1Somewhat like Monsieur Jordain, I have been doing cognitive semiotics long before I knew the term. Even in my early work, I argued against the brand of French structuralism of the time that the “text” alone could not bring us to scientific Paradise. We might come a little closer to the goal, if we were willing to take on a series of similar texts to which we applied the same model, in order to determine whether it worked for them all and gave the same results. But, even so, experiments would also be useful. Semiotics, even in my early conception, had to be able to account for results from psychology, sociology, and other disciplines which are relevant to the object studied. This does not mean, however, that psychological, sociological and other already well-known methods and models can take the place of semiotics. When we make our own experiments, we can define them in more properly semiotic terms, and we can relate the kinds of psychic and social structures and abilities studied within other sciences to specifically semiotic structures and abilities. This means that semiotic theory cannot be dispensed with. On the contrary, it becomes even more important than before. For, whatever Hjelmslev and Greimas meant by the curious dictum that the theory should be arbitrary, at the same time as it was adequate, it is difficult to see how such a feat can be realized. There is a limit to the arbitrariness of a theory that should be relevant, and thus adequate, to the experiential facts of semiosis.
2In many papers, including those presented at the conference of the Association for Visual Semiotics in Liège in 2015, authors have put forward the claim, not only that pictures, but perceptual reality itself, is profused with abductions and/or arguments. The latter term may no doubt be construed in different ways, but the former one was coined by Peirce, and within Peircean theory, abduction is a kind of argument, and, in Peirce’s sense of the latter term, it has to be made up of propositions/predications (Peirce’s dicisigns), and the latter must be made up of terms (Peirce’s rhemes). On the other hand, all philosophical theories seem to come together to deny that there could be anything similar to a proposition in pictures, let alone in perceptual reality. Eco’s heroic attempts to make pictures into something similar to language with double articulation is now recognized, even by Eco (2000) himself, to have failed – so there is not only nothing similar to phonemes in pictures, but neither are there any words, and thus no terms. As for perception, the case seems at least as hopeless. This all seems to confirm the conventional wisdom that pictures and percepts cannot formulate any propositions. On the other hand, even if we suppose this judgement to apply to the more familiar kind of pictures found in newspapers and picture books, its generalization renders quite mysterious the proven usefulness of all kinds of scientific and instructive pictures, from Blissymbolics and Isotype to machine drawings and the IKEA type of instructions for putting furniture together.
3Elsewhere, I have observed that the picture is evidently incapable of affirming anything, if one defines affirmation as something that is done by using language (Sonesson 1996; 2011b; 2014a). We have to start by acknowledging the difference in nature of the semiotic resources at the disposal of the picture and those used by the verbal argument. If we define an affirmation as a verbal construction, then it is trivially true that the picture cannot affirm anything. However, if the assertion is more generally defined as a transaction, by means of which a specific property is assigned to a particular entity, then it is possible for the picture to make affirmations in the specific way of a picture. However, even this may seem impossible, if it is true that pictures, as has often been said, merely reproduce the world of our experience. Nevertheless, in posing a similarity, an iconic sign, such as a picture, establishes a distance between the sign and reality, which is also an area of freedom, allowing for a comment or a perspective, in the sense in which Bakhtin (1990) said that not only language, but also a painting contains the point of view of the other on the object. Contrary to Bakhtin, however, I see no reason to deny a point of view also to the photograph, though it must be conceived to be of a more global nature than in a hand-made picture, where the point of view can be different for different parts of the picture (see Sonesson 1989b; 1999; 2015b). It is in the nature of the iconic sign to posit at the same time its resemblance and its dissimilarity to the object depicted: by the first stroke, the sign creates the expectancy of an identity that, by the second stroke, it must necessarily disappoint.
