ISSA Proceedings 2002 – The Conceptual Basis Of Visual Argumentation – A Case For Arguing In And Through Moving Images

• by: László Tarnay - University of Pécs

1. The problem of explicating images as arguments
Argumentation theory in the past decades have evolved basically from what was once called new rhetoric. This genetic trait has considerably determined both the methodology and the scope of the theory. It determined its methodology in the sense that the definition of argument always already implies that an argument is something that can be made explicit; that is it is explicated formally as a step within a chain of reasoning. This requirement imposes propositionality on anything to be assessed as argument. Let us call this the requirement of propositionality. No wonder that those forms of communication which do not bear propositionalty on their sleeves like pictures, music or smell should fall outside the scope of argumentation theory. But not entirely. Undoubtedly  there is a growing interest in analyzing images (first and foremost advertisements or cartoons) as explicit arguments or as potential sources for retrieving arguments in certain contexts (especially when they are used with an identifiable intention to persuade). In these cases images, sounds or other non-verbal objects (henceforth we sample out images as a paradigm case) are treated as texts or conveyors of texts. It is in this manner recent approaches to extend argumentativity to visual objects (e.g. Groarke, 1996) tend to see a continuity rather than a rupture between the verbal and the visual (forms of argumentation): they are looking for a general level at which verbal and non-verbal forms of argumentation can be equally described and compared. This general level is expressed in a meta-language and it is only in the latter that images can be said to fulfill the requirement of propositionality.

Yet, with the studies of intermediality on the horizon, it seems to be a more tenable alternative to drive a wedge between image and text, instead of giving full vent to their convertibility. The main reason is that there is an equivocation in the explication of images as arguments. For, when images are ‘translated’ into a meta-language to compare with verbal arguments, they are taken to represent arguments, viz. they are see as intrinsically argumentative. The theorist’s aim then naturally is to recover ‘those’ arguments in his meta-language. On the other hand, images can be used as a whole as arguments, just like any object can. For example, Sperber & Wilson describe a situation when Mary wants Peter to mend her hair drier not by asking him openly but by leaving it lying around. They call such cases as instances of ostensive-inferential communication, viz. communicating without a code. (Sperber & Wilson, 1968, 30) It can hardly be said that the hair drier represent an argument. Rather it is meant to elicit some inference in the target person, viz. it is used to persuade him (or her) or to make him (or her) perform a particular action. The theorist’s aim can only be to reflect the cognitive effect achieved by the hair drier left lying around, and the thus recoverable arguments are external, and not intrinsic, to the object in question. If we accept that images can be used in this manner, they can fulfill the requirement of propositonality but indirectly: they cannot be said to ‘translate’ into the meta-language.

The difference between the internal and the external question of argumentativity regarding the visual may be compared to the difference between the description or report of a performative act and the making of it. One can describe or report the meaning of an image like a traffic sign as, say,
1. You should turn left.

but it is totally different from actually using that sign for some corresponding purpose. that latter may be expressed as an imperative (2):
2. Turn left!

(2) differs from (1) in that the latter can be true or false, but not the former. (1) is a deontic statement, whereas (2) is a performative. We do not claim that the deontic/performative distinction is the same as the internal/external argumentativity of the visual, but it can be used to illustrate it. Another analogy in a similar vein is the difference between asserting that one has some pain or other sensory experience and the corresponding phenomenal consciousness of it. Or still going further, and following Wittgenstein, we could venture to say that the performative use of images is similar to aspect seeing like, for example, hearing a tune as sad or joyous. The latter two analogies may even be better in revealing that the translation or verbalization of the visual cannot capture the performative uses of images. For there are no proper linguistic forms to express phenomenal experience or aspect, which would be analogous to (2). Forms like (3) or (4) appear to be ungrammatical in the Wittgensteinian sense.
3. Taste it bitter!
4. See it vertical!

