Alhassanain(p) Network for Heritage and Islamic Thought

Collective Speech Acts

Anthonie Meijers[1]

Table of Contents

Introduction 3

A note on groups 4

Speech acts as communicative acts 5

Are collective speech acts reducible? 7

The analysis of collective speech acts continued 12

Consequences for Searle’s taxonomy of speech acts 15

The construction of social reality 21

Conclusion 22

REFERENCES 23

Endnotes 25

Introduction

From its early development in the 1960’s, speech act theory always had an individualistic orientation. It focused exclusively on speech acts performed by individual agents. Paradigmatic examples are ‘I promise that p’, ‘I order that p’, and ‘I declare that p’. There is a single speaker and a single hearer involved. In his bookSpeech Acts, for example, Searle’s analysis of promising starts from the following description: “Given that a speaker S utters a sentence T in the presence of a hearer H, then, in the literal utterance of T, S sincerely and non-defectively promises that p to H if and only if the following conditions 1-9 obtain [etc.]” (Searle 1969: 57).Though this focus may initially have been due to Searle’s methodological approach of starting his analysis of speech acts with clear-cut examples, it has led in the end to an unnecessary and undesirable bias in speech act theory.

In this paper I will extend the traditional analysis of speech acts by focusing on collective speech acts. These are acts performed by collective agents or addressed to collective agents. Our language is full of collective speech acts. Examples are: “we promise to be back in time”, “we believe that this is the wrong approach”, “the Security Council appeals to Israel and Hezbollah to stop fighting”, “I want the orchestra to start with the first part”, “we urge the government to reconsider its policies”, and so on. Surprisingly, these collective speech acts have escaped the attention of almost all philosophers of language thus far.[2]

Collective speech acts are not just interesting in themselves because they are omnipresent in everyday language. They are also crucial for understanding an issue that belongs to the very core of Searle’s philosophical project: the nature of social reality. InThe Construction of Social Reality Searle defends the view that language is constitutive of social institutions. Institutional facts require symbolic representations that are publicly understand­able. I will argue that this claim can be made more precise using the concept of a collective speech act.

A note on groups

Groups do not form a single category. As a matter of fact, there are many types of groups: the fans in a soccer stadium, the team playing, the members of the club, the visitors older than 21 years, the executive committee of the club, a family attending the match, etc. There is also not one way of classifying groups. This can be done in terms of the physical and / or the intentional relationships between the members of the group. A family is bound by biological kinship, fans by the physical boundaries of the stadium and by their desire that their team will win, a team by their common goal and the arrangements that have been made for the match (their strategy, division of labour), an executive committee by the institutional rules of the club, and so on.

Rom Harré (1997: 200) makes a useful distinction between three types of groups, which I will follow roughly here:

taxonomic groups or classes. An example is the visitors in the stadium older than 21. Members of the group only share a particular characteristic, but they do not have ‘internal’ relations. If one of them drops out, it will not affect the others.

crowds. For example, the crowd entering the stadium. Members share a common goal, but do not have beliefs about each other given that common goal, or rights and obligations.

structured groups. Members are either physically or intentionally related. An example of the former is a family, of the latter a team playing soccer. Team members share a common goal, they have beliefs about each other, and they have rights and obligations. The members of structured groups are internally related, i.e., the loss of one of the members will affect the others.

For the purpose of this paper I will restrict most of the analysis of collective speech acts to structured groups of the intentional type. The main reason is that speech acts, beingacts , are performed byagents and if the notion of a collective agent makes sense it will refer to members of a group that are intentionally related. They share intentions, goals, beliefs, or desires. On the other hand, it is hard to understand how groups of the other types (family, crowd, taxonomic) could be characterized as collective agents at all, since they need not share any intentional states. Speech acts may beaddressed to those crowds or taxonomic groups, an issue to which I will come back when I discuss the reducibility of collective speech acts.

Speech acts as communicative acts

There is a sense in whichall speech acts are collective acts, and it is important to clarify this sense before starting the analysis of collective speech acts. Searle’s speech act theory and most subsequent theories conceive of speech acts as communicative acts. They are the minimal units of communication (Searle 1969: 16). The model of communication on which Searle’s theory is based is Grice’s model. According to Grice (1957, 1989), for the speaker to mean something by uttering X is to intend that the utterance of X willproduce some effect (belief, desire) in the audienceby means of the recognition of the speaker’s intention. Unlike Grice, Searle defines this effect as the hearer’s understanding of the utterance of the speaker. In doing so, he makes room for the well-known distinction between the illocutionary effect and the perlocutionary effect of speech acts.

The resulting account of communication shares with many other accounts - such as the traditional sender-receiver model, the information processing models in cognitive science, or the radical translation / interpretation model - the idea of communication as aone-way transfer between speaker and hearer. The speaker produces some effect in the hearer by means of his intentions, or sends a message via a communication channel that needs to be decoded by the hearer, or utters a sentence that needs to be interpreted from the hearer’s third person point of view. A dia­logue or conversation, then, is accounted for by applying this one-way model reci­procal­ly. It is a sequence of one-way transfers, where speaker and hearer change roles.

