Dialectic and rational arguments in philosophy

Socratic dialogue is modeled on dialectic, and for that reason it is a central part of Western philosophy. In the previous post, I pointed out that, historically speaking, dialectic contrasts with three other argumentative styles — rhetoric, scholasticism, and mathematics. Unlike rhetoric, dialectic is not about persuasion for its own sake, but the pursuit of stable conclusions (as we saw in selections from both Gorgias and Phaedo). Unlike scholasticism, the dialectician attempts to resolve disputes through engagement (i.e., the method of disputation), not through deference to written authority in the form of scripture. And unlike mathematics, dialectic investigates the worthiness of its premises (i.e., what I called the ‘collapse-and-consequence’ model), instead of treating premises as axiomatic.

Last time, I suggested that these three historical contrasts help to hone in on a particular feature of concept of dialectic, which is that dialectic is a form of second-order rational persuasion. I suggested that the constitutive point of dialectic is to convince people that some passages of thought or speech are rational, and to resolve disputes in that minimal sense of creating directed change towards a state of intellectual common ground. I called this ‘persuasionism’. A vital part of the persuasionist thesis is the idea that dialectical arguments occur in the context where they are directed towards change in mental state (what Gilbert Harman calls a “change in view”), leading to resolution of dissonance. I argued that the persuasionist theory is superior to the purity thesis, i.e., someone who thinks the collapse-and-consequence model is sufficient to characterize dialectic, and that no reference to effective perspective change is strictly necessary.

The persuasionist thesis says that dialectic involves a directed change in view accomplished by means of demonstrating the rational defensibility of a passage of thoughts in light of potential challenges. One might wonder whether demonstrating defensibility of some train of thought actually counts as “persuasion”. But a moment’s reflection shows it clearly does. As a matter of definition, to persuade just is to cause someone to believe or act in some directed fashion that they did not before. When you subject a set of reasons to potential objections, you leave the set of reasons altered — stronger, if all goes well for the defender of those reasons. This means that in the process of demonstrating defensibility, you have produced a change in view about the status of the arguments as being more reasonable than they seemed at the outset, all other things equal. And my suggestion is that this sort of directed change is not an accident or an irrelevant side-effect, but rather is a part of the dialectician’s stance of attempting to direct a change in view during the course of presentation of argument. Notably, though, it is an attempt at mutual persuasion between defender and opponent; that is to say, it is as a joint enterprise with reciprocal expectations. Hence, when the dialectician fails to persuade their good-faith interlocutor of the rational qualities of their passage of thought, they thereby gain some reason to regard those passages of thought as irrational under some description.

In the rest of this post, I provide reasons to think that persuasionism makes the most sense of dialectic in philosophy. First, I’ll make a brief remark on the consequences of persuasionism on meta-philosophy. I suggest, briefly, that is persuasionism is conducive to productive philosophy. (Indeed, I think it is even more conducive than the purist’s alternative, which I think is worse than sophistry; but I will not argue this point in this post.) Second, I will consider some attempted refutations, based on the idea that I am excluding some kinds of argument as examples of dialectic.

1. On meta-philosophy. When I say that dialectic is not just an autodidactic exercise of getting ideas clear in isolation — of studying logical implications and entailments, or (Harman again) “what follows from what” — my emphasis is on the word “just“. Dialectic involves the study of such entailments, but is not reducible to that study. I offer two reasons. First, as we have seen in the previous post, Socrates himself thought he was attempting rational persuasion. Indeed, one of the characteristic tropes of Socratic argument is his willingness to throw the whole game away, if only a good answer can be given to a master question (which he then shows cannot be done).

But second, even in a parallel world where our Hellenistic heroes thought they were just making ideas clear independently of their audience’s convictions, it is still a fact that people can do a lot of things with all sorts of side-effects, and some of those side-effects might actually be the thing that makes the activity essentially worth doing. Sometimes, a practice has a function, and that function occurs independently of the ways the practice is conceived; it, instead, has to be recovered by examination of intuitively valenced presuppositions. And that fact makes it possible to engage critically with the tropes in Socratic dialogues, to separate the stuff Socrates thought he was doing well from the stuff that he actually was doing well. Which is just to say that contemporary critical thinkers could probably do without Socrates’s leading questions, for example, or Plato’s noble lies, even if for whatever reason Plato and Socrates in our parallel world had decided these  ideas were essential parts of their whole philosophical package. Revisionism is the price we sometimes pay for rational reconstruction.