With (3), (4) and their likes we get as far as possible from the fulfillment of the requirement of propositionality. Then the crucial question for the present argument concerning the visual is the following: Can we still do argumentation theory without the requirement of propositionality? In other words, can images and other non-verbal objects persuade, or rather be used to persuade, in a way that is essentially (i.e. intrinsically) non-propositional? This seems to constitute a non-sequitur, since reasoning is most often, if not always, taken to be a chain of arguments leading from premises to some conclusion. It seems to be hard to imagine what a non-propositional chain of arguments would look like. One could certainly try to avoid this problem by saying that ‘argumentation aims at the adherence of minds’ (Perelman & Olbrecht-Tyteca, 1971, 14), and since adherence is to values and topoi, images can be seen as a means other than verbal to bring about such an adherence. It is acknowledged that when the theorist makes explicit an argumentation conveyed by (a sequence of) images, he or she does a work of hermeneutic reconstruction: he or she constructs an argumentation (premises and conclusion) on the basis of:
a. the hierarchy of meanings associated with, or conveyed by, the images (logos),
b. the rhetorical situation they occur in, or the intention (ethos), and
c. the emotional effect evoked (pathos). It is in this sense that ‘some topical arguments can be manifested both in verbal and visual communication’ (Kjeldsen, 1998, 457). To repeat, propositionality is not a property of the images, but of the meta-language in which the arguments associated with, or elicited by, them are made explicit. More loosely put, it is ‘instantiated’ visually.

One basic problem with this all-inclusive rhetorical approach to visual argumentation is that it widens the door of rhetoric too much; e.g. an act of shooting can be an instance of argumentation, if done with proper intention in a proper context. A more interesting point not unrelated to the previous one, however, is that it abstracts away from the specific traits of the visual, their perceptual quality, the mode of articulation and the mode of (perceiving) their physical substance. It has become a common contention in the study of visual metaphors that metaphors or analogies are ‘grounded in perceptual similarity’ (Veale, 1999, 39). This fact has not received enough attention by argumentation theorists despite the claim that the figurative aspect of both the verbal and the visual can be covered by an all-inclusive rhetorical approach like the one above, one can hardly see, however, how the figurative force of the visual could be captured in a rhetorical approach to visual argumentation with almost all its attention devoted to the topoi; so much so that it is said to be difficult to achieve a change in topical hierarchy (of values), if the opposition in the viewer is too strong: it would require the much more explicit and clear verbal argumentation.
The appeal to verbal argumentation we contend is a false start here. The argument proposed in this paper takes its start from the fact that the power of the visual is actually so strong that it might even overcome the most fierce resistance of existing opinion. And this power is independent of the emotional effect that visual images are often said to cause. It is indeed acknowledged by the all-inclusive rhetorical approach that images may well have a greater emotional effect that texts. For take e.g. advertisements or political campaign posters, which are most often the target of visual argumentation theory. No doubt that these are designed in a way to achieve the maximum effect (as great as possible) on the (change in the) adherence of minds. Yet what we claim here is not that certain images achieve a significant effect because they act upon our emotions; rather it is vice versa, they are emotionally laden because they operate at a relatively low level of psychological or cognitive operation. The point is that ‘associations between concepts are automatically recognized and noted’ (Veale, 1999, 39), and thus it precedes, if not grounds, all ‘higher-level’ conscious construction of arguments and metaphors.
The all-inclusive rhetorical approach bypasses this inherently perceptual aspect of images when it makes the retrieval of arguments dependent on the topical hierarchy of values. In other words, it trades the formal pragmatic aspect (logos: topoi) for the structural or compositional aspect of the visual by associating the latter to a strictly semiotic, hence outdated, analysis. But there are certain ‘compositional’ features that are there to drive or guide the eye (or any other sense organ) in exploring the images for possible arguments. By compositional we do not mean simply the non-linear, non-discursive aspects emphasized in (Gilbert, 1997); we mean visual cues (of depth, motion, distance, etc.) which are directly perceived as it is theorized in ecological psychology after the pioneer work of J. J. Gibson. These features we argue are intrinsically perceptual, and they are perceived and processed at a relatively low level. In the remaining part we will refer to them in short as the genuinely visual. It remains to see, however, how this low level processing of the genuinely visual could be identified and characterized.

