Symposium Introduction
When we argue, we argue for something and with someone. With arguments, we advance reasons for why we should believe or do something, and we do that in order to bring someone to agree with us on that. We argue to justify and persuade. But an important insight that Matthew McKeon’s new book is about is that arguments also portray that we’ve successfully reached those conclusions. Arguments are intentional directions of our attention and our judgments based on the reasons that make them up. Consequently, with arguments it seems we make not only claims about our premises and our conclusions, but how those inputs support that output.
A few consequences await us once McKeon has identified this extra claim to arguments. One is that our intellectual virtues must have a self-regarding element to them—we must be able to see and endorse our reasoning as we are doing it and going over it. We must follow the argument when we see the support it provides, and especially when we recognize it. This is the core of intellectual integrity and honesty.
With this symposium, three leading scholars of argumentation theory (John Casey, Hubert Marraud, and Tomáš Kollárik) engage with McKeon’s new and challenging book.
1.21.26 |
Response
Inference Claims and Weighed Reasons
Matthew McKeon’s proposal in Arguments and Reason-Giving has the merit of taking as a starting point for the analysis of argument the epistemic concept of reasons, thus departing from the inherited conception that takes as a starting point the logical concept of inference. The novelty may be obscured because McKeon, as do many informal logicians, does not always distinguish between premises and reasons (for criticisms, see Snoeck Henkemans 2000, 460; Juthe 2019). Thus, McKeon equates premises and reasons when he says that whoever presents an argument dialectically claims that the premises, collectively, are reasons for believing the conclusion (3, 19, 46, 77, 159; emphasis added). Trudy Govier (2010, 1), for example, also fails to distinguish premises and reasons when she affirms that in the argument
Marijuana should not be legalized. That’s because sustained use of marijuana worsens a person’s memory, and nothing that adversely affects one’s mental abilities should be legalized.
the conclusion “marijuana should not be legalized” is supported by two reasons: first, sustained use of marijuana worsens a person’s memory, and second, nothing that adversely affects one’s mental abilities should be legalized. However it is obvious that in Govier’s example there are two premises or, depending on the model of argument chosen, a premise and a warrant, but not two reasons, if “A single reason is the smallest amount of information that by itself lends some measure of credence to a position” (Blair 2012, 148). Confusion can be avoided with a little care in formulations—e.g., saying that reason-giving uses of arguments involve the user’s judgment that the premises, collectively, are a reason for believing the conclusion.
McKeon’s book consists of two parts: the first part presents a model for reason-giving use of argument, which is used in the second part to discuss the topics of formal validity, persuasion, argumentative rationality, and intellectual honesty and integrity. My commentary is on the first part of the book.
In Arguments and Reason-Giving Matthew McKeon sets out to study “our uses of arguments to advance their premises as reasons for believing their conclusions, i.e., as reasons for believing that their conclusions are true” (1). It is, therefore, a pragmatic model of argumentation, focused on offering epistemic reasons. This choice places some limits on his research.
First, McKeon restricts his research to theoretical argumentation about what to believe, as opposed to practical argumentation about what to do. “I primarily focus on theoretical reasons and ignore practical, pragmatic, or explanatory reasons” (95). While this is a respectable choice, it has the drawback of leaving out of the discussion the rich literature on normative reasons (see Álvarez & Way, 2024). That biases McKeon’s theoretical proposal somewhat, as we shall see.
Second, within epistemic reasons, McKeon limits his attention to beliefs with a given epistemic status: “the book does not discuss uses of arguments to advance their premises as reasons for believing that the conclusion is plausible, compelling, reasonable, nontrivial, helpful, plausibly deniable, or possible in principle” (1).
McKeon calls such uses of argumentation “reason-giving uses of argumentation.” According to his description, this is an intentional use of arguments: “One uses an argument in a reason-giving way only if one intends the premises to be reasons for believing the conclusion” (1). As a consequence, one who uses an argument in this way commits oneself to the premises expressing a reason to believe the conclusion. McKeon calls this commitment an “inference claim,” borrowing the term from David Hitchcock. McKeon contends that whoever uses an argument in this way asserts the inference claim, as opposed to merely implying it or implicating it (102). In fact, belief in the inference claim seems to be the only essential thing in reason-giving use of arguments (3).