2. On excluded cases. Most of this post derives from a spat I had with the author over at Siris blog, who seems to be a purity theorist. In our exchange, he argued that the persuasionist view of dialectic excludes a few cases of rational argumentation. 1) It seems to exclude cases where we apply the collapse-and-consequence model through habit. 2) It seems to exclude practice arguments, e.g., as when the student makes use of natural deduction. 3) It seems to exclude cases involving a stimulating exchange of reasons for exploratory purposes. But these examples are not on equal footing. So, my view is that (1) is not an argument at all, (2) is rational argument but not dialectic, and (3) is an unobvious kind of dialectic.

Habitual processing. I reject the idea that arguments are, or can be, merely habitual passages of thought. For a person to suggest that habitual passages of thought are not directed at change in view, is for that person to fail to attend to the internal point of view, and in particular to neglect the intuitive force of argumentation. Intuitively, there seems to be a difference between mere regularities and rules, and rational arguments are about rules, so regular habits of thought are not themselves arguments.

The point can be made in part by appealing to the philosopher’s ego. If merely habitual orderings of thought counted as philosophical arguments — if it were even possible to follow the quick turnabouts in collapse-and-consequence model into habits — then it would turn philosophy into something even worse than sophistry. Indeed, it would collapse the study of rational argument into the study of the psychology of reliable heuristics, or the study of computational processing. It is a rare philosopher who is eager to make themselves Turing-incompatible in this way.

Perhaps the purity theorist would consider it a strength of their view that they think they can rationally argue as a function of personal habits. And, indeed, much of logic feels like habitual or schematic, once it is mastered. And if they could get away with that, then to be sure, “persuasion” would drop out of the analysis. But the only *rational* way you can get away with the habitualist’s conviction is by finding some independent means of calibrating your passages of thought by placing them into an orderly rule-like quasi-sentential (propositional or imperative) structure. And it is difficult to see how habits or mere regularities could have that rule-like character — a man who “argues” with himself habitually is not engaged in inference, hence not arguing rationally at all. In that sense, the approach from habits is going to founder on the question, “What makes this rational?”, and one does not even have to be a persuasionist to suspect that it is a mistake. But even if we come up with an adequate causal account of rules (as, indeed, we might), there is the remaining requirement of needing to account for the ‘following‘ part of ‘rule-following’, which is an intentional activity that seemingly requires both identification of rules and calibration of them.

Practice arguments. A different argument proceeds by observing that, when we are doing proofs in natural deduction, we aren’t trying to persuade anyone of anything. From premises, we are given the task of showing their consistency. Sure enough, this does not look like rational argument.

In this case, I think it would be useful to remember that philosophical argument is not all dialectic. The geometric or analytical method, of deriving consequences from axioms, is one method in philosophy, though it is not a Socratic method. So, one might insist (correctly) that the geometric method has got all the bells and whistles of a rational methodology, and that this is being ignored in a conversation about dialectic. And then one might notice that practice arguments have the form of analytical arguments.

This argument has my blessing, though it is not of first importance in a conversation which is meant to be about the merits of rational argument insofar as it has been conceived of through the Socratic approach. It also reminds us that we ought to notice that a presumptive dichotomy, between dialectic and rhetoric, is a false one. The mathematician is not just doing rhetoric.

Bullshit sessions. The author of Siris also asserts, plausibly, that the persuasionist view of argument seems to make no sense of ‘stimulating thought’ exchanges, where the aim is apparently to open oneself to exchange, not to create a directed change. I agree these contexts are not obvious attempts at rational persuasion; it is easier to say that they are attempts to explore the space of reasons. In bullshit sessions, for example, rational people can take on points of view “for the sake of argument”.

But appearances are deceiving, because the difference has got to do with whether or not the attempts at change are built to last. I submit that in these cases, participants are attempting to persuade others into the view that it is rational to regard some perspective as appropriate in a context, not to persuade people that it is rational to hold the positions are true. The attempt is still to show that, in a contest of reasons, one comes out stronger, even if the contest is local, and comes to an end when the sun goes down. So they still fit with the persuasionist model of dialectic.

Dialectic and rational persuasion

In an early essay (2006), I suggested that legitimate philosophical argument shall sometimes have some specific sophistic qualities. For example, legitimate pedagogical simplifications, which I dubbed “sophiboles”, or wise exaggerations. The point of that essay was meta-philosophical, to direct attention at the ways in which philosophers, scientists, and sophists are on the same spectrum, despite the fact that they have distinctive aims. Call this the ‘Isocratic thesis’. The motivating case was of Galileo, whose defense of heliocentrism was (rightly) phrased in a bold, realist form for the purposes of popular instruction, though a more conservative probabilistic statement would arguably have been more appropriate at the time. At any rate, the Isocratic thesis was a provocative meta-philosophical statement of purpose. Its central aim is to deflate excessive philosophical pretensions while acknowledging the distinctiveness and productivity of philosophical activity.