2. The identification of a lower bound of visual argumentation: from ecology to the theory of blends
The suggestion is then that the genuinely visual be defined as a lower bound of argumentation precisely in the sense that it serves as a condition of possibility for producing and retrieving arguments. No doubt that the identification of such a low level is highly controversial at least for two reasons. First and foremost because there exists a long tradition of an inferential theory of perception, often attributed to Helmholtz, that considerably determined both the hermeneutics of art and the theory of perception (See e.g. Rock, 1983, 1997). Briefly, the claim is the operation of our sense organs (or whatever it is that computes and processes sense data) can be described as an inferential activity under the level phenomenal consciousness (The strongest version that our eyes ‘argue’ can be found in Bonfantini, 1987). Certainly there is a difference between the claim that they could be described inferentially and that they actually compute inferences. Yet the adherents of the inferential theory have not made much of this distinction. Maybe the reason is analogous to why the all-inclusive rhetorical approach neglects the aforementioned ‘compositional’ features of the visual: any concession that there might be something non-inferential in the processing of the visual would be tantamount to admitting that the requirement of propositionality does not apply unrestrictedly to visual images. In other words, it would be the acknowledgement of a lower bound to argumentation. The retrieval of arguments should not be confined to higher – semantic and pragmatic – level of processing, but it should be grounded on certain ‘automatic’ processes. But what are these automatic processes? Well, the appeal to direct, or lower level, processing as a lower bound of argumentation avoids the pitfalls of describing inferences. For if it can be proved that there are certain features which are directly perceived, the description of these features as arguments (added premises) can never be taken to mean that the image in question represent arguments. Clearly, the description belongs to a meta-language in which it makes explicit the conditions of possibility of using the image for some argumentative purpose.
The second reason why the identification of low level can appear controversial is methodological. The retrieval of arguments from images trades on the – in our mind most problematic – aspect of verbal argumentation: the reconstruction of missing premises. It constitutes the problem of making explicit. While it is an adage in verbal argumentation that it is seldom, if ever, the case that all the premises are explicit (note the need for principles of bridging), the reconstruction of the missing premises has always certain given ones to start with (it never starts from nil). Visual arguments, however, in most cases have to be recovered in their entirety (We say most cases only to exclude the ones when the image is accompanied with some verbal explanation or inscription). One can easily formulate a kind of slippery slope by saying that if one premise can (and in fact should) be reconstructed, why not reconstruct all the premises? It leads to a reductio ad absurdum of visual argumentation to the effect that any image could be interpreted argumentatively in some way.

Many would interject at this point that visual arguments presuppose visual hermeneutics, the recognition of figures, scenes etc. which would considerably restrict argument reconstruction. No doubt that it would, but appealing to such a hermeneutics would not in itself help with the reductio in question. For it is functionally equivalent with the pragmatic move in the all-inclusive rhetorical approach above to draw upon the topical hierarchy of values as a condition on argument retrieval. But while verbal argument analysis (the identification of topoi) has a semantics to start with, there is no such semantics of images other than the one ‘grounded in perceptual similarity’. Without clarifying what this similarity is, the appeal to pragmatic factors remains circular: an image is argumentative if there is a certain hermenuetics that its use makes accessible. But the accessibility of the hermeneutics rests with its intended argumentativity. Any approach that disregards the grounds in perceptual similarity is bound to make authorial intention and hermeneutical interpretation interdependent, that is, pragmatically given.

We find this second methodological problem of argument retrieval analogous to the traditional problem of the potential narrativitity of images. While otherwise narrativity and argumentativity are complementary (and many times exclusive, see Parret, 1986), the narrative and argumentative interpretation of images face the same problem of sequentiality: how can a sequence of steps (be them narrative or argumentative) recovered from the depiction of a single step (still image)? The methodological answer is of course that it can be done by drawing upon, viz. extracting, the missing steps. This is already implied in the instruction given by Lessing that painters should try to depict the ‘fertile moment’ (the one  immediately preceding the climax of the action to be represented) in order to enable the viewers to recover the entire story. Lessing, no doubt, wants the viewers to replicate the authors cognitive processes. Disregarding the problem of cognitive symmetry, narrative reconstruction follows the same model as argumentative interpretation. Both run the risk that the recovered sequentiality is nothing but the property of the cognitive process itself, and not the property of external events (story) or arguments (premises-conclusion). Without grounding the interpretation in the image itself, the circularity cannot be avoided.

