended to look and be like an ordinary part of my bookshelf and to induce visitors to believe it is an ordinary part of my bookshelf. But the tunnel door doesn’t depict an ordinary part of my bookshelf, since I don’t intend it to induce this effect by means of recognition of my intention.

16A deeper reason that the analysis of speaker meaning ought to be applied to depiction is its general nature. There is no distinctively linguistic or verbal element in the analysis of speaker meaning except for the word ‘utterance’, which is used in an extended sense to cover any kind of action, including dropping handkerchiefs and drawing pictures as well as writing and talking (Grice, 1969, 92). This means that the analysis should be expected to apply not just to language but to all communication, including pictorial representation. Different kinds of representation – like depiction and description – can then be defined as different kinds of speaker meaning.

17So the analysis of speaker meaning should be applied to specifying how depiction is representational. As Catharine Abell writes: “… in the case of depiction ‘the maker A, means picture Y to depict an object, Z’ is roughly equivalent to ‘A produced Y with the intention of inducing a belief about [or other effect concerning] Z in the observers of Y in virtue of those observers recognizing this intention’” (Abell, 2005a, 59; see also Novitz, 1977, for a similar account). So the Mona Lisa, for example, depicts Lisa because Leo intended to induce an effect – such as the belief that Lisa smiled – in his audience, by means of recognition of his intention.

18But exactly how to apply the analysis of speaker meaning to the analysis of depiction is not obvious: simply conjoining resemblance with speaker meaning is no better than merely conjoining it with representation. The sentence ‘this sentence is thirty-five letters long’, for example, resembles itself and may be written with the intention of inducing in audiences the effect of believing that it is thirty-five letters long, by means of recognition of this intention. But, like the phrase ‘this phrase’, the sentence does not depict itself, because its resemblance to itself is incidental to its representation of itself. The problem of insufficiency is still unresolved.

2.2 The intended effect in Grice’s analysis

19A common way to analyse a particular kind of meaning or representation is by specifying the intended effect in the analysis of speaker meaning. Assertions, for example, are those utterances which are intended to induce beliefs by means of recognition of intention, whereas commands are those utterances which are intended to induce actions by means of recognition of intention. Similarly, jokes are utterances intended, by means of recognition of intention, to induce amusement. This suggests depiction might also be analysed by an appropriate specification of the intended effect in the analysis of speaker meaning.

20To take a more exotic example, fiction can be analysed by substituting imagination as the intended effect in the analysis of speaker meaning (Currie, 1990, 18-51). So, for example, it is fictional in The Adventures of Sherlock Holmes that Sherlock Holmes is a detective because the reader is intended, on the basis of recognition of that intention, to imagine that Sherlock Holmes is a detective. If there is some effect on the audience that is distinctive of depiction, as imagination is an effect distinctive of fiction, then that could be combined with the analysis of speaker meaning to provide an analysis of depiction.

21Robert Hopkins (1998, 50-51) suggests that audiences of depictions experience them as resembling what they represent. This suggests substituting experienced resemblance for the intended effect in the analysis of speaker meaning as follows:

(6) Something depicts another if and only if the former is intended to induce the former to be experienced as resembling the latter by means of recognition of this intention.

22So the Mona Lisa, for example, is supposed to depict Lisa because Leo intended it to be experienced as resembling Lisa, by means of recognition of his intention.

23One clarification. An experience of resemblance does not require both the thing that resembles and the thing that is resembled to be present. If somebody is familiar with a company logo, for example, then whenever they see that logo their experience is likely to represent it as having the property of resembling the instances of the logo that they have seen in the past (Peacocke, 1987, 385). So experiencing a resemblance is not just experiencing similar things simultaneously. Rather, it’s having an experience which represents something as having the property of resembling another, which may or may not be present.

24By specifying experienced resemblance as the intended effect in the analysis of speaker meaning, this analysis seems to provide for a non-incidental connection between resemblance and representation. Although the sentence ‘this sentence is thirty-five letters long’, for example, resembles itself and is intended to induce an effect in its audience by means of recognition of intention, the intended effect is not for it to be experienced as resembling itself. This suggests that substituting experienced resemblance for the effect in the analysis of speaker meaning escapes the counterexamples to straightforwardly conjoining resemblance with representation.

