Two Views of Speech Acts: Analysis and Implications for Argumentation Theory

: Argumentation theorists need to command a clear view of the sources of the obligations that arguers incur, e.g., their burdens of proof. Theories of illocutionary speech acts promise to ﬁll this need. This essay contrasts two views of illocutionary acts: one, that they are constituted by rules, the other, that they are constituted by paradigmatic practical calculations. After a general comparison of the two views, the strength of the pragmatic view is demonstrated through an account of the illocutionary act of making an accusation. It is shown that the essential conditions of ACCUSING revealed by conceptual analysis are just what is practically necessary to manage a routine, but complex, communicative problem. The essay closes with remarks on the implications of the pragmatic view of speech acts for argumentation theory generally.


Introduction
[This essay was left incomplete at the untimely death of the lead author in 2017. I have edited the manuscript for clarity and length; any extended interventions appear within square brackets-J. G.] Argumentation theorists need to command a clear view of the sources of the responsibilities and obligations arguers incur. These include the arguer's responsibilities for the truthfulness of what she says, for the relevance of argumentation, and for such probative or dialectical obligations as she may undertake, etc. Such obligations determine important norms related to the quality of the arguments which may be demanded of an advocate, they fix limits for what may be demanded of an arguer, they play critically important roles in determining the pivotal issues in much argumentation, and they are related to the persuasive forces that at least some kinds of argumentation may acquire. Indeed, if Ralph Johnson (2000) is correct, understanding arguers' dialectical obligations is required to comprehend the rationality of argumentation.
Arguing is a communicative activity, even when one argues with oneself, and some of the critically important obligations arguers incur are engaged in connection with the communicative acts-the speech acts-performed in the course of argumentation. It is apparent that obligations are engaged in some speech acts. PROMISING is frequently analyzed as an example of an act in which the promisor generates an obligation to do what she says she will do. However, the phenomena of commitment, in which a speaker undertakes a responsibility or an obligation, can be observed across a wide range of speech acts. In seriously saying and meaning something, a speaker commits herself to the truthfulness of what she says. Additionally, in a wide range of speech acts performed in and by seriously saying something-acts belonging to the class J. L. Austin called illocutionary acts-speakers incur various argumentatively significant obligations. This makes a subset of illocutionary acts of paramount interest to students of argumentation-a fact recognized

Preliminary Comparison of the Two Views
Regarded in fairly general terms, the rule-constituted view and the pragmatic view of illocutionary acts share certain important similarities and are, at the same time, marked by several very deep differences. It will be useful to begin with a broad comparison of the two views.
First, both views start with J. L. Austin's seminal identification of the illocutionary act. Austin (1962), it will be recalled, identified three broad classes of what he calls speech acts: locutionary acts or acts of saying things, illocutionary acts or acts performed in saying things, and perlocutionary acts or acts performed by saying things. The major discovery in this taxonomy turns out to be the illocutionary act. According to Austin, in performing an illocutionary act, a speaker says a few appropriate words, e.g., "I promise to be home by seven", speaking in conformity with the appropriate conventions and therein producing a promise, which has the illocutionary force of committing her to do what she has said she will do. Subsequent work shows Austin's conception of illocutionary acts to be deeply flawed, and both of the approaches in question amend it. However, both take Austin's work as their starting point.
In this connection, we should notice that the rule-constituted view stays closer to Austin's original conception than does its pragmatically oriented cousin. Austin's work on speech acts started with work on what he called performatives, i.e., acts on the order of CHRISTENING, MARRYING, etc. Acts of this kind are plainly constituted by conventions. In many Western societies, persons marry each other by following a conventional rule which requires that each says "I do" in response to a question along the lines of "do you take so and so to be your lawfully wedded spouse?". Answering "I do" in the appropriate circumstance dramatically alters the relationship between these parties. When Austin turned his attention to the broader array of speech acts that includes PROMISING, APOLOGIZING, COMMANDING, ADVISING, PROPOSING, and ACCUSING, etc., he thought of these illocutionary acts as being similar to the performatives with which his inquiry had started, and so he conceived of illocutionary acts in general as acts constituted by conventions.
When Searle (1969) initiated a research program elaborating Austin's work, he (and others) retained the idea that illocutionary acts are constituted by rules, and he included Austin's performatives within the broad category of the illocutionary act. Searle does not retain the idea that illocutionary acts are strictly conventional acts (although he is not very clear about the status of the rules that he alleges constitute illocutionary acts). Searle, it will be recalled, works with a distinction between rules that govern acts, e.g., rules that regulate parking cars in municipal lots, and rules that constitute acts, e.g., the rules for playing football or chess. The former "regulate antecedently or independently existing forms of behavior", whereas constitutive rules create or define new forms of behavior" (p. 33). In Searle's view, the illocutionary act of PROMISING, to use the only example of an illocutionary act that he explores in depth, is constituted by rules closely analogous to the rules that constitute games. He holds, moreover, that the rules constituting a kind of illocutionary act are encapsulated in semantics of the corresponding illocutionary verb; thus, the rules that constitute PROMISING inhere in the semantics of the verb promise (this latter point has been much criticized and so will claim little of our attention here; see Stampe 1975).