In neither case does using an argument in a reason-giving way require that one believes that the premises are true. What it does require is that one believes that the premises one advances are reasons for believing the conclusion in the intended sense. (4)
McKeon maintains, however, that the inference claim is not a component of the argument, which he defines, following the logical tradition, as a compound of premises and conclusion (19). Even more, McKeon argues that this is the prevailing view among informal logicians (14). Although I agree with him that the inference claim is not a part of the argument, I disagree that this is the predominant view among informal logicians. What most informal logicians hold is that the inference claim is not a premise of the argument, but from that it cannot be inferred that it is not a part of the argument, unless one assumes that the only parts of an argument are the premises and the conclusion. My position, dubbed “holism” is, however, that the inference claim is not a part of the argument, which entails recognizing the existence of contextual factors that are relevant to evaluating the logical properties of arguments (see Leal and Marraud 2022, Marraud 2019 and 2024). I don’t know whether or not McKeon realizes that his view commits him to some sort of holism.
McKeon emphasizes as one of the main novelties of his proposal the explanation of the content of the inference claim in terms of epistemic reasons (14). This allows him to associate the inference claim with the use of the argument, rather than with the argument itself, and to distinguish a variety of operative inference claims, depending on the different notions of reasons for belief involved and the epistemic amount of the reason adduced (5).
While I agree with McKeon that, typically, one who argues that C because P is committed to P being a reason for C, I disagree that this entails an authorization to infer C from P. That is, I reject the following tenet:
I will take the import of an inference claim associated with an argument to merely be that the conclusion follows from the premises(s). Of course, if the premises are reasons for believing the conclusion, then the conclusion follows from them in some sense (105).
Reasons mix badly with logical inferences. In the theory of reasons (a branch of metaethics) “favoring” is used to designate the relation between a reason and that for which it is a reason.
. . . the favouring relation, the relation in which features of the situation stand to action or to belief when they are reasons for doing one thing rather than another or for believing one thing rather than another (Dancy 2004, 79).
I think that the assimilation of “P favours” to “C follows from P” takes away from McKeon’s model of reason-giving arguments uses a significant part of its potential.
In short, the main difference between reasons and logical inferences is that the former, but not the latter, are weighed (Lord & Maguire, 4). It is possible that McKeon overlooks this because of his decision to limit his research to epistemic reasons, leaving aside normative reasons—for which weighing is salient. The fact that reasons are weighed implies that although the fact that viruses possess the ability to replicate is a reason to believe that they are living beings, it does not authorize one to conclude that they are, because it can be counter-argued that viruses depend on the host cell for all biochemical activities that allow them to multiply and spread. Thus, there can be (good) reasons for and against believing that a claim is true. To draw a conclusion it is necessary to combine and compare reasons for and against it. The conclusion is primarily the conclusion of a network of reasons, not of an isolated argument. Assimilating “favoring” and “following” makes it difficult to account for this basic fact in our argumentative practices.
References
Álvarez, Maria, & Jonathan Way (2024). Reasons for Action: Justification, Motivation, Explanation. The Stanford Encyclopedia of Philosophy, Edward N. Zalta & Uri Nodelman, eds., https://plato.stanford.edu/archives/fall2024/entries/reasons-just-vs-expl/.
Blair, J. Anthony (2011). Groundwork in the Theory of Argumentation. Dordrecht, Heidelberg, London and New York: Springer.
Dancy, Jonathan (2004). Ethics without Principles. Oxford: Oxford University Press.
Govier, T. (2010). A Practical Study of Argument (7th ed.). Belmont, CA: Wadsworth.
Juthe, André (2019). Reconstructing Complex Pro/Con Argumentation. Argumentation 33(3), 413–454. https://doi.org/10.1007/s10503-018-9467-9
Leal, Fernando, & Hubert Marraud (2022). How Philosophers Argue: An Adversarial Collaboration on the Russell-Copleston Debate. Cham: Springer.
Lord, Errol, & Barry Maguire (2016). An Opinionated Guide to the Weight of Reasons. In Errol Lord & Barry Maguire (Eds.), Weighing Reasons (pp. 3–24). Oxford: Oxford University Press.
Marraud, Hubert (2020). Holism of Reasons and its Consequences for Argumentation Theory. In C. Dutilh Novaes et al. (Eds.), Reasons to Dissent: Proceedings of the 3rd ECA Conference (pp. 167–180). London: College Publications.
Marraud, Hubert (2024). The Logical Perspective in Pragma-dialectics. Topoi 43, 1247–1258. https://doi.org/10.1007/s11245-024-10065-4
McKeon, Matthew (2024). Arguments and Reason-Giving. New York: Oxford University Press. https://doi.org/10.1093/oso/9780197751633.001.0001
Snoeck Henkemans, A. Francisca (2000). State-of-the-Art: The Structure of Argumentation. Argumentation 14, 447–473. https://doi.org/10.1023/A:1007800305762
-
1.21.26 |
Reply
Response to Hubert Marraud
I am grateful for the opportunity to respond to Hubert Marraud’s insightful considerations of the first part of my book, which advances an account of reason-giving uses of arguments. I address Marraud’s considerations in the order that he advances them.