25But the following counterexample shows that it doesn’t. Suppose I sincerely write: ‘I intend this sentence to induce itself to be experienced as resembling itself by means of recognition of my intention’. Since I do intend the sentence to induce itself to be experienced as resembling itself by means of recognition of my intention, the analysis predicts that I have not only written that sentence, but also depicted it. But although the sentence is sincerely intended to be experienced as resembling itself by means of recognition of my intention, it is not a depiction of itself. On the contrary, it’s a paradigm example of a non-depictive, descriptive, representation.

26In this case, the fact that the sentence is intended to be experienced as resembling itself is incidental to the way that the sentence represents itself in the same way that the fact that ‘this sentence is thirty-five letters long’ resembles itself is incidental to the way that it represents itself. The problem with the analysis is that it specifies the distinctive feature of depiction by specifying what depiction is used to do. But depiction is distinctive, not because it is used to do something special, but because it achieves what it is used to do in a special way. So depiction cannot be analysed by specifying the intended effect in the analysis of speaker meaning.

27Four objections. First, perhaps the counterexample can be avoided by denying that things can be experienced as resembling themselves, and thus denying that I can sincerely intend my sentence to be experienced as resembling itself by means of recognition of my intention. But since there are pictures that represent themselves (Walton, 1990, 117-121), denying that things can be experienced as resembling themselves is not an available response. One might just deny that this particular sentence can be experienced as resembling itself. But there is no principled reason to do so, nor to think that this would solve the underlying problem.

28Second, perhaps the argument tacitly relies on the false premise that nothing can be both a depictive and a descriptive representation. Pattern poems, for example, often represent their subjects both through the conventional meaning of their words and the arrangement of those words into patterns that resemble what they represent: John Hollander’s (1981, 31) pattern-poem Pattern Poem, for example, makes its point by both describing itself as making a point and depicting itself as making a point, in virtue of resembling itself in respect of being point shaped. I don’t want to deny that such pattern poems are both descriptions and depictions of their subjects.

29But the argument doesn’t rely on the premise that nothing can be both a depiction and a description. I don’t deny that the sentence ‘I intend this sentence to be experienced as resembling itself by means of recognition of my intention’ is a depiction merely because it is a linguistic description, but because both the resemblance of the sentence to itself and the fact that the sentence is intended to be experienced as resembling itself are incidental to its representation of itself: the sentence refers to itself only because of its use of the demonstrative expression ‘this sentence’ to refer to itself.

30So the fact that the sentence is a paradigmatically linguistic representation doesn’t exclude the sentence from being a depiction, and the claim that linguistic representations cannot also be depictive representations is not a premise of the argument. The fact that the example is an explicit and paradigmatically linguistic representation plays only a heuristic role: it makes it more obvious than another example would that the sentence’s being intended to be experienced as resembling itself and its resemblance to itself are incidental to the way it represents itself.

31Third, it might be objected that the argument proves too much, because it undermines the analysis of fiction as well. Suppose I say ‘I intend you to imagine that it was a dark and stormy night, by means of recognition of this intention’. It might be argued that, despite meeting the conditions of the analysis of fiction, what the sentence says is not fictional, because a person sincerely uttering it would be telling the truth: the sentence would be a factually correct account of that person’s intention. So explicitly stating that the conditions of the analysis of fiction are met appears to produce an analogous counterexample to its sufficiency.

32But the sufficiency of the analysis of fiction may be defended by arguing that my utterance of ‘I intend you to imagine that it is a dark and stormy night by means of recognition of my intention’, as well as accurately reporting my intentions, really does make it fictional that it is a dark and stormy night. The example is not one in which an utterance of non-fiction meets the conditions of the analysis of fiction, but is an example in which someone meets the conditions of the analysis of fiction by simultaneously speaking the truth and explicitly creating a fiction.

33The following example illustrates the plausibility of this response. The first sentence of If on a Winter’s Night a Traveller, by Italo Calvino, is ‘You are about to begin reading Italo Calvino’s new novel, If on a Winter’s Night a Traveller’ (Calvino, 1982, 1). Like a sentence that explicitly states the conditions of the analysis of fiction are met, this opening is at once a statement of fact, since normally the reader is about to begin reading the book, and an explicit announcement that the author is engaging in fiction, since it states that you are reading a novel.

