Click here and press the right key for the next slide (or swipe left)

also ...

Press the left key to go backwards (or swipe right)

Press n to toggle whether notes are shown (or add '?notes' to the url before the #)

Press m or double tap to see a menu of slides

\title {Naturalising Joint Actions \\ }

\maketitle

Naturalising Joint Actions

\def \ititle {Naturalising Joint Actions}
\def \isubtitle {}
\begin{center}
{\Large
\textbf{\ititle}: \isubtitle
}

\iemail %
\end{center}
Readings refer to sections of the course textbook, \emph{Language, Proof and Logic}.

\section{Collective Goals}

\section{Collective Goals}

In virtue of what do actions involving multiple agents ever have collective goals?

Some actions involving multiple agents are purposive in the sense that
among all their actual and possible consequences,
there are outcomes to which they are directed
and the actions are collectively directed to this outcome
so it is not just a matter of each individual action being directed to this outcome.
In such cases we can say that the actions have a collective goal.
To illustrate, suppose that Ayesha takes a glass and holds it up while Beatrice pours prosecco. Ayesha might say, truthfully, ‘The goal of our actions was not to soak Zach's trousers in sparkling wine but only to fill this glass.’ Suppose this statement is true not because, or not only because, the goal of Ayesha’s and Beatrice’s individual actions was to fill the glass: the goal is one to which their actions are collectively directed. What then could make Ayesha’s statement true?
Here are two sentences:

The tiny drops fell from the bottle.

- distributive

The tiny drops soaked Zach's trousers.

- collective

The first sentence is naturally read *distributively*; that is, as specifying something that each drop did individually. Perhaps first drop one fell, then another fell.
But the second sentence is naturally read *collectively*. No one drop soaked Zach's trousers; rather the soaking was something that the drops accomplised together.
If the sentence is true on this reading, the tiny drops' soaking Zach's trousers is not a matter of each drop soaking Zach's trousers.
Now consider an example involving actions and their outcomes:

Their thoughtless actions soaked Zach's trousers. [causal]

- ambiguous

This sentence can be read in two ways, distributively or collectively. We can imagine that we are talking about a sequence of actions done over a period of time, each of which soaked Zach's trousers. In this case the outcome, soaking Zach's trousers, is an outcome of each action.
Alternatively we can imagine several actions which have this outcome collectively---as in our illustration where Ayesha holds a glass while Beatrice pours. In this case the outcome, soaking Zach's trousers, is not necessarily an outcome of any of the individual actions but it is an outcome of all of them taken together. That is, it is a collective outcome.
(Here I'm ignoring complications associated with the possibility that some of the actions collectively soaked Zach's trousers while others did so distributively.)
Note that there is a genuine ambiguity here. To see this, ask yourself how many times Zach's trousers were soaked. On the distributive reading they were soaked at least as many times as there are actions. On the collective reading they were not necessarily soaked more than once. (On the distributive reading there are several outcomes of the same type and each action has a different token outcome of this type; on the collective reading there is a single token outcome which is the outcome of two or more actions.)
Conclusion so far: two or more actions involving multiple agents can have outcomes distributively or collectively. This is not just a matter of words; there is a difference in the relation between the actions and the outcome.
Now consider one last sentence:

The goal of their actions was to fill Zach's glass. [teleological]

Whereas the previous sentence was causal, and so concened an actual outcome of some actions, this sentence is teleological, and so concerns an outcome to which actions are directed.