I have come to appreciate that the Isocratic thesis concerning philosophy can be distinguished from a more modest thesis about the nature of dialectic. Dialectic, a species of rational argument, is a practice of reducing premises and inferring consequences; Socratic dialogue is the exemplar. But what, exactly, is going on in Socratic exchanges — how does it work, and why do philosophers think it worth doing?

I assume that everyone can agree that the Socratic exchanges have a characteristic rational structure: first, some set of hypotheses are derived from some foundational principles, and then (and only then) consequences are inferred. (Kenny, A New History of Western Philosophy, p.128) In the Phaedrus, Socrates refers to these as processes as generalization and division, and attributed to the dialectician’s toolkit. Call this the “collapse-and-consequence” model of dialectic, and suppose that it is the answer to the how-question. But then there’s still the question of what the philosopher thinks they’re trying to accomplish — what makes their activity productive. An alluring answer would to achieve a greater mutual understanding of some topics and/or decreased confusion, under a guise. Key to this idea is that argument is a conversational effort at reconciliation between two or more divergent perspectives. Hence, the later Hegelian formulation of dialectics as “thesis, antithesis, synthesis”. From these scholarly data-points, one might believe that the aim of dialectic is to be rationally persuasive (under some description), and call that conviction the ‘persuasionist’ interpretation of dialectic.

It is not uncommon to find folks who endorse the persuasionist view today. It is a core element in Trudy Govier’s excellent introductory textbook in critical thinking, A Practical Study of Argument (Seventh Edition): “When we use arguments in the sense of offering reasons for our beliefs, we are responding to controversies by attempting rational persuasion.” (2) Also, it is held in argumentation theory as a pragma-dialectical theory, according to which argument is intended to resolve disputes by diagnosing the appropriateness of the competing standpoint through a model of critical discussion. Resolution, here, can be read as persuasive success or rational cogency.

(I cannot help but also note that Catarina Dutilh Novaes almost endorses something like the persuasionist argument in Dialectica (2015). In that article, she argues that deductive inference as such can be understood as a kind of lopsided debate. “The claim is that, rather than for mono-agent mental processes, (deductive) logic in fact comprises norms for specific situations of dialogical interaction, in particular special forms of debates… It will become apparent that the conception of logic as tightly connected with debating and disputing rather than with thinking has been quite influential throughout the history of logic, even though if it is now mostly forgotten.” (588) Mind you, her account is not as striking as an account of persuasionism, since it applies to a form of inference that has been willfully contorted away from its initial dialogical purpose, in order to provide means for the “built-in Opponent” to get on with their part during the business of disputation. On first blush, I worry that Novaes’s conception is too agonistic, or post-Gricean, to qualify as persuasionist. But I could be wrong about that, and it would be interesting to find out.)

The persuasionist thesis is not the only game in town. It contrasts with the ‘purity thesis’. Purity theorists are sufficiently impressed with the answer to the ‘how-question’ as to think that it tells us everything we need to know about dialectic. The purity theorist holds instead:

  1. Historically, dialectic is defined in terms of pure structure of reasons (the answer to the above-mentioned ‘how question’), without reference to powers of persuasion or any other communicative aims. Hence,
  2. Some kinds of arguments do not seem to involve persuasion at all, e.g., practice arguments, are excluded by the persuasionist. Since persuasionism fails to explain parts of the denotation of the concept of rational argument, it fails as an account.

The purity theorist is wrong on both counts. (1) only makes sense so long as it relies on a provincial and distorted historiography (which Novaes alludes to in the pull-quote in the above parenthetical). Also, (2) is only distinctive so long as it involves eccentric ideas about what counts as an argument in philosophy. In the rest of this post I will only suggest some counter-examples to (1), and will say something about (2) on some other occasion.


For present purposes, my point of departure is the medieval conception of dialectic. For scholars in the Middles Ages, dialectic was conceived of as a method of disputation. “As the art of discussion– disciplina bene disputandi–dialectic deals with substances”. (Paul Vignaux, Philosophy in the Middle Ages, p. 21) At the time, dialectic was regarded with suspicion, in contrast with the scholastic method, whose point is to make credulous (usually theological) inferences based on the authority of scripture. The method of disputation sure sounds like it has all the airs of an attempt to engage in mutual, rational persuasion (under a description).