In contrast, the view that illocutionary acts are constituted by underlying practical calculations grants that Austin's performatives are constituted by conventions but denies that the larger array of illocutionary acts is either constituted by conventions or for the most part by rules of any sort. Strawson initially articulated this view. In his famous essay on "Intention and Convention in Speech Acts", he maintains: It seems . . . clear that . . . there are many cases in which it is not as conforming to an accepted convention of any kind (other than those linguistic conventions which help to fix the meaning of the utterance) that an illocutionary act is performed. It seems clear, that is, that there are many cases in which the illocutionary force of an utterance, though not exhausted by its meaning, is not owed to any conventions other than those which help to give it its meaning. Surely there may be cases in which to utter the words "The ice over there is very thin" to a skater is to issue a warning (is to say something with the force of a warning) without its being the case that there is any statable convention at all (other than those which bear on the nature of the locutionary act) such that the speaker's act can be said to be an act done as conforming to that convention (Strawson 1964, pp. 443-44).
As Strawson includes within the scope of the term "convention" all social and institutional rules which constitute games and institutionalized procedures, his argument involves a rejection of the claim advanced by Searle that illocutionary acts, such as games, are constituted by rules. This position has been elaborated in an array of important essays in the philosophy of language, to which we will recur over the course of the present discussion. These subsequent developments regard central kinds of illocutionary acts as constituted not by rules, but by paradigmatic practical calculations about what can be done by the speaker in conjunction with saying and meaning things.
A second broad area for comparison lies in the use that each view makes of H. P. Grice's analysis of utterance-meaning. Shortly after the publication of Austin's discovery of the illocutionary act, H. P. Grice offered an analysis of what is essential to the production of utterances that have meaning in the sense associated with the symbols, gestures, etc., used in communication. This analysis of utterance-meaning had an immediate impact on studies of illocutionary acts. Without taking up all the details, important as they are, the connection between Austin's work and Grice's work is easy to grasp. Although Grice's conception of utterance-meaning is designed to embrace a broad array of gestures, signals, and figures, etc., and is not limited to utterances using linguistic structures, still, it is fair to say that Grice's work on utterance-meaning offers the skeleton of the primary communicative act of seriously saying and meaning something-the act to which we ordinarily refer in so-called indirect speech reports of the form speaker (S) says that p, where p is what S means by her utterance. One would expect, then, that Grice's analysis of utterance-meaning would illuminate at least part of the structure of acts performed in and by saying something, i.e., Austin's illocutionary acts. Since Grice's analysis, though not without its difficulties, is quite robust (Avramides 1989), both the rule-constituted and the pragmatic views of illocutionary acts have tried to incorporate Gricean insights into their accounts, and both have also introduced major modifications into Grice's initial analysis, albeit their amendments are strikingly different.
According to Grice's (1969) initial analysis of utterance-meaning, it will be true that a speaker (S) has said that p, only if S produced an utterance (u) with the following complex intention (or at least acting as if she were speaking with the following): S's primary speaker-intention (I 1 ): S intends 1 that A respond (r) that p (or at least acts as if S intends 1 that A r that p); S's second speaker-intention (I 2 ): S intends 2 that A recognize S's primary speakerintention (or at least acts as if S intends 2 that A recognize I 1 ); S's third speaker-intention (I 3 ): S intends 3 that A's recognition of I 1 provide A with at least part of A's reason for ring that p.
According to this analysis, Mary will have said that Uncle Bill is ill, only if she produced an utterance semantically equivalent to "Uncle Bill is ill," and if she deliberately (I 2 ) gave her addressee to believe that she is speaking with the intention (I 1 ) that her addressee believe that Uncle Bill is ill, and if she at least acts as if she intends (I 3 ) that her addressee's recognition of her primary speaker-intention provide her addressee with a reason for believing Uncle Bill is, indeed, ill.
Grice's initial analysis provides important insight into the primary communicative act of seriously saying and meaning something, but subsequent considerations show that it is not quite complex enough. Counterexamples generated by Strawson, Stampe, and others indicate that S not only deliberately gives her addressee to believe that she is speaking with her primary (I 1 ) intention but that she also openly intends (new I 3 ) that her addressee recognizes her secondary speaker-intention (I 2 ), and this complex recognition of S's intentions is to provide A with a reason for responding as S primarily intends.
To return to the two views of speech acts, Searle (1969) elaborates his rule-constituted view in terms of Grice's analysis of seriously saying things. Searle maintains, however, that Grice errs in two important respects. First, he holds that Grice's analysis "does not show the connection between one's meaning something by what one says, and what that which one says actually means in the language" (p. 43). Second, Searle holds that Grice mistakenly takes perlocutionary effects as the primary response aimed at in a speaker's communicative efforts, whereas a proper analysis would recognize that a speaker's illocutionary efforts can only directly achieve such communicative outcomes as involve an understanding of what the speaker is saying and the intentions with which she is speaking. Other ends such as inducing belief, alerting persons to danger, and securing sympathetic assistances-aims which do animate speakers-lie, according to Searle, outside the scope of the speaker's immediate illocutionary act.