I have criticized this model for not being able to capture what is in my view essential for communication: that it is a form ofcooperation (Meijers 1994, 2002). Communication is not a series of mono­logical speech acts, where speaker and hearer act independently of one another. There are forms of communication that are like that (for example, advertisements). But even in these cases, the utterances make use of conventional means in order to be under­stood, where these conventions originate in prior cooperation or coordination. Communi­cation in a full sense, as in dialogue, is different from monological utterances. A single speech act is embedded here in a larger conversational framework, where speaker and hearer cooperate in order to reach understanding with respect to the matter being communicated. This understanding cannot simply be analyzed as thehearer’s under­standing of the speaker’s utterance. ­Grice’s theory is in that sense only part of the story. The result of dialogue is more properly described asshared understanding. If communi­cation is successful, speaker and hearer share understanding with respect to the speaker’s speech act. As Charles Taylor said, “Communication doesn’t just transmit information (...). It brings about the acknowledgement that some matter isentre nous, is between us” (Taylor 1980: 295).

This is obviously a large issue that needs careful discussion (see again Meijers 1994, 2002). But we need not go into the details of the analysis here. What is important for the discussion of collective speech acts is that there is a sense in whichall speech acts are collective acts. As communicative acts they are performed together, and require collective intentionality. The intention to perform a promise, a request or an order, is embedded in the collective intention ‘we intend to communicate’. In addition, a successful speech act requires ‘uptake’ or successful communication and the result, shared understanding, introduces again a collective dimension in the analysis.

In the discussion below I will take this collective dimension of speech acts for granted. It applies toall speech acts performed in dialogue, not to a particular type of speech act. My focus in this paper will instead be on a particular type of speech act, namely those speech acts which involve collective agents, either as speakers or as hearers. This type is different from speech acts which involve only individual agents, and which have been the exclusive focus of speech act theory thus far. I will call the latterindividual speech acts, to distinguish them fromcollective speech acts which involve collective agents. There are three forms of collective speech acts:

those in which a collective speaker addresses an individual hearer (for example, ‘we expect you to be in time’);

those in which an individual speaker addresses a collective hearer (for example, when the coach says to the team ‘play your own game and not your opponent’s’);

those in which a collective speaker addresses a collective hearer (for example, when the prime minister says to the secretary general of the UN: ‘we accept resolution 1701 and will respect a ceasefire’).

Collective speakers are restricted to structured groups of the intentional type. They have to be intentional agents. Collective hearers, on the other hand, can be either crowds or structured groups of the physical or intentional type. We may ask a crowd to disperse, tell the passengers waiting at an airport that all flights have been cancelled, or order an army to withdraw.

Are collective speech acts reducible?

It is a well-known fact that in uttering a particular sentence, more than one speech act may be performed. The assertion “it will be a formal dinner”, for example, may also be the directive speech act “don’t forget to dress properly”. Searle’s analysis of indirect speech acts has generated a lot of interest in this phenomenon. The fact that we can perform more than one speech act in uttering a single sentence, has relevance for the analysis of collective speech acts. It is an interesting question whether upon closer analysis collective speech acts are nothing but a collection of individual speech acts. The idea is that instead of performingone speech act, a collective speaker makes several speech acts, where the number will depend on the number of individuals making up the collective. The ‘we’ is conceived as a set of ‘I’s. Similarly, a speech act addressed to a collective hearer, for example a crowd, may be a set of speech acts, where the set is defined by the individuals making up the crowd. The obvious advantage of such a reductive strategy is that we don’t have to introduce a new category of speech acts, since the existing types can do the job. This is an issue that every account of collective speech acts will have to address. Are collective speech acts reducible?

Let me start with an example. Imagine a case in which the management team of NASA makes the following speech acts at a press conference: “we believe that the shuttle is safe enough to launch”, “we intend to launch the shuttle later today”, and “we will do everything possible to make sure that the shuttle returns safely”. A reductive approach will have to rewrite these collective speech acts as a combination of individual speech acts, where the ‘we’ is replaced by ‘I’s. One way to do this is to rewrite the speech act

(a)  we believe that the shuttle is safe enough to launch

as a conjunction of the speech acts

(b)  I believe that the shuttle is safe enough to launch &

(c)  I believe that the shuttle is safe enough to launch &

       and so on, for all the members of the management team.

The question, then, is: Is this conjunction of individual speech acts equivalent to the original collective speech act?

To understand why they are not equivalent we have to analyze the conditions of success of speech acts. Searle distinguishes four types of conditions of success: the propositional content condition, the preparatory condition, the sincerity condition and the essential condition (1969: 57ff, 64ff). The propositional content condition specifies conditions on the type of content that can be part of a particular type of speech act. As we will see, there are propositional content conditions that are specific to collective speech acts. This is a strong indication that the equivalence relation between collective speech acts and a conjunction of individual speech acts does not hold.

Imagine that there are dissenting opinions in the management team of NASA with respect to the safety of the shuttle (which in fact happened in the past). If the management team want to launch it, they will have to make up their mind with respect to the aircraft’s safety. In such a situation it can very well be that the management team as a collective body takes the position that the shuttle is safe, while there are members who do not share this view. These members, then, are not willing to perform the individual speech act “I believe that the shuttle is safe”, while they are willing to perform the collective speech act “we believe that the shuttle is safe”. More generally, collective speech acts involve collective intentional states (beliefs, desires, intentions), whose content cannot be unconditionally attributed to the participants outside the context of their collective action. As independent individuals, they may believe, desire or intend differently.