34Fourth, it might be objected that the argument proves too much because it undermines the analysis of speaker meaning in general. Suppose I sincerely say ‘I intend you to believe that it’s raining by means of recognition of my intention’. It follows from the analysis of speaker meaning that the person has said that it’s raining. But intuitively, runs the objection, by uttering that sentence I would not mean that that it’s raining. Rather, I would mean that I intend you to believe that it’s raining by means of recognition of my intention.

35But the sufficiency of the analysis of speaker meaning may be defended by arguing that by uttering ‘I intend you to believe that it’s raining by means of recognition of my intention’ as well as explicitly stating my intentions, I also mean that it’s raining. As for the analysis of fiction, the response becomes plausible if one considers explicit performatives. ‘I hereby declare that there is a cat on the mat’, for example, is at once a declaration that there is a cat on the mat and a report that the speaker declares that there is a cat on the mat (Schiffer, 1972, 64-68).

36Note that the argument applies not only to experienced resemblance, but to any attempt to analyse depiction by specifying the intended effect in the analysis of speaker meaning. Whatever effect is specified, it will be possible to write a sentence sincerely announcing the intention of the writer to induce that effect by means of recognition of that intention. That sentence would meet the conditions of the analysis, but it would fail to be a depiction. So to provide an analysis of depiction, the analysis of speaker meaning has to be combined with a feature distinctive of depiction not by specifying the intended effect in the analysis, but in some other way.

2.3 The salient feature in Grice’s analysis

37To see how resemblance and the analysis of speaker meaning can be combined to analyse depiction, consider the following two examples. First, suppose a shopkeeper antecedently knows that I want cigarettes. I pass the shopkeeper fifty dollars in order to induce him to give me cigarettes, by means of his antecedent recognition of my intention that he give me cigarettes. Second, suppose I pass the shopkeeper eleven dollars and fifty-five cents, the exact price of the brand I prefer, intending him to infer from the fact that I pass him that amount that I intend the money to induce him to give me those cigarettes by means of recognition of my intention.

38Because in the first case the shopkeeper already knew that I wanted the cigarettes, it does not seem that by passing him the fifty dollars I meant that he should give me the cigarettes. But in the second case, because the shopkeeper inferred what I wanted, it does seem that I meant by the eleven dollars and fifty-five cents that I want the cigarettes (Grice, 1969, 94). To reflect this difference, the analysis should be altered to the following (Avramides, 1989, 47):

(7) A person means something by an utterance if and only if the person intends that:
a. the utterance has a certain feature
b. an audience recognise that the utterance has that feature
c. the audience infer at least in part from the fact that the utterance has that feature that the person intends:
d. that the utterance produce an effect in the audience
e. and that that effect be produced at least in part by means of the audience’s recognition of intentions (a)-(e).

39So, for example, my passing the shopkeeper eleven dollars and fifty-five cents means he should give me the cigarettes because it has the feature of being exactly the price of my brand of cigarettes, and because the shopkeeper infers from that feature that I intend him to give me the cigarettes by means of recognition of my intention (see Harman (1974) for defence of the self-reflexivity in (e)).

40In the case of linguistic representation, the feature from which the audience is intended to infer the speaker’s intentions is the conventional meaning of the utterance or the fact that speakers have uttered it before with the same meaning. My utterance of ‘it’s raining’, for example, means that it’s raining because I intend you to infer from the conventional meaning of ‘it’s raining’ in English that I intend to induce you to believe that it’s raining by means of recognition of my intentions (Schiffer, 1972, 12). Convention connects words with what they represent by allowing audiences to infer the communicative intentions of speakers who conform to the convention.

41But whereas words are connected to what they represent by convention, depictions are connected to what they represent by resemblance. So it’s natural to mark this difference by inserting resemblance into the analysis of meaning in the same place that allows for the role of the conventional meanings of words, as follows:

(8) Something depicts another if and only if it is intended that:
a. the former resembles the latter in a certain respect
b. an audience recognise the former resembles the latter in that respect
c. the audience infer at least in part from the fact the former resembles the latter in that respect that it is intended:
d. that the former produce an effect in the audience
e. and that that effect be produced at least in part by means of the audience’s recognition of intentions (a)-(e).

