The uses of argument in communicative contexts

This paper challenges the view that arguments are (by definition, as it were) attempts to persuade or convince an audience to accept (or reject) a point of view by presenting reasons for (or against) that point of view. I maintain, first, that an arguer need not intend any effect beyond that of making it manifest to readers or hearers that there is a reason for doing some particular thing (e.g. for believing a certain proposition, or alternatively for rejecting it), and second that when an arguer is in fact trying to induce an effect above and beyond rendering a reason manifest, the effect intended – the use to which his or her argument is put – need not be that hearers “do” what the stated reasons are reasons for “doing.” Where the actual or intended effect of making a reason R for “doing X” manifest is something other than “doing X,” I call it an oblique – as opposed to a direct – effect of making that reason manifest. The core of the paper presents an overview or map of the main categories of effect which arguments can have, and the main sub-types within each category, calling attention to the points at which such effects can be indirect or oblique effects. The purpose of that typology is to make it clear (i) how oblique effects can come about and (ii) how important a role they can play in the conduct of argumentation.


1
There is, undoubtedly, an important connection between argument and argumentation, on the one hand, and persuasion, on the other. One of the meanings listed for 'argument' in the Merriam-Webster Dictionary is "discourse intended to persuade"; and the first sense of "persuade" in that same dictionary is rendered as "to move by argument, entreaty, or expostulation to a belief, position, or course of action." More to the point, within the literature on argumentation it is not uncommon to find 'argument' defined by reference to concepts of persuading or convincing 2 -for example, van Eemeren and Grootendorst (1984, p. 43 and all of chapter 3), Johnson and Blair (1977, p. 3), Johnson (2000, p. 168), and even myself. 3 Indeed, the emphasis on persuasion in the rhetorical analysis of argumentation goes back at least to Aristotle's definition of rhetoric as "the faculty of observing in any case the available means of persuasion" (Rhetoric I, 2, 1355b 26-27), and in the Aristotelian rhetorical tradition arguments (or logoi) are treated almost as exclusively as instruments of persuasion.
However, as I will try show below, the mere fact that a reason or an argument has been presented in a communicative context can effect changes in that context, and can affect the behavior and commitments of participants in that context, even when the reason or argument presented does not persuade any participant to do what it supplies a reason for doing. Effects of the mere presentation of an argument -effects that an argument can have even though it does not persuade -will be called its non-persuasive effects. The purpose of this paper is (a) to call attention to the existence of non-persuasive effects, (b) to sketch a map of the main categories of effect which arguments can achieve and the main types within those categories, calling attention to points at which the effects categorized can be non-persuasive effects, (c) to suggest that a full understanding of the uses of argument is impossible without paying careful attention to non-persuasive effects and their consequences.
In what follows, I recognize three types or levels of effect that the presentation of an argument can bring about.
(a) a primary effect, which consists in making it manifest to participants in a communicative context (i) that there is a reason for doing something and (ii) what one such reason is, (b) secondary effects, which consist of holding (or not holding) certain cognitive, conative 4 or evaluative attitudes when these are induced by argument, (c) tertiary effects, which consist of two sorts of consequences that may flow from secondary effects: (i) changes in what I call the context availability of propositions and (ii) the performance or non-performance of overt actions when it is a direct result of the secondary effects produced by argument.
For example, suppose I persuade you that it will rain tomorrow by offering you reasons for believing that conclusion. Then a) the primary effect of presenting those reasons is to put them "out there" in such a way that you can understand what they are and in such a way that they might have an effect on you, b) the secondary effect is inducing you to believe that it will rain tomorrow, c) one tertiary effect is to make "it will rain tomorrow" available as a premiss in an argument I intend to offer later -perhaps an argument for canceling a tennis match planned for tomorrow.
The use to which an argument is put -the point of making it -most frequently lies in its tertiary effects. And the secondary effects from which such tertiary effects flow will often, I maintain, be non-persuasive effects.

I. Preliminaries: Argument, Persuasion, and Communicative Context
A transactional context may permit only one-way communication or it may permit twoway communication. "Two-way" communication makes possible interactions in which parties produce messages that are responses to the messages of other parties. I will call a context that provides little or no scope for responses non-interactive; 9 I will call contexts that provide scope for response interactive.
Transactional contexts can be two-party or multi-party (i.e., involve more than two parties). These distinctions are represented in the following table, in which four "basic types" of transactional context are identified.

Non-interactive Interactive Two party
Private communication Dialogue

Multi-party Broadcast Forum
See Appendix A for a categorization of additional key features of transactional contexts.

3
In English, we can speak of persuading a person to do something and we can also speak of persuading someone that something is the case. A necessary condition of persuasion in both instances is that we induce someone to do something by means of some sort of communication. 10 To persuade someone to give up smoking is to induce him by means of some sort of communication to give up smoking. And to persuade someone that it will rain tomorrow is to induce him by means of some sort of communication to believe or to accept the proposition that it will rain.
Of course, we can persuade others not to do some particular thing as well as persuade them to do it -as when I persuade you not to quit your job just yet or persuade you not to accept a certain proposition.
For purposes of this paper, my remarks about persuasion will emphasize the fact that through persuasion we induce others to do, or not to do, various things.

II. The Existence of Non-Persuasive Effects 4
Because I don't want to get drawn into a debate about how the word 'persuasion' should be defined, 11 I will introduce a notion of normal persuasion by argument by formulating two requirements which an instance of persuasion much meet in order to be considered "normal." Non-persuasive effects will be said to occur when an argument induces someone to do something, but their doing it is not the result of "normal persuasion." For "normal persuasion" by argument to occur, a person must accept the argument offered to him -by which I mean that a person must accept the premisses of that argument, must do what the argument gives him a reason for doing, and must do it on account of and in response to that very reason. For example, Tom induces Harry to repay a debt within the week by offering Harry reasons for doing so -saying, perhaps, "You gave me your word that you would repay me when I need the money, and I really have to have it this week" -and as a result Harry repays the debt immediately for those very reasons (i.e., he does so because of his promise coupled with Tom's need to have the money this week). This is in line with an idea I've suggested in earlier papers; namely, that arguments are invitations to inference. Accepting such an invitation is a matter drawing the inference which the argument invites.
Normal persuasion has the following two features: a) what S is induced to do is the very thing which S has been given reasons for doing, b) S does (or doesn't do) what he was invited to do (or not to do) because he makes the inference he was invited to make.
Non-persuasive effects of presenting arguments are effects that occur even though the argument presented is not accepted in the sense just indicated. 12 5 Here are three examples of non-persuasive effects.

Example #1
Smith induces Jones not to believe the proposition that it will rain tomorrow (i.e., to give up that belief) by offering Jones reasons for believing the proposition that it will not rain tomorrow -perhaps by calling attention to the fact that the CBC weather person says that the probability of precipitation tomorrow is less than 10%. Notice that although Smith induces Jones not to believe the proposition that it will rain, the reason offered was a reason for believing a different proposition (the contradictory of the first proposition).
But perhaps any reason for believing a proposition to be false is eo ipso a reason for ceasing to believe it. If so, then by giving Jones a reason for believing it's not raining, Smith has by the very fact given him a reason not to believe the proposition that it is raining. In that event, the example cited above could turn about to be a case of normal persuasion after all. 13 But now consider the possibility that Smith's argument had its effect even though Jones was not induced to believe the proposition that it won't rain. Perhaps in Jones' estimation the reason Smith offered for believing "It won't rain" counterbalanced reasons Jones already had for believing "It will rain" -for example, Jones may have known that an equally reputable weather forecaster had predicted heavy rains for tomorrow. In such instances, I submit, we don't have normal persuasion by argument because Jones does not make the inference which Smith's argument invited him to make -an inference that would require Jones to conclude that it won't rain tomorrow. Smith's argument has modified Jones' attitude toward the proposition "it will rain tomorrow" even though Jones didn't make the inference which Smith's argument invited him to make.