- also ambiguous

Like the previous sentence, this sentence has both distributive and collective readings. On the distributive reading, each of their actions was directed to an outcome, namely soaking Zach's trousers. So there were as many attempts on his trousers as there are actions. On the collective reading, by contrast, it is not necessary that any of the actions considered individually was directed to this outcome; rather the actions were collectively directed to this outcome.
Conclusion so far: two or more actions involving multiple agents can be collectively directed to an outcome.
Where two or more actions are collectively directed to an outcome, we will say that this outcome is a *collective goal* of the actions. Note two things. First, this definition involves no assumptions about the intentions or other mental states of the agents. Relatedly, it is the actions rather than the agents which have a collective goal. Second, a collective goal is just an actual or possible outcome of an action.
Now there is a key difference between the ideas these last two sentences convey.
Because we are familiar with the idea of common effects, it is obvious how two or more actions can have an outcome collectively. You hold the bottle in one place and I hold the glass in a another, too far away, and sure enough Zach's trousers get soaked as a common effect of our actions. If you had moved the bottle slightly further towards me or if I had moved the glass slightly further towards you, everything would have been fine. This is all that is required for soaking Zach's trousers to be an outcome that our actions have collectively.
By contrast, it is less obvious, I think, how two or more agents' actions might come to be collectively directed to a single outcome.
So this is our question: in virtue of what can two or more actions involving multiple agents have a collective goal?

In virtue of what do actions involving multiple agents ever have collective goals?

As what Ayesha and Beatrice are doing---filling a glass together---is a paradigm case of joint action, it might seem natural to answer the question by invoking a notion of shared (or collective') intention. Suppose Ayesha and Beatrice have a shared intention that they fill the glass. Then, on many accounts of shared intention,
the shared intention involves each of them intending that they, Ayesha and Beatrice, fill the glass; or each of them being in some other state which picks out this outcome.
The shared intention also provides for the coordination of their actions (so that, for example, Beatrice doesn't start pouring until Ayesha is holding the glass under the bottle). And coordination of this type would normally facilitate occurrences of the type of outcome intended. In this way, invoking a notion of shared intention provides one answer to our question about what it is for some actions to be collectively directed to an outcome.
Are there also ways of answering the question which involve psychological structures other than shared intention? In this paper we shall draw on recent discoveries about how multiple agents coordinate their actions to argue that the collective directedness of some actions to an outcome can be explained in terms of a particular interagential structure of motor representations. Our actions having collective goals is not always only a matter of what we intend: sometimes it constitutively involves motor representation.

Preliminary: Just One Agent

\section{Preliminary: Just One Agent}

\section{Preliminary: Just One Agent}
Let me start by stepping back and consider an individual action. Consider what Ayesha and Beatrice do---their attempt to fill Zach's glass by one pouring prosecco while the other holds a glass---but now imagine that one person does both parts of the action.

In virtue of what do actions involving just one agent ever have collective goals?

- intention

A standard answer involves appeal to intention. For an intention identifies an outcome, coordinates some of the agent’s actions around that outcome, and coordinates them in such a way that, normally, the coordination would facilitate the occurrence of the outcome. Thus someone's intention to wash the dishes might, for example, constrain the order of her actions so that that the prosecco glasses are done before greasy pans. In this way, appealing to an intention concerning an outcome and its role in coordinating actions can sometimes allow us to explain in virtue of what several actions are collectively directed to that outcome. Are there other facts in virtue of which multiple actions involving just one agent can be collectively directed to an outcome?

- motor representation

Some philosophers have argued that some actions have collective goals in virtue of motor representation rather than intention. It will be helpful to outline these arguments since my aim today is an attempt to generalise such arguments from one agent to multiple agents.

motor representations

can

represent outcomes

Let us start by reviewing some almost uncontroversial facts about the coordinating role of motor representations for two or more actions involving just one agent. Some motor representations represent outcomes to which actions could be directed, outcomes such as the grasping of a glass or the pressing of a switch rather than mere bodily configurations and joint displacements \citep{hamilton:2008_action, cattaneo:2009_representation}.