4Using the Husserlean term of predication, leaving aside such near-synonyms (in one of their senses) as statement, proposition and affirmation, I feel confident to claim that the picture can make predications in the manner of pictures: (a) by posing an identity and/or a likeness (an iconicity); (b) by introducing a dissimilarity and/or a divergence; (c) by transferring our acceptance from a to b. Alternately, the picture predicates (a) by installing an identity and/or a likeness (an iconicity); (b) by posing a second identity and/or a likeness (an iconicity) in a relation of contiguity or factoriality (relation of part to whole) with a (an indexicality); and (c) by transferring the accession of a and b to the relationship between the two (see Sonesson 2011b). It will be noted, however, that Husserl (1939) talks about perception as being a pre-predicative experience (which Merleau-Ponty translated to “expérience antéprédicative”). Indeed, even in his study of “pictorial consciousness”, where he starts out considering the picture as being closely akin to perception (a kind of “presentation”, that is, something which is directly experienced), but ends up approaching it to linguistic signs (“re-presentation”, that is, something which has at least an expression distinct from a content), certainly without identifying the latter two, he never broaches the idea of pictures containing predications. The case of Charles Sanders Peirce is clearly different, as Frederik Stjernfelt (2014) has recently reminded us: to Peirce (and to Stjernfelt), pictures (and notably diagrams; see Stjernfelt 2007; 2011) must just as well contain predications (“dicisigns”) as meanings conveyed by means of language.
5A by now classical take of this issue within cognitive psychology and, later on, cognitive science, is the idea of there being a difference between two ways in which information may be “coded” in the mind. For a long time, a distinction has been made between pictorialism (represented by Stephen Kosslyn 1994; Kosslyn, Thompson & Ganis 2006) which holds that mental pictures really are (or resemble) pictures; and descriptivism (represented by Zenon Pylyshyn 1985; 2003) which claims that mental images are in fact propositions (i.e. language-like structures). Jerry Fodor’s idea of a “language of thought” also seems to imply a descriptivist conception. As Evan Thompson (2007) points out, however, this distinction concerns the way mental images are instantiated in the brain (“subpersonally”). According to Thompson, nevertheless, both conceptions agree that mental images are experienced as pictures; but, in fact, this distinction is not very obvious when reading what Kosslyn and Pylyshyn write on the matter.
6In any case, the whole point of recognizing an “iconic code” in memory seems to be to deny to this code any statement function (see Thomas 2014). This is reminiscent of Nelson Goodman’s (1968) idea that pictures, unlike language, are dense and replete: that they can be divided into any number of units in any arbitrary way (see Sonesson 1989a). Husserl and Thompson, however, claim that mental images are not experienced as pictures, which makes this into a moot point, whatever we take to be true about pictures. Indeed, among other things, mental images are not located on a surface; they are not flat; and you cannot turn them over to inspect them (see Thomas 2014). Still, these differences do not necessary imply that mental images, unlike real pictures, contain propositions, although the fact that they are stored in memory could be taken to mean that, unlike real pictures, they convey only a rudiment of perceptual meaning. Whether we are talking about mental images or real pictures, however, the impact attributed to them, as in the recent case of the dead Syrian child on the beach, is more on the level of emotions than cognition. The part of “comment”, in Bakhtin’s sense, incorporated into the picture would be emotional and not propositional. Even the psychologist Rudolf Arnheim (1971), who makes a forceful plea for the existence of visual thinking, does not go as far as to talk about a statement function.
7In the first paper in which they apply rhetorical terms to pictures, “La chafetière” by Julien Key, which combines cat parts and coffee pot parts into a whole, Groupe µ (1976) ends up denying that this is a metaphor, mainly because both parts of the comparison are present, unlike what happens in a verbal metaphor. It might be argued that this only shows that pictorial rhetoric is more at the level of (explicit) comparisons (similes) than (implicit) metaphor. Rather than saying “coffee pot = cat” or even “The coffee pot is a cat”, the picture affirms that the coffee pot is similar to a cat in some respects. This then explains that Groupe µ (1992) abandons all further talk of metaphors and uses instead the term “pairing” (“couplage”). However, since some of these pairing involve the merger of the two items involved in a more or less complete way, nothing changes on the descriptive level.
8At first, this result may seem surprising. In an earlier publication, Groupe µ (1970, p. 107) would seem to propose a theory of metaphor (perhaps unintentionally, since they have never returned to it) which might be thought to make cases like the cat-coffee pot into ideal metaphors. The latter, we are told, “extrapole, elle se base sur une identité réelle manifestée après l’intersection de deux termes pour affirmer l’identité des termes entiers. Elle étend à la réunion des deux termes une propriété qui n’appartient qu’à leur intersection”. This conception of metaphor appears to posit a partial identity, which is transmuted into a total coalescence. On the face of it, the pictorial metaphor should thus be more perfect as such than the verbal counterpart, since it reproduces the semantic overlap of the content on the level of expression. In other terms, the cat-coffee pot is iconic in its very form, merging the expression planes of the two signs just as it merges the content planes. We could easily interpret this in terms of statements: we go from a demonstrated “The coffee pot is in some respects similar to a cat” to “The coffee pot is a cat”. It will be noted that this result is expressed in terms of statements, that is, in terms of predication: it does consist in ascribing cat properties to the coffee pot, and vice versa.
