THE METAPHYSICS OF SHARED ACTION: AN ANALYSIS BASED ON THE IDEA OF COMMITMENTS

In this paper I consider two examples of social interaction and try to figure out what is distinctive of each example by pursuing an examination of the nature of the commitments present in each situation. I discuss the ideas of pre-commitment and joint commitment and pursue an understanding of their roles in joint or shared action. The main argument of the paper is that the metaphysics of shared action is better understood through the proper kind of commitments that are constitutive of the action.


Thiago Monteiro Chaves
The Metaphysics of Shared Action: an analysis based on the idea of commitments intuitio ISSN Porto Alegre Vol.10 -Nº.1 Julho p. 107-122  1983-4012  2017   alone and having a car travel with someone.If I am a smoker, I surely will not restrict my freedom to smoke in the car if I am alone, but if I am accompanied I may bind myself to some rule of politeness.I must take the other's desires, intentions and beliefs into consideration.But shouldn't I take other's desires, intentions and beliefs into consideration when I am having a lonely car trip?Surely I do if I don't want to have a car accident.But merely coordination does not seem to be enough to capture the elements we want to capture when we talk about two friends having a car trip together or a business man travelling with a professional driver.A hint from Margaret Gilbert4 will help us here.When two people are travelling together, or doing anything together, they will promptly say that "we are travelling".By saying "we are travelling" when two friends are travelling together, we mean something different from what we mean when we say something about the drivers in São Paulo who are jammed in traffic."They" in "they are jammed in traffic" do not grasp the strong collective meaning which is grasped in "they are travelling together" in the friends' situation.There is a sense in which the coordinated actions of the drivers in São Paulo's traffic are taking place (individually) in parallel.The action of the group of friends and the action of the business man with his professional driver are taking place in a much more shared way5 .The fact that individual actions are coordinated is not enough to characterize the sharedness present in joint action.).But is it not the case that the silence of the class is a signal of agreement?And is it not the case that when someone says "Ok, let´s do it" I may infer from that sentence that she has agreed on something?
I do not think we can clearly make the distinction between tacit and express agreements appealing to Hobbes's distinction.Granted that there is a difference between inferring one's intention when one expresses oneself with words and inferring one's intentions when one expresses oneself by being in silence.But in both cases we are assuming that signalling one's intentions (through words or silence) is prior to the establishment of the agreement.I think that the failure to notice this has led some philosophers to put too much weight on the distinction between tacit and express agreements.I tend to think that this distinction does not even exist, since an inference of one's intention to agree has logical primacy on the it without thinking carefully about it, or that actors make them out of the blue.On the contrary, I mean that the actors do not make any agreement conceived as an act of the mind 7 .
With these preliminaries, I am able to present my argument concerning the kind and role of commitments in the two examples with which we opened our discussion.Contrary to the interpretation of agreements present in macro-social coordination, as exemplified by the lone drivers, the agreements present in the two inaugural examples must be conceived as actual agreements, as acts of one's mind.
These actual agreements seem to establish some sort of sharedness element which is absent in mere coordination of individual actions, as depicted by the lone (coordinating) drivers.So the question I propose to investigate is: can the distinction between the two situations exemplified by the inaugural examples be grasped by a difference between the commitments to these actual agreements?