A further indication that the equivalence relation does not hold is that in the example “we intend to launch the shuttle today”, the reduction of the collective speech act to a conjunction of individual speech acts of the type “I intend to launch the shuttle today” does not make sense. The launching of the shuttle is a collective act which cannot be carried out individually and thus cannot be the content of an individual intention (similarly, I cannot have the individual intention to perform Monteverdi’s Vespers of St. Mary, since it takes a choir and basso continuo to do that). Here the individualist may argue that there are ways of rewriting the collective speech act that do not violate the propositional content condition. For example, one might rephrase the speech acts as a conjunction of individual speech acts of the type “I intend that we launch the shuttle today”, or “I intend to do my part in launching the shuttle”. Though this seems to avoid the problem in a number of cases, we will see below that such a reformulation violates other conditions of success of speech acts.

Both examples show that the propositional content condition makes a reduction of collective speech acts difficult and possibly restricted to cases where the propositional content of collective speech acts can satisfy the propositional content condition of individual speech acts. The other conditions for the success­ful performance of speech acts point in the same direction. Preparatory conditions are conditions that need to be fulfilled in order for the speech act to be performed properly. They concern, among other things, the physical, mental and social status of the speaker and the hearer. Collective speech acts obviously require a collective agent to be in place. A conjunction of individual speech acts, on the other hand, requires a number of individual agents to be in place. These require­ments are the sameonly if a collective agent can be conceived of as a summation of individual agents. Or, put differently, only if a straightforward reduction of collective agency is possible. I will address this issue below.

Searle’s third condition, the sincerity condition, specifies the psychological state expressed by a speech act. An assertion counts as the expression of the belief that the propositional content is true, a promise counts as the expression of the intention to do what is promised. The collective promise “we will do everything possible to make sure that the shuttle returns safely”, thus counts as the expression of a collective intention to do everything possible to makes sure that the shuttle returns safely. More generally, collective speech acts express collective intentional states, whereas individual speech acts express individual intentional states. The reductionist strategy works, then,only if collective intentional states can be reduced to (combinations of) individual intentional states (whatever their content is). I will argue below that such a reduction is not possible.

Finally, the essential condition specifies the illocutionary point of a speech act, i.e. what the speaker wants to accomplish in performing the act. The point of an order is to get the hearer to do what is ordered, the point of a promise is to place the speaker under the commitment to do what is promised. A collective speech act is in that sense is different from a collection of individual speech acts. Take again the case of the collective promise “we will do everything possible to make sure that the shuttle returns safely”. Here the manage­ment of NASA places itself under the commitment to do what is promised. They do this as a collective agent, even as an institutional body, andnot as a collection of individual persons. The commitment to carry out the promised action thus remains in place even if individual members of the management are replaced by others (unless the promise is explicitly withdrawn).

The attempt to reduce a collective promise to a conjunction of individual promises can only succeed if a collective agent is nothing more than a summation of individual agents and if a collective intention can be reduced to a collection of individual intentions. This was also the outcome of the discussion of the preparatory condition and of the sincerity condition. It is time now to address this issue directly: (i) Are collective agents nothing more than collections of individual agents and (ii) are collective intentional states nothing more than collections of individual intentional states? A negative answer will refute the reductive strategy, and will have the consequence that collective speech acts will have to be regarded assui generis acts. These questions are obviously large questions which have been discussed extensively in the literature. I believe that the second question is the more fundamental one. Collective intentional states already have a ‘we’ as a subject, and the minimal form of a collective agent is two individuals sharing an intentio­nal state (a physical or biological relation is not enough for agency, though it may result in a structured group).

That collective intentionality is an irreducible, primitive phenomenon has been argued by many philosophers, including Margaret Gilbert, Raimo Tuomela, and John Searle. In Searle’s view, there is a “deep reason” why collective intentionality cannot be reduced to individual inten­tio­nality. “No set of ‘I-Consciousnesses’, even supplemented with [mutual] beliefs, adds up to a ‘We-Consciousness’. The crucial element in collective intentionality is a sense of doing (wanting, believing, etc.) somethingtogether [my italics], and the individual intentionality that each person has is derivedfrom the collective intentionality that they share” (Searle 1995: 25). Many species of animals, according to Searle, have this “biologically primitive” capacity for collective intentionality, i.e., the capacity for cooperation andsharing intentional states.

I believe that Searle is right. The awareness of doing something together cannot be ‘synthesized’ out of the awareness of individual agents doing something individu­ally. Additional arguments can support the non-reductionist position. I have argued for the irreducibility of collec­tive intentio­nality on the basis of an analysis of the commit­ments involved (Meijers 2003). Take the example of the launching of the shuttle. Suppose that members of the management team agree to do this. Since launching the shuttle is a complicated affair, they will intend to do thistogether . Such an intention brings about commitments among the agents who make up the collective agent. There will be a division of labour in which they are committed to do their part. Not only that, they are also entitled toclaim that others do their part as well. The collective intention thus creates inter-individual commit­ments and normative relationships among the members of the team. The crucial point for the present discussion is that no set of individual intentions can ever generateinter-individual commitments. An individual inten­tion may eventually generate a commitment to oneself to do what one intends to do (but such a commitment may be nothing more than the original intention), but it can never generate commit­ments to others that I do that, let alone a claim by others that I do that. Collective intentions, on the other hand, create such inter-individual commit­ments and claims. From the fact thatwe intend to do something together, we are both committed towards each other to do our part and I can claim that you do your part and you can claim that I do mine. Seen from this perspective, then, the irreducibility of collective intentionality stems from the impossibility to generate inter-individual commitments on the basis of individual intentionality.