9Oddly enough, Groupe µ (1970, p. 108) then goes on to claim that this is the same thing as to say that a metaphor is made up of two synecdoches, one of which is generalising, e.g. from birch to flexibility, while the other is particularising, e.g. from flexibility to girl. But of course, this is a quite different conception: it amounts to forming a single class made up of all things which possess in common the property of being flexible, neglecting all further attributes. This theory therefore only accounts for the intersection, not for the union.
10According to another well-known theory of metaphor, usually attributed to Max Black (1962; cf. Goodman 1968, p. 68; Ricœur 1975, p. 109), the two terms brought together by the metaphor, called “vehicle” and “tenor” (in the terminology of I.A. Richards employed by Black), must persist in a state of “tension” and “resistance”, which will naturally be the case if one sign is substituted for another sign which is less befitting to the context; but, as I have argued elsewhere (Sonesson 1989a; 2015a), this is even more clear if, at the root of the construction, there is a wilful miss-classification of something into a category to which it does not belong. Where the Black/Goodman tradition supposes a non-coincidence of the two categories brought together, the theory formulated in Groupe µ (1970) appears to posit a partial identity, which is transmuted into a total coalescence. However, if a union of the two sets of semantic features is to result, the metaphorical connection has to be symmetrical, which means that, contrary to Tversky’s (1977) judicious observation, “A tree is like a man”, and “A man is like a tree” would be the same metaphor. But this is clearly unsatisfactory; so, let us suppose instead, that in the metaphor, from a partial overlap, meaning is extended to include the complete set of features of the reference point, but not those of the subject of comparison. In the case of the metaphor “A man is like a tree”, the man would not only acquire roots, as Tversky suggests, but he would be more generally tree-like, while there is no hint at all that the tree should be seen as man-like (of course, because of the context, or because of peculiar presuppositions prevailing in a given sociocultural Lifeworld, the “feature space” can be so constrained, as to focus on particular features of that space, as, in this instance, the “roots”, on Tversky’s interpretation). However, the union of the overlapping parts of the two feature sets, and the remaining features of the reference point is simply identical – to the reference point! The metaphor would be a simple identity; and since this is obviously untrue, we are led back to the first metaphor theory considered above.
11Considered as a statement (a predication), in any case, the pictorial figure, whatever its nature, is problematic, because its direction is underdetermined. The real problem with “la chafetière” is that it is indeterminate between the statement “The cat is a coffee pot” and “The coffee pot is a cat”. It is easy to see, using linguistic examples that there is an appreciable difference between the relations posited between the king and the lion in the two statements “The lion is the king of the jungle” and “The king fought like a lion”. Or, to take an example more famous in recent linguistic literature, there is a world of difference between saying “The surgeon is a regular butcher” and “The butcher works like a true surgeon”. What really stands in the way of any talk of pictorial metaphor is thus not the explicit presence of both elements involved in the comparison, nor the impossibility of formulating pictorial meaning in terms of statements, but the fact that the comparison, as far as the pictorial meaning goes, is necessarily symmetrical. Labels (or even general expectancy) may take care of this issue, however, as we shall see (in 1.3.). Mutatis mutandis, this applies also to perceptual reality. There is perhaps not much risk that we shall ever get to see something similar to the cat-coffee pot in perceptual reality, but if we consider the kind of monsters that mediaeval people expected to find in other parts of world, those made up of half a body of a dog or a fish and half a human body, we should probably be at a loss to decide whether we a facing a man with a dog’s head, or a dog with a human trunk, and so on. As I have suggested elsewhere (see Sonesson 2014b), pictures do not dispose of any regular devices such as “theme/rheme”, “given/new”, etc., which serve to establish what linguistics call an “information structure”, i.e. a hierarchy of importance, from some particular point of view, in the meaning presented. Going beyond the picture and perceptual reality itself, however, there is such a device, which is the act of attention.