II. Commitments in two examples of Social Interaction
In order to say what I think is the difference between the situation in which a group of friends travels together and the situation in which a business man travels in the company of a professional driver, I need to introduce some game-theory vocabulary.I will first analyse the second situation.My question is what is the kind of commitment involved in the agreements that generate the joint action.I think everyone will agree that the second example is a situation which can be accounted for in strategic terms.The business man wants someone to get him to somewhere, and the professional driver wants to get paid for a job.The "general sense of common interest", to quote Hume 8 , seems enough to generate the shared action.
But as we have seen, there is a crucial difference between this situation and the situation of the lone (coordinating) drivers.The lone drivers do not fulfil the condition necessary for an action to be labelled joint.Remind our definition of a joint or shared action as an action two or more people carry out together, in a manner that they can say about it that "we are doing such and such", and we can say about them that "they are doing such and such".I think no one would say of the lone drivers which are coordinating with each other on traffic rules that they are doing such and such.Do the lone drivers together form a we?Only if they join together in order to do something (like a protest, or to form a convoy, etc.).Do the lone drivers form a they?In a loose sense, yes, as in the sentence "they are jammed in traffic".But consider the following scenario.One driver stops at a gas station.After that, another driver who is driving on the same 7 I am here avowedly making reference to the tradition of spontaneous order of which Hayek is a major representative: "It would be no exaggeration to say that social theory begins with-and has an object only because of-the discovery that there exist orderly structures which are the product of the action of many men but are not the result of human design" (HAYEK, F. road stops at the same gas station.An observer at the gas station, who saw the latter driver overpassing the former some minutes ago, is asked by the manager if they were driving together, since the first driver forgot to pay for the gas and the second driver is still there.The observer answers that she does not know, but she "guess they are not".The observer may have got her opinion from many observations, like the fact that the drivers did not seem to be together because one overpassed another in a cold manner.In this important sense, they are not driving together.So, although the example of the business man travelling accompanied by the professional driver may be treated strategically, it is not clear if the mere coordination of actions explain what is socially shared in the situation.My guess is that the element of sharedness present in the business man example is captured by a proper sort of social contract established between the business man and the professional driver.This social contract is a special kind of convention, but not of the same form of the conventions of traffic rules which emerge (or at least could emerge) spontaneously as in the lone (coordinating) drivers' example.In binding oneself to the prior agreement, the business man carries out a certain decision in order to be able to make another decision in a posterior time (the same applying to the professional driver).This sort of contract or agreement depends on the mechanism of selfbinding or pre-commitment9 .I will try to elaborate this claim within the following analysis.Eventually, commitments understood as self-binding and as joint commitments will be distinguished.
It is undisputed that the business man and the professional driver both think of their agreements as binding.The driver may have proposed some price for his or her services; the business man may have promptly accepted or tried to negotiate.But they must have reached an agreement and surely find it binding.If this were not true, why would they bother to even think of making contracts?In order to understand what is going on here, we may assume that prior to the agreement or negotiation, both the business man and the driver are in a situation which they rank as less preferable than the situation in which they form an agreement.Let me call this situation the state of nature.It is important to note that the preferences of the actors in the state of nature relative to the example of the lone drivers may generate a convention in Lewis's sense10 , but the preferences of the actors in the state of nature relative to the example of the business man cannot generate a convention.This fact must not be surprising since we saw that the agreements involved in the example of the lone drivers are spontaneous, as opposed to actual agreements.The example of the business man seems to be an instance of the prisoner's dilemma, and even though agreements are possible and binding by a mechanism of pre-commitment or self-binding.
An interesting way of putting this view is to ask if the preferences of the business man and the driver in the state of nature permit a strategic equilibrium to occur with a Pareto-optimal combination of strategies.A combination of strategies is an equilibrium if each agent has done his best given the strategies followed by other agents11 .A combination of strategies is said to be Pareto-optimal if there is no other combination which is weakly preferred by any agent and strictly preferred by at least one agent12 .If a combination of strategies is not Pareto-optimal, it is said to be Pareto-inferior 13 .If an equilibrium that is also a Pareto-optimal combination cannot occur, so a convention in strategies W,W constitutes an equilibrium.This cannot happen in a convention, and that is precisely one of the characteristics that differentiates a social contract from a convention: in the former, the Paretooptimal combination of strategies cannot be an equilibrium; on the other hand, in a convention the Paretooptimal combination of strategies constitutes always a kind of equilibrium.In this situation, it pays for each agent to choose P, if the other is also choosing P. Eventually both agents will choose P, since it is an equilibrium and a Pareto-optimal combination of strategies.An instance of this matrix could be Hume's example of two men pulling the oars of a boat.If both men are not pulling the oars in a coordinated way, they are in a worst situation than if they pull the oars in a coordinated way.A convention to row in an efficient way will emerge spontaneously since the equilibrium point is Pareto-optimal.
If Hobbes's state of nature is a prisoner's dilemma as depicted by the first matrix, the perpetual conflict inherent of that situation is due to the dominance of the strategy of keeping war.External coercion happens to be the only solution for achieving peace, since contracts, although possible, are not binding.
Note that a social contract in Hobbes's terms cannot be a convention because although the agents would prefer achieving peace (through a certain kind of commitment to follow a social contract), each of them would prefer lone deviation from the social contract (which is the dominant strategy in the prisoner's dilemma game).In a convention the opposite just happens to be the case: each agent prefers general conformity to conformity of all but himself.Lone deviation is a poor strategy in a coordination game when a convention is occurring.This intuition was first captured by Hume when he thought of conventions as a "general sense of common interest" 16 .
But what if some sort of equilibrium that happens to be a Pareto-optimal combination of strategies be achieved in the state of nature relative to a social contract?If we put time in the story, this could be the case.And that is what Hobbes seems to have in mind when he disapproves what the Fool has to say about keeping one's promises.In his words: agreement itself.So it seems, actually, that the manner agreements are made cannot fully explain what sort of sharedness element is present in our inaugural examples and absent in the picture of the lone (coordinating) drivers.We must pursue that difference in the kind of agreements and the role they play in social interactions of the sort we are considering.Can the sharedness element that is present in the inaugural examples and absent in the description of the lone drivers coordinating on traffic rules be found in the kind of agreements that were made in each situation?I think it can.In the inaugural examples, the agreements are taking place as attitudes of the actors.I mean by attitude any sort of stance one's mind has towards something.Believing and intending are paradigmatic examples.As an attitude, an agreement takes place in time and space.In the picture of the lone (coordinating) drivers, agreements are not attitudes in the mind of the actors, but they are a result of a complex interrelation of the actors' attitudes.Let me call the first kind of agreement, which I think is the important sharedness element present in the first couple of situations, an actual agreement.And let me call the second kind of agreement, which I think is present in the situation of the lone (coordinating) drivers, a spontaneous agreement.In calling an agreement spontaneous I do not mean that the actors makeThiago Monteiro ChavesThe Metaphysics of Shared Action: an analysis based on the idea of commitments intuitio

Matrix 1: Preferences in the state of nature relative to a social contract
Lewis's sense cannot emerge spontaneously from a state of nature, and a state of nature relative to a social contract must be distinguished from a state of nature relative to a convention.In order to see why this is the case, consider the preferences of two agents in a Hobbesian state of nature.In a Hobbesian state of nature, men find themselves in a permanent state of conflict.It has been widely assumed by commentators that the choices and preferences of men in this state of nature match those of a prisoner's dilemma matrix 14 .Let us assume that each person in the state of nature has two courses of action: seeking peace (P) or keeping war (W).And let us assume that the preferences for each person hold as follows: a > b > c > d, transitivity being the case.So, in a Hobbesian state of nature we have the matrix 15 number 1:The strategy of keeping war strictly dominates the strategy of seeking peace, so whatever one agent does, it pays for the other to choose war.What makes this state of nature a state of stable conflict is the fact that, although the agents would prefer the pair of strategies P,P; the Pareto-inferior pair of