By contrast, the pragmatic view of speech acts builds on Grice. Grice's analysis raises the troublesome question of how A's recognition of S's complex speaker intentions could provide A with a reason for responding as intended (MacKay 1972). Dennis Stampe has made an important contribution by providing a highly plausible and well-defended account of how Gricean reflexive speaker intentions do their work (Stampe 1967(Stampe , 1975). Stampe's account turns on the truism that a person is responsible for her intentional acts. Accordingly, when S deliberately and openly gives A to believe that she is speaking with the primary intention (I 1 ) of inducing A to, e.g., believe that p, she openly takes responsibility for her primary communicative effort. By thus openly accepting responsibility for her primary communicative effort, Stampe argues, S warrants a presumption of veracity on behalf of what she says. Her addressee is entitled to reason that S has manifestly put herself in a position from which she cannot subsequently deny responsibility for trying to get A to believe that p, and this is something, A may reasonably suppose, S would not do, as a prudent and reasonable agent, were she not advancing a proposition that she sincerely believes, on the basis of some acceptable effort to ascertain its truth and rational adequacy.
Stampe's account of how Gricean speaker intentions work opens important possibilities for the pragmatic explication of illocutionary acts. Strawson suggests that some of the differences among illocutionary acts may arise from complexity in how the speaker hold her primary intention, and to this important idea we may add the possibility of complexity in the speaker's primary intention itself. So, in the case of simple belief-inducing utterances, as might be the case with stating something, S does little more than say that p, thereby generating a presumption of veracity on behalf of what she says. However, in the complicated case of warning, we might plausibly conjecture that S would deliberately and openly speak with the primary speaker-intention of alerting A to some impending danger, and in this case, the content of the presumption that she is speaking truthfully would relate specifically to her self-imposed responsibility for causing A to be alarmed.
It is this possibility of mapping a larger pattern of speaker responsibilities and obligations, including probative obligations, that warrants special interest in the pragmatic view of illocutionary acts on the part of students of argumentation. Unlike rule-constituted views, this account does not invite sheerly taxonomic speculation. Instead, it is a conjecture that can be best be explored by the careful explication of specific kinds of illocutionary acts.

The Pragmatic Constitution of ACCUSING
Our attention now turns to explicating the illocutionary act of making an accusation.
[This act is noteworthy for argumentation theory in that it is a clear case of an illocutionary act that creates significant probative responsibilities, i.e., the accused's responsibility to defend his conduct, and the accuser's burden of proof.] Although a pragmatic view hopes to penetrate to the practical calculations that underlie and constitute our concepts of speech acts of this kind, our inquiry starts with an analysis of the concept itself. Our analysis focuses on the semantics of the verb "accuse" and its cognates. By starting (in Section 3.1) with a strictly conceptual analysis of ACCUSING, we may (given several substantive assumptions) arrive at a picture of the components of that act, which are practically necessary and sufficient for potentially efficacious, paradigmatic performances of that speech act in favorable circumstances. From that picture, we can adduce (Section 3.2) a plausible conjecture about the practical constitution of ACCUSING and eventually (Section 3.3) explain in practical terms the necessity for those conditions, which are essential to our conception of this illocutionary act.

Analysis of the Ordinary Act of ACCUSING
To unearth the pragmatics of the paradigm act of ACCUSING, which undergirds and constitutes our concept of this kind of illocutionary act, our investigation begins with an analysis of our ordinary conception of this speech act. This analysis enables us to determine what is conceptually necessary for a speech act to qualify as an illocutionary act of the kind we call "accusing". On the substantive assumptions that we do commonly in truth identify appropriate speech acts as accusations and that, as well-established practices, those speech acts are coherent and potentially successful, identifying the truth conditions for our concept of ACCUSING will at the same time identify the means, which are practically necessary and (ceteris paribus) sufficient for the performance of this kind of act in paradigm cases.
Two conditions are conceptually necessary and sufficient to make an accusation: (1) the accusation-making statement with the accuser's implied negative evaluation; (2) the intention with which the accuser openly speaks.

The Accusation-Making Statement and the Accuser's Implied Negative Evaluation
If it cannot be reported in truth that the speaker said that the accused did thus and such, the speaker will not have made an accusation. Were Smith to say "I accuse Jones of breaking Aunt Maude's lamp; mind you, I am not saying that he broke it" in an apparent effort to speak seriously and literally, we would be inclined to suppose that she did not know the meaning of the word "accuse". Correspondingly, a person cannot be accused of doing one thing by a speaker saying that he did some related thing. Suppose one member of a band of conspirators says that another member has been talking to a policeman. If saying this makes an accusation, then the second has been accused of talking to a cop. He has not, however, been accused of betraying the conspiracy; to make the latter accusation, the first must say that the second gave their cause away.
The accuser must also imply that she believes what the accused did is wrongful, blameworthy, or reprehensible in some sense. That implication cannot be cancelled and an accusation still have been made. Mary cannot accuse Jack of damaging her good name by saying "your jokes are ruining my reputation, though I see nothing wrong with that", despite the second clause being a perfectly candid expression of her beliefs. She could say this and mean exactly what she has said-she might be musing about the state of her relationship with Jack. However, clearly, she cannot say this and thereby accuse Jack of ruining her reputation. Likewise, the description of what the accused is alleged to have done, as set out in the accusation-making statement, cannot be readily modified in ways that vitiate an implied adverse evaluation of the accused's behavior. A person cannot be accused of justifiably or rightly doing something. The accuser's evaluation of the accused need not be expressed by what the speaker says in making her accusation. That is, what the accused is alleged to have done may, but need not, be described as wrongful or unjustifiably harmful in the accusation-making statement.