There is an ontological side to this as well. Though I agree with Searle that collective intentionality is irreducible, I find his conception of collective intentionality problematic from an ontological point of view. Let me explain. Searle formulates two conditions of adequacy for an account of collective intentionality: (i) it must be consistent with the fact that society con­sists of nothing but individuals, and (ii) it must be consistent with the fact that all intentio­nality, whether individual or collective, could be had by a brain in a vat (Searle 1990: 407). These two conditions, which state the obvious for Searle since they are presented as “facts”, result in an individualistic conception of collective intentionality. We-intentions are the intentions ofindividual agents. These agents are capable of individualand of collective intentionality. Ultimately, it is therefore possible that asingle individual agent has a collective we-intention, for example the intention “we intend to launch the shuttle today”.

This is a very unfortunate consequence of Searle’s theory in my view. As we have seen, inter-individual commit­ments are part and parcel of collective intentionality. These commitments existbetween individuals. They cannot be accounted for in terms of the intentional states of asingle individual, because that would effectively amount to the reduction of relations to the intrinsic properties of one of the relata. I take it to be evident that this cannot be done. The question, then, is: Can Searle’s theory be repaired to avoid this unacceptable consequence? I believe it cannot, because the consequence is strongly related to Searle’s internalist conception of the mind. It would mean, among other things, giving up the second condition mentioned above. This condition, however, is not a peripheral item of the theory, but one of the key tenets of Searle’s philosophy. As such it is a condition forall accounts of intentionality, including collective intentionality.

An externalist conception of collective intentionality, in which genuine relationsbetween individuals play an important role, avoids these problems.[3]In this conception collective intentions need to have a foundationin re . There has to be another agent ‘out there’, so to speak, for collective intentionality to be possible. If no other agent is in fact ‘out there’, there is no collective intentionality. Of course, individual agents may have the false belief that their intentions are shared and they may actas if there is such collective intentionality. But the beliefs and the intentionality in question will then be just that: false beliefs andas-if intentionality.

Whatever the outcome of the debate on the ontology of collective intentio­nality is, the main result of our discussion for this paper is that collective intentionality cannot be reduced to individual intentionality and that consequentlycollective speech acts are acts suis generis . They form a separate class of speech acts, and cannot be reduced to combinations of individual speech acts.[4]

The analysis of collective speech acts continued

The previous sections indicated already a number of ways in which collective speech acts are different from individual speech acts. In this section I will explore more differences between individual and collective speech acts and in the next section I will discuss the consequences for Searle’s taxonomy of speech acts.

As we have seen already, speech acts are communicative acts and as such they have certain characteristics. They are performedin the act of speaking, they are addressed to a hearer and they require uptake or understanding by the hearer in order to be successful. There is an important asymmetry between individual and collective speech acts in that a collective speaker (say the management of NASA) cannot literally utter sentences. Only indivi­du­als can do that. Collective speech acts are therefore performed by individual speakers on behalf of the group. The analysis of the intentions involved in collective speech acts has to reflect that.

In his pioneering article “Group Speech Acts”, Justin Hughes (1984) analyzed the intentions involved in collective speech acts and the conditions that have to be fulfilled for these speech acts to be successful:

“The conditions, then, of a group speech act are as follows [for a group G, speaker S and utterance X]:

there exists a group (G), this group has an illocutionary intention, and X conveys that illocutionary intention.

S (believes that he/she) knows the illocutionary intention of G and that X conveys this illocutionary intention.

G does not object to S uttering X on its behalf and if G intends for any specific individual(s) to utter X, it intends for S to utter X. S (believes that he/she) knows this.

#2 and #3 are (the) reasons S utters X.”(Hughes 1984: 388)

These conditions are preparatory conditions for collective speech acts. They concern the individual speaker who will speak or write on behalf of the group.

The first condition is obvious and simply states that there has to be a group, that the group has to have an illocutionary intention, and that the utterance has to convey this intention. The second condition, however, introduces a new element in the analysis of speech acts. There is an interesting asymmetry between individuals and groups in an epistemic sense. Individuals usually don’t have to make an extra effort to know their illocutionary intentions when performing a speech act. They just know it. Groups, on the other hand, do not have the type of epistemic access that individuals have. In order to know what a group’s intention is, a conscious effort need to be made by the speaker who acts on behalf of the group. Usually (s)he needs to consult other members and there has to be some procedure to decide on the group’s intention. The condition states that the speaker knows or believes to know what the intention of the group is, and also that (s)he knows that the utterance of X conveys this illocutionary intention. The third condition is about the delegation of the speech act. Notice that the actual speaker need not be a member of the group for the collective speech act to be successful. Hughes makes a useful distinction here between the ‘intender’ of the speech act (the group) and the actual speaker. In a collective speech act they may be different, while they are always the same in individual speech acts. The condition allows for the explicit delegation of the speech act, as well as for situations in which there is no such explicit delegation (for example, when all members can speak on behalf of the group). The final condition states that the reasons for performing a collective speech act should be internally related to the group’s illocutionary intentions. If it is done for other reasons, for example self-interest of the speaker, the speech act does not count as acollective speech act.