42This analysis is the result of specifying that resemblance is the feature of the picture from which audiences infer depicters’ intentions, just as conventional meaning is the feature of linguistic utterances from which audiences infer speakers’ intentions.

43By specifying the role of resemblance in depictive representation, this analysis avoids the counterexamples to analysing depiction simply by conjoining representation with resemblance. The sentence ‘this sentence is thirty-five letters long’, for example, is not counted as a depiction of itself because, although it resembles and represents itself, the intentions of people who utter it to induce effects in their audiences by means of recognition of their intentions will normally be inferred not from the sentence’s resemblance to itself, but from its conventional meaning in English. So the analysis captures the non-incidental role of resemblance in depictive representation.

44Moreover, a counterexample to this analysis cannot be produced by a linguistic restatement of the relevant conditions. Even if I say ‘I intend that this sentence resemble itself in shape, that you recognise this resemblance, and that you infer from this resemblance that I intend to induce you to believe that it’s raining by means of recognition of my intention’ I cannot be sincere because those inferences cannot be made from the sentence’s resemblance to itself in shape and so I cannot rationally or successfully have the intentions I declare I have using that sentence. So the analysis avoids the counterexamples to the insufficiency of resemblance for representation.

45It might be objected that analysing depiction in terms of intention reverses the correct order of discovery, since the intentions of a depiction’s perpetrator are discovered from knowledge of what the depiction represents, rather than what it represents being discovered from knowledge of its perpetrator’s intentions. Rather than discovering that the Mona Lisa represents Lisa, for example, by discovering that Leo intended to induce a certain effect in us by means of recognition of his intention, we discovered that Leo intended to induce this effect in us by means of recognition of his intention because we know that the Mona Lisa depicts Lisa.

46This objection is only relevant if the analysis is intended as an epistemic reduction of depiction to intention. But the analysis isn’t intended as an epistemic reduction of depiction to intention, so whether we discover what pictures represent by knowing their perpetrators’ intentions or vice versa is irrelevant. The analysis is intended as a metaphysical reduction of depiction in terms of intention. But the analysis plausibly succeeds as a metaphysical reduction, since intention is plausibly metaphysically prior to depiction: intentions might have existed without depictions, but depictions could not have existed without intentions.

47But the analysis is committed to an epistemic priority of resemblance over depiction, because it requires the communicative intentions of a depiction’s perpetrator to be inferred from the resemblance of the depiction to what it represents, and this is open to question. It might be argued that the resemblance of hidden figures to what they represent, for example, is inferred from antecedent knowledge of what the figures represent, rather than what the figures represent being inferred from what they resemble (Lopes, 2005b, 167). If this is the case then – contrary to the analysis – there is no epistemic priority of resemblance over intention.

48But although what hidden figures represent cannot be recognised immediately, this need not be because what they resemble is inferred from what they represent, since what hidden figures represent cannot be recognised immediately either. Moreover, it must be possible to recognise what hidden figures resemble independently of what they represent, since it would still be possible to recognise the resemblance of a figure to what it represents even if it ceased to represent that thing. It might be difficult, for example, to notice a figure hidden in the midst of a painting of clouds. But it must be possible, for the resemblance could still be pointed out even if the clouds were real.

49Three clarifications. First, the analysis requires not merely that the picture is intended to resemble what it represents, but that it is intended to resemble what it represents in a certain respect. And the analysis requires not merely that the audience recognise that the picture resembles what it represents, but that they recognise that it resembles what it represents in this specific respect. For the Mona Lisa to depict Lisa, for example, the analysis requires that Leo had in mind a specific respect in which he intended the Mona Lisa to resemble Lisa. Moreover, it requires us to be able to recognise that the Mona Lisa resembles Lisa in this specific respect.

50But it sometimes seems possible to recognise a resemblance without recognising the respect in which the resemblance obtains. It may be possible to recognise a similarity between two faces, for example, without being able to recognise the respect in which that similarity consists (Armstrong, 1978b, 98). And it may be possible to recognise a similarity between a portrait and the face it represents, without being able to recognise the respect in which the portrait and the face it represents are similar. If this is correct, then the portrait may depict the face without the artist or audience recognising in what respect it resembles the face, and so the analysis should be weakened.