We have already proposed that low level processing should be understood in the sense of direct perception in ecological psychology after the pioneer work of J. J. Gibson. We cannot recapitulate the whole history of the debate between the Gibsonian theory of perception and the inferential approach. The debate has flamed up most recently with growing empirical evidences which seem to underscore either the one or the other. It culminated in approaches to reconcile the two theories (See most recently Norman, 2002). It also seems to settle on the issue of the division of labor of two visual processing systems, the dorsal and the ventral ones. Without going into details at the neurobiological level, the crux of the matter is the relation of the two systems. Are they functionally distinct? Do they have access to different types of information? Do they differ in the way they operate? That is, do they constitute two different modes of processing? Or are they rather structurally different? In the light of currently available data, it seems that both systems have access to all kinds of visual information, which explains – together with the plasticity of the brain – why one can take over a task assigned to the other in case the latter should be impaired. On the other hand, they show considerable difference in the types of information processed: the dorsal is responsible for the perception of real and short range apparent motion, and possibly, depth (linear perspective and motion parallax), for it can very fine discriminations in time, while the ventral system is slow in time, but processes distance, shape and color, and in general, is very good at observing details. It is this fact that explains why categorial thinking and phenomenal consciousness are most often associated with the ventral system. So, they constitute partially distinct pathways with different processing capacities, but still with the ability to take over certain functions. Furthermore, and not with the least importance, the capacity to draw inference or to deliberate is also assumed to be essential ventral, whereas the dorsal system is characterized as a means of direct perception (especially of motion).

For our argument here, however, the most important question is whether there exist cases of rivalry between the two systems. At first sight, the division of labor seems to exclude rivalry. Yet, since we know little of how the different types information are integrated, if they are, after the two systems have done their share of processing, we should be very cautious in our answer. Thus, when it comes to the question of identifying low level processing as a lower bound of argumentation, basically we have two choices; either single out the type information processed by the dorsal system as the condition of possibility of all argumentative interpretation (of visual character), or concede that at least certain information carried by the ventral system occurs at this level. We do not want to settle this issue here. We would like, however, to appeal to one particular dominant theory in cognitive science, namely, the theory of blends, or Conceptual Integration Networks (see e.g. Fauconnier & Turner 1998, Hofstadter, 1995), which makes use of so-called image-schemas operating at a ‘low-level of description’ and ‘serve both as selectional filters and basic structure combinators for input spaces’ (Veale, 1999, 42). Furthermore, and more importantly, it is such image-schemas that makes it possible to recruit ‘perceptually-grounded conceptual blends’ so much so that concepts which otherwise have nothing in common become related by means of a bridge-relation, or in fact a mediating blend which connects concepts with common ‘perceptual (i.e. appearance-related) properties’ (Veale, 1999, 45). In other words, metaphoric relations are made possible by resemblance-relations through mediating blends. It is visually given resemblance, or iconicity in short, that gives way to higher-order inferences and reasoning.

The perceptual grounding of Conceptual Integration Networks constitutes, in our mind, that lower bound or low level processing that can lead us out of the hermeneutic circle of topological hierarchies applied in visual argumentation theory. It also explains how the requirement of propositionality is bypassed when establishing a framework for visual argumentation. No wonder that certain approaches to visual arguments, like Groarke, 1996, try to extract a coherent propositional structure from images which contain the depiction of physical incongruities, looking for a direct mapping between the elements (tenor and vehicle), instead of searching for mediating blends. Whether blends have a propositional structure, or they could be made explicit propositionally, does not influence our argument here. For it is not them (the blends themselves) that matter but that they are presented visually, or rather, they are perceptually cued. Were it not so, and this is the very basic of our argument here, visual argumentation in any sense would be impossible. This is not to deny the relevance of other pragmatic factors, like hierarchies of values, but to contend that they are not sufficient to identify visual arguments. The perceptual grounding also explains how and why metaphoric relations can become a source for higher order reasoning.

3. Three modes of visual argumentation
In the rest of the paper we identify and describe briefly three different modes of what could be called visual argumentation. It is important to emphasize that we do not claim that these modes constitute visual argumentation in themselves. The most we can say at this stage of the research is that they constitute modes in which the visual appears to be translatable, or transferable, into the verbal. This may be very strong, maybe even self-evident, in the first mode. The modes can be ordered from the purely textual to the genuinely visual. In the purely textual mode images are nothing but the visualization of verbal arguments. Or vice versa, they appear to be entirely verbalizable. Classical allegories belong to this mode, which is also the one on which the all-inclusive rhetorical approach to visual argumentation capitalizes. For example, in Daumier’s drawing ‘The New Aerodynamics’ cited and analyzed in Groarke, 1996, we see Europe as a woman impersonating Peace resting on the tip of a bayonet. In Groarke’s interpretation, the picture says that ‘European peace is not stable because it rests on armement’ (Groarke, 1996, 109). This interpretation is typical of visual argumentation theory that ‘extracts’ propositions from images. The approach seems to be justified by the allegorical quality of the drawing. We do not want to deny the relevance of such interpretations. In fact, we propose that allegorical representations should be singled out as a first mode of visual argumentation when the image is meant to translate the verbal.