Example #2
Larry, Moe and Curly are debating whether Al Gore should have been declared winner of the US presidential election in 2000. Larry is trying to mount an argument based on the claim that Gore's plurality over Bush in Palm Beach was under-reported by 6000 votes, and Moe in inclined to buy Larry's argument because he takes Larry's word about a 6000 vote undercount. Curly then claims that Larry can't be right, because in the year following the election several newspapers examined the Palm Beach ballots and found that the Palm Beach plurality was under-reported by only a few dozen votes. Neither Larry nor Moe have ever heard of such a review and moreover are not willing to take Curly's word about it. In other words, they don't accept Curly's argument or its conclusion -they are not even sure whether Curly's argument actually has merit, because Curly's account of the details seems pretty sketchy. (This differs from the previous example, in which Jones knew that Smith's argument had merit. 14 ) Even though Larry and Moe don't accept Curly's reason for saying that Larry can't be right, the mere possibility that Curly's argument might turn out to be a good one induces them to defer judgment on Larry's earlier claim about a 6000 vote undercount. As a result, the argument Larry was attempting to mount loses its force for the time being.

Example #3
Larry, Moe and Curly are still discussing the US presidential election. Larry has been holding that the Supreme Court never should have stopped the Florida recounts that were in progress. Curly steps in with an argument against Larry's position which Moe finds quite persuasive. Larry thinks Curly's argument is "a lot of nonsense" -he is quite convinced that its premisses are false (they consist of claims about the authority and the responsibility which the US constitution gives to the Supreme Court) and he's quite sure that they don't really support Curly's conclusion either. (This example differs from the preceding one, in which Larry and Moe simply weren't sure about the merits of Curly's argument.) But Larry's mastery of the facts of constitutional law and his debating skills are limited, so has no idea how to show Curly's argument is defective, especially in the face of Moe's respect for Curly's supposed "expertise" about matters of constitutional law. Larry thinks it would be counterproductive to challenge Curly's credibility in the face of Moe's respect for him. As a result, Curly's argument induces Larry to abandon his position on this particular point, even though Larry doesn't accept that argument, and is quite convinced that it is without merit. This is an example of how an argument which does not induce a participant in a discussion to believe its conclusion can nevertheless induce that participant to modify what he is willing to say within that discussion -in other words, it's an example of change in a conative or evaluative attitude toward certain speech acts which can have a profound effect on subsequent developments within a discussion 6 The presentation of arguments can, of course, influence hearers and readers in a myriad ways -from boring them, causing them to daydream about some example mentioned in the argument, inducing Big Mac attacks, and so on. If the concept of non-persuasive effect is to help us understand the use of arguments in communicative contexts, that concept needs to be spelled out in more detail and the scope of its application needs to be pinned down. I attempt to achieve those objectives with the mapping and analysis attempted in Parts III and IV of this paper.

III. Primary and Secondary Effects
The uses of argument -the purposes realizable by argument-are limited to what can actually be achieved by offering reasons for doing something. This is so because a purpose is an intended effect. To understand the uses of argument, therefore, we must understand what arguments can actually achieve. As I indicated in the introduction, I distinguish between primary, secondary and tertiary effects of arguments.

The primary effect 7
The primary effect of an argument is to make a reason for doing something apparent or manifest -and in such a way that it is apparent or manifest (a) what the reason is (b) what the reason is a reason for and (c) that it is a reason (that it seems to call for doing what it is offered as a reason for doing).
Someone who presents an argument can fail to achieve that primary effect if his presentation is unclear, muddled, cryptic, insufficiently articulated, and so on. In such cases, the reason he is attempting to float will not be manifest (i.e., "readily perceived by the eye or the understanding").
If the primary effect is achieved, secondary effects become possible -and these may be either persuasive or non-persuasive effects. Whether these secondary effects actually occur depends on whether and how an argument receiver reacts or responds to reasons that have been made apparent. Before looking at how secondary effects occur, I will first give a more detailed account of what they are.

Secondary effects 8
Secondary effects are changes in cognitive, conative or evaluative attitudes induced by the presentation of an argument. To make use of this concept the more manageable, I restrict the secondary effects that are of interest to two types. Type I secondary effects of an argument A are attitudes which are induced by A and also meet the following two conditions. a) Where A consists of a reason for holding or not holding a cognitive, conative or evaluative attitude toward a propositional content p, Type I secondary effects of A include only related attitudes toward p, or toward the contradictory of p. 15 b) Where an argument contains reasons for holding or not holding some conative or evaluative attitude toward a non-propositional content O, Type I secondary effects of the argument include only related conative or evaluative attitudes towards O.
For purposes of these restrictions, a pair of attitudes will be considered to be related attitudes only if both attitudes are of the same general type (i.e., both are cognitive or both are conative or both are evaluative) and either (i) they are competing attitudes toward one and the same thing or (ii) adopting one attitude toward a given object is an obvious motivation for adopting the other attitude toward the same object (so that if reasons offered as reasons for believing p induce acceptance of p, acceptance of p will count as a Type I secondary effect of those reasons).
A Type II secondary effect of an argument A is an intention to perform (or not to perform) a speech act concerned with argument A, its premisses or its conclusion. For example, if your argument for believing p makes me unwilling to assert not-p, my unwillingness to assert not-p will count as a Type II secondary effect of that argument. Or if your argument for p contains q as a premiss and your presentation of that argument induces me to present an argument against q, my intention to present an argument against q will count as a Type II secondary effect or your argument.

9
I recognize three types of cognitive attitude toward propositions a) doxastic attitudes, which consist of full belief that a proposition is true, and a series of attitudes which are incompatible with full belief but are alternatives to it-presuming a proposition to be true, having some inclination to believe it, suspecting it is true, considering the matter of its truth a live option, being agnostic toward it, holding it false, and so on. b) acceptance attitudes, in a sense of accept in which to accept a proposition is to be prepared to use that proposition as a premiss or starting point in our reasoning (see, for example, Cohen 1992, esp. chapter 1) 16 c) degree of confidence in a proposition -as measured by the odds at which we would be willing to bet that it is true.
Acceptance is not a doxastic attitude. Acceptance is concerned with the role we are prepared to accord a proposition in our reasoning, doxastic attitudes toward propositions are concerned with whether or not those propositions are true.
Since people are usually prepared to use propositions which they believe as starting points in their reasoning, arguments that induce belief typically induce acceptance as well. But although acceptance often follows in the wake of belief, sometimes it does not: it can be perfectly reasonable to refuse to use something which we believe as a premiss in a particular context, and it can be perfectly reasonable to use something we don't believe as a premiss (as when we reason from simplifying assumptions that we know to be false). See Pinto 2003a and2003b and Appendix B on these points.