Marraud objects to the book’s construal of the premises of a dialectical argument advanced as individual reasons for believing the conclusion. Marraud favors Blair’s view of the connection between premises and reasons according to which premises taken together are advanced as a single reason for adopting a certain proposition attitude toward the conclusion or for doing what it describes (e.g., see Blair 2003; in my 2024, I discuss and defend Blair’s view). I am not sure what turns on this. It seems more than just a terminological disagreement regarding “reason” and “premise,” but it is hard to see the substance of the different treatments of premises qua reasons for believing the conclusion in our talk about dialectical arguments. For example, consider the following from Snoeck Henkemans on the structure of argumentation in an article that Marraud cites.
Reasoning is serial if one of the reasons supports the other. If reasoning is linked, each of the reasons given are directly related to the standpoint, and the reasons work together as a unit. When each reason separately supports the standpoint (to some degree), the reasoning is convergent. A complex argument can combine all of these types of argument structure. (2000, 447)
This seems to reflect the book’s construal of the premises of a dialectical argument advanced as individual reasons for the conclusion and not collectively as a reason for the conclusion. How is Snoeck Henkemans’s account of the structure of argumentation damaged by equating premises with reasons?
Moving on, Marraud claims that my view that inference claims are not parts of arguments commits me to some sort of holism understood as entailing the existence of contextual factors that are relevant to evaluating the logical properties of arguments. I look forward to learning more about such holism from Marraud’s work. In the meantime, I am not sure about Marraud’s claim. I take it to be standard in informal logic to take arguments to be premise-conclusion complexes. The argumentative import of an inference claim is that it communicates how an arguer is using their argument. Accordingly, the primary import of an arguer’s inference claim is pragmatic, not logical. A central theme of the book is that it isn’t arguments that are good or bad, but rather reason-giving uses of arguments as instruments of persuasion, justification, inquiry and so forth.
Finally, Marraud rejects what he takes to be a tenet of the book, put as follows.
I will take the import of an inference claim associated with an argument to merely be that the conclusion follows from the premises(s). Of course, if the premises are reasons for believing the conclusion, then the conclusion follows from them in some sense (105).
Marraud takes this to passage to imply an incorrect assimilation of “P favours C” to “C follows from P,” understanding the former as follows.
. . . the favouring relation, the relation in which features of the situation stand to action or to belief when they are reasons for doing one thing rather than another or for believing one thing rather than another (Dancy 2004, 79).
I have two responses.
First, the passage that Marraud cites is from the discussion in Chapter 3 aimed at answering the question: How does an arguer’s statement of their argument convey the associated inference claim? The text immediately preceding what Marraud cites is: “In order to separate my account of inference claims in terms of reasons for belief from the issue of how they are conveyed by statements of arguments, unless otherwise specified in what follows . . .” Accordingly, what Marraud cites is not intended to be a tenet. Rather, my taking the import of an inference claim to be that the conclusion follows from the premises is a rhetorical move in Chapter 3 which amounts to bracketing my favored view of inference claims in terms of reasons for belief for purposes of answering Chapter 3’s question. On any plausible view of the content of an inference claim, it isn’t true unless the conclusion follows—in some way—from the premises. I thought that this rhetorical move would help ensure that I wasn’t begging any questions in pursuing the aims of Chapter 3.
Second, as suggested in the passage on p. 105 that Marraud cites, the book adopts Blair’s normative conception of argument in so far as the premises of an argument count as reasons for believing the conclusion only if the conclusion follows from them. This is thoroughly discussed in the book, in Chapters 5 and 7. Unless I am misreading Dancy, P favours believing that C only if C follows—in some sense—from P. Hard to imagine a counterexample. That the earth is the third planet from the sun doesn’t favor believing that McKeon teaches at MSU because the latter doesn’t follow from the former. The book explicitly rejects the converse. A main job of Chapter 5 is to defend my rejection that if C follows—in some sense—from P, then P favours believing that C. As I discuss, if correct, this deflates some criticisms of classical logic. E.g., that P & ~P ∴ Q is a classically valid argument does not commit the classical logician to the claim that for some P, Q, “P & ~P” favors believing Q. So, it is consistent to accept the validity of the argument while denying that an argument that instantiates the pattern can be successfully used in a reason-giving way. In short, I don’t see how the book assimilates “favoring” and “following.”
References
Blair, J. A. (2003). Towards a Philosophy of Argument. In J. A. Blair, R. H. Johnson, H. V. Hansen, & C. W. Tindale (Eds.), Informal Logic at 25: Proceedings of the 25th Anniversary Conference [CD-ROM]. Ontario Society for the Study of Argumentation.
McKeon, M. W. (2024). In Defense of a Normative Concept of Argument. Argumentation 38, 247–264.