Hughes’ conditions contribute to the analysis of the complex phenomenon of a collective speech act. Most of them just add to the list of prepara­tory conditions that Searle already distinguished. There is one point that deserves special attention. It concerns the way in which groups develop their collective aims, beliefs, desires or intentions. For Hughes, “the ideal group is characterized by active, open, and free exchange among all members of the group. In this interchange the utterances of individuals are directed to all other persons of the group”; at the heart of such a group there is a “consensus mechanism” (1984: 385 and 384). This mechanism takes care of the formation of the group’s illocutionary intentions, or, more generally, its collective intentional states. Deviant cases are cases where the consensus mechanism is marginalized, usually because of urgent situations: an SOS message on behalf of the group without prior deliberation, or an offer for a cease-fire made by a gang leader during a fight between groups, without holding a referendum first. But even in these cases there may be a rudimentary consensus mechanism at work, for example based on body language. But it appears, according to Hughes, “that [in these cases] we recognize group speech acts as such if we believe that there is some other basis by which the utterer ‘knows’ (is justified in believing he knows) the group’s intention” (1984: 384).

Hughes’ analysis suffers, in my view, from taking a particular type of group, an ideal demo­cratic group, as the general paradigm for groups (pars pro toto ), thereby taking a particular type of mechanism for the formation of collective intentional states, the consensus mechanism, as a general mechanism. As a matter of fact, the situation of a really free exchange of ideas that leads to a consensus in which power relations among the members of the group do not play any role, seems to me rather exceptional. It is even doubtful whether this can be used as a paradigm on which real group processes can be modelled. As we have seen in Section 1, there are several types of groups, and probably several types of developing collective views. A family will be different in this respect from a company.

Groups in which agents are not only allowed to speak on behalf of the group, but are also authorized to act on behalf of the group, are especially interesting from the point of view of speech act theory. Imagine a structured group of the intentional type, in which the leader of the group has been authorized to do whatever (s)he thinks is necessary in a specific situation, for example a negotiation. The leader will utter speech acts such as “we believe the proposal is unacceptable, or “we will withdraw from the negotiations if the political situation does not improve”. Or, to take another example, imagine that president George Bush utters the speech act “the US government believes that the axis of evil includes stem cell biologists”, without consulting his ministers. His utterance is never­theless a collective speech act made by the government of the US, because it is not an option for the members of his government qua members to deny that this speech act expresses a collective belief of the US Government to that effect. Why is this so?

Contrary to Hughes’ idea, that in these cases we recognize group speech acts as such if we believe that there is some basis by which the uttererknows the group’s intentional states about the subject, I believe that we encounter here a phenomenon that is similar to the one Searle (1989) described in his analysis of performatives. In a performative, for example “I hereby promise you to come tomorrow”, the speaker makes it the case that s/he is promising bydeclaring that s/she is promising. Similarly, an authorized speaker can make it the case that his utterance expresses a group belief bydeclaring that it expresses that belief. When George Bush says that the US govern­ment believes that the evil of axis includes stem cell biologists, this belief hasde facto become the collective belief of the US government. In that respect these collective speech acts are declarations, just as performatives are declarations. They are declarations in addition to being assertives, directives, commissives, or expressives.

Consequences for Searle’s taxonomy of speech acts

Speech acts differ in very many respects.[5]The style of the performance, the force of the illocutionary point, the extra-linguistic institutions needed for the speech act, or the relation to the rest of the discourse, are among the many properties of speech acts that can be used to bring order in the apparent chaos. Searle’s taxonomy is constructed on the basis of three main qualities of speech acts: their illocutionary point, their direction of fit, and their expressed psychological state. The resulting categories are substantially different from each other and very useful in practice. Searle distinguished five types of speech acts that are well-known ever since: assertives, directives, commissives, expressives and declara­tions. A key question for this paper is: Do collective speech acts fit Searle’s taxonomy, i.e., can we distinguish the same types of collec­tive speech acts on the basis of these qualities? If so, that would show that collective speech acts are not only different from individual speech acts, but also similar in important respects.

Assertives are the first category of speech acts. Their illocutionary point is to commit the speaker in varying degrees to the truth or falsity of the propositional content of the speech act; the expressed psychological state is a belief in that content; and the direction of fit is ‘words to world’, i.e. what is asserted in the speech act needs to fit the state of affairs in the world. For example, in my assertion “it is safe to launch the shuttle”, I commit myself to the truth of the proposition that it is safe to launch the shuttle, the speech act expresses my belief to that effect, and the speech act is supposed to reflect this state of affairs in the world. Assertives can be performed by collective agents. In such a speech act a collective speaker is committed to the truth of the propositional content, the speech act expresses the collective belief in the content and the direction of fit is words to world.