12Unlike most philosophers, Peirce clearly thinks there are propositions (statements) or, to use his terminology, Dicisigns, all over, also in perceptual reality itself. If this only means that perceptual reality may be conceived as structured like a proposition, as in the case of Husserl’s pre-predicative reality, there may not be any problem. But does Peirce also mean to say that percepts themselves are assertions, in the sense of a subject taking responsibility for manifesting a particular perspective on the world at a particular place and time? Stjernfelt (2014, p. 55, p. 103) pokes fun at Austin’s “famous quip” according to which Peirce, in spite of his 66 divisions of signs, did not distinguish between a proposition and the assertion of it. Although Peirce certainly made this distinction, and, true to his spirit, added more variants, it seems that he did not attend very often to the difference, and Stjernfelt certainly does not. If indeed we generalize the notion of proposition or Dicisign to the perceptual world, to pictures, diagrams, gestures, and the like, as Stjernfelt clearly does in his book, what would we then take to be equivalent to the assertion of such a proposition? The answer could be relatively straightforward, as long as we forget about the perceptual world itself. But then, the plot thickens.
13One of the most classical examples involving pictures from Peirce’s writings is the portrait with a label, also mentioned by Stjernfelt (2014: 67). Is this an assertion, or only a potential proposition? No doubt the label functions as an index here, but this is presumable already true about the proposition. Again, Peirce even admits that photographs (but no other pictures, apparently) already in themselves contain both parts of the Dicisign, i.e. both subject and predicate. According to Peirce, a Dicisign is made up of two terms (sinsigns), one of which is an index (what/where it is) and the other an icon (what properties it has). And photographs are in themselves at least as much indices as icons, according to Peirce. But is not the notion of index used in different ways in these two cases? Does this mean that they are asserted propositions, or only potential ones? This all seems rather unclear. These are special cases, however. It seems rather natural to say that an item of spy photography, or a photo-finish, must be taken as assertions, i.e. as asserted predications (see Sonesson 1989b, 2015b): this would be the case with a picture conveying the proposition “Soviet missiles are being emplaced in Cuba” which was asserted in 1962; it would also apply to any particular pictorial assertion of the proposition “This horse won” as long as the assertion took place directly after a particular race, and the demonstrative pronoun pointed truthfully to the horse which really won the race. But these are special cases. So is the case mentioned by Stjernfelt (2014: 183ff), the Zapruder film, which could be conceived to be a pictorially formed assertion, to the extent that it is used to show whether or not president Kennedy was shot by more than one person.
- 1 According to the Greimas school narrativity is at the basic level of all semiotic resources, which (...)
14It is curious that Peirce, and Stjernfelt following in his tracks, think that a picture would ordinary necessitate a label to become a proposition (or at least a quasi-proposition), but that a photograph can dispense with labels and still be a (quasi-)proposition. Of course, photographs are indexical, even in the sense of being causally, and not simply “existentially”, connected to their object (see Sonesson 1989b; 2015b), but this only serves to illustrate the ambiguities of such terms as indexicality (and iconicity), which has already been profusely discussed (see Sonesson 1989a, 1989b; 2010; 2011a). Elsewhere I have argued that even static pictures, whether photographs or paintings, may tell a story (Sonesson 1997a), which implies that, in some sense, they are made up of quasi-propositions1. Thus, for instance, a picture may assign the property of falling from the sky to Icarus, and that of ploughing to a farmer, and so on. From this point of view, there is no difference between a painting and a photograph. There is a difference, of course, in the truth conditions, or, more exactly, with respect to our Lifeworld experience, or, to use a Peircean term, our collateral knowledge. We know that photographs, at least as they were before the Photoshop revolution, have a somewhat more direct relationship to the perceptual world than paintings. But this has no bearing on their quasi-propositional character.