10
The bulk of the empirical research on persuasion appears to be concerned with changes in conative and evaluative attitudes -see for example O'Keefe 2002, chapters 1-6, in which 5 of the 6 research areas covered are concerned with persuasive effects on conative and evaluative attitudes.
I will mention here one type of conative attitude and a few types of evaluative attitude which seem to me to have obvious bearing on the use of arguments in communicative contexts. a) establishing or undermining intentions to act, or a willingness to act, in a given way (these are conative attitudes) i) a very large number of attempts at persuasion are attempts to induce people to perform, or not to perform, an overt action of some kind; such attempts frequently work, if they do, by inducing or inhibiting intentions to perform actions of those kinds. ii) many arguments appear to be aimed at preventing or discouraging other participants from offering particular arguments or objections; again, arguments can have such effects only if the make other participants unwilling to offer those arguments or objections.
b) inducing or undermining evaluative attitudes of trust, respect or liking for a participant (these are evaluative attitudes) i) there is considerable empirical evidence that factors such as the credibility of an arguer (a function of trust and respect afforded him or her) and whether hearers like an arguer affect how messages containing arguments are processed (O'Keefe 2002, chapter 6) and have an impact on the persuasive effect of messages (O'Keefe 2002, 181-199).
ii) participants who have credibility can often gain acceptance for the propositions they put forward without having to offer arguments for accepting those propositions, merely by giving assurance of the truth of those propositions.

Necessary conditions for normal persuasion 11
Think about what must happen if presenting an argument is to result in normal persuasion. a) receivers must not ignore the argument, but must pay some sort of attention to it. 17 b) they must understand the argument -i.e., they must understand what reasons are being presented and must understand what they are intended to be reasons for. c) they must accept every premiss that is crucial to the reasons presented. d) it must seem to them that the reasons "call for" doing what they are offered as reasons for doing. e) they must not decide that the reasons presented are undermined or overridden by other considerations. f) they must not "postpone a decision" about whether to do what the reasons are reasons for doing -on the grounds, for example, that they have reasons for postponing a decision about whether to do that thing which outweigh the reasons for doing it.
All but one of these are things that it lie within their power to do or not to do -the exception being its seeming to them that the reasons presented call for doing what they are offered as reasons for doing. 18 Of course, an arguer can try to present an argument in a way that will increase or maximize the chances (a) that receivers will pay attention to the argument, (b) that they will find it acceptable and/or (c) that they will not postpone a decision about whether to do what the reasons are reasons for doing. As I see it, tailoring presentation of an argument to maximize such chances is part of what the "rhetoric" (as contrasted with the "logic") of argument is about. For purposes of this paper, I ignore that rhetorical dimension, despite the fact that it has enormous practical importance.

The verdicts that mediate secondary effects 12
Typically, the effect that an argument has on the cognitive, conative and evaluative attitudes of a receiver depends on what the receiver makes of the argument after considering it. 19 I recognize four possible outcomes such assessment can yield.
O1) accepting the argument: doing what the reasons presented are reasons for doing, and doing it both for those reasons and in response them.
If an argument has been considered but not accepted, I say that it has been declined. 20 I recognize three "motivated" ways of declining an argument. O2) bracketing the argument: failing to accept it because one is unsure about a crucial feature or dimension of the argument -perhaps because there is insufficient reason to accept one or more crucial premisses, or because one isn't sure how strongly those premisses support the conclusion or because one wants additional information before making up one's mind (and therefore postpones a decision about whether to accept it). O3) rejecting the argument as defective: refusing to accept it because one is convinced that the argument has a disqualifying defect -perhaps because one or more crucial premisses is false or because on reflection one concludes that the reasons presented don't really call for doing what they are supposed to be reasons for doing. O4) refusing the argument in light of additional considerations: declining to accept it because one is convinced that additional information undermines, 21 overrides or counterbalances the reasons provided in the argument.
Let me call attention to seven possible ways in which arguments can produce nonpersuasive effects. I make no claim that the following list is exhaustive.

How arguments can produce non-persuasive effects 13
Consider non-interactive contexts first.
E1) A bracketed argument for adopting a certain attitude toward not-p can have a secondary effect on competing attitudes toward p. This is because a bracketed argument makes me aware that there might be a good reason for adopting that attitude toward not-p. If I currently hold a competing attitude toward p, and I treat the possibility that the original argument might turn out to be a good one as a "live option," I may come to hold my current attitude in abeyance. Thus a bracketed argument for believing not-p may induce me to withdraw assent from p -this is what happens in Example #2 of section 5 above. Or a bracketed argument consisting of reasons for me to return home may induce me to question the decision I made yesterday to stay put and lead me to hold that decision in abeyance. Or again, a bracketed argument for distrusting a person may induce me to hold my trust of him in abeyance. Whether or not such things happen will depend on the strength I think the bracketed argument would have, were it to prove sound, as compared to the strength I attribute to the reasons I have for my current attitude toward p. E2) Most arguments refused as defective or refused in light of additional information are unlikely to induce a change in cognitive attitudes and hence unlikely to have a secondary effect. However, when an argument is rejected in light of additional information that the receiver thinks is counterbalancing (but not undermining or overriding), it may well have a secondary effect, prompting receivers to withhold assent from something they previously assented to -this is what happens in Example #1 of section 5 above.
14 Next consider interactive contexts. Both of the points just made about non-interactive contexts apply here. In addition, additional non-persuasive effects become possible because new elements are introduced in interactive contexts.
In two-party interactive contexts, the receiver of an argument now has the possibility of speaking, and arguments that are declined can create an intention to speak. E3) bracketed or dismissed arguments can induce a desire or intention to present my reasons for bracketing or dismissing it or induce a desire to request backing for a premiss which appears arbitrary. E4) arguments rejected as defective can induce a desire or intention to show that crucial.
premisses are false or that the reasons presented are insufficient to warrant doing what they are offered as reasons for doing. E5) arguments refused in light of additional information are in effect recognized as creating a presumption in favor of doing what they purport to offer reasons for doing, and therefore as shifting the burden of proof. Arguments declined for this sort of reason typically induce a desire or intention to present a rejoinder consisting of the counter-evidence that undermines, overrides or counterbalances the reasons presented in the original argument.
In multi-party interactive contexts, beliefs about what other parties believe, accept and expect become factors which can affect what individual parties do and are willing to do. E6) arguments which are bracketed, discounted, refused or dismissed by one party may nevertheless have persuaded others. Individuals who decline to accept a particular argument, yet think it has persuaded others, may not know how to counter its effect on others and may therefore become unwilling to defend positions they would otherwise defend -this is what happens in Example #3 of section 5. E7) in many transactional contexts, parties have a general expectation that whenever an argument has been presented for not assenting to a proposition, no one should use that proposition in an argument unless the argument for not assenting to it has been rebutted 22 -i.e., shown to be defective or else undermined or overridden by counterevidence. Where such a general expectation is in effect, almost any argument that is not ignored or dismissed by everyone out of hand will create an expectation of rebuttal. In such circumstances, those who bracket, discount or refuse arguments they are unable to rebut are likely to become unwilling to do what those arguments "forbid."