Yet even in classical allegories like Daumier’s which are a kind of visualization of some text or verbal argument, one can trace elements which have an intrinsically visual character, viz. they do not wear propositionality on their sleeves but are directly (or indirectly) perceived. Such is for instance the perception of planes, the ‘cues’ of gravitation, depth or shape. To consider these genuinely visual elements as co-constituents of visual argumentation (together with verbalizable elements) is to switch to another mode – let us call it mixed – in which essentially tropological: it consists in making an essentially creative attempt to combine incongruous elements within a blend, viz to see the figure of Europe as a blend. For the construction of blends is motivated by incongruities or even contradictory properties (of the different input spaces, say, of peaceful rest and restless armament). In this case the drawing gathers its force not only from the fact that the woman is resting on a bayonet; actually, such an interpretation overlooks the allusion in the title to flying objects. The force of picture is due to the incongruity between two states: lying and floating. The question of Europe’s personification as a woman (representing Peace) can only be answered within the blend. The meaning of instability could not be created with other representations of Europe (e.g. its map), nor with some flying object. Or at least the visual impact would have been much diminished (Cf. Veale, 1999, 44). What we have in the blend instead is a kind of aspect change, or double-think, in that we see a human figure both lying firmly on the ground and floating on the tip of a bayonet at the same time. This feeling can be explained by recalling the dorsal/ventral divide of visual processing. There is little doubt that we perceive a figure drawn on a flat surface as standing or lying on the ground. In fact with a lying figure the chances are greater for seeing the surface as horizontal, while with a standing one they are more balanced, viz. the figure could be ambiguously standing on and in front of a plane. Note that the upper part of the bayonet may well be seen as pointing to the sky, and hence being diagonal to the ground, viz. to the same plane on which the woman is taken to be lying. To see the figure as resting on the bayonet would require that we take the surface to be both horizontal and vertical at the same time. An impossible visual manoeuvre.

Add also the fact that we tend to see masses like a human figure as gravitating to the ground, and you get a neat example when our visual processing systems vacillate between alternative ‘strategies’. Surely we could construct a blended space in which lying and floating are merged like levitation. But then we would not be able to account for the presence of the bayonet. To achieve that, we may activate our knowledge of acrobatics and see the drawing as an incredible circus performance. And certainly we could go on in recruiting elements from within and from without the blend, or making use of its rich internal structure. But maybe it is enough to demonstrate how much visual argumentation trades on perceptual resemblance or iconicity instead of propositional knowledge. If we are right, the argumentational meaning  ‘extracted’ from the picture is much more than the simple observation that peace is unstable. It can be taken, for instance, to allude to a kind of  somnambulism like sleepwalking, when acrobatic acts are performed unaware and to aim at awakening the peoples of Europe from that torpor.
It is important to note once again that the ‘rivalry’ in visual processing is not one between possible inferences. We do not infer that the figure is lying or floating, or that the bayonet is pointing to the sky; we perceive them this or that way directly. It is this aspect of direct perception which can account for the fact that ‘the visual is more powerful than the verbal’ (Groarke, 1996, 106).

That not all visual representations are mere allegories and the visualization of verbal arguments is attested by the ‘thickness’ or density of the visual medium vis a vis the articulation of the verbal, a fact emphasized in (Barthes, 1977). Any account of visual argument would have to clarify how images can be articulated (Let us note in passing that what Barthes and some other visual hermeneutics foregrounded in the first place is the specific quality of sensory, and thus visual experience, like Ivan’s beard in Eisenstein’s Ivan the Terrible, that will resists all efforts of verbalization. We would say, accordingly, that it is beyond the lower bound of visual argumentation). As we have seen, visual images could be articulated in two ways; either by retracing some linear order within the image structure analogously to narrative understanding, and re-describing it accordingly, or by constructing ‘bridges’, mediating blends, grounded in perceptual similarity. Against the first, we argued that visual processing is qualitatively different from verbal – argumentative or narrative – understanding: it is distributed among at least two different pathways, it involves rivalry between them, it can result in continuously changing aspects, incongruencies or ambiguities. Such perceptual incongruencies or ambiguities are ‘solvable’ only by constructing blends. Yet blends should not be taken as a way to translate arguments. They can give way to arguments, but they are not arguments themselves. We called the second mode of visual argumentation mixed precisely because it makes use of both verbal or textual and visual capacities. It can be compared to higher order reasoning in that it aims at creating new concepts. In this sense it could also said to reach an upper bound of argumentation: it provides new input spaces (viz. new premises) to build new blends out of blends (giving thus fuel for further argumentation). Yet it could not be over-emphasized that the construction of blends, and thus the second mode, is conditioned by direct perception of the genuinely visual which constitute the lower bound of visual argumentation. Now the logical question is: Is there a distinctly visual mode of argumentation at this low level?