15There is a paradox to the Peircean account: on the one hand, a drawing needs a label to become a proposition; on the other hand, a photograph is a proposition even without a label. Moreover, it is not clear whether this difference concerns a proposition or an asserted proposition. Just like a drawing of Peirce, a photograph of him would seem to need a label in order to embody the proposition “This is Peirce” (including that Peirce had such a nose, such a beard, etc.) – that is, it is an asserted proposition. Perhaps the photograph can work as an asserted predication, if there is contiguity in the actual world between the photograph and the subject, not only in the past. But even if this is true, it does not seem to be restricted to the case of photographs. To have the picture on the bird cage (whether a drawing or a photograph) work this way, it must have a label in addition. Normally there would be a picture of the bird with a label affixed below it on the case, and together these two items will help to identify some of the birds inside the cage as being of the species so described. This does not seem to be very different from seeing the photograph of Peirce with his name below it, and then identifying him, not in the cage, on not at any place close-by, but at the next conference. Of course, we latter-day semioticians will have to make do with recognizing Peirce from one picture to another. Whether the picture of Peirce or the bird is a photograph or a drawing would not seem to be of any consequence in the present case.
16In conclusion, a picture may need a label to appear as an asserted proposition (see section 1.2), but this seems to apply to photographs just as to drawings and paintings. This is not to deny that there is a difference between drawings and photographs, as they are conceived in our Lifeworld experience, or, to use a Peircean term, in our collateral knowledge (even though, in our present socio-cultural Lifeworld, the genres are getting ever more difficult to tell apart). Nevetheless, this does not seem to have any obvious bearing on the quasi-propositional character of both genres. On the other hand, collateral knowledge is needed to identify the spy photograph or the photo-finish as such.
17If we define an assertion/predication as a verbal construction, then it is trivially true that the picture cannot assert anything. However, if the predication is more generally defined as a transaction, by means of which a specific property is assigned to a particular entity, then it is possible for the picture to make predications – not, to be sure, in the way of language, but in the manner of pictures. This can work if there is a standard of comparison, and a picture can offer that to the extent that it is both similar and different from what it depicts (see Sonesson 1997b; 2011b). On the face of it, the case of perceptual reality is more problematic. Perception can no doubt be described in terms of a property being ascribed to an object. Nevertheless, no matter how close the structure of perceptual experience can be to a predication, this experience is not subject to being stated, except in particular cases such as shop windows, installations, performances – and pictures. If we start out from these particular cases, however, we may be able to define some constraints which serve to make percepts into statements: to transform pre-predicative reality, in Husserl’s terms, into predications.
- 2 As was pointed out by an anonymous reviewer, comparativity in this sense should not be confused by (...)
18Even if pictures are able to predicate, there are two important differences between a picture and perceptual reality: first, a picture, as suggested above, allows for a comparison between itself and that piece of reality which it invokes, but percepts cannot be compared to anything else (this may be seen as a generalization of the ability of Dicisigns to take on truth values); in the second place, a picture involves a frame, which also means that it has at least an elementary mechanism for shedding parts of reality which are not relevant, and for organizing reality within the frame in terms of focus and margins, whereas perceptual reality has no determinate limits (it has ever more outer horizons), and its focus is vague and/or continuously shifting. Let us call the first difference the comparativity requirement, and the second the framing requirement. Then it could be said that not only a picture, but also a shop window as well as an artistic “installation” fulfil the framing requirement, and if they consist of objects that are arranged in a way that is clearly different from that of ordinary life, they also fulfil the comparativity requirement2. The ordinary Lifeworld does not fulfil any of these requirements, however.
19Yet there is a way in which the experience of perceiving could be said to correspond to predication, but, as we will say, to a predication lacking the predicative function. That is to say that it has, in some respects, the same structure as a predication, or, in other terms, a statement, but there is nobody around to state it, nor any specific time or place when the predication is made (there is of course a time and place of the perceptual experience, and even a person having it, but that is something quite different – more like the time, place, and person the predication is about). To avoid all ambiguities, let’s say that perceptual reality may be structured into propositions, but that these are not asserted. This could explain why Husserl (1939) calls this pre-predicative experience. Nevertheless, Husserl does not seem to explain the difference further. In fact, if we prove deeper into the logic of perception, any piece of perceptual reality turns out to be structured as a whole matrix of propositions – or perhaps we should say, of potential propositions.
- 3 A point made more explicitly by Sokolowski (1974), but before that also observed, at a more general (...)