Snoeck Henkemans, A. F. (2000). State-of-the-Art: The Structure of Argumentation. Argumentation 14, 447–473.
1.26.26 |
Response
Abstract Arguments, Dialectical Arguments, and Inference Claims
Introduction
Matthew W. McKeon’s Arguments and Reason-Giving impresses with its rigorous analysis, broad engagement with literature, and thorough justification of its theses. These traits suggest McKeon would welcome critical questions over praise. My commentary examines two issues: 1) the abstract versus dialectical argument distinction, and 2) McKeon’s account of how an inference claim is conveyed to an argument’s recipient.
1
1.1
McKeon distinguishes between abstract arguments, defined as premise/conclusion pairs, and dialectical arguments. Dialectical arguments are, in essence, abstract arguments in use—whether someone utters them, thinks them, or writes them down (McKeon, 51). This idea evokes an image as if authors of arguments reach into an abstract pool of arguments and clothe these abstractions in unique situations in time and space with sounds and ink, according to the rules of a given natural language. In the following lines, I will attempt to describe the problem with this understanding of the relationship between abstract and dialectical arguments.
1.2
McKeon states on p. 51 that abstract arguments consist of propositions independent of the mind (50). He continues: “If you give a dialectical argument, then you express an abstract argument whose premises you offer or advance as reasons for believing the conclusion” (51). I take this to mean that every dialectical argument is an abstract argument in use. I assume we can agree that sometimes we use invalid arguments. Abstract arguments either contain only valid arguments (VA) or also invalid arguments (IA). By VA, I mean there is a support relation between premises and conclusion, based on monotonic or non-monotonic reasoning; by IA, such a relation is absent.
1.3
If the pool of abstract arguments contains only VA, how do people sometimes use IA? But if it also includes IA, what defines it then? It is reasonable to assume that abstract arguments are relatively, not absolutely, valid or invalid. This means they are valid in some frameworks (e.g., modus ponens classically) but invalid in others (e.g., L3). If VA are relative and not absolute, then whether the conclusion strictly follows from the premises depends on the context of use, contrary to McKeon’s assumption (80).
Again, let validity–invalidity of abstract arguments be absolute. Then, if all abstract arguments are VA, it is unclear how it is possible that we sometimes produce invalid arguments. However, if there are also IA, then it is unclear what defines the classes of abstract arguments. Alternatively, abstract arguments are valid or invalid relatively. In that case, however, they do not have the property of the conclusion strictly following from the premises independently of the context of use. For example, in a setting where it is assumed that the principle of bivalence does not hold, an argument of the form modus ponens may not be valid. The same argument would be valid in classical conditions.
1.4
A pragmatic view avoids these problems: whether statements express an argument depends solely on intentions, not on logical connections between propositions or factual relations between what they refer to. “Peter is unmarried because he’s a bachelor” is an argument in this sense, just as “This glass is fragile because that coffee is black” or “Fabius will not die at sea because he was born under the star Sirius” are arguments. Analyzing these relative to a framework like PL1 is modeling the basis for inference from a specific framework’s perspective—an interpretive effort driven by a methodological goal.
1.5
Two propositions from different books aren’t an argument unless linked by intent. Imagine opening two books by different authors from different times and finding a proposition in one logically implying another in the second. It seems intuitively wrong to say we’ve found an argument, even an abstract one. For example, we may know nothing about direction of support—what if the logically implied proposition is actually false and therefore constitutes a good reason to reject the first one? What we’ve found in that book is just a logical connection between two distinct propositions. But if a coded message links “Big Brother is watching” in one book as a reason for “It’s time to pack up” in another, that would be an argument, because the statements would be linked by the intent of whoever directed us to them.
2
2.1
If a speaker presents premises as reasons for believing a conclusion, then, according to McKeon, they use an argument in a reason-giving way. Setting aside my doubts about McKeon’s understanding of “using” abstract arguments, let us turn to Chapter 3, which I see as the book’s central theoretical contribution. In its Preamble, McKeon writes:
In Chapter 3, my aim is to deepen understanding of how an inference claim matters to a reason-giving use of argument. Toward this end, I defend two theses:
(I) A reason-giving use of argument isn’t successful unless the corresponding inference claim is true.
(II) An inference claim is conveyed by one’s statement of an argument by means of assertion as opposed to being merely implied or implicated.
2.2
My further critical remarks focus primarily on thesis II. The inference claim (IC) is not among the argument’s premises but is the claim that the premises “are collectively reasons for believing the conclusion, i.e., for believing that it is true” (77). In a reason-giving use of argument (RGUA), the speaker must not only believe the IC (98), but the IC must also be true for the RGUA to succeed (83–84). Beyond belief and truth (or acceptability), RGUA requires that expressing the IC be the primary, not secondary, point of stating the argument (100).
