6. Philosophy, certainly in Greek discourse, viewed dialogues and dialectic as a primary mode of advancing philosophical thought. Since the advent of the printed page, and especially with the rise of printed philosophical journals and books as the primary mode of communication of finished philosophical thinking, philosophy has lessened its emphasis on oral exchanges (see #5 above for how philosophy might revert to something like a quasi-oral mode). Equally important is the fact that even in the oral tradition philosophy has had a primarily propositional form. Philosophical positions are given as propositions to be defended, defences are given as arguments consisting of sequences of propositions, and the attitudes taken towards the positions and arguments are propositional attitudes such as believing that, knowing that, or speech acts such as warrantedly asserting that, and so on. This propositional approach has been enormously fruitful, especially when augmented with the resources of logic that were developed in this century, and what follows is an attempt to suggest supplementation, rather than a wholesale replacement. Aside from its success, there is an obvious practical reason why this orientation has persisted. It is both difficult and expensive to reproduce anything but straightforward prose within the printed page.
7. Compare that constraint with the already available ability to transmit and freely reproduce graphical images electronically. These images moreover need not be static but can easily be both convincingly three-dimensional and dynamic. Now note the fact that 50% of our brain function is devoted to processing visual information. Why deny oneself as a philosopher the opportunity to make use of that cognitive capacity?
8. Distinguish two, connected, uses to which propositions are put in philosophy: (a) As representational devices conveying information, (b) as constituents of arguments. I contend that at least the first of these uses, and perhaps the second, are seriously hindered by the restriction to propositional modes.
9. First take some commonplace examples of (a). Airline safety instructions that are given on those cards placed in the seat pocket in front of you are purely pictorial. Rather than presenting what you need to know in a dozen different written languages, they successfully convey what needs to be known in a non-propositional format. (This is not just knowing how to get out of the aircraft, it involves knowing that there are five emergency exits on the aircraft.) International road safety signs are similar.
10. What of knowledge that is of philosophical interest? In the philosophy of science, there are enormous advantages in having access to non-propositional representations. Take theories. These are standardly represented either as sets of formulae in some formally specified language, or as set-theoretic objects. (Formal models in that rather narrow sense in which logic construes models). Rather than representing these in some formal language such that solutions to, say differential equations, represent the state of the system at some time, a graphical representation generated by computational solutions to the theory can provide a real-time dynamic display of how the system evolves over time (see, e.g., Humphreys 1991, 1995, Hartmann 1996). This provides an enormous increase in our understanding of what the consequences of the theory are. The same thing is true of running solutions to game-theoretic problems such as the iterated prisoners' dilemma. Optimal solutions that are not evident from case-by-case inspections of alternatives can become quite clear when multiple runs are made (see, e.g., Axelrod 1984).
11. The fixation on propositional representations is not so firmly established in science. The two well-known books by Edward Tufte (Tufte 1983, 1990) are fine examples of how both scientific and everyday information can be represented economically and effectively via visual devices. Statisticians have long taken advantage of graphical means of displaying information. Nowadays, pictorial representations are almost forced on practitioners of certain sciences, such as astrophysics, because of the staggering quantity of data that must be presented.
12. These examples are, of course, cases where philosophy, via philosophy of science, is studying the output of another discipline (as are some of the other examples below, such as aesthetics). Yet at the very least, unless philosophy moves to incorporate these non-propositional modes of display, it will no longer be the philosophy of X, because X will have largely abandoned the kinds of representations that philosophers are discussing.
13. More directly philosophical, certain sorts of possible worlds can now literally be seen. Not just simple ones such as those where the inverse square law of gravitation is replaced by an inverse cube law, but those that are generated by evolutionary algorithms such that the resulting structure is one that is initially unimaginable by us. Of course there is a translation from the formal theory into the code that generates the graphical image and those who are fixated on ``the linguistic turn'' would argue that all we have here is a translation from one representational scheme into another. But to emphasise that would be to miss the point. The information that is available directly from the images and the inclusion of a temporal dimension goes well beyond what our inferential powers can generate from the propositionally represented theory. Moreover, by using realistic representations, such as graphical images, one avoids one of the most difficult problems of the propositional mode, which is that you cannot compare a proposition directly with the fact(s) that it represents. The correspondence relation between words and the world has always been mysterious. Not so with images. There one can have a direct comparison between the thing imaged and the sensory evidence available from the thing itself. There will be real isomorphisms, not the artificial ones holding between semantic models and some other abstract representation of the system.
14. The increasingly strained emphasis on languages of thought within certain areas of cognitive science can itself be seen as a consequence of this philosophical tradition of propositional representations. Although many other kinds of representation can be (inefficiently) coded into propositional forms, they need not be, and there is no imperative to conceive of all cognitive capacities as requiring manipulation of symbolic structures in the linguistic tradition. We do often literally just see that something is the case.
15. The advantages in other areas of philosophy of having non- propositional representations are almost too obvious to need stating. In aesthetics, the direct presentation of paintings, dance, and music can serve as examples of philosophical theses far better than can their representations via descriptions or some notational system. Even areas such as philosophical investigations of humour can benefit from the availability of non-propositional examples.
16. Regarding (b), the advantages of moving away from a
propositional orientation are at the moment less dramatic. We are
all familiar with the use of devices such as Venn diagrams and
truth trees in logic, but there is also a contemporary move towards
argumentation through diagrammatic structures, for example Shin
(1995, Chapter 6); Hammer (1995); Baigrie (1996).