Related to these differences in the mode of expression for an accuser's beliefs, there is an important difference in the way she may be presumed to hold them. In making an accusation, a speaker purports, not merely to believe, but to believe with some certaintyeven to know-that the accused did the deed alleged. The accusation-making statement must be categorical. An accusation cannot be made by saying, e.g., that Jones probably broke Aunt Maude's lamp. On the other hand, the accuser does not have to purport to be certain in her evaluation of the accused. She may be, but there is nothing odd about saying, e.g., "I accuse Jones of breaking the lamp; I want to know if there is any justification for the harm he caused".
While a statement describing what someone did and an implied evaluation of that conduct are necessary, they do not suffice to make an accusation. Speaking quite seriously and meaning just what she says, Smith might protest "Jones broke Aunt Maude's lamp, and I don't think that there is any justification for what he did. However, I am not accusing him of breaking that lamp. I am not going to argue about it, and I am not interested in what he has to say. I just want him to know that his actions did not go undetected". [Thus, it is necessary to add an additional condition to the analysis: the intention with which the accuser speaks.] 3.1.2. The Intention with which the Accuser Speaks A natural and theoretically attractive view would be that the intention with which an accuser says, e.g., that Jones broke the lamp, is the intention that the addressee believe that Jones did break the thing. However, this intention cannot be necessary to making an accusation. Where accusations are addressed directly to the accused, it must be assumed that the accuser believes that her addressee already knows what the accuser is telling him.
With what intention, then, does the accuser necessarily speak? She speaks with, or as if with, the intention of securing certain communicative responses from the accused. Specifically, the accuser wants, or at least purports to want, to secure an answer from the accused-whether an admission that the accused did what he is alleged to have done, an attempt to justify or to excuse what he did, or an attempt to deny the allegation. This intention resembles the purposes speakers have in asking questions and raising objections. Questions are asked and objections are raised in order to get someone to say something, i.e., to obtain an answer. Likewise, when an accusation is made, the speaker at least purports to want an answer from the accused.
Like questions and objections, accusations can be pressed. Except in cases where questions are used figuratively or abusively, it would be odd to ask a question and not stay around for the answer. While not automatically irrational, one would be forced to look for some special explanation to make sense out of the speaker's apparently contradictory conduct. Likewise, the behavior of someone who makes an accusation and does not stay around for an answer is mildly odd; it seems that she is first acting with one intention and then with a contrary plan in mind. The fact that in such cases an explanation for apparently odd behavior must be found indicates that at least acting as if an answering response were sought is an essential part of making an accusation.
Of course, it is not necessary that an answer be sought as a direct and immediate conversational sequel to the accusation. Nor need the accuser necessarily purport to want the answer herself; she might want some third party to try to secure the answer, perhaps hoping the accused will be eventually punished.
The (purported) intention of securing an answer from the accused distinguishes ACCUSING from other speech acts that use similar locutionary means. By saying, e.g., "you destroyed Aunt Maude's lamp," Smith might be ACCUSING Jones, CONDEMING his conduct, CRITICIZING his behavior, REPRIMANDING him for what he did, or REBUKING him. It seems that in performing each of these kinds of speech acts a person speaks with a distinct, more-or-less overt intention. Thus, in condemning someone's conduct, a statement of what the offender did would be produced with the intention of telling the addressee what one is going to hold against the offender or otherwise punish him for doing. In criticizing conduct, one says what was done in order to call attention to the merits or defects of that act or its products. In reprimanding a person, one states what that party did so that he will know what one expects him to avoid doing in the future; reprimands seem to be close to warnings in this respect. Additionally, in rebuking a person, one describes his conduct with the intention of shaming him. The full exposition of these differences would require a lengthy discussion. However, enough has been said to make it apparent that ACCUSING differs from neighboring speech acts, at least partly, in the accuser's (purported) effort to secure an answering response from the addressee.
There are various strategies for inducing confessions that resemble, but are not, instances of ACCUSING. In order to rule out these cases, it is necessary to recognize that in ACCUSING a speaker deliberately and openly gives it to be believed that she is speaking with the intention of securing an answer from the accused; moreover, the accuser must at least act as if her communicative efforts are calculated to provide the accused with reason to respond to her charge. An outraged wife, who punishes her husband by castigating him, may both state what an awful thing he did and intend that her husband confess, but his recognition of this intention is incidental to her attempt to browbeat him into submission. This type of case can be excluded by noticing that the accuser must intend 2 that her addressee recognizes her primary intention to secure an answer to her allegations. There are also cases in which a speaker tells a third party that, e.g., Jackson has been stealing funds, expecting Jackson to overhear the conversation and to recognize that the speaker wants him to confess. Here, the speaker does not openly manifest her intention to secure a confession, and such cases can be ruled out by requiring that the accuser openly intend 3 that the addressee recognize her secondary intention 2 . There are also cases in which the speaker openly confronts a culprit with a statement of the offense, telling the latter to turn himself in and confess or the speaker will herself go to the authorities. Such cases fail to be accusations in that it is the speaker's threat and not her openly manifest intentions that is to provide a reason for the offender to confess.