2.3
McKeon insists these conditions (belief, value, primacy) are met only if we assume the IC is asserted when presenting the argument. Alternative hypotheses fail to satisfy one or more of these conditions. McKeon rejects Hitchcock’s view that the IC is conveyed by mere implication based on an illative expression’s meaning (103–105), Grice’s conversational implicature (105–108), and Grice’s conventional implicature (108–111). McKeon defends his stance—that the IC is asserted—via semantic multidimensionality, where uttering one sentence can assert multiple propositions:
This enables the analysis of the meaning of a speaker S’s utterance, p, so q, in terms of two propositions expressed by the sentence S utters: p and q, q follows from p. Here, what is expressed comprises more than one proposition. Since part of what is expressed in stating an argument is the inference claim asserted, the inference claim must be true if what is stated in the statement of the argument is true (112).
2.4
I won’t delve into the details of McKeon’s justification or his critique of rival views. Instead, I focus on a category mistake central to formulating one RGUA condition (Value), critiquing alternative explanations of how the IC is conveyed, and defending his view that the IC is asserted. This category mistake lies in treating arguments as bearers of truth-value:
(Value) If the statement of an argument used in a reason-giving way is true or acceptable, then so too is the inference claim it conveys (100).
2.5
On p. 113, McKeon explicitly states that “statements of arguments in their reason-giving sense are truth evaluable.” If this were so, arguments would stand in the same relations as propositions: some consistent, others contradictory or contrary, and some logically following from others. This is, of course, unacceptable. Arguments and propositions are closely related but distinct kinds of objects. McKeon overlooks this fundamental difference, enabling his critique of opponents and defense of his own view. On p. 112, for instance, he treats an argument as a statement comprising multiple propositions—premises, conclusion, and IC. His critique of Hitchcock and conventional implicature hinges on the idea that an argument’s author may not be aware of their statements’ logical implications (103, 110). Yet, this critique consistently applies to McKeon’s own solution, as the alleged assertion of the IC is logically implied by removing conjunction from the complex statement, he takes an argument to be. But this veers into details I meant to avoid. The key is that an argument is neither a conjunction nor a set of propositions. It consists at least of propositions and a relation between them, differing so fundamentally from logical connectives that we speak of validity or convincingness, not truth.
2.6
McKeon’s main reason for rejecting rival explanations of how the IC is conveyed is that they don’t guarantee the author’s belief in the IC, its truth, or its status as the primary aim of stating the argument. The problem is that no absolute guarantees are possible. Even an explicit IC could be a lie. Understanding arguments and assessing participants’ beliefs and commitments in argumentative situations always involves context-sensitive interpretation, replete with non-monotonic reasoning and resting on background assumptions. If someone says “p, thus q,” I assume charitably they are sincere and believe what they say. I assume they know the conventional meaning of “thus” and what it logically implies. And I conclude via inference to the best explanation that their use of “thus” with p and q reflects their belief that p supports q.
-
1.26.26 |
Reply
Response to Tomáš Kollárik
Thanks to Tomáš Kollárik for his thoughtful review of Arguments and Reason-Giving. I am grateful for the opportunity to respond to his critical reflections. Kollárik focuses on two issues: 1) the abstract versus dialectical argument distinction, and 2) my account of how an inference claim is conveyed to an argument’s recipient. In what follows, I reply starting with 1).
Kollárik correctly describes the book’s distinction between abstract and dialectical arguments. Abstract arguments are defined as ordered pairs: a set of propositions, and a proposition (the book ignores arguments studied in multiple-conclusion logic). Intuitively, the former is the set of premises, and the latter is the conclusion. Dialectical arguments are, in essence, abstract arguments in use—whether someone utters them, thinks them, or writes them down.
Kollárik sees two problems with this picture, which I put as follows. First, if there are valid as well as invalid abstract arguments then it is unclear what defines the classes of abstract arguments. The force of this consideration turns on the idea that abstract arguments are only clearly defined based on a logical connection between premise-set and conclusion. Second, abstract arguments are valid or invalid relatively. Accordingly, the conclusion of a valid abstract arguments doesn’t strictly follow from the premises independently of the context of use as I claim in the book. The force of this consideration turns on the idea that because validity is a theory-relative concept that an abstract argument is valid is not independent of how it is being used dialectically. I now respond to each problem.
First, what clearly defines the classes of abstract arguments is set-theoretic and independent of criteria of validity/invalidity. For example, relative to domain of propositions, e.g., {P, Q, R} it is fairly straightforward to determine the class of arguments as characterized above. E.g., <{P}, Q>, <{P, Q}, R>, <{R}, R>, <∅, Q>, . . . . Don’t we know how to carry on without appealing to criteria for validity? Seems so to me.