There is, however, a complication with respect to the notion of a collective belief. In what sense can a collective agent be said to have a genuine belief and what would that add to the beliefs which the individuals making up the collective already have? For this discussion it is important to make a distinction betweenbelief andacceptance. [6]To believe a propo­sition is to be committed to the truth of that proposition, where only epistemic reasons count as valid reasons for an agent to believe that proposition. Acceptance, on the other hand, is a different cognitive attitude. To accept a proposition is to adopt that proposition as a valid premise in one's practical reasoning about future actions, where in addition to epistemic reasons practical considerations play a role. For example, for reasons of prudence an agent may accept in her practical reasoning the proposition that the Dow Jones Index will go down. In such a situation the agent may even believe in an epistemic sense that the Index will go up, but she plays it safe and acts under the assumption that the Index will go down.

Given this distinction, there are two readings of a collective belief: we may interpret it as a genuine belief, or we may take it as the acceptance of a proposition. I have argued for the second view.[7]Let us review the options before us. If we take it as a genuine belief, only epistemic considerations can play a role; in matters of truth practical considerations do not matter. This means that if the management of NASA asserts: "it is safe to launch the shuttle", the members of the managementqua members are committed the truth of that proposition. And if they are committed qua members, they are also committed as indivi­duals. Otherwise we would have a contradiction: you cannot in one role believe that X is true and in another role believe that X is false. A collective belief in this reading, then, is nothing more than the conjunction of the individual beliefs of the members of the group. Consequently, if one of the members changes her view and starts believing that the shuttle isnot safe to launch, the earlier collective belief ceases to exist. The question, then, is: What does the notion of a collective belief add to our analysis here? I am tempted to say that it adds nothing. We can account for such a collective belief in a reductionistic way, because we can rewrite this belief as a conjunction of individual beliefs. The group’s commitment to the truth of the proposition, however, does add something to the individual commitments to its truth.

In the second reading a collective belief is interpreted as the collective acceptance of a proposition. Let us take again the situation in which there are dissenting opinions within NASA's management about the shuttle launch, but that they agree as a group that it is safe to launch the shuttle. At a press conference the chairman will then say: "we believe that it is safe to launch the shuttle ". Though this is literally said to be a belief, the expressed psychological state is really the acceptance of a view, for the decision taken is based on epistemicand practical considerations. For example, they may have reasoned that they cannot postpone the launch indefinitely, that further upgrading of the shuttle will only marginally improve the situation, that funding may be in danger if they will not launch the shuttle shortly, and that therefore the majority view in the management will be followed. In this situation membersqua members of the manage­ment will accept as their joint view that the shuttle is safe to launch, while some members will individually have reservations and believe differently. They accept it, and in their acceptance non-epistemic reasons play a role. My conclusion is that either we have a reductive notion of collective belief as the conjunction of individual beliefs, or we have an irreducible notion of collective belief which turns out to be the acceptance of a view upon closer analysis.[8]The expressed psychological state of a genuine collective assertion should be described as the collective acceptance of a proposition rather than a collective belief.

Directives are Searle's second type of speech acts. Their illocutionary point it is to get the hearer to do what is specified in the propositional content of the speech act. The expressed psychological state is the speaker's wanting that the hearer does this, and the direction of fit is world to words. The speech act is satisfied if the hearer does what the speaker wants him to do. Collective agents can perform such directive speech acts. For example, when the parliament asks a minister to be present at a meeting. The illocutionary point here is to get the minister to be present, and the expressed psychological state is a collective wanting or desire to that effect. Things are different when a collective agent is involved as thehearer of the speech act and the speaker is an individual agent. For example, when a coach urges his team to play defensively. The expressed psychological state then is an individual wanting or desire.

Commissives form the third category of speech acts and their illocutionary point is to commit the speaker to some future course of action. The expressed psychological state is the intention to do what is promised, and the direction of fit is world to words. Collective agents can perform collective promises. For example, when the government promises to the parliament to improve airport security. In doing so, the government places itself under the commitment to improve airport security, the expressed psychological state is the collective intention to do what is promised, and the speech act is satisfied if the world changes in a way that the speech act specifies. Similar to the case of directives, if the speaker of a collective promise is an individual agent and the hearer a collective, the expressed mental state is an individual intention.

Expressives are a type of speech act whose illocutionary point it is to express the psycho­logical state specified in the propositional content. The sincerity condition or expressed psychological state is again the one mentioned in the propositional content, while expressives do not have a direction of fit. Collective agents are able to perform expressive speech acts. For example, when the government formally expresses regret for an action by one of its ministers. The point of such a speech act is to express this feeling of regret, and the expressed psychological state is the government’s collective feeling of regret with respect to the action of the minister. Notice that agents need not actually have these (collective) psychological states, the speech actcounts as the expression of these states. This is a general point that applies to all types of speech acts.

Declarations are the fifth and last type of speech acts Searle distinguishes. Their illocutio­nary point is to bring about a correspondence between the propositional content of the speech act and reality, i.e. to make true in the world what is said in the content of the act. The expressed psychological state is twofold, namely a belief and a desire, and the direction of fit is both words to world and world to words. Declarations can be performed by collective agents. For example, when the jury utters the verdict: “guilty”, the illocutionary point of the jury is to declare the subject to be guilty. Due to this act, he or she changes status from suspect to guilty person. Or, more precisely, he or she is both made a guilty person by this act and described as such. The expressed psychological states are the collective acceptance of the proposition that the suspect is guilty and the desire to change the suspect’s status into that of a guilty person. As we have seen with the other types of speech acts, collective declarations can have collective agents as speakers and as hearers. An example of the latter is when a soccer team is declared to be the best team of the year by a panel of journalists.