Well, the logical and empirical answer is yes. Indeed, the idea taken over from ecological psychology that a considerable part of perception is direct, and not inferential, there is the source of a third – genuinely visual – mode for the visual to appear to be translatable into the verbal. It is the most often quoted case of staging narratives in film. Take the classical Hollywoodian movie, for instance. Almost all approaches in film theory, the theory of the moving image, agree that it presupposes transparency of its medium, instead of foregrounding its physical substance. It is claimed that Hollywoodian film makers arrived at certain ‘thumb-rules’ in order to provide a realistic effect (See e.g. Anderson, 1996). All these rules serve the common goal to create the sense of continuity through shots. Fortunately enough, we do not have to develop a full argument how these rules are parasitic on ecological and psychological laws operating in real life situations, since it has been done by ecological film theorists. It has been elaborated in details in (Anderson, 1996). What should be emphasized here is the fact that the rationale for the thumb rules is nothing less than rendering the scenes as credible as possible. If the ecological approach is right, then we have a clear case which appears to be a kind of visual argumentation based especially on the persuasive (ethos) and emotional (pathos) elements. That is it is not dependent on the represented topoi or the ‘content’ of the images that the rhetorical approach makes pains to extract. But the ecological account of classical film can also be characterized by a lack of reference both to visual tropes and topological hierarchies of values. This way it constitutes a counterpart both to the theory of blends and the pragmatically motivated standard view of visual argumentation (following the new rhetoric).

4. Perspectives
It is not simply for architectural reason that we conclude with the re-formulation of the second mode in which the visual serves as a source for discovering emergent structures, or new concepts, by constructing mediating blends. We think that the most part of (moving) images belong to this mode which can be summarized with the adage that image is thought and thought is image. We also think that just as there is always a way to go beyond pure allegory, there is no purely realist movie, an ideal target of ecological analysis. The second mode is a mixed, or impure, mode in that the medium is neither totally transparent (purely perceptional), nor is it subservient of some verbalizable argument. Instead, it presupposes medium-consciousness as it is parasitic on associative mechanism and ideology. The term ‘intellectual movie’ was already used by Eisenstein to highlight that fact that films are made to cause a particular effect (both intellectual and emotional) on the viewers. This it is meant to achieve by means of juxtaposing distinct, often incongruous or even contradictory, elements both within the frames and through editing. No wonder that this technique has been seen as the visual counterpart of verbal argumentation (Kjeldsen, 1998, 458). But we should be cautious in taking mise-en-scene, editing, disposition to be the counterpart of logos (speech) in visual art. For one reason, because we have seen that they are the very means by which continuity (suture) is realized in classical film. For another, if they are revealed as conveyors of thought, they are dependent on conceptual integration or blending. In other words, topology is always already tropology in visual art.