Regarding the second problem, it is worthwhile to highlight that an abstract argument in use, i.e., a dialectical argument, is an abstract argument whose premises are given as reasons for believing the conclusion (51). Suppose in logic class I put a disjunctive syllogism on the board and ask my students to give a proof of its validity in classical logic and give a counterexample in a paraconsistent logic. Kollárik is right to think that I am using the disjunctive syllogism (for pedagogical purposes). However, the disjunctive syllogism so used isn’t a dialectical argument as I am not giving its premises as reasons for the conclusion (even if a student should be persuaded that conclusion is true in light of its premises). In short, the relevant use of an abstract argument that qualifies it as dialectical argument is a reason-giving use of it. That an argument is (formally) valid is independent of whether it is or can be used in a reason-giving way. Accordingly, in judging an abstract argument as valid according to a logical theory does not thereby qualify the argument as dialectical. In short, the pragmatic view advocated by Kollárik with its emphasis on arguer intentions aligns with the book’s conception of dialectical arguments, but is not germane to the nature of abstract arguments.
I now turn to 2) my account of how an inference claim is conveyed to an argument’s recipient. Kollárik has two concerns here. The first concerns its reliance on the claim (Value): If the statement of an argument used in a reason-giving way is true or acceptable, then so too is the inference claim it conveys (100). The problem Kollárik sees is that this involves the category mistake of treating arguments as bearers of truth-value. Kollárik’s second concern is that the book’s claim that an inference claim is asserted by an arguer in their statement of an argument is undermotivated since in an argumentative exchange there can be no guarantee of an arguer’s belief of their inference claim. An interlocutor must infer this from contextual factors, which may be uncertain. I now respond starting with Kollárik’s first concern.
The book distinguishes between (i) abstract arguments, (ii) dialectical arguments, i.e., reason-giving uses of abstract arguments, and (iii) statements of arguments used in reason-giving ways. I agree with Kollárik that to think (i) and (ii) are truth evaluable is to commit a category mistake. However, that (iii) are truth-evaluable, as Kollárik rightly notes the book claims, does not imply that either (i) or (ii) are truth-evaluable. From the book’s characterizations of (i) and (ii) it should be clear that I agree with Kollárik that an argument is neither a conjunction nor a (non-ordered) set of propositions. However, I say that in stating an argument that one uses in a reason-giving way one asserts that the premises are true, that the conclusion is true, and that the former constitute a reason in the intended sense for believing the conclusion. I don’t see how this invokes a commitment to arguments being truth evaluable.
Regarding Kollárik’s second concern, I don’t see why the fact that there can be no guarantees that an arguer believes their inference claim undercuts my view that an inference claim is conveyed by assertion. To borrow from Kollárik’s example, suppose A says to B “p, thus q.” B assumes that A’s utterance is felicitous. How is A’s inference-claim belief conveyed to B? I say it is by means of assertion and not by implication or implicature. B may conclude from A’s utterance via inference to the best explanation that A’s use of “thus” with p and q reflects A’s belief that p supports q. But a premise of B’s inference then is that A asserts their inference claim rather than merely implying or implicating it. B’s inference is problematic if instead a premise is that A’s utterance merely implies or implicates A’s inference claim because there is no guarantee that A thereby believes it.
John Casey
Response
Reason-Receiving and the Asymmetry of Argument
Matt McKeon’s (2024) Arguments and Reason-Giving offers a normative account of the reason-giving use of argument whose goodness is grounded in the intention, accuracy, sincerity, and integrity of the one giving the reasons. Reason-giving, in this context covers reasoning with oneself and reasoning with others. Broadly speaking, however, the norms of reasoning with oneself ground the norms of reasoning with others.
In this short comment I’ll be concerned with reasons-receiving, since, in argumentation at least, there is no reason-giving without reason-receiving, and reason-receiving is a thing on a different order from reason-giving. I think an account of the one is not an account of the other. Let’s call this the problem of the asymmetry of argument: from my point of view, my giving of reasons is direct, voluntary, and pointed, rather more so than your receiving of them. What’s more important, I think, is that the nature of reason-receiving strongly bears on the nature of reason-giving, or it ought to. After a short review of some of the relevant parts of McKeon’s text, I’ll illustrate this case with two internally connected issues of reasons-receiving: (1) closure and (2) meta-argument.