20If, like Husserl, we start out from perception, we might want to vary the different ways of perceiving the cube, which was indeed a recurrent example in his work. There are indeed many acts of perception that are still the perception of a cube, and even, more specifically, the perception of this identical cube. Most notably, of course, the cube may be seen from different sides, from different perspectives, only a part of which may appear in a peep-hole, and so on. In fact, Sokolowski (2000, p. 17ff) extended Husserl’s observations pointing to a threefold division: the cube has sides, some of which are seen entirely, while others are more or less hidden away; it also has different aspects, because it may be perceived from different points of view; and it has profiles, because it is possible to revisit some of the aspects over and over again. Husserl collapses the two later cases, calling both of them noemata or adumbrations (“Abschattungen”)3. Taking a cue from the two possible translations into English of the German term used by Husserl (“Würfel”), I pointed out that the cube may be seen immediately as something more peculiar, in this case, a die (see Sonesson 1989, p. 34ff). In doing this, I added a variation to the original pristine cube, which was now not simply a cube anymore (although, in this case, still remaining a cube). Since a cube can only be seen as a die in a given culture, in which dice are used, this provides a cultural dimension to phenomenological experience. As we now know from posthumous publications of late writings, Husserl was very much aware of this cultural-historical dimension. In later works, Husserl talked about sedimentation as the way meanings accrue to objects through experience. In the final lecture given by Husserl (1954), history as such, and the particular historical moment, were made into the theme. But the posthumous papers show these preoccupations to have been very much at the centre of phenomenology (see Steinbock 1995; Welton 2000). In any case, given all the sedimentations which may be culturally layered onto a die, the quantity of (potential) propositions applying to is immensely greater than those appended to a cube.
21While the particular noema by means of which I presently see the cube only contains three of its sides in different perspectival deformations, I immediately see it as a cube, complete with its six sides, not as some strange object I hypothesize to be a cube. Through an act that Husserl calls reflection, the phenomenologist, the psychologist, and the aesthetically minded contemplator may choose to attend to the acts of consciousness and their corresponding noemata instead, thereby transforming them into new objects with their own noemata. In normal consciousness however, the act will only give a particular modification to the perception of the object, a tinge of meaning: some parts of the object appear more specified, others only roughly outlined. What is merely sketched out in one noema may be brought out in a number of others, and the knowledge that we can always go further in the exploration of the object is part and parcel of our perception of the object, as expressed in the Etc. principle. Whereas retentions of already seen sides are the basis for further exploration, protentions may be specified or rejected when the earlier unseen sides come into view (see Husserl 1939, 1962).
22Gurwitsch (1957, 1974), who approaches this Husserlean conception to the ‘‘spontaneous phenomenologies’’ of the Gestalt school, has pointed to the ‘‘Gestalt-coherence’’ with which the mutually confirming noemata form the object of perception. Criticizing Husserl because he seems to consider the object itself as a separate instance, an “X” which is the bearer of the noemata, Gurwitsch (1974, p. 254) tells us that the perceived thing is “nothing else than the internoematic system itself, i.e., the system of multiple adumbrational presentations and of the properties and qualities exhibited in those presentations”. Similarly, the predication (“X is red”, and so on) which Husserl conceived to be a “synthesis”, an adjunction of new properties, is really an “analysis”, an explicitation of what is already contained in the horizons of the perceptual thing. If so, the meanings of the cube cannot be described, as Husserl maintains, as “X is blue”; rather, it should be something like “this cube (which, apart from obligatory cube-properties, is blue, worn on the edges, rather big for a cube, etc.) is blue”. In other words, a lot of propositions are co-given, but most of them are adumbrated back into the whole, while one single proposition is picked out of the manifold by means of the act of attention. One of the interesting things that Stjernfelt (2014, p. 55ff) unearths concerning Peirce’s Dicisign doctrine is the idea that every assertion contains two different signs having the same object. In more classical philosophical lingo, this means that the predicate and the subject have the same reference. This is reminiscent of, but less precise than, Gurwitsch’s proposal to see the predicate as being picked out from within the subject.
- 4 Gurwitsch probably means to say that this is the structure of predication, but such an organization (...)
23Even though I agree with John Drummond (1990) in criticizing Aron Gurwitsch’s idea that the perceptual object is made up of all its possible noemata as parts, I think we can still agree with Gurwitsch (1974, p. 254ff) that a perceptual predication must be conceived to be an analysis, starting out from the manifold of perception, rather than an synthesis. Thus, unlike a predication, perceptual experience starts out from the whole and goes on to particulars, that is, it narrows down the perceptual focus. It always has a theme, a thematic field, and a margin (see Gurwitsch 1957, 1964, 1985; Arvidson 2006). I therefore cannot agree with Stjernfelt’s (2012) formulation according to which “[p]erception already involves intricate logical inferences performed within the visual material”, a phrase which is reminiscent of Helmholtz’ “unconscious inference”. When more closely considered, then, the structures of predicative and pre-predicative experience are not really the same4.