V. The Uses of Argument
The nature of tertiary effects 15 Tertiary effects consist of two sorts of consequences that flow from secondary effects: a) making a proposition "available" or "unavailable" for subsequent use in the communicative context in which the argument occurs -I call this context availability. b) the performance or non-performance of overt actions arising from changes in intentions, when those changes are themselves secondary effects.
Context availability Speaking roughly, context availability is a matter of whether a proposition is available for use as a premiss in arguments and objections. The context availability of a proposition depends on two factors: (a) whether participants in the transactional contents are prepared to use it as a premiss in their inferences; that is to say, whether they accept it -this is a consequence of the fact that arguments are invitations to inference. (b) whether anything stands in the way of the asserting (or appealing to) the propositionsince a proposition that can't be asserted or appealed to is unavailable for use as a premiss in an argument.
Acceptance and assertability are individually necessary and jointly sufficient conditions of context availability.
Implications of the fact that acceptance is a necessary condition of context availability Three points are worth noting in this regard.
a) Since participants may be prepared to use a proposition in inferences about one matter or for one purpose, but not be prepared to use it for a different matter or for a different purpose, availability will always be availability for one or another specific purpose. b) In multi-party contexts, a minority of participants can fail to accept a proposition which the majority of participants accept. This raises a question about whether a proposition should be judged unavailable if the preponderance of participants accept it and only a few participants fail to accept it. c) Arguments that deprive propositions of acceptance make them unavailable; arguments that induce acceptance of propositions make them available only if they are also assertable.
Though there can be belief without acceptance, and acceptance without belief, in general people are prepared to use propositions which they believe as premisses in their inferences. In fact, it is safe to assume that someone who believes a proposition will accept it for any given purpose in the absence of a specific reason not to use it for that purpose. 23 In light of these considerations, it is easy to see that TE1) when belief or acceptance are persuasive effects of argument, a tertiary effect of those arguments will frequently be that propositions become context available TE2) non-persuasive effects which deprive propositions or belief or acceptance will deprive them of context availability as well (as for example do the non-persuasive effects listed under E1 and E2 in section 13).
Assertability as a necessary condition of context availability What people are prepared or permitted to say in one transactional context they may not be prepared or permitted to say in another transactional context. And what people are prepared to use in their "private" thinking and reasoning is not limited to what they are prepared or permitted to say in a more or less open forum.
As a result, propositions which are accepted may not be available for use as premises in arguments because no one is prepared or permitted to "put them in play" in open discourse. The forces that filter out "epistemically acceptable" propositions are various. Social disapproval of the expression of certain opinions, reticence about matters considered "personal" or "private," loyalty to friends, family or country, obligations to employers or clients, fear of reprisal -these and a host of similar factors can inhibit what discussants are prepared to say and thereby eliminate or reduce the chance that certain arguments get made. Moreover, these factors can, and in my experience sometimes do, have a profound effect on the course of argumentation in public debate and deliberation and therefore on its outcome.
In addition to forces like those just mentioned, participants are sometimes limited in what they can say by rules and conventions which the participants mutually acknowledge as binding on their communicative interaction. 24 I want to call attention to one such convention or "rule" that participants in a discussion sometimes mutually acknowledge, namely If a party to the discussion has objected to a proposition, then that proposition may not be appealed to as a premiss in an argument until the objection has been dealt with.
When such a convention is in force, it gives rise to several important consequences TE3) an argument for not-p will render p at least temporarily unavailable, even if that argument fails to induce anyone to withhold acceptance of p, until effective rejoinder is made to it. TE4) if those who decline an argument for not-p and continue to accept p are unable or unwilling to mount a rejoinder to that argument (as in Example #3 of section 5 ), p may remain permanently unavailable. TE5) an argument that prevents a possible objection to p from being made -perhaps by rendering one of the premisses of that objection unacceptable or unassertable -can be responsible for securing the context availability of p -even where the secondary effect of that argument is a non-persuasive effect (as in E7 of section 14). 25 The effects of changes in evaluative attitude on context availability Since testimony and the citation of competent authorities are crucial sources of data in communicative contexts, factors which affect the force of such testimony have a crucial impact on context availability. Evaluative attitudes of trust and respect toward those sources are required for their testimony to have force, and accordingly arguments which induce or inhibit trust and respect have an impact on context availability. TE6) alterations in attitudes of trust and respect can render the content of testimony available or unavailable as data in a communicative context.
Trust or loss of trust in a source can be a persuasive effect of argument. But, as we saw, under certain conditions loss of trust or respect can also be a non-persuasive effect of argument (see E1 in section 1313 ).
Performance or non-performance of overt actions as tertiary effects 16 Where the secondary effect of an argument has been to create or inhibit intentions to perform overt actions, and as a consequence of such a secondary effect an overt action is performed or not performed, the performance or non-performance of the action is a tertiary effect of the argument.
It is worth distinguishing between two types of effect that fall within this category: a) the overt action is a speech act and its performance or non-performance is an event that transpires within the communicative context in which the argument occurs. b) either the action is not a speech act, or its performance or non-performance occurs outside the communicative context in which the argument occurs.
Of particular interest are cases of the first type. Here is a partial list of such effects: TE7) inducing someone to point out the defects in an argument that has been presented (E3 and E4). TE8) inducing someone to present a rejoinder to an argument that "shifts the burden of proof" (E5). TE9) preventing someone from advancing a position or forestalling an argument or an objection by making participants to assert the propositions necessary to articulate that position (E7 of section 14). TE10) forestalling an argument or an objection, by making participants unwilling to assert crucial premisses (E7 of section 14).
Typically, arguments that induce rebuttal -as in TE7 and TE8 -are arguments that have not had a normal persuasive effect on the rebutter. And, as we saw above (E6 and E7 in section 14), an unwillingness to continue maintaining a proposition can be a non-persuasive secondary effect of both bracketed and discounted arguments. In short, all four of the tertiary effects just listed can be the consequences of non-persuasive secondary effects.

The import of tertiary effects 17
Consider first the import of changes in context availability. Creating or securing the context availability of propositions for various purposes enables argument and counter-argument to more forward by supplying the premisses on which arguments, objections and counterarguments will hinge. For the most part, context availability is made possible by persuasive secondary effects. But we have seen (TE5 above) that context availability can be secured by nonpersuasive effects that forestall objections.
Rendering propositions unavailable has equally crucial impact on the ebb and flow of argumentation. It can undermine the force of an argument or objection that has been presented, forestall arguments and objections that might be presented in the future (closing off entire courses of development that might otherwise be possible), or reduce the likelihood that certain possible arguments and objections will be seen as sound. As well, it can deprive participants of reasons for trusting and respecting sources of testimony, with the result that the content of that testimony is no longer accepted and therefore unavailable (T6). Quite often propositions are rendered unavailable as a result of arguments that persuade participants to accept or believe their contradictories. But, as we saw in sections 5 and 13 above, withdrawing assent from a proposition, or ceasing to trust a source of testimony, can just as easily be a non-persuasive secondary effect.
Consider next the import of arguments whose tertiary effect is the performance or nonperformance of speech acts in the communicative context in which those arguments occur. Arguments that induce rebuttal (TE7 and TE8) carry the flow of argumentation forward by bringing new considerations to light; arguments that discourage the open adoption of positions (TE9) or inhibit the presentation of arguments and rebuttals (TE10) can profoundly affect the direction in which argumentation unfolds. And as we've seen, all four of these effects can be consequences of non-persuasive secondary effects.
There can be little doubt, I think, that the tertiary effects I've just catalogued have a profound impact on the ebb and flow of argumentation in a communicative contexts, and that very frequently they are achieved as a result of non-persuasive effects.