51However, the specific respect must be mentioned in the analysis since if the abundant conception of properties is a necessary truth, it’s also a necessary truth that everything resembles everything, and so it’s also a necessary truth that a picture resembles what it represents (if the picture and what it represents exist). In this case, the intention that a picture resembles what it represents would be problematic – it would be like having an intention that one plus one is two, or that bachelors are unmarried men. Moreover, the audience could easily recognise that the picture resembles what it represents, but it would be difficult for them to infer anything from this about the artist’s intentions.

52So if the sparse conception of properties is correct, then the analysis can be simplified and the counterexample above avoided by omitting to require that the resemblance be in a certain respect. But if on the other hand it’s the abundant conception of properties which is correct, then the analysis cannot be simplified in this way, and the purported counterexample must be addressed head on. But this is not so much a problem for the analysis as a problem for the abundant conception of properties, for a proponent of the abundant conception of properties must argue that recognition of resemblance, unless it is vacuous, is always recognition of resemblance in a certain respect.

53Second, though it contains the word “recognition”, the analysis is not a version of the recognition theory of depiction. According to recognition theories, something depicts another only if the former causes (usually non-veridical) recognition of the latter. But according to analysis (8), something depicts another not only if it is intended that an audience recognise the latter itself, but only if it is intended that an audience recognise the resemblance of the former to the latter. And analysis (8) requires the recognition of the resemblance of the former to the latter be veridical. (See Newall (2011, 42-94) for an extremely helpful comparison of the recognition and resemblance theories.)

54Third, the analysis leaves open the nature of the intended effect mentioned in clause (d). It is open to the proponent of experienced resemblance, for example, to hold that it is a necessary condition of depiction that experienced resemblance be the intended effect. But I’m doubtful that experienced resemblance or any other kind of experience is necessary for depiction, especially in view of my stipulation in the introduction that depiction may occur in any medium, and includes allegorical fiction. Moreover, if the intended effect in the analysis is allowed to be an action, belief, imagining or similar attitude, then the parallel with the analysis of speaker meaning is more perfect.

2.4 Abell’s analysis of depiction

55In arguing that depiction should be analysed in terms of resemblance and speaker meaning, I’m agreeing with Catharine Abell, according to whom something depicts an object O if “Its maker intended it to resemble O in a certain visible respect(s) and thus to bring O to viewers’ minds, and intended that these resemblances have this effect in part because viewers recognise this intention ...” (Abell, 2009, 208). So the Mona Lisa depicts Lisa, according to Abell, because Leo intended the Mona Lisa to resemble Lisa in a visible respect and thus to bring Lisa to viewers’ minds, in part because viewers recognise this intention.

56The purpose of this section is to clarify the differences between Abell’s analysis and my own. I agree with Abell about the sufficiency of the condition quoted in the last paragraph, which differs only superficially from analysis (8) in the last section. I also agree with Abell that there are counterexamples to the necessity of that condition and analysis (8). But I disagree with Abell about how to accommodate counterexamples to the necessity of her sufficient condition and analysis (8) and so also disagree with her about the final analysis of depiction. This section explains Abell’s analysis and how I disagree with it. My own version will not be finalised until chapter nine.

57The first issue I disagree with Abell about is how to address counterexamples to the necessity of an audience. It’s possible to depict something without intending to induce an effect in an audience: it is possible, for example, to depict by doodling in margins, drawing preparatory sketches and tracing patterns in fogged-up glass, but not intend that the doodles, sketches or traces ever reach an audience other than oneself. In some of these cases, one can argue the intended audience of the doodle, sketch or tracing is simply the doodler, the sketcher or the tracer, but this response, I will argue in section 3.3, is not available in every case of this kind.

58The second issue I disagree with Abell about is how to address counterexamples to the necessity of intention. Abell argues that an intention to depict, or to respond to a communication problem, is necessary for depiction (Abell, 2009, 205). Nevertheless, Abell argues that what an artist depicts may not be what the artist intends to depict. If, for example, a drawing inadvertently turns out to look more like a fat man than like a svelte man then, according to Abell, the drawing could turn out to depict a fat man even though it was intended to depict a svelte man (Abell, 2009, 194). In contrast, I’ll argue in section 3.2 that there are no counterexamples to the necessity of intention.

