Hello - welcome to OR. I’m your host, Doug Kulchar (@dkulchar)
The overriding point I have to make in this newsletter is threefold. First, that rhetoric is probabilistic. Second, and following from the first, artificial intelligence is a rhetorical project. Third, and most speculatively, by getting clear on what the first two mean, we can understand rhetoricity - that is, rhetoric as an agent-neutral, multiply realizable inferential process - as a fundamental property of complex systems, living and artificial. That is, rhetoricity is something near to not quite the ‘meaning’ of life or the universe, but rather to its operation.
This is heady stuff. Eventually, some version of this will become my dissertation. But dissertations are strange, often arcane documents. Rhetoric belongs to the people. So there will be plenty of time for us to get lost in the weeds, but as an opening salvo, I want to setup the core concepts we’ll be working with - the rest will come later. So: let’s go.
Last time, we left off with the idea that rhetoric is a provisional, pragmatic sort of skill: when we do rhetoric, we’re tinkering, slapping things together, seeing what works, and just trying to get the job done - that job being, of course, to persuade someone of something. But, of course: what does that mean, anyway?
If we take a familiar point of departure, we know that to be “persuaded” of something implies a change of some kind - from ~P to P, from P to Q, or perhaps simply a shift in our degree of belief. While this is a neat, tidy way to think about things, it’s not particularly difficult to begin listing counterexamples - what about when a person isn’t necessarily persuaded of P, but complies with an action associated with P for some extrinsic reason? or, conversely, when a person comes to believe P, but to no material consequence? Can one be caught up in the moment and be persuaded, and then only later revert to one’s prior belief?
We don’t even need to make a ruling here to realize that a given definition of “persuasion” brings with it a whole theory of mind, subjectivity, language, and so on. That is, "persuasion" is a concept with a number of logical dependencies. Everyday experience suggests that persuasion happens through words, of course - people share information and communicate using words as a medium - which we might characterize as the substance of whatever is said. But we rely also on gesture, tone, personality, and so on - all of these things, we might classify as ‘style.’ Roughly speaking, we might think of style as a kind of coefficient: knowing when and how to 'multiply' stylistically whatever substance we wish to convey changes the output. What is said gets filtered through how we say it, which may increase or decrease the chances of persuasion. This is then filtered once more, by the audience. In characterizing pathos, one of the three artistic modes of persuasion, Aristotle notes "the judgements we deliver are not the same when we are influenced by joy or sorrow, love or hate."1 Delivering some information - say, a piece of bad news - will be received differently if our audience is in this or that mood: a good mood might soften the impact of the bad news, while a bad one might lead to an entirely different reaction, perhaps making a bad mood worse.
Persuasion, insofar as it makes use of emotions, requires first an agreement about what emotions are. Depending on what our purpose is, a greater or lesser degree of explication is required. If the goal is provision of relatively simple advice - as the saying goes, one catches more flies with honey than with vinegar - then taking the emotions covered by the metaphorical 'honey' and 'vinegar' as primitives is probably perfectly fine. One can certainly get by with a working definition. But if we want to explicate 'persuasion' - words mixed with other things like emotions - some story of how the two are connected is required. Given that Aristotle goes on to discuss how persuasion may also be effected by the speaker's character (ethos) and the logical character of what is said (logos), rhetoric emerges not as a simple theory of ornament in speech, but as something requiring a full-bore semantic theory: 'how to be convincing,' yes, but the production of conviction depends on an understanding of *how* words mean, such that they can be convincing or not, which itself depends on an understanding of who hears the words.
We might compare the present case to the similarly complex phenomenon of perception. When I see an object, photons enter my eyeballs, stimulate some kind of electrical impulse that does...uh, something or other…to some parts of my brain, and this is what, objectively speaking, 'takes place'. Let's call this series of physical events an A-process2. To describe an A-process, we don't need any concept of vision or consciousness - it is ruthlessly asubjective. The A-process, however, is not the same as the collection of colors, movements, and shapes which we would call 'seeing something.' Let's call this everyday experience of things a B-process. "Making a cake" is a B-process, while "mixing flour, sugar, eggs, etc, in such-and-such measures, and exposing them to so many degrees Fahrenheit for so many minutes" is an A-process. Philosopher of mind David Chalmers describes this in terms of ‘supervenience’: one property supervenes on another if the facts about the former determine the latter entirely3. If I have a substance in a jar which you know to be molecules composed of 1 Oxygen and 2 Hydrogen atoms, it would make no sense to ask whether the substance is “water” - knowing the former is the definition of the latter. What matters for Chalmers is that conscious experience (here, a B-process) does not seem to be exhaustively explained by the physical facts of neural activity (here, an A-process) - there seems to be something extra which is not explained by the lower-order physical facts. If the something extra is logically conceivable - that is, if the existence of that something implies no contradictions or absurdities - then that B-process cannot be reduced to the A-process. If we can imagine a creature identical to humans in every way, but without conscious experience, then conscious experience can’t be considered fully explained by physical facts.
Along similar lines, certainly persuasion is within the sphere of conscious experience. At the same time, I think persuasion is clearly neither exclusively an A- nor a B-process. Physical facts are involved in persuasion, but there is some remainder. Rhetoric may have certain neural or bodily correlates, but they aren’t the whole story4. But let us fall back on the original observation: persuasion, basically, is about some sort of change: P to Q. Whatever might be involved in persuasion, it becomes contracted into the bottleneck designated by ‘persuasive speech’:
The situation is somewhat like Chess: an unexpected move is made by our opponent, and it seems that they’ve left a piece undefended. Did our opponent make a mistake in hanging a knight? Or is it bait, part of a multi-move combination which would destroy our position? We don’t know. Strictly speaking, all we know is that the move was made; the rest must be inferred. The full complex of A- and B-processes which go into a Chess move (including motor control, rules of the game, principles of sound chess, etc.) would exhaustively explain what the move was and why it was made - if only we had that information. But we don’t, so we guess based on available knowledge and hope for the best5.
The central difficulty, though, is that the tools by which we would set out to investigate persuasion are also subject to this bottleneck. The tools themselves are the thing we wish to investigate, almost as if we needed a hammer to drive a nail…into the handle of that same hammer. This problem of recursivity is where we’ll pick up next time.
Aristotle, Rhetoric 1.2
I’m borrowing a version of the argument made in Chalmers’ 1996 book The Conscious Mind concerning supervenience. We’ll get there in a second.
A potentially interesting parallel with Chalmers’ argument re: supervenience is G.E. Moore’s “open question argument” in metaethics.
Determining the exact relationship (if there is one) is a tantalizing prospect, but also brings along the potential for a small-minded reductionism, or a scientifistic fundamentalism - a kind of I Fucking Love Neurorhetoric approach, whereby rhetoric “really is” such-and-such or an epistemically squishy rhetoric is reduced to the Hard Logic And Facts of natural science methods. For a more reasoned, less polemic approach, see Jack and Appelbaum, “‘This is Your Brain on Rhetoric’: Research Directions for Neurorhetorics”, in RSQ 40.5.
I’m not good at Chess, but I’m always interested in playing; find me at chess dot com username: pharmakon