We have seen, then, that a speaker S will have accused some party P of doing X if the following conditions are satisfied: (1) S says that P did X; (2) S implicates that she believes it was wrong of P to X; (3) S deliberately and openly gives it to be known that she intends that P answer in the way of a denial that P Xed, of an admission of having Xed, or of a justification or an excuse for having Xed; and (4) S intends or acts as if she intends that conditions (1)-(3) provide P with a reason for answering as indicated in condition (3).
The logical sufficiency of these conditions regarding the truth of a report that an accusation has been made is strongly suggested by the unequivocally accusatory character that certain tag-like expressions give to an utterance of the form "he did thus and such".
Where a person seriously and literally says, e.g., "you broke the lamp", she may be making an accusation, but the illocutionary identity of her utterance is not apparent from what she says. Nor does her utterance necessarily become more clearly accusatory if she also makes known her adverse evaluation of the lamp's destruction. However, if a person says "you broke the lamp, surely you don't deny doing that?" and implies an adverse evaluation of what she says was done, her utterance quite explicitly makes an accusation. A speaker, it seems, could not say and imply all that and still deny having accused somebody of breaking the lamp. From this evidence, it seems that the conditions we have adduced as necessary to make an accusation are also sufficient.

The Practical Rationale Underlying ACCUSING
Why, it should now be asked, is the accused supposed to respond to the charges leveled against him? What reason is he expected to have for answering his accuser? Additionally, how is he supposed to arrive at that reason, given what is done in making the accusation? These questions address the pragmatic rationale underlying ACCUSING. They inquire into the practical calculations that constitute this illocutionary act.
In response, I will argue that the accused is expected to answer the allegation because he is under an obligation to do so; accusations characteristically issue from attempts to impose that obligation on the accused or on his representatives. This is a plausible speculation about ACCUSING in that (1) it is continuous with an account of the way obligations are incurred in a wide range of human affairs, and (2) it affords a practically satisfying view of the rationale underlying ACCUSING.

The Obligation the Accused May Have to Explain His Conduct
One commonly incurs an obligation to explain one's behavior where one has caused harm or otherwise given offense. In general, we believe that a person responsible for causing harm has the burden of rectifying the unhappy state of affairs he brought about, insofar as it is reasonably in his power to do so. We include among the redressable damage resulting from a person's behavior those beliefs, attitudes, and doubts, etc., which may foreseeably arise in parties aggrieved by his conduct. Thus, if Jones breaks Aunt Maude's lamp, Smith may be angered, outraged, upset, or disturbed, etc., by the fact that Jones destroyed the treasured possession. These states of disequilibrium are themselves harmful. If the offender has something to say in the way of an explanation or an apology that would redress that harm, then he will, under certain circumstances, have an obligation to provide his explanation or apology.
The concept of obligation has been analyzed by Warnock (1971), who argues convincingly that at least some kinds of obligation are incurred where: (1) it is foreseeable that others will suffer or will continue to suffer harm in the event the obliger does not act; (2) others are counting on his acting in order to avert, prevent, ameliorate, or rectify that harm; and (3) he must so act in order to avoid speaking or having spoken or even having acted falsely. A person thus naturally incurs an obligation to explain his behavior where: (1) his conduct causes another to be angry, upset, or disturbed, etc.; (2) the party taking offense is forbearing retribution, revenge, and so forth, while relying on the offender for an explanation of his conduct; and (3) responsive to the requirements of veracity, it can be said that the offender committed the offense. When these conditions are satisfied, an offender will often acknowledge that he does have an obligation to explain his behavior. However, it does happen that parties believed to have given offense sometimes, predictably, refuse to acknowledge that they have this burden. In that event, those taking offense may try to impose an obligation to explain the deed on the unresponsive suspect.
An obligation can be imposed in the face of resistance by actions which are calculated to ensure that the conditions under which it would obtain are satisfied. We impose assignments taxes and all sorts of obligations on persons who would resist or evade the imposition. My claim is that ACCUSING is calculated to ensure that the obligation to explain behavior is in place in spite of a rather specific deficiency that may be created, or be anticipated, in the conditions under which that burden naturally would be incurred, viz., the deficiency that is created when an offender denies that he did whatever gives offense. Suppose Smith asks Jones why the latter broke Aunt Maude's lamp, purporting thereby to have an explanation coming from Jones. Jones may well be disinclined to respond candidly to the question; he may doubt whether an angry Smith will treat him fairly or he may have something to hide. Disinclined to try to explain the matter, Jones may respond, e.g., "I don't owe you an explanation for that, I did not break your lamp". By saying that he did not break the lamp, Jones generates a presumption of veracity on behalf of his denial. As long as that presumption stands, it is also to be presumed that Smith's beliefs about Jones' conduct are not the truth; the obligation to explain why Jones broke the lamp, which Smith believes Jones has, would then be void.
Accusations are designed to remedy this deficiency by impugning the accused's conduct. The accuser, that is, tries to call the accused's conduct into question by generating a presumption of veracity, which would counteract the accused's (anticipated) denial. Even though a person did not do what he is believed to have done, he still may fall under an obligation to explain his conduct if a responsible effort to determine the truth could not but be expected to find that he committed the offending act. Exploiting this possibility, the accuser would have it presumed that she has made a responsible effort to determine the truth of his allegation, so that even if the accused denies the charge, he would nevertheless fall under an obligation to at least explain the facts that give rise to the impression that he committed the offense. If Smith secures the presumption that she has made a responsible effort to determine that Jones broke Aunt Maude's lamp, then Jones' denial would merely avert the burden of explaining why he broke it and would still leave him the burden of explaining the facts that cause Jones to be upset.