18.
The point, purpose or use of an argument is the effect which an arguer intends to achieve by presenting the argument. Most often, the point lies in one or more specific tertiary effect that is achieved as a result of some secondary effect -sometimes as the result of a persuasive secondary effect, sometimes of a non-persuasive secondary effect. But there is no reason to suppose that the arguers who aim at those tertiary effects suppose, in every case, that they will achieve their objectives by means of "normal" persuasion. That is to say, there's no reason to suppose that in every case arguers intend to achieve their objectives by inducing others to make the inferences which their arguments invite.
Moreover, arguments are frequently offered simply to call attention to a reason for doing something -for example, a reason that has yet to be canvassed in the context at hand. In such cases there may be no intention or expectation of convincing anyone to do what the reasons presented are reasons for doing. The point of such argument is simply to make a reason for doing something clear or manifest-not to show that there is conclusive reason doing it or to show that it would be unreasonable not to do it. In such cases, the arguer does not aim at anything beyond the primary effect of argument.
In light of these considerations, it should be apparent that persuasion is just one among many possible purposes that argument may have.
Of course, the salience and the importance of non-persuasive effects and objectives will vary from one transactional context to another. Their frequency in and importance to the context in which they occur will depend on the purposes that shape what happens in that context, on the sophistication of participants and on the complexity and richness of the argumentation which ensues.
In a thirty-second television ad promoting a candidate during an election, any arguments presented are likely to be aimed at inducing a belief or attitude that might result in a vote for the candidate. Those who craft such ads have considerable sophistication, but probably don't presuppose much sophistication on the part of their audience, are unlikely to have any objective other than persuasion in mind, and will probably avoid persuasive techniques that involve complex argumentation or argumentation that is rich in texture.
In other contexts -criminal trials, for example, where the principal goals of prosecution and defense are to induce judge or jury to render a particular verdict -persuasion is an overall goal, but the argumentation used often exhibits complexity and richness, and there is scope for tertiary effects achieved through non-persuasive means. 26 Finally, the argumentation which transpires in many deliberative bodies, in debates among researchers, or in the published literature of an academic discipline typically serves more complex purposes, involves participants who are capable of a great deal of sophistication in their reasoning and frequently exhibits considerable complexity and richness.
What transpires in well-conducted communicative interchange of the latter sort is, I think, ill-served by the idea that persuasion is the heart and soul of argument and argumentation. What those interchanges exhibit is the gradual emergence or manifestation of a complex interplay of reasons which grows and changes along three dimensions: first, the number of propositions, or the amount of data, manifestly relevant to the issues at hand grows over time; second, the "quality" of data improves -imprecision, ambiguity, oversimplification in the summarizing of results are all decreased, and questionable data is eliminated from further consideration; finally, the reasonableness of adopting various positions on the issues at hand shifts over time as the costs and benefits of holding those positions comes more clearly into view through the exchange or argument and counter-argument.
The upshot of this process is seldom a meeting of minds, and where a meeting of minds does occur it is frequently the case that what parties come to agree on is not something that any of them had in mind at the start.

Appendices
Appendix A: Key features of communicative contexts I take the key features of a transactional context to be the following. a) Acknowledged purposes. There will usually be one or more purposes which the participants jointly understand their communicative interaction to have. When they exist, such purposes should be considered a key feature of a transactional context. Such purposes need not be shared by all the parties to the exchange, but must be acknowledged by those who willingly participate in it. (For example, in such "one way" contexts as advertisements or pamphlets, the producer of the ad or pamphlet typically has persuasive purposes in mind. Any reader or viewer who understands the nature of an ad or a pamphlet knows that, even though it's rarely a purpose of the reader or viewer to persuade or be persuaded. In such contexts, persuasion is an acknowledged purpose of the transactional context.) For one way of classifying types of "dialogue" by reference to their purposes, see Walton and Krabbe 1995. b) Participant roles. The participants in a transactional context usually have differing roles within that context -the roles frequently being a function of the acknowledged purposes that structure communicative interaction. The roles may impose certain obligations on participants, and almost always structure which sort of messages are appropriate for communicators who have a speaking role. Examples of roles that occur in different contexts are: doctor and patient; teacher and student; confessor and penitent; judge, jury, defense counsel, prosecutor; proponent and respondent (at different phases in an argumentative dialogue); labor, management and mediator (in contract negotiations), and so on. Roles rooted in acknowledged purposes of the context and recognized by the participants should be considered key features of a communicative context. c) Rules and conventions. Typically, the participants recognize formal or informal rules or conventions governing their interaction. Of particular interest are the rules or conventions that further the acknowledged purposes of the communication. There are almost always either rules or informal practices governing who may speak and when, and there are sometimes rules that the filter the content of messages (for example, rules against hearsay evidence in judicial proceedings, or rules that prohibit accusing another member of parliament of lying, etc.). Any such rules or conventions which are recognized by the participants and which further the acknowledged purposes of the communicative content should be considered key features of the communicative context.
Of course, further features of transactional contexts -such as the medium in which communication takes place (face-to-face oral interchange, telephone, print, electronic, handcrafted letter, etc.), the genre to which the messages belong (letter, books, conversational turns, formal speeches, etc.), the immediacy of message transmission (from face-to-face dialogue to the production of the books that comprise the literature on a topic) -are also important to understanding and appraising what is transpiring in a communicative context.

Appendix B: Belief and acceptance
What is acceptance?

1.
The word 'accept' is sometimes used in a sense in which accepting a proposition is the same thing as believing it. As I am using the term 'accept,' however, accepting a propositionbeing prepared to use it as a premiss in my inferences -is not the same as believing it. If I believe a proposition at a time t and do not believe it at a later time t 1, either I have forgotten something or I have changed my mind. But I can use a proposition as a premiss at a time t and refuse to use it at a later time t 1 without forgetting anything and without changing my mind. It is possible to do this -and it is often reasonable to do this -because use of a proposition as a premiss can be appropriate on one occasion but not on another, even though neither the doxastic nor the epistemic status of that proposition has changed.
How can the use of a proposition be appropriate on one occasion but not on another. Here are three examples. a) Suppose I'm quite convinced that cigarette smoking "causes" cancer -and suppose that it is reasonable for me to hold this belief. But suppose I've been asked to review the evidence supporting this belief, and to try to estimate how great a degree of confidence we (or I) should have in the claim that smoking "causes" cancer (or significantly increases the risk of cancer). In reviewing and assessing the evidence I'm going to have to reason from that evidence to a conclusion about the effect of smoking on cancer risk, and in that reasoning I'm going to have to draw on a variety of background of assumptions which it is reasonable for me to make. But I cannot, on pain of begging questions, draw on the proposition that smoking causes cancer -if I did, the claim being assessed would have a degree of credibility of 1, since it is entailed by the data plus one of the background assumptions. That is to say, I should not be prepared to use it as a premiss in this context, even though I am prepared to use it as a premiss in most contexts. b) We are often called upon to make judgments based on evidence, where it is inappropriate to use evidence of certain kinds. Thus members of a jury are supposed to render their verdict based solely on the evidence presented in court: they are barred from tapping any information available to them from other sources. In making decisions about hiring or about accepting students into a university program, it is often inappropriate (and sometimes illegal) to use certain kinds of information -even information that consists of propositions which I know or have good reason to believe are true. For example, in hiring it may well be inappropriate to extrapolate predictions about likely success on the job which are based on race or gender. In university admissions, it may well be inappropriate to base a decision on information that is not in a candidate's official dossier (especially where a referee has extensive personal knowledge about one of the candidates but none of the others). Or again, if I'm reasoning as a physicist or a biologist, it is quite likely inappropriate to permit my religious convictions -no matter how reasonable -in influence the judgment I make qua physicist or qua biologist. 27 c) There are contexts where it is appropriate to use as a premiss a simplifying assumption that I know to be false-e.g., Newton's inverse square law for most calculations concerting gravitational force. But such simplifying assumptions won't be appropriate candidates for premiss in all contexts.