and

trigger planning-like processes

Relatedly, motor representations can trigger processes which are like planning in some respects. These processes are planning-like in that they involve starting with representations of relatively distal outcomes and gradually filling in details, resulting in motor representations that can be hierarchically arranged by the means-end relation \citep{bekkering:2000_imitation,grafton:2007_evidence}. Some processes triggered by motor representations are also planning-like in that they involve meeting constraints on the selection of means by which to bring about one outcome that arise from the need to select means by which, later, to bring about another outcome \citep{jeannerod_motor_2006, zhang:2007_planning, rosenbaum:2012_cognition}.
Consider this case. An agent fills a glass by holding it in one hand, holding the bottle in the other, bringing the two together and pouring from the bottle into the glass.
[*demonstrate].
It’s a familiar idea that motor representations of outcomes resemble intentions in that they can trigger processes which are like planning in some respects. These processes are like planning in that they involve starting with representations of relatively distal outcomes and gradually filling in details, resulting in a structure of motor representations that can be hierarchically arranged by the means-end relation \citep{bekkering:2000_imitation, grafton:2007_evidence}. Processes triggered by motor representations of outcomes are also planning-like in that they involve selecting means for actions to be performed now in ways that anticipate future actions \citep{jeannerod_motor_2006,zhang:2007_planning,rosenbaum:2012_cognition}.
Now in this action of moving a mug, there is a need, even for the single agent, to coordinate the exchange between her two hands. (If her action is fluid, she may proactively adjust her left hand in advance of the mug’s being lifted by her right hand \citep[compare][]{diedrichsen:2003_anticipatory,hugon:1982_anticipatory, lum:1992_feedforward}.) How could such tight coordination be achieved? Part of the answer involves the fact that motor representations and processes concerning the actions involving each hand are not entirely independent of each other. Rather there is a plan-like structure of motor representation for the whole action and motor representations concerning actions involving each hand are components of this larger plan-like structure. It is in part because they are components of a larger plan-like structure that the movements of one hand constrain and are constrained by the movements of the other hand.
But how is any of this relevant to our question about in virtue of what several actions can have a collective goal?
Recall that intentions ground the collective directedness of some actions to an outcome by representing that outcome, coordinating the actions, and coordinating them in such a way that, normally, the coordination would facilitate the occurrence of the outcome. We have seen that a parallel claim is true of motor representations. Like intentions, some motor representations represent outcomes to which several actions could be collectively directed, outcomes such as the washing of a glass or the emptying of a bottle. Like intentions, some motor representations also play a role in coordinating several actions by virtue of their role as elements in plan-like hierarchies. And, like intentions, some motor representations coordinate these actions in a way that would normally facilitate the occurrence of the outcome represented. So, arguably, anyone who accepts the standard story about intention grounding the collective directedness of several actions to an outcome should also accept a similar story about purposive action and motor representation.

Parallel Planning

\section{Parallel Planning}

\section{Parallel Planning}
A representation or plan is \emph{agent-neutral} if its content does not specify any particular agent or agents; a planning process is agent-neutral if it involves only agent-neutral representations.
Practical vs theoretical reasoning: ‘The mark of practical reasoning is that the thing wanted is \emph{at a distance} from the immediate action, and the immediate action is calculated as a way of getting or doing or securing the thing wanted’ \citep[p.\ 79]{Anscombe:1957ln}. See also \citet[p.\ 1]{millgram:2001_practical}: ‘Practical reasoning is reasoning directed towards action: figuring out what to do, as contrasted with figuring out how the facts stand.’
Some agents each \emph{individually make a plan for all the agents' actions} just if: there is an outcome; each agent individually, without discussion, communication or prior arrangement, plans for that outcome; and each agent’s plan specifies roles for herself and all the other agents.
Our planning is \emph{parallel} just if you and I are each planning actions that I will eventually perform and actions that you will eventually perform, where the resulting plans non-accidentally match.
Two or more plans \emph{match} just if they are the same, or similar enough that the differences don't matter in the following sense. First, for a particular agent's plan, let the \emph{self part} be those steps concerning what will be the agent's own actions and let the \emph{other part} be the other steps. Now consider what would happen if, for a particular agent, the other part of her plan were as nearly identical to the self part (or parts) of the other's plan (or others' plans) as psychologically possible. If the agent's self part would not be significantly different, let us say that any differences between her plan the other's (or others') are not relevant for her. Finally, if for some plans the differences are not relevant for any of the agents, then let us say that the differences don't matter.