Let us end this paper with a brief examination of a clip from a movie which sums up the very basic of our argumentation. In the first scene of The Sweet Hereafter by the Canadian director Atom Egoyan the camera pans in a continuous shot parallel with a plane onto which a shadowy texture is overcast. It takes some 40 second till it settles on the frame of a family all naked and asleep in a more than ideal position. The music accompanying the pan of the camera also enhances the idyllic quality of the scene. Yet, the continuity of the shot is by no means unproblematic. Just as in Daumier’s litograph the lying and floating positions of the woman are incongruously superimposed on each other, here we are kept in uncertainty as to the position of the plane with which the camera runs parallel. First we may have the feeling that it is vertical (in fact, we seem to see a fence-like row of wooden panels), later it slowly dawns on us that it is horizontal (a wooden floor), which is then corroborated by the figures lying on the floor. Thus, we are subjected to going through a continual change of aspects. Yet, although the last frame shows an idyllic scene of love and sleep, the fact that the viewer has been offered a series of incongruous visual cues seems to mar the idyllic quality of final picture. Now there are at least three different ways in analogy to our three modes to formulate what we have seen. First, it seems plausible that the moving images ‘argue’ that there is something wrong with this family. And indeed, the rest of the film seems to be nothing but the making explicit of the uncanny quality of the first shot. We are confronted with consecutive scenes of bribing, violence, crash, and even incest. But on second thought, we become aware that the previous interpretation is the result of higher order cognitive operation in which the idyllic family becomes a trope for imminent danger. That is, the interpretation of the whole film is dependent on the construction of a blend in which the incongruity of the vertical and the horizontal planes is ‘resolved’, or (elements from) the two planes are somehow merged.
And last but not least, we should be reminded that the incongruity of the vertical/horizontal is the result of direct perception of invariants in the optic array; that is, we do not infer that the plane is vertical or horizontal. We cannot but see it as this or that according to our ecologically determined capacities. Thus we reach the lowest level, or the rock bottom, of visual argumentation. By incorporating the incongruous perceptions of the plane into a full range of arguments we definitely verbalize the seen. This may be just as long as the extracted arguments are said to explicate how certain images are used to convey meaning. What we have been trying to show among other things in this paper is that the extracted arguments are not represented by the (moving) images.

REFERENCES
Anderson, J. (1996). The reality of illusion. carbondale/Edwardsville: Southern Illinois UP.
Barthes, R. (1977). Rhetoric of the image. In: Image, text and music. Trans. by S. Heath. Glasgow: Fontana/Collins.
Bonfantini, M. (1987). La semiosi e l’abduzione. Milano: Bompiani.
Fauconnier, G. & Turner, M. (1998). Conceptual integration Networks. Cognitive Science, 22, 133-187. [Expanded web version: http://www.wam.umd.edu/~mturn/WWW/blending.html]
Gibson, J. J. (1986). The ecological approach to visual perception. Hillsdale, N.J.: Erlbaum. [Original work published in 1979]
Gilbert, M. (1997). Coalescent argumentation. Mahwah: Erlbaum.
Groarke, L. (1996). Logic, art and argument. Informal Logic, 18, 105-126.
Hofstadter, D. R. and the Fluid Analogy Research Group (1995). Fluid concepts and creative analogies: computer models of the fundamental mechanisms of thought. New York: Basic Books.
Kjeldsen, J. E. (1999). Visual rhetoric – From elocution to Inventio. In: F. H. van Eemeren, R. Grootendorst, J. A. Blair, & C. A. Willard (Eds.), Proceedings of the Fourth International Conference of the International Society for the Study of Argumentation (pp. 455-460). Amsterdam: SIC SAT.
Norman, J. (2002). Two Visual Systems and Two Theories of Perception: An Attempt to Reconcile the Constructivist and Ecological Approaches. Behavioral and Brain Sciences, 24, XXX-XXX.
Parret, H. (1986). Le sublime du quotidien. Paris/Amsterdam/Philadelphia: Hadés-Benjamins.
Perelman, C. & Olbrecht-Tyteca, L. (1971). The New Rhetoric. A treatise on argumentation. London: University of Notre Dame Press.
Sperber, D. & Wilson D. (1968). Relevance. Communication and cognition. Cambridge: Basil Blackwell.
Rock, I. (1983). The logic of perception. Cambridge, MA: MIT Press.
Rock, I. (1997). Indirect perception. Cambridge, MA: MIT Press.
Veale, T. (1999). Pragmatic forces in metaphor use: the mechanics of blend recruitment in visual metaphors. In: C. Nehaniv (Ed.), Computation for metaphors analogy, and agents (pp. 37-51). Berlin/Heidelberg: Springer Verlag.

You May Also Like

ISSA Proceedings 2006 ~ The Criminal Abduction Paradox

ISSA Proceedings 2006 ~ A Discussion Of Habermas’ Reading And Use Of Toulmin’s Model Of Argumentation

ISSA Proceedings 2006 ~ The Rules Of The Critical Discussion And The Development Of Critical Thinking

ISSA Proceedings 2006 ~ An Analysis Of Preschool Hebrew Speaking Children’s Arguments From The Perspective Of The Pragma-Dialectical Model

ISSA Proceedings 2006 ~ Putin’s Terrorism Discourse As Part of Democracy And Governance Debate In Russia

ISSA Proceedings 2006 ~ Cultural Differences In The Persuasiveness Of Normatively Strong And Normatively Weak Expert Evidence

Comments