However, such efforts to impose an obligation to explain behavior by impugning a person's conduct contain a potential and crucially important source of embarrassment to the accuser. Calling someone's behavior into question causes him harm; it would be odd to speak of harmlessly impugning a person's conduct. When Smith's allegation raises a question about the veracity of Jones' denial, we recognize that Jones has suffered an injury about which he is likely to complain. This harm is a potential source of difficulty for the accuser because it is liable to be interpreted as a form of retribution or revenge, while the obligation she is trying to impose will only hold if she refrains from such injury to the accused. Were Smith just to impugn Jones' conduct, Jones could evade the obligation to explain his behavior by complaining that Smith's allegation damages Jones' good name.
To avoid this potential embarrassment, I hold, the accuser also tries to have it presumed that she is making an effort to treat the accused fairly and, thus, is forbearing more serious injury while affording the accused an opportunity to explain, justify, or muster an excuse for his behavior. In my view, then, accusations are designed to impose an obligation on the accused by impugning his conduct while having it presumed that the accuser, nevertheless, is exercising forbearance in an effort to treat the accused fairly.
Whatever other merits this conjecture may have, it at least provides an account of how the accused would come to owe an explanation for his conduct-an account which is broadly continuous with a larger view of how obligations are incurred across a wide range of human endeavors. The obligation to explain behavior, we have seen, conforms to the larger analysis of obligations offered by Warnock, and its imposition by an accuser is, on my account, simply a matter of seeing to it that deficiencies in the conditions under which that burden naturally arises do not render it void.

The Accuser's Practical Problem and Its Solution
A second argument for the plausibility of my view is that it attributes a coherent practical calculation to the accuser. That is, it posits a workable solution to a problem that accusers recurrently face.
There can be little doubt that where a person is believed to have committed an offense, it is often necessary to contend with the fact that he may have a stronger reason to falsely deny having done the deed than he has for speaking the truth. Accordingly, where one wants or needs to afford an offender the opportunity to explain his behavior, one may well need means for imposing that burden that are impervious to his denial.
The stringent test of whether my conjecture posits a workable practical strategy is: Can that calculation be implemented in favorable circumstances by doing no more than is minimally necessary to make an accusation? We have seen reason to believe that the accused will have an obligation to explain his conduct, even though he denies the allegation, if: (1) it is to be presumed that the accuser has made a reasonable effort to ascertain the truth about the accused's behavior; (2) it may be supposed that those beliefs disturb or upset the accuser; and (3) the accuser is forbearing retribution, revenge, etc. I consider now how these conditions may be brought about by doing no more than is logically necessary to make an accusation.
To make an accusation, we have seen, a speaker must at least say that the offender did whatever it is that causes the accuser to be upset or disturbed. Generally, according to Stampe's account of seriously saying and meaning something, where a speaker says that P, she incurs responsibility for having made a reasonable effort to ascertain P's truth, and by incurring that responsibility, she generates a reason to presume that he has made the required effort. The primary presumption needed to counter a denial by the accused is simply part of the presumption of veracity; in principle, this presumption can be engaged by making an appropriate statement.
The analysis also shows that an accusation will have been made only where it is evident that a speaker judges, perhaps tentatively, that what she believes the accused did is in the nature of a wrongful act. Where it may be supposed that a person believes something was done, which that person judges, albeit tentatively, to be on the order of a wrong or an offense affecting herself or her fellows, it may reasonably be inferred that this person is upset, disturbed, or angered, etc., as a consequence of those beliefs. Thus, the accuser's implied evaluation of the accused's conduct can reasonably be expected to secure the supposition that he is disturbed or upset by what he alleges was done.
But how do the conditions necessary to make an accusation afford reason to believe that the accuser is exercising forbearance when her statements impugn the accused's conduct? Recall that the analysis shows that the accuser must, at a minimum, openly give it be to believed that she wants an answering response from the accused. By making that intention apparent, she can have it presumed that she is making an effort to treat the accused fairly and, hence, is waiting retribution or punishment on the accused's explanation. Here, as is everywhere the case when a person gives it to be known that she is trying to secure a certain response, the speaker overtly incurs responsibility for her effort and for its foreseeable consequences. When the accuser gives it to be known that she wants an answer from the accused, she makes herself vulnerable to criticism for unfairly causing harm to the accused should she receive an apparently adequate answer and nevertheless persist in her resentment of the accused's behavior. Thus, it may be inferred that, rather than inescapably subject herself to such criticism, the accuser is following a course of forbearance.
In short, the conditions shown in Section 3.1 to be necessary to make an accusation, enable a speaker to deal, at least in favorable circumstances, with the prospect that an offender may falsely deny having committed an offense. By giving it to be known that she wants an answer from the accused, where it is evident that the speaker is upset, angered, etc., by what she believes to be the accused's conduct, a speaker can see to it that the conditions under which the accused has an obligation to explain why he acted as believed are satisfied. Additionally, by saying that the accused committed the offense, the speaker can ensure that if the accused tries to deny the offense, the latter still will be under an obligation to answer for his conduct. Since the problem that speakers can resolve in this way is in fact a recurrent practical problem that accusers encounter, it is plausible to suppose that ACCUSING is, at base, calculated to cope with the possibility of false denials.