The picture that emerges from our discussion is that collective speech acts roughly fit Searle’s taxonomy of individual speech acts. The same types of speech acts can be distinguished using the three qualities of illocutionary point, expressed psychological state and direction of fit. I said ‘roughly’ because we encountered a problem with respect to the expressed psychological state in the case of assertives. For an individual assertive this may be a belief or an acceptance, for a collective assertive this is always the acceptance of a proposition. The problem can be solved by describing these two psychological states on a higher level of abstraction in the taxo­nomy. For example, both can be described as the ‘assent to a proposition’. By definition, then, this would cover the individual speaker’s commitment to the truth of the proposition, and the collective speaker’s commitment to the practical value of a proposition in practical reasoning. In my view such a solution is to a large extent a cosmetic, because it conceals rather than solves the initial ambiguity.

We are confronted now with a puzzle. On the one hand, collective speech acts are largely similar to individual speech acts in terms of the three main qualities of speech acts and the taxo­nomy based on them. On the other hand, we have found that collective speech acts are really different from and irreducible to individual speech acts (see Section 4). They aresuis generis acts. As we have seen the main reason for their irreducibility is that collective agents and collective intentionality cannot be reduced to individual agents and individual intentionality. The resulting puzzle shows, in my view, that an important quality of speech acts is missing, which should have been included in the construc­tion of Searle’s taxonomy. It is thekind of agency involved in speech acts. This quality is orthogonal to the illocutio­nary point and the expressed psychological state of a speech act, in that these two qualities allow for an individual and a collective reading. It is no surprise, then, that the resulting taxonomy of speech acts allows for individual as well as for collective agents, both as speaker and as hearer.

The addition of an extra quality or dimension to the basis on which the taxonomy of speech acts will be constructed has a rather fundamental impact. It converts Searle’s two-dimensional matrix into a three-dimensional one, as is shown below:

The third dimension of ‘kind of agency involved’ can have a number of values, given the typology of groups I have discussed in Section 2. At one extreme there is only an individu­al agent, at the other there is a structured group of the intentional type, with several other types of groups in between.

One may wonder whether Searle’s taxonomy already includes elements of collective agency, and thus a hidden third dimension after all. A case in point is Searle’s category of declarations. Though there may be exceptional cases of individual declarations - for example, when a speaker defines a term – most declarations require extra-linguistic institutions and thus collective intentionality and collective agency. And in many cases a speaker acts as the authorized delegate of an institution. Take, for example, the standard example of declaring somebody husband and wife. Though an individual civil servant performs this declaration, the official acts and can only act on behalf of an institution in such a case. It seems therefore more appropriate to say that the state has declared them husband and wife, rather than the civil servant. That makes the act a collective speech act. Or, to give another example, if a head of state declares war to another nation, it is not that particular individual who declares war, but the state he represents. Again, this seems to be clearly a collective speech act.

The construction of social reality

On Searle’s view (Searle 1995: Chapter 1), there are three building blocks of social reality: the imposition of function on entities that do not have that function prior to the imposition, collective intentionality and the distinction between constitutive and regulative rules. In a trivial (because definitional) sense every collective speech act creates a social fact, then, since it essentially involves collective intentionality

Social reality consists to a large extent of institutions and institutional facts. Institutional facts, according to Searle, are created by the collective imposition of status functions on entities according to the constitutive rule “X counts as Y in context C” (Searle 1995: Chapters 2 and 3). These status functions exist only because they are collectively imposed and cannot be achieved solely in virtue of the physics of these entities. Language is fundamental in this analysis in two ways: (i) there has to be asymbolic representation of the status function Y, since this function goes beyond the physics of the X; this symbolic representation has to be conventional and publicly understandable; (ii) the imposition is collective; this means, among other things, that communication will be involved in the process of the collective imposition of status functions.

Searle’s claim that language is “essentially constitutive of institutional reality” (1995: 59) can also be phrased in terms of the previous analysis of collective speech acts. The collective imposition of function obviously requires a collective subject, the rule “X counts as Y in context C” has the form of a declaration, and since the Y term needs to be symbolically represent­ed in a publicly understandable way, it will usually involve a linguistic act. Collective declara­tions have exactly the requisite properties. When the chairman of a jury says: “The suspect is guilty”, an institutional fact is created by a collective declaration and new deontic phenomena such as powers, rights, obligations and duties will have been created. Collective declarations, then, are essentially constitutive of institutional reality.

Conclusion

Up to now, speech act theory has had a blind spot for those speech acts that are performed by collective agents. The paradigm case has been an individual speaker addressing an individual hearer. I have shown in my paper that speech acts performed by collective speakers are suis generis acts. They cannot be reduced to individual speech acts. The analysis of their propositional content condition, their preparatory conditions, their sincerity condition, and their essential condition has shown important differ­ences with individual speech acts. Collective speech acts also involve preparatory conditions that are specifically related to their delegation and author­ization. The conse­quence of my analysis is that Searle’s taxonomy of speech acts has to be modified because an important classifying principle, the type of agency involved, is missing. Searle’s claim that language is essentially constitutive of institutional reality can be rephrased on the basis of the previous analysis: collective declarations are the stuff institutional reality is made of.[9]

REFERENCES

Bratman, M. (1993), “Practical Reasoning and Acceptance in a Context”,Mind 102, 1-15.