24Percepts contain “judgeable contents”, in Gottlob Frege’s sense, not assertions or even statements, propositions or predications. Nevertheless, to summarize what has been said above. I think they can be made into statements – asserted predications, by means of:
-
The comparativity constraint. A picture allows for a comparison between itself and that piece of reality which it invokes, but percepts cannot be compared to anything else. However, the act of attention may be directed to something which we expect to perceive, and something else may be substituted for what was expected.
-
The framing constraint. A picture involves a frame, thus having an elementary mechanism for shedding parts of reality which are not relevant, and for organizing reality within the frame in terms of focus and margins, whereas perceptual reality has no determinate limits (it has ever more outer and inner horizons), and its focus is vague and/or continuously shifting.
25Nevertheless, attention may be fixed on a particular part of perceptual reality, so that outer and inner horizons become irrelevant for the time being. Both the framing constraint and the comparative constraint may be instantiated by socially instituted devices. A shop window, for instance, is a place which by definition is meant to draw intention to what is contained in it, and the same is of course true of a theatre scene. This is an instance of the framing constraints. On the other hand, a place which is meant to contain a certain class of object, but which in reality contains something else, such as a perambulator left on a parking lot, draws attention to itself. This is a case of the comparative constraint. But, more generally, the act of attention, as it is directed to different things in the environment, may be said to operate the assertion of a predicative structure. This goes to show that the devices of elocutio may gain a statement function and thus be more on the inventio/dispositio side.
26It makes sense, I believe, to think of the act of attention as being as kind of proposition. Taking my inspiration from Gurwitsch’s (1964) ideas about the “field of perception”, later reconceived by Sven Arvidson (2006) as the “sphere of attention”, I have suggested that the gaze may function as an assertive device (Sonesson 2012). When first writing about attention as a device for asserting a perceptual act, I was actually thinking about the assertion being made for the perceiving subject itself (Sonesson 2012). It is, however, a curious fact that human beings are alone among all animals in possessing the white of the eyes, the presence of which in other fellow human beings is what allows us to see more clearly than any other animal what another person is looking at. Many animals can follow the direction of the head of the other to discover what it is looking at. However, it seems that only human beings can follow the gaze directly, since such observation is facilitated by the unique morphology of the human eye, notably by the exposed white sclera (see Kobayashi, Hiromi & Kohshima 2001). This is of course what makes joint attention possible.
27It would be more to the point to say that the act of attention as such is what constitutes a proposition (or, better, a quasi-proposition, to differentiate these two terms better than Stjernfelt does in his book), but that only the act of attention which is attended to by another subject (in the sense of person) makes up a quasi-assertion. Attention is what transforms a piece of perceptual reality into an asserted predication. That is, basically, it is the gaze of a subject which makes the predication into an assertion. Nevertheless, it is also the observed gaze of the other which serves to comply this function. That is, there are Ego-assertions and Alter-assertions. Thus, it would seem that also when it comes to other than linguistic means of making assertions, human beings are privileged.
28Pictures are in some ways similar to language, and in other ways they are not. We have suggested here that one thing which pictures may share with language is the capacity of asserting proposition, though this must of course not be understood to work in the way it does in language. We understand a proposition to mean simply the assignment of a particular property to a thing appearing in the picture. Such a proposition can be considered to be asserted when a particular subject can be understood to take the responsibility for this property being attributed to the thing at some particular time and place. Perceptual reality is even less similar to language than pictures are, but there are certain cases when even perception may appear to work in similar ways to language. At the level of propositions, it is easier to think of similarities between language and other semiotic resources than on the level of assertions. Yet, there may be cases when even perceptual reality may be conceived of as containing asserted propositions. This may happen because perceptual reality, like the picture, is constraint by a standard of comparison, or because it is somehow framed of from the rest of reality. More generally, however, such framing may be operated by the act of attention. The gaze is no doubt the primary source of assertions, not only to the gazing subject, but also to the one observing the first subject. This opens up a vast domain of semiosis, about which we know very little so far.