The big thought of McKeon’s Argument and Reason-Giving is that the norms of argument can be derived from an investigation into the nature of reason-giving. When I really mean to give reasons, in other words, certain things must be the case. First, I must do it intentionally; I mean that my premises are reasons to believe the conclusion. Second, your reason-giving must involve an inference claim, the idea that the premises are reasons for the conclusion. This distinguishes your activity from mere assertion, among other things. Third, in a reason-giving use of argument, one believes the inference claim. One thinks the reasons given really are reasons for the conclusion; this makes it authentic. At the center, then of the reason-giving notion of argument is a conception of inference. Arguing, so construed, is a voluntary, intentional, and purposive activity. Now what about the addressee? Following Pinto (2001), McKeon conceives of argument as an arguer inviting an addressee to perform the inference that they themselves have performed. If the addressee performs the inference, or something like the inference, then we have success. Provided all involved do this with honesty and integrity (explained in Chapter 8), we have authentic reason giving. All of this, I think, is certainly salutary for our practice of argument. I worry, however, that it’s not enough.
The first feature of McKeon’s account is both strongly voluntarist and, for lack of a better term, atomistic. I mean just this argument in this way. This makes sense, the collection of speech acts involved in uttering reasons seem to be things under one’s direct and voluntary control. If you’re going to have advice for what to do, it’s best if someone can do it. It seems to me, however, that the voluntarism about arguing McKeon defends can only be a general voluntarism, if that. This means that a person can intend to give reasons in general. A means to give B reasons to conclude that p. B has some role in interpreting A’s reasons. B’s interpretation of A’s reasons can go in a couple of directions, direction McKeon seems to exclude. The trouble is, for me at least, I don’t think he can exclude them.
Let’s illustrate this by playing on a standard distinction in argumentation theory, that between commitments and beliefs. There’s a strong tradition, running from Hamblin (1970) through Walton and Krabbe (1995, 23), Hitchcock, and Pragma-dialectics, where arguments are not about beliefs (which Hamblin thought “too psychological” (264)) but rather about commitments (what L. J. Cohen (1995) called “acceptances”) expressed in speech acts. The virtue of commitments is that they’re public, trackable, voluntary, and so you can take on and drop them at will, and finally, they’re subject to the closure property. Beliefs, by contrast, are not voluntary in the same way and they’re not subject, qua beliefs, to closure. By taking on a commitment, I become thereby committed to other things logically (broadly speaking) associated with it. So, to use their example, if I commit to making dinner, I’m committed to making food. If I commit to p and q, then I’m committed to p and also to q.
What does this mean for reason-giving? While it is true that in the giving of reasons, I may mean that my interlocutor be persuaded by some specific inference claim about which I might be very clear. But the fact is that the logical nature of commitments means that I bring along quite a lot more than that. To repeat, the rational offering of reasons for McKeon is strongly intentional—I mean just these reasons connected in this inferential claim. Be that as it may, our inferential claims, under closure, are multi-directional, and they involve us in obligations well beyond what we might imagine. In arguing for p, for instance, we may find ourselves tied to deeper commitments, what Walton and Krabbe called, “dark side commitments” (1995, 11). What is interesting is that it’s in arguing that such dark-side commitments come to light. And this can only happen if our argumentative intention is broad rather than narrow. This is why I say that the argumentative intention is a general rather than a specific one.
My second point is related to this. Let’s take arguer A. In making the case for p, they’re by necessity making additional argumentative moves. Call this meta-argument. These meta-argumentative moves accompany the core argumentative inference being invited. Let’s look at a simple case. Arguer A makes the case for p to B. Given that Addressee B is going to interpret the meta-argumentative backdrop, arguer A, in making their case, would seem to be on the hook for it. A’s making the case for p, in other words, might in itself (or maybe atomically) meet all of the criteria of intellectual integrity outlined in Chapter 8, but it might fail in meta-argument. This failure can only make sense if reason-giving is understood in terms of reason-receiving.
In sum, reason-giving, for me at least, is a chaotic, complicated, multi-layered, and also fundamentally asymmetrical affair. A’s giving reasons and B’s receiving them are not the same sorts of activities: A’s giving of reasons can be directed in ways that B’s receiving of them are not. But they’re bound up together such that there’s no giving without receiving. Now I don’t find that anything McKeon has done here explicitly denies that. And there’s certainly a case to be made, as McKeon has done, the real responsibility is in the giving and that the receiver is just another giver. I leave that question to him.
References
Cohen, L. J. (1995). An Essay on Belief and Acceptance. Clarendon Press.
Hamblin, C. L. (1970). Fallacies. Methuen.
McKeon, M. W. (2024). Arguments and Reason-Giving (1st ed.). Oxford University Press. https://doi.org/10.1093/oso/9780197751633.001.0001
Pinto, R. C. (2001). The Relation of Argument to Inference. In F. H. van Eemeren, R. Grootendorst, J. Wenzel, & J. Woods (Eds.), Argument, Inference and Dialectic (Vol. 4, pp. 32–45). Springer Netherlands. https://doi.org/10.1007/978-94-017-0783-1_4
Walton, D. N., & Krabbe, E. C. W. (1995). Commitment in Dialogue: Basic Concepts of Interpersonal Reasoning. State University of New York Press.