The Necessary Features of This Speech Act Explained
The preceding speculations about the nature of ACCUSING not only provide a plausible conjecture about the practical strategy that constitutes this illocutionary act (Section 3.2), they also serve to explain the conditions that are conceptually necessary to make an accusation (Section 3.1). Moreover, the explanation provided is both continuous with a broadly Gricean theory of statements and borne out by the kinds of criticism that can be leveled at defective accusations. A discussion of these claims will first treat the accusation-making statement and then will take up the accuser's overt desire for an answering response.
The analysis shows, it will be recalled, (1) that an accuser must say that the accused did the offending deed and (2) that it must be evident that the accuser judges, perhaps tentatively, that what the accused is alleged to have done is wrong. We will consider in turn the explanation our account provides for each of these facts.
The necessity for an accusation-making statement is explained by the accuser's practical need to have it presumed that she has made a reasonable effort to determine the truth of his beliefs about the accused's conduct. As per conjecture, an alleged offender will have an obligation to explain his behavior if the accuser believes the culprit did something that is in the nature of an offense. In order to see that this burden is in place, even if the offender denies responsibility for the offense, the accuser needs to make known the relevant belief such that she manifestly takes responsibility for having made a reasonable effort to determine its truth. As we observed in Section 2, Stampe's account of the speech act of saying something seriously provides a good reason to hold that wherever a person gives it to be believed that she believes that p, and thereby openly takes responsibility for the truth of p, it will be true that she has said that p. Thus, in conjunction with a broadly Gricean analysis of statements, our hypothesis explains the logical necessity for an accusation-making statement in terms of the fact that to impose on an offender the obligation to explain his conduct one must do those things that make it true that one has said he committed the offense.
This account of an accuser's responsibility to the truth not only conforms to our general account of presumptions of veracity but can also be corroborated by the judgments we do make concerning the consequences of an accusation. We would not consider an accusation harmless just because there was little or no chance that it would be believed, but we would be inclined to regard one in that light if it failed to impugn the accused's conduct, e.g., if the charge made were so preposterous or so trivial that it raised no question about the accused's behavior. Similarly, we would not be inclined to overlook a false accusation because it stirred up a beneficial discussion of issues. In either event, one would have to contend with the harm the thing does to the accused. Additionally, it is, I think, a commonplace fact that where a person has been falsely accused of something, our concern lies primarily with the harm he suffers and less so with those who might have believed the accusation. Of the latter, we are inclined to believe that they should have recognized that in making the accusation, the speaker only produced an allegation.
With the support of a few rather unobjectionable premises, our hypothesis also explains the remaining features associated with accusation-making statements.
First, the fact that an accuser must at least purport to be certain that the accused acted as alleged can be explained by the fact that an offender will have an obligation to explain his behavior, should he deny committing the offense, only if a reasonable effort to ascertain the truth cannot but lead to the conclusion that he did what he is said to have done. Since, as our account suggests, the accuser would have it believed that she cannot find grounds for doubting that the accused committed the offense, she must at least act as if she is certain that the latter did the offensive deed.
Second, the fact that it must be evident that the accuser believes, perhaps tentatively, that what the accused did is wrong can be explained by the fact that an obligation to explain behavior is in force only where what one person does causes another to be upset, angry, or disturbed, etc. Generally, the fact that a person has done something is not by itself sufficient cause for another to be upset. It must also be the case that what he did is wrongful or reprehensible, or at least there must be some possibility that what was done is in the nature of a wrong. Thus, where, e.g., one person is upset because another turned out the lights, one might ask "I don't see what you are so upset about, there's nothing wrong in what he did?". If an accuser is to impose an obligation to explain behavior, it must be evident that she is upset or angry about what was done, and that will be evident only if it is also plain that she believes there is at least some possibility that what the accused did is wrong.
Third, the fact that an accuser need not be certain that what the accused did is wrong can be explained by the fact that one may be upset by a person's conduct even though one is not certain that what this person did is reprehensible. Believing that Jones broke Aunt Maude's lamp, Smith may be upset by the possibility that Jones inflicted this damage carelessly or maliciously. Thus, to impose on Jones the obligation to explain the damage, Smith need not even act as if she were certain that breaking the lamp was wrong.
To return to the analysis, recall that it further shows that (3) an accuser must give it to known that she wants an answering response from the accused. This is explained on our account by the accuser's need to provide reason to believe that, in spite of impugning the accused's conduct, she is forbearing retribution and punishment. To generate such reason, I have suggested, an accuser engages a presumption of fairness by giving it to be known that she wants an answer to his charge. This account has two very substantial virtues.
First, within the limits of commitments one can undertake conversationally, an openly manifest desire for an answer responding to the accuser's charge seems to be, not only an efficacious means for generating a presumption of fairness, but also to be the bare-bones minimum commitment capable of engaging that presumption. In many cases, of course, an accuser would explicitly commit to treat the accused fairly, pledging that commitment in language which spells out her intention to treat the accused fairly, to afford him an opportunity to clear up the matter, or to defend his good name, etc. However, the heart of such elaboration is the accuser's commitment to seeking an answering response from the accused. Minus that commitment, she could, in conjunction with impugning his conduct, hardly purport to be trying to treat him fairly.