59To accommodate these possibilities, Abell argues that “Just as we are able to interpret sentences uttered without communicative intentions because they employ linguistic conventions with which we are familiar, we are able to interpret pictures produced without communicative intentions because they employ stylistic conventions with which we are familiar” (Abell, 2009, 204). So just as, for example, the sentences in my diary have meaning even though I don’t intend them to reach an audience because of their conventional meaning in English, the doodles in my margins depict because they fall under certain stylistic conventions, which usually govern communication.

60I have two objections to this response. First, it’s possible to depict without intending to reach an audience or to conform to a stylistic convention. A painter, for example, may choose not to exhibit his or her paintings, precisely because the paintings are stylistically innovative and so don’t conform to any stylistic conventions. Likewise a diarist, in the linguistic case, may write in non-literal language, the intended meaning of which isn’t the conventional meaning of the words. To accommodate these cases, I will argue in section 3.3 that it is a necessary condition of depiction not that it induces an effect in an audience, but that it does so if it reaches an audience of a certain type.

61Second, using convention to accommodate audienceless and unintentional depiction risks failing to distinguish between depiction and onomatopoeia. The word ‘woof’, for example, resembles but does not depict a dog’s bark. But just as sentences now have their sentence meaning because of conventions arising from what they were used to speaker mean in the past, some onomatopoeic words have their meaning because of conventions arising from what they were used to depict in the past. So allowing that merely falling under stylistic conventions is sufficient for depiction risks including not only audienceless and unintended depiction, but also onomatopoeia.

62However, Abell’s appeal to convention may avoid this problem because falling under a stylistic convention, according to her analysis, is not quite sufficient for depiction. Instead, something depicts an object O if its maker produced the resemblance at issue “... and intended thereby to adhere to a stylistic convention extant in his or her community, characterized by resemblance in the respect(s) at issue, and thus bring O to viewers’ minds” or else had “… the intention of responding to a communication problem to which a conventional solution is extant in his or her community, which is characterized by resemblance in the given respect(s)” (Abell, 2009, 204-205).

63The first condition does exclude onomatopoeia, since though onomatopoeic words fall under stylistic conventions extant in their communities, they need not be intended to fall under these stylistic conventions. But the first condition also excludes depiction in the absence of an intended audience, since it requires the maker to intend to bring the object to viewer’s minds, and unintentional depiction, since the maker must intend the resemblance at issue and intend thereby to adhere to the stylistic convention which is characterised by that resemblance. Abell’s example of an inadvertent picture of a fat man, for example, does not satisfy the first condition.

64The second includes depiction in the absence of an audience, since it doesn’t require the maker to intend to bring the object to viewer’s minds, and unintentional depiction, since makers may produce the resemblances at issue with the intention to respond to communication problems, while inadvertently producing a resemblance in the wrong respect. But in this case, onomatopoeia is included, since onomatopoeic utterances are intended to respond to communication problems to which conventional solutions are extant, and since those conventional solutions are characterised by resemblance in the respects which are responsible for explaining the origin of the onomatopoeia.

65The Gricean program in the philosophy of language employs a strategy of divide and conquer by analysing speaker meaning in terms of intention, and sentence meaning in terms of convention and speaker meaning. By invoking conventions in the analysis of depiction, Abell hopes to employ a parallel strategy to respond to counterexamples to the necessity of intentions for depiction. In chapters four and five I will also argue for analysing some kinds of depiction in parallel with the analysis of sentence meaning – but I will deny that a convention which has its origin in resemblance is sufficient for depiction, and so exclude onomatopoeia.

66The third issue I disagree with Abell about is how to respond to the problem posed by the intentionality of depiction. Three features of depiction are symptomatic of its intentionality. The first is the apparent possibility of depicting non-existents. The second is the possibility of depicting something without depicting anything in particular. The third symptom is the possibility of depictive misrepresentation. All three symptoms of the intentionality of depiction are problematic for the platitude that depiction is mediated by resemblance, but the difficulties are raised most strikingly by the problem of the depiction of non-existents.

67Abell argues that depicting non-existents is possible because such depictions would resemble what they represent, if what they represent had existed. Although Santa’s portrait, for example, does not actually resemble Santa, because Santa does not exist, it seems plausible that Santa’s portrait would have resembled Santa if Santa had existed, because if Santa had existed, both he and his portrait would have shared the property of being partly red. In general, depictions of non-existents don’t resemble what they represent, but they would have done if what they represent had existed, so the caveat appears to accommodate the depiction of non-existents (Abell, 2009, 212).