2.
We need to distinguish (i) actually using a proposition as a premiss and (ii) being prepared to use it as a premiss. For example, someone could come to accept a lemma in a proof, but refuse to accept the proof as a whole because she refuses to make the inferences in which that lemma functions as a premiss. This example also makes it clear that one can come to accept something "as a conclusion," and that in communicative contexts reasons are often offered precisely for accepting something as an intermediate conclusion on which further conclusions may be based.
Hence we need to distinguish between believing a proposition and being prepared to use it as a premiss. We also need to distinguish between believing a proposition and accepting it as a conclusion. In short, we need to distinguish between belief and acceptance.

3.
Actually using a proposition as a premiss or accepting it as a conclusion always occurs in a context of reasoning and/or inquiry. This opens up the possibility that I may be prepared to use a proposition as a premiss in one sort of context but not in another -in other words, it opens up he possibility that acceptance, unlike belief, is context-relative.
And indeed, acceptance of a proposition is or at least can be context-relative, because it can be conditioned by and contingent upon the following sorts of factor (i) the purposes for which we're reasoning in a context, (ii) the practical implications of reaching a false conclusion in that context, (iii) the cost of failing to reach a conclusion in that context, and (iv) the relative costs of various ways of reaching a conclusion in that context.
Moreover, whether it is reasonable for an individual to accept a proposition on a given occasion depends on contextual factors because it is sensitive these four sorts of factor. For examples, see Pinto 2003b.
Sometimes an individual accepts a proposition at a given time and in a certain context because he has decided to do so in light of factors like (i)-(iv). But people can and usually probably don't accept propositions as a basis for reasoning as the result of any actual or explicit consideration of such factors -their acceptance is rooted in an unreflective reliance on belief. That is to say, what usually explains the fact that they are disposed to use a given proposition as a premiss is simply the fact that they believe it to be true -and their use of that proposition at that time is not mediated by any actual consideration of factors like (i)-(iv). In Cohen's account, acceptance is voluntary, whereas belief is not. That idea is consistent with the idea that acceptance is often rooted in an unreflective reliance on belief. To see that this is so, consider the following. Someone like Cohen can maintain that, even when an individual unreflectively accepts what he happens to believe, the individual always "could have done otherwise" -always could have stopped and reflected on whether use of this proposition in his reasoning would be appropriate in the context at hand in light of factors like (i)-(iv). Therefore the individual could have decided not to use it as a premise. In other words, Cohen can maintain that one is always in a position to decide what one uses (or is prepared to use) as a premise, even though one is not (as Cohen maintains) in a position to decide what to believe. We can say, then, that even when a person unreflectively accepts what he happens to believe, there is an "implicit" decision to accept it.
Though there need be nothing wrong with accepting a proposition on the basis of an unreflective reliance on belief, a thorough appraisal of the reasonableness of accepting a proposition would have to attend to factors of types (i)-(iv). Here's a way of conceiving the situation: Appraising the reasonableness of an inference requires appraising the reasonableness of the explicit or implicit decision to accept the propositions which function as premisses in that inference (see Pinto 2003b, which approaches the matter in this way). At the beginning of his account of acceptance, Cohen says (1992, 4) …to accept that p is to adopt a policy of deeming, positing or postulating that p -i.e., of including that proposition or rule among the premisses for deciding what to do or to think in a particular context, whether or not one feels it true that p.' Clearly, then, Cohen recognizes the context relativity of acceptance -though at a later point (p. 16) he says …acceptance, unlike opinion, may be relative to context, although it is certainly possible to treat some premisses as being acceptable in any context whatsoever [italics mine].
But to my mind Cohen fails to highlight or develop this most interesting and most import feature of acceptance as a basis for and as a target of our reasons and reasoning -the fact that it is context-relative in ways that belief is not. His focus is on other issues -for example, the voluntariness of acceptance versus the involuntariness of belief (about which I think he is right), and what he takes to be a difference in the way deductive closure applies to belief and to acceptance (about which I think he is wrong).

3.
Cohen's account of belief won't do. It is unilluminating to base one's account of belief, s he does, on the idea that someone believes a proposition if and only if he or she is disposed to "feel" it is true. For to say "I feel that p is true" is just another way of saying "I believe that p is true." Moreover, I believe that Cohen is mistaken in supposing that "feeling that p is true" has something essentially to do with "feelings" in anything like the sense in which pleasure and pain are feelings. Admittedly, he distinguishes "credal feelings" from emotional feelings (1992, p. 9), but I don't think this helps very much; moreover, the concept of credal feelings (which he appears to apply to a range of feelings) is mentioned only in passing and is not well-developed in Cohen 1992.

4.
Harman (1986,47) equates "belief in a proposition P" with "full acceptance of P"something which he says has two features: one allows oneself to use P as part of one's starting point in further theoretical and practical thinking [one] takes the issue to be closed in the sense that one is no longer investigating whether P is true.
He contrasts belief or "full acceptance" of P with tentatively accepting P "as a working hypothesis," which has the first of these features but not the second (p. 47).
He says (p. 46) that to accept something as a working hypothesis is to "try it out," to see where one gets by accepting it, to see what further things such acceptance leads to. Accepting a particular working hypothesis is fruitful if it allows one to make sense of various phenomena; if it leads to solutions to problems, particularly where there are independent checks on these solutions; and if it leads naturally to other similarly fruitful hypotheses.
This account does not recognize or make provision for two points I consider of utmost importance: a) that even when one "takes the issue [of the truth of P] to be closed," whether or not one is prepared to use P as a "starting point" in further theoretical and practical inquiries often is, and always should be, contingent on particulars of the context of the inquiries in which P might be used b) that accepting something as a working hypothesis in order to "try it out" is just one of many kinds of context in which people can and do accept propositions which they do not "fully believe" as starting points for inquiry. Moreover, as I will try to show in a moment, this kind of context is not the most interesting kind of context of this sort.
Point (a) is illustrated by the three examples cited in section 8 above. With regard to point (b), tentative acceptance of a working hypothesis is, in one respect, among the least interesting kinds of reasoning from propositions we do not "fully believe." For in this kind or case -as in suppositional reasoning -I do not "embrace" the conclusions derived from tentative (or suppositional) starting points. My "conclusions" will have the tentative or suppositional status of the "starting points" from which they've been derived (at least until they are independently confirmed). But when I reason from a simplifying assumption or from something that I consider probable but don't actually believe, the status my "conclusion" has a "byte" that a conclusion drawn from a mere supposition does not have. When I reason from the inverse square law (which I don't believe) to a conclusion about the gravitational force exerted by the earth on the moon at a certain time (for instance, that it is n), I will may well come to believe that the force at that time is, for all practical purposes, n. And even if I don't quite believe it, I will make design decisions based on it -something a wise person will not do with a conclusion deduced from a mere supposition. When a judge reasons from the assumption that plaintiff was injured by respondent -not because she sure he was, but because that proposition is supported by the "balance of probabilities" -to the further conclusion that the plaintiff is entitled to damages in a certain amount, she will issue a judgment to that effect, and force the respondent to comply with it. (And, interestingly enough, the judge will probably believe that plaintiff is entitled to damages in that amount.) For these reasons, I don't think Harman's account of "full belief" can shed much light on what, in the final analysis, the crucial differences between belief and acceptance are.