1.9.26 | Matthew McKeon
Reply
Response to John Casey
Thanks to John Casey for his insights and critical commentary. I am grateful for the opportunity to respond. I’ll address Casey’s main considerations in the order that he presents them.
A central claim of the book, put generally, is that arguing qua reason-giving is inherently intentional in so far as it involves an arguer advancing premises they intend to serve as reasons of some sort for the conclusion. No argumentative intention, no arguing. Casey is correct in his assessment that my conception of the argumentative intention makes it specific: it attends to the status of the premises as reasons for the conclusion. Casey worries that the argumentative intention so construed is too specific since it ignores the commitments/obligations the arguer incurs by their act of reason-giving. Argumentative exchanges between arguers and their addressees can bring arguers’ commitments to light, which may count as news for arguers. According to Casey, this can only happen if our argumentative intention is broad rather than narrow. This is why he says that “the argumentative intention is a general rather than a specific one.” However, I am not seeing the motivation for so generalizing the argumentative intention. As Casey notes, on my view the rational offering of reasons is strongly intentional—I mean just these reasons connected in this inferential claim. How, exactly, does this fail to reflect the relevance of arguer’s commitments to argumentative exchanges between arguers and their addressees? Yes, my act of reason-giving invokes commitments that I may not understand, or that I may be surprised by. But why does this mean that “a person can intend to give reasons in general”? My commitments are not advanced as reasons for the conclusion, although they may help explain why I have reasons for the conclusion.
This raises another point in Casey’s motivation for the generality of argumentative intention that I am not sure about. I’ll use Casey’s example to elaborate.
What, exactly, is B’s role in A’s act of reason-giving? That is, why does B’s interpretation of A’s reasons matter to A’s reason-giving? On my view, B’s interpretation of A’s reasons plays no role in determining that A gives B reasons for p. This aligns with the idea that whether or not B understands A to be arguing for p is irrelevant to whether A so argues. If an aim of A’s reason-giving is to persuade B, then yes B has some role in interpreting A’s reasons, which may involve uncovering and assessing A’s commitments. But unlike A’s premises, A’s commitments are not advanced as reasons for believing the conclusion.
Suppose A’s argument instantiates: P⊃Q, P, so Q. The status of A’s premises as reasons to conclude Q turns on the correctness of modus ponens. Accordingly, A is committed to the correctness of modus ponens. However, that modus ponens is correct is not a reason in addition to the premises that A gives for believing the conclusion. Casey may think otherwise. If so, he may face a version of Carroll’s paradox.
I now turn to Casey’s second main point, which is that the fact an arguer’s meta-argumentative moves in support of their conclusion might fail as interpreted by their addressees only makes sense if reason-giving is understood in terms of reason-receiving. I take it that the central idea is that the arguer has a dialectical obligation to be accountable to their interlocutor’s assessments of their arguing. However, what I am not seeing is why this only makes sense if reason-giving is understood in terms of reason-receiving.
As Casey acknowledges, “A’s giving reasons and B’s receiving them are not the same sorts of activities.” Perhaps, receiving reasons and accepting the reasons received should also be distinguished. At any rate, I think it is important to distinguish between (i) an arguer’s meta-argumentative moves, (ii) an addressee’s interpretation of them, and (iii) an addressee’s evaluation of them. It is far from obvious to me that (i) is partly constituted by (ii) or (iii). Hence, I don’t understand why the aforementioned arguer’s dialectical obligation only makes sense if reason-giving is understood in terms of reason-receiving.
To wrap up, I take argumentation to be a dialogical exchange in which there is reason-giving and reason-receiving. Casey’s commentary highlights how my conception of arguing (but not argumentation) is arguer centric. Put simply, one argues if and only if one advances premises one intends to be reasons to conclude some claim p. My book simplifies and unpacks “reasons” in terms of “reasons for belief.” The reception of one’s reason-giving is irrelevant to whether one so argues. Accordingly, arguing qua reason-giving is not characterized in terms of its aims whether that be to convince, confuse, justify, anger, or whatever. Bringing such aims into the conception of arguing may make it more addressee centric. Such a conception of arguing or reason-giving may be one that Casey favors according to which whether there is reason-giving may turn on dimensions of reason-receiving. A locus of disagreement here may be my view that reason-receiving is irrelevant to whether or not one argues, i.e., whether or not one gives reasons to conclude some claim p. What is relevant is that one, in fact, gives reasons in the sense that one intends for p. This central idea of my book links the epistemology of reasons for belief to the pragmatics of argument by showing how different types of reasons that one may advance for a claim generate different ways of arguing for it that thereby necessitate different norms for arguing.