Is the medieval conception of dialectic viable in light of other contrasts? I think so, because it follows naturally from the Platonic canon, in its discussion of the contrast between dialectic and rhetoric. Two books are pertinent here, Gorgias and Phaedrus. In Gorgias, the distinction is made in the following way: while rhetoric is directed at persuasion for its own sake (or for the sake of money), depends on the science of human character, is fickle and fleeting. Meanwhile the Socratic method, hence dialectic, is better at persuasion because it is built to last, trading as it does on insight and methodological rigor. Persuasion has a potential role in both forms. The purist’s alternative is to say that dialectic is about pure structure — the collapse and consequence model while rhetoric is about mere persuasion. But this flattens out the nuances of what is said in the above-mentioned passages which most directly bear on the question.

The purity theorist would probably reply by invoking the Phaedrus, where Socrates seems to describe dialectical methods in terms of their potential to allow personal growth and self-education. Hence, the serious pursuit of the dialectician is to “find a congenial soul, and then with knowledge engrafts and sows words which are able to help themselves and him who planted them, and are not unfruitful, but have in them seeds which may bear fruit in other natures, nurtured in other ways — making the seed everlasting and the possessors happy to the utmost extent of human happiness.” (89) We might struggle to make sense of the fruit-seed metaphor in connection with the idea of persuasion. This does not sound much like persuasion, since it lacks a sense of directed acquisition of judgments. But this is a one-sided observation. The persuasionist thinks that argument can occur in contexts of mutual persuasion, which involves an assumption of reciprocal effort with convergent ends. The fruits found by the student may very well differ from the seeds planted by the teacher, but do we think Socrates would have us say that the teacher is released from prizing the fruit once it has grown? One or more parties are expected to recognize the attractiveness of the “winning” argument, if one is found.

In fact, I shall suggest that the persuasionist’s point is entirely resonant in Phaedrus.  Socrates concludes the dialogue with the following remarks. He begins with a fairly clear, though elliptical, reference to the dialectician’s method, and concludes that it is the most effective means of teaching or persuading. “Until a man knows the truth of the several particulars of which he is writing or speaking, and is able to define them until they can be no longer divided… he will be unable to handle arguments according to rules of art, and far as their nature allows them to be subjected to art, either for the purposes of teaching or persuading…” (Phaedrus, 90, emphasis mine) And we have already seen from the Gorgias that he is not inclined to find many resources to distinguish between education and persuasion. The dialectician’s work has clear import to the art of persuasion, if there is one.

It can be objected, at this point, that the collapse-and-consequence model does not say anything at all about the aim of mutual persuasion. Indeed, to be sure, is fairly well established in the canon that dialectic, unlike rhetoric, is not about persuasion of a positive first-order intellectual programme. Instead, I would like to say that dialectic is about clearing the intellectual ground and inferring what follows — its point is to persuade us that some passages of thought are rational, not to compel belief in a particular proposition. So the ‘persuasionist’ label is in a way misleading, since one can try to persuade someone of something without engaging in dialectic. But this is not to say that persuasionism is incorrect; it is only to deny naive persuasionism. Instead, we should say that dialectic — a project of reducing premises and inferring consequences — is a second-order attempt at persuasion. It is not an attempt to convince people about the true and the false, but about the rationality and irrationality of passages of thought and speech through characteristic means.

So, if the villainous brainwasher from Netflix’s Jessica Jones, Killgrave, were to say to me, “Socrates is a mortal”, I would likely be persuaded of the claim’s truth, but it would not be an argument. Meanwhile, if someone gives me the Socratic mortality syllogism, I will be persuaded of the rationality of the conclusion given the premises. And if the same or similar results were given in the collapse and consequence model, it would be dialectic. The point of rational argument is to direct a change in view, allowing an alternative perspective to observe the inferential train as a rational passage of thought or talk. It does not require that they agree with the conclusion or premises.

So I do not quite see the historical appeal of the purity thesis. I am sure there must be a more compelling case, and I have seen some people try to defend it. But I suspect, for the moment, that the problem here is somewhat ironic: namely, that the teachers of classics sometimes resort to sophiboles outside of the classroom. It is a fact that, during the teaching of a freshmen level course, dialectics and rhetoric are simplified for pedagogical purposes. In that context, I can imagine a harried instructor drawing a strong line between them, with ‘persuasion’ on the rhetoric side, ‘pure reason’ on the dialectic side. It is, at best, a wise exaggeration — though I confess I am not convinced of its wisdom.