5.
Despite this, something that Harman says about tentative acceptance applies, I think, to most kinds of "acceptance without belief." What he says may help to throw light on the difference between belief and acceptance.
Harman writes (p. 47) Tentative acceptance is not easy. It takes a certain amount of sophistication and practice to be able to investigate an issue by tentatively accepting various hypotheses. Ordinary people, and even most scientists, are quick to convert tentative acceptance into full acceptance in a way that seems overly hasty to critical reflection....
I am inclined to think that, like tentative acceptance, the sorts of "acceptance without belief" I've described also require "a certain amount of sophistication and practice." Harman explains why tentative acceptance is difficult in the following terms (p. 49) The same practical limits that keep one from operating in accordance with the foundations theory [i.e., revising one's beliefs whenever the evidence on which they were originally based is undercut] also keep one from never ending inquiry and always tentatively accepting one's conclusions as mere working hypotheses.
In the midst of inquiry one must keep track of which tentatively things depend on others, so that ongoing revision can be neatly accommodated. One needs to remember what reasons there are for and against various possible outcomes of inquiry, where this can involve also the reasons for various possible intermediate conclusions.
And most interestingly of all he adds (p. 50) If one had unlimited powers of record keeping and an unlimited ability to survey reasons and arguments, replies, rebuttals, and so on, it would be rational always to accept things tentatively as working hypotheses, never ending inquiry. But since one does not have such unlimited powers of record keeping and has a quite limited ability to survey reasons and arguments, one is forced to limit the amount of inquiry in which one is engaged and one must fully accept most of the conclusions one accepts, thereby ending inquiry. Tentative acceptance must remain a special case of acceptance. It cannot be the general rule. [Italics mine.] I am inclined to generalize this and say that for somewhat similar reasons acceptance without belief must remain a special case of acceptance; it cannot be the general rule. But the emergence of such sophisticated forms of acceptance drives a wedge between the concepts of belief and acceptance. 28 If I may put this somewhat differently, to reason on the basis of propositions we don't fully believe -or on the basis of propositions that we know to be false -requires us at least temporarily to adopt a sophisticated mindset, a mindset which may sometimes be difficult to maintain for very long. Though such a mindset makes our cognitive lives more complicated and more demanding in some respects, it promises considerable payoffs when we are forced -as we always are -to operate under conditions of uncertainty with limited cognitive resources and limited time.
It is essential to our concept of belief that it is a factor in what Dennett has called the intentional explanation of action, and it is also essential to that concept that in its role as mediating action it can be the starting point and the end-point of reasoning (or the input and output of "information processing"). 29 But the criteria for identifying the content of beliefcriteria which anchor that concept within the network of our practices -are intimately tied to the idea of assertive utterances which are "sincere," serious and successful, 30 criteria that are pretty much independent of the role of belief in intentional explanations of action and of the role of belief as a starting point and end-point of our reasonings.
As I see it, the emergence of those sophisticated mindsets that I call "acceptance without belief" means that what is picked out by the criteria for belief (the content of sincere, serious and successful assertive utterance) no longer coincides with the input and output of the reasoning that mediates action. Hence the need to disentangle the concept of belief from the concept of acceptance, even though -in most contexts most of the time -what we believe coincides with what we accept.