68However, Santa may be depicted both as mostly red or as mostly green. But had he existed, he would have been either mostly red or mostly green, but not both. So either the depiction of Santa as mostly red or the depiction of Santa as mostly green would fail to resemble Santa in respect of colour, even if Santa had existed. In general, non-existent objects may be depicted as appearing in many different, often inconsistent, ways. If there is a fact of the matter about how the non-existent object would have looked, some depictions of it wouldn’t have resembled that object; if there’s no fact of the matter, then there’s no fact of the matter about whether they’d resemble it or not.

69Abell may argue that there are facts of the matter about how inexistent objects look, and that pictures which don’t depict them as they would look are misrepresentations, the problem of which has to be solved differently. The problem of misrepresentation is that if a depiction misrepresents what it depicts in some respect, then it will fail to resemble what it represents in that respect. If a tree is depicted as blue, for example, then the depiction fails to resemble the tree in respect of colour. Likewise, if Santa is depicted as partly green instead of partly red, Abell may argue that this is possible for the same reason it is possible to depict a green tree as blue.

70A picture misrepresents something as having a property, according to Abell, only if it resembles that thing or would if that thing existed and resembles things that have that property or would if they existed (Abell, 2009, 212). So, for example, it’s possible for a picture to depictively misrepresent a green tree as blue, because the picture may resemble the tree in respect of shape and simultaneously resemble something blue in respect of colour. So as long as it’s possible for a marked surface to simultaneously resemble a tree and blue things, it’s likewise possible for the surface to depict the tree as blue, as long as the surface is marked with the relevant communicative intentions.

71But this strategy cannot accommodate all examples of depictive misrepresentation, because some depictive misrepresentations don’t resemble what they represent in any intended respect. Suppose, for example, that the police are completely misinformed about the appearance of a dangerous criminal. The police believe that the criminal is brunette, but he is blonde; the police believe he is bearded, but he is shaved; the police believe that he is tall, but in fact he is short; and so on (Kaplan, 1968, 198). If the police drew a wanted poster of this man, then it would resemble someone who is brunette, bearded, tall, and so on, but not resemble the man in any intended respect.

72By invoking counterfactuals to address the problem of non-existents, Abell hopes to exploit the idea that although apparent depictions of non-existents don’t resemble any actual thing, they may resemble a merely possible thing. In chapters eight and nine I will also argue for addressing the problems of depicting non-existents, depicting non-particulars and depictive misrepresentation by exploiting resemblances towards mere possibilia, but without invoking counterfactual conditionals. Chapters eight and nine will argue this approach is a general solution to the problem of intentionality, so there is no special problem for the platitude that depiction is mediated by resemblance.

73So Abell and I agree that depiction should be analysed by resemblance in combination with communicative intentions, but disagree about how purported counterexamples to this analysis should be accommodated. In the case of counterexamples due to absence of the relevant intentions, I disagree with Abell’s appeal to convention. In the case of counterexamples due to the depiction of inexistents and depictive misrepresentation, I disagree with Abell’s appeal to resemblance in counterfactual respects and her claim that even a misrepresentation must resemble what it represents in at least one relevant respect. These disagreements are all about the details, not the spirit, of the analysis.

2.5 Conclusion

74Resemblance in depictive representation is intended to enable audiences to infer the communicative intentions of a depiction’s perpetrator. The Mona Lisa depicts Lisa, for example, because Leo intended audiences to infer from the fact that the Mona Lisa resembles Lisa that he intended an effect in them by means of recognition of his intentions. By specifying the non-incidental role of resemblance in depiction, this analysis avoids counterexamples to the sufficiency of resemblance for representation, and so supports the platitude that whereas descriptive representation is mediated by convention, depiction is mediated by resemblance.

75By defining both depiction and description as kinds of intentional representation, the analysis also supports a strong analogy between depiction and description. Both depictions and descriptions, according to the analysis, depend for their meaning on the intentions of their perpetrators to induce effects in their audience, by means of recognition of these intentions. The difference is merely in the feature from which these intentions are inferred: in the case of description, they are inferred from the conventional meanings of words, whereas in the case of depiction, they are inferred from the resemblance of pictures to what they represent.

Acheter

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search