Cohen, J. (1992),An Essay on Belief and Acceptance . Oxford: Oxford University Press.

Crosby, J. (1990), “Speech Act Theory and Phenomenology”, in A. Burkhardt (ed)Speech Acts, Meaning, and Intentions; Critical Approaches to the Philosophy of John R. Searle, Berlin: De Gruyter, 62-88.

Gilbert, M. (2002), “Belief and Acceptance as Features of Groups”,ProtoSociology 16, 35-69.

Grice, H.P. (1957) “Meaning”,Philosophical Review 66, 377-388.

------ (1989)Studies in the Way of Words , Cambridge, Mass.: Harvard University Press.

Harré, R. (1997), “Crews, Clubs, Crowds, and Classes”, in J.D. Greenwood (ed.)The Mark of the Social, Lanham/London: Rowman & Littlefield, 199-211.

Hughes, J. (1984) ‘Group Speech Acts’,Linguistics and Philosophy 7, 379-395.

LePore, E. and Gulick, R.v. (eds.) (1991)John Searle and His Critics, Cambridge, Mass,: Basil Blackwell.

McDowell, J. (1991) “IntentionalityDe Re ”, in LePore, E. and Gulick, R.v. (eds.)John Searle and His Critics, 215-225.

Meijers, A. (1994)Speech Acts, Communication and Collective Intentionality: Beyond Searle's Individualism, Leiden.

----- (2002) 'Dialogue, Understanding and Collective Intentionality', in G. Meggle (ed.)Social facts and collective intentionality , Frankfurt: Hänsel, 225-255.

----- (2002) “Collective Agents and Cognitive Attitudes”,ProtoSociology 16, 70-85.

----- (2003) “Why Accept Collective Beliefs? Reply to Gilbert”,ProtoSociology 18, 377-388.

----- (2003) “Can Collective Intentionality be Individualized?”,American Journal of Economics and Sociology 62, 167-183.

Searle, J.R. (1969)Speech Acts; an Essay in the Philosophy of Language, Cambridge, UK: Cambridge University Press.

----- (1979)Expression and Meaning; Studies in the Theory of Speech Acts, Cambridge, UK: Cambridge University Press.

----- (1983)Intentionality, an Essay in the Philosophy of Mind, Cambridge, UK: Cambridge University Press.

----- (1989) “How Performatives Work”,Linguistics and Philosophy 12, 535-558.

----- (1990) “Collective Intentions and Actions”, inIntentions in Communication , P. Cohen, J. Morgan and M.E. Pollack (eds.), Cambridge (Mass.), 401–415.

----- (1991) “Reference and Intentionality”, in LePore, E. and Gulick, R. v. (eds.)John Searle and His Critics, 227-241

----- (1995)The Construction of Social Reality, New York: Free Press.

Taylor, C. (1980), “Critical Notice onLinguistic Behaviour by Jonathan Bennett”,Dialogue 19, 290-301.

Wray, K. B. (2001), “Collective Beliefs and Acceptance”,Synthese 129, 319-333.

Endnotes

[1]Eindhoven University of Technology, P.O Box 513, 5600 MB Eindhoven, the Netherlands. E-mail a.w.m.meijers@tue.nl.

[2]The only exception to my knowledge is Hughes (1984). I owe this reference to Margaret Gilbert. Hughes’ paper has been ignored in the subsequent literature. Much earlier, at the beginning of the 20^th century, Adolf Reinach developed in his phenomenology a proto-theory of speech acts for the legal domain, which included a discussion of speaking on behalf of a group. See Crosby (1990).

[3]See for a more elaborate account Meijers (2003). See also the discussion on intentionality de re between McDowell (1991) and Searle (1991).

[4]An exception may be those cases in which an individual speaker addresses a collective hearer, where the hearer is a crowd or a structured group of the physical type. Imagine that there is a fire in a building and that the porter orders the inhabitants via the loudspeakers in the rooms to leave the buil­ding as soon as possible. This collective speech act can probably be analyzed as the conjunc­tion of a number of individual speech acts, though even in this case there is the complicating factor that the speech act will become part of the public domain and that the hearers will have common knowledge about it. They know that the others know that they know that the others know [etc.] that the speech act has been performed.

[5]In his paper “A Taxonomy of Illocutionary Acts” (Searle 1979: Chapter 1), Searle lists at least twelve respects in which speech acts can differ.

[6]The distinction goes back to Jonathan Cohen (1992) and Michael Bratman (1993).

[7]See Meijers (2002) and (2003), Gilbert (2003) and Wray (2001). The view has become known as 'rejectionism' with respect to collective beliefs in the literature.

[8]See also the discussion in the literature mentioned in the previous footnote. Volumes 16 and 18 of the journal ProtoSociology are partly devoted to this issue.

[9]I would like to thank SavasTsohatzidis for stimulating comments on an earlier version of this paper.