6.
A number of points that flow from what I just said need to be sketched here: a) In communicative contexts, reasons offered in support of a "conclusion" are readily conceived of as reasons for accepting that conclusion as a basis for action or further reasoning in the context at hand. b) If such reasons for acceptance become effective as reasons for belief, it is because acceptance in the context at hand "spills over" into a person's belief system. The mechanisms that account for such "spilling over" -that account for a person's becoming prepared to use a proposition in other, somewhat different contexts -are not in my opinion well understood. Empirical research on persuasion effects may already shed some light on this phenomenon, and may in the future shed more light.
It strikes me as much easier to evaluate reasons when we approach them as reasons for accepting a proposition in the context at hand, because contextual factors of the four types described earlier usually render quite determinate the standards that are appropriate for such evaluation. Where the context of discussion or reasoning is indeterminate, standards for appraisal will be very hard to pin down.
Notes 1 This paper grew out of the idea that was planted in my mind by Jean Goodwin in a conversation I had with her in June of 2002 -the idea that arguments are often used for purposes other than persuasion. I've since learned that Goodwin had presented something very much like that idea in Goodwin 1999. 2 Several dictionaries list "convince" or "convincing somebody of something" as one of the meanings of 'persuade' -see the entries for 'persuade' in Encarta and in Random House.
However, there is a view, sometimes attributed to Campbell and to Whately, according to which persuading should be contrasted with convincing on the grounds that the former is based on an appeal to emotion or will and the latter on an appeal to reason. That view will play no part in what I say.
See also the discussion of persuading and convincing in van Eemeren & Grootendorst (1984, p. 48), which contains a faint echo of the idea that persuading, but not convincing, involves an appeal to "the will." However, these authors say that for their purposes "[the] difference between persuading and convincing is not important." It is tempting to suspect that the attempt to make a sharp distinction between persuading and convincing may be connected with a point about usage that is no longer in favor. Some have held that while persuade can be followed by 'that,' 'of,' and 'to,' convince can only be followed by 'that' or 'of.' See the usage notes on 'convince' in Random House and in Encarta, both of which reject this grammatical point. 3 In Pinto 1995(reprinted in Pinto 2001. 32) I defined argument as "a set of statements or propositions that one person offers to another in the attempt to induce that other person to accept some conclusion." 4 For purposes of this paper, conative attitudes will be limited to desires and intentions. 5 O'Keefe's view is "that a paradigm case of making an argument-1 involves the communication of both (1) a linguistically explicable claim [he's thinking of the conclusion of the argument-1] and (2) one or more overtly expressed reasons which are linguistically explicit." 6 Offering reasons can be a matter of giving an explanation as contrasted with presenting an argument. In general, when I offer reasons for doing something as part of an argument, the reasons that I offer are reasons for the person whom I am addressing to do what the reasons are reasons for doing. In explanations, that is rarely if ever the case. For example, if I give you reasons for me to do something, I am typically explaining to you what I did or am about to do that thing. But if I give you reasons for you to do something, I am typically presenting you with an argument for doing it. In what follows, I will confine my self to cases in which reasons offered are reasons for the addressee to do something.
It is perhaps worth noting that even explanatory reasons can play a crucial role in carrying a discursive interchange forward to a successful conclusion. Consider what happens in negotiations, when A states her reasons for rejecting an offer that B has made. A is offering an explanation rather than an argument, but presenting the reasons for rejection is crucial to the process of negotiation (and in some jurisdictions failure to provide them will be considered evidence of bad faith bargaining), since they make it possible for negotiators to move toward an agreement. Only by knowing A's reasons for rejecting earlier offers can B know how to make subsequent offers which are more likely to be accepted. (Moreover, if even A's explanation is false -doesn't give the "real" reason why the offer was rejected -it has an impact on the future course of negotiations, since it tends to limit the reasons A can give in response to future offers that B makes.) 7 I will consider a context communicative only if the "sender" is not identical with the "audience" -something I jot down to remind myself to do something, a "note to self," or a diary intended for the writer's eyes only will not count here as communicative contexts. 8 Which may be may be a "remark" that someone makes in a conversation, or a speech, or a written communication such as a letter, book, article or academic paper, and so on. 9 The possibility of non-linguistic responses -applause, booing, throwing tomatoes -will not, for my purposes, render a transactional context interactive. And only permitted linguistic responses will render a context interactive. For example, in some contexts (e.g., a university lecture) heckling is "against the rules" and its occurrence doesn't render the context interactive. In other cases, heckling is tolerated or encouraged, and its occurrence would render a context interactive in such cases. The distinction between interactive and non-interactive transactional contexts is not terribly sharp, but it is no less useful on account of that. 10 A useful account of necessary conditions for "paradigm cases" of persuasion -with which I largely agree -can be found in O'Keefe 2002 (1-6). In O'Keefe's account (p. 4), "paradigm cases of persuasion are those in which the effects are achieved through communication (and perhaps especially through the medium of language)." 11 Someone might try to argue that the very idea that there are non-persuasive effects induced by the presentation of an argument involves an oxymoron. It might be maintained, for example, that "by definition" inducing people to do something by means of an argument just is persuading them to do it. 12 This is intended as a first pass at explaining what a non-persuasive effect is. Later on I will severely restrict the class of "secondary effects" to which the labels 'persuasive' and 'nonpersuasive' can apply. See section 10. 13 On this reading, Smith has given Jones a reason not to accept that it's raining, and the result is that Jones does not accept that it's raining for, and in response to, that very reason. 14 I am assuming that an argument has merit if its premises are acceptable and the premises call for acceptance of the conclusion in the absence of countervailing evidence. Jones can concede that Smith's premises are acceptable and call for his conclusion, but refuse to draw the conclusion on the basis of countervailing evidence. 15 As for example when an argument for accepting p leads me to withdraw my acceptance of notp 16 It is perhaps worth pointing out that acceptance is not a doxastic attitude since it is neither full belief nor an alternative to full belief. One can both believe a proposition and also accept it, whereas doxastic attitudes other than belief are defined as attitudes toward a proposition that we hold when we do not fully believe the proposition to be true. 17 It was Jean Goodwin who pointed out to me the crucial importance of getting a receiver to pay attention to an argument one offers. See Kauffeld 1998 (especially the section on "proposing") and Goodwin 2002 for a remarks about some of the ways in which speakers provide hearers with incentives to listen to what speakers have to say. 18 Peter Loptson is right when he points out in his reply to this paper that what I imply in this sentence is incorrect: understanding an argument is, as Loptson claims, not something that lies directly within my power to do or not to do. 19 In the empirical literature on persuasion, there has been extensive discussion of how receivers "process" messages that contain arguments, and of the factors that influence how they process them -see O'Keefe 1996 and O'Keefe 2002, Chapter 6, for a summary and discussion of some of that literature. Despite controversy over how such processing is best understood, it is apparent that receivers accept or decline to accept arguments on the basis of some kind of assessment of the argument and the issues surrounding it. 20 An argument that is completely ignored is neither accepted nor declined. 21 Considerations that undermine aren't, as such, reasons for not doing X. Suppose, for example, that someone has offered the fact that Mary is a Canadian citizen as a reason for believing that Mary is able to speak either French or English. The additional information that Mary is only 12 months old undermines that reason (since quite a few 13-month-olds don't speak any language yet). But it isn't a reason for refusing to believe that Mary speaks French or English, since (a) quite a few 13-month-olds have begun to speak and (b) the fact that this reason doesn't "float" does not preclude the possibility that the addressee has other reasons for believing the proposition in question. For more on the distinction between undermining and overriding, see Pinto 2201, 14, 28 and 102-103. 22 There are two varieties of rebuttal: (i) showing that an argument is defective or (ii) showing that it is undermined or overridden by additional information. See the distinction between "Type I" and "Type II" criticisms in Pinto 2001, 103. 23 Reasons for not using propositions that we believe are easy enough to understand (see Pinto 2003b and Appendix B for some examples.) So are reasons for accepting propositions we don't believe (see Pinto 2003a and2003b for examples). 24 For example, in criminal trials there are proscriptions which prevent evidence of certain sorts from being presented -thus under most circumstances evidence of prior bad acts is inadmissible, as in general is the presentation of any evidence which in the opinion of a judge would be unduly prejudicial. It is clear, I think, that in many cases precluding such evidence -and therefore eliminating the possibility of lines of argument which might be based on it-has a decisive effect on the outcome of a trial. 25 For example, A offers an argument for accepting a proposition q, several other participants (B, C and D) who are not persuaded by that argument to accept q nevertheless don't know how to rebut the argument and consequently forego any attempt to object to q. A's argument prevents B, C and D from objecting to p even though that argument hasn't persuaded them of anything. 26 For example, a defense attorney who rests his case on "reasonable doubt" may be able to raise reasonable doubt by means of arguments whose conclusions are not accepted by the jury, but are effective because the reasons they embody counterbalance the reasons available to support a guilty verdict.
27 See Harman on acceptance as a member of a group" (1984,(50)(51)(52)(53)(54)(55)(56)(57)(58)(59)(60)(61)(62). 28 On the other hand, if we should come to agree that Cohen (1992, section 16: pp. 86-100) is right that the proper attitude toward scientific theories is usually one of acceptance without belief, the "special case" will not be all that uncommon! 29 As Fodor, and other like-minded functionalists would have it. 30 That is to say, our best evidence for what a person believes is usually what that person assertively utters -indeed, a fine-grained account of the content of people's beliefs is all but impossible except on the basis of an examination of their assertive utterances. But to use utterance as a clue to belief, we must learn to identify and discount cases of prevarication (or insincere speech), cases of "unserious" or playful speech (joking, irony) and cases in which someone fails to say what he or she intends (unsuccessful speech). The use of our concept of belief is anchored in the network of practices that enable us to identify insincerity, playfulness and misspeaking, and is impossible in abstraction from those practices. I developed these ideas in a paper entitled "Believing" that was delivered at the annual meeting of the Canadian Philosophical Association in 1982, but which has never been published.