How to get from individual to joint planning-like processes?

Consider how the trick has been turned in discussions of shared intention.
In discussion of joint action and related phenomena, much emphasis has beeen given to interconnected planning. For example, Bratman writes that shared agency consists, at bottom, ...

shared intentional agency consists, at bottom, in interconnected planning agency of the participants.’

(Bratman 2011, p. 11)

Interconnected planning is planning where facts about your plans feature in my plans & conversely.
I don't think there's anything wrong with the idea of interconnected planning as such. But clearly it isn't something we can appeal to in trying to understanding motoric planning-like processes concerning others' actions. After all, interconnected planning requires that facts about other's plans feature in plans; this is fine where intentions are concerned since they are propositional attitudes and so can be arbitrarily nested. But since motor representations are not nestable (you can't have motor representations representing motor representations), interconnection seems like the wrong it.
Is there an alternative? To see that there might be, consider a simple activity like moving a tbale.
Suppose you and I are tasked with moving this table through that door. In doing this, must my plan take into account facts about your intentions as well as about the weight of the table, width of the door etc? This case has some special features: (i) there is a single most salient route for the table given our objective; (ii) there is a single most salient way of dividing up the roles between us. I suggest that, in this situation, neither of us needs to form a plan involving the others' intentions. The situation makes this redundant. All we have to plan is how two people in our situations should move the table through the door.
To a first approximation, then, what the situation seems to call for is not that our plans are interconnected but rather that we each make a plan for the table-moving action as a whole.
This is inspiration for the view that we might understand how to generalise from individual planning to multi-agent planning by reflecting on parallel rather than interconnected planning ...

shared intentional agency consists, at bottom, in interconnected planning agency of the participants.’

(Bratman 2011, p. 11)

parallel planning

You plan my actions as well as yours, and I do likewise.

In parallel planning, I plan all of our actions and you do the same.
I want to suggest that shared agency sometimes requires only parallel, and not interconnected planning.
Some of you are probably already thinking that the very idea of parallel planning is incoherent, and I will face up to this objection.
But first I want you to suspend disbelief and consider how parallel planning could enable us to coordinate our actions and our plans ...
Suppose you and I are parents about to change our baby's nappy.
This involves preparing the baby and preparing the nappy.
You're holding the baby and I'm nearest the pile of clean nappies, so there's a single most salient way of dividing the task between us.
Preparing the baby is, of course, a complex action ...
Now there are relational constraints on how the baby and nappy should be prepared; how you clean constrains and is constrained by how I prepare the clean nappy (because we don't want to get pooh on it).
\textbf{How do we meet these relational constraints?} The fact that I have a plan for the whole thing and so do you, and the fact that these plans are identical or similar enough that the differences don't matter means that your plan for your actions is constrained by your plan for my actions, which is my plan for my actions. So thanks to our parallel planning---to the fact that we each plan the whole action---your plan for your actions is indirectly constrained by my plan for my actions; and conversely. So: in parallel planning, we meet these relational constraints not by thinking about each other's intentions but by planning each other's actions.
There's just one tiny problem. In supposing that we both make a plan for the whole action, I'm implying, of course, that we each make plans for actions that are not our own. And this seems incoherent, unless perhaps we (the agents performing the action) are irrational or ignorant. It seems incoherent because the elements of plans we make are intentions; so, apparently, in making a plan for your action I would end up intending your actions. But I can't intend your actions, I suppose. What can we do?
Let me repeat the problem. Here I seem to face a dilemma. [First horn: not intention] It seems that I can't intend it in the ordinary sense because then I'd be intending its parts, and some of its parts are actions that, I know, you will eventually perform. So, apparently, I'm blocked from intending the plan [Second horn: not belief] On the other hand, if I merely entertain a belief about the plan --- if, for instance, I merely believe that the plan