Second, the power of that minimal commitment is reflected in the force excuses have as responses to an accusation. [In an extended argument omitted here, Kauffeld argues that the strength of the accuser's "minimal commitments" is demonstrated by the fact that should the accused successfully put forward an excuse, the accuser is obligated to walk-back the accusation.] To satisfy the demands of fairness, she may thus have to reduce the charge to negligence or carelessness. The accused, as it were, gets out of the fire into the pan.

Conclusions and Implications
If the preceding explication of ACCUSING is correct, 1 several observations about the constitution of this and similar illocutionary acts are in order. Some thoughts about the development of cultures supporting argumentation can also be offered.
First, it should now be apparent that ACCUSING is constituted by practical calculations, not by rules. 2 The account I have given for ACCUSING makes no appeals to rules that might be thought to constitute this kind of speech act, beyond those rules of syntax and semantics, which enable speakers to say things. Where it might be thought to invoke something like a rule, as in Warnock's analysis of obligation, the rule in question (if it is a rule at all) is not specific to the speech act of ACCUSING. That is not to say one cannot formulate rules for making accusations; in fact, the rules governing judicial proceedings in courts of law do have rules for making accusations. However, these are regulative rules, not constitutive rules. They govern a kind of action that can occur independent of the rules.
[We should thus be wary of continued reliance on the rule-constituted view of speech acts to provide a framework for the study of argumentation.] Second, and of particular relevance to argumentation theory, it should now also be apparent that in connection with ACCUSING-and with other illocutionary acts, though by very different routes-that a speaker incurs a specific argumentative burden-a burden of proof. As a consequence of the accuser's commitments, the accused is in a position to demand that the accuser substantiate her charge if the accused purports to be innocent.
[ACCUSING thus has the capacity to rope two perhaps unwilling participants into an exchange of arguments governed by specific norms: the accused obligated to explain his conduct, the accuser obligated to be fair and if her accusation is denied, provide arguments in support. In short, the account of ACCUSING given here provides an explanation of the genesis of argumentatively important responsibilities as well as the practical value of discharging them.] This brings us to my first point about the pragmatics of illocutionary acts and larger argument cultures. In order to have cultures that support productive argumentation in public and institutional forums, one needs appropriate norms, rules, structures for recognition and reward, and supporting modes of education and training. A careful explication of the pragmatics of those illocutionary acts that are constituted on the basis of underlying practical calculations (and I would reserve the term "illocutionary" for those), puts us in a position to address design questions related to shaping supportive cultures for argumentation. Our own study of ACCUSING enables one to better appreciate why the courts need rules that assign the burden of proof to parties bringing an accusation. Where accusations are outside the courtroom and in the absence of rules, the accused can, by protesting his innocence, impose a burden of proof on the accuser, but this requires a complicated maneuver beyond the competence of many. So, establishing an institutional rule that allocates the burden of proof to the accuser ensures fairness in a practice that otherwise is very vulnerable to abuse. I suggest that a pragmatic view of speech acts puts us in a better position to design rules and rewards, etc., than a rule-constituted view that does not show us what is practically necessary for communication to work. A rule-constituted approach may claim to provide a systematic account of the different kinds of functions that language might be used to perform (i.e., utterance forces) and the conditions under which an utterance can be used to perform those functions. However, there is no claim to identifying the ends-means practical calculations that enable the act to thus function; one just gets the conditions under which the rule can be invoked, the procedure for invoking it, and what the rule mandates. This masks the pragmatics [of what can be challenging interactions].
Second, accounts of illocutionary acts in terms of the pragmatics of commitments enable argumentation theorists to comprehend how discursive practices accommodate variations in circumstances and in argumentative strategy. Scott Jacobs has identified a propensity for rule-constituted views of speech acts to gloss over "the way in which the functions of utterances are subtly fitted to the particular circumstances of their placement and finely attuned to the nuances of their expression" (Jacobs 1989, p. 351). Such a tendency inheres in any rule-constituted view of communicative acts. Rules are formulated for the standard case, and the protasis for a rule rather rigidly fixes its range of application. Searle's account of illocutionary acts would have its best chance of success in a world in which every serious utterance was expressed with a linguistically articulated illocutionary force, e.g., "I hereby promise that I will be home by seven," As Strawson points out, that would be an intolerably rigid world, far different from our own. When we see illocutionary acts constituted pragmatically, we see a world in which enormously more variation and flexibility is possible and exercised. Illocutionary force depends upon the criticisms speakers openly risk, and such criticisms vary with the commitment speakers undertake by openly manifesting their intentions. This is a kind of process open to enormous adaptation and variation. [Kauffeld next recaps the argument he made elsewhere (Kauffeld 2009) about the structure of Martin Luther King's "Letter from a Birmingham Jail,"] King here openly undertakes a commitment and generates a rationale for audience participation in his address; we have no canonical illocutionary act for what he is doing, but I think King performed one. If we recognize the pragmatic constitution of illocutionary acts, we will be in a better position to understand and deal with the variation and nuance of practices necessary to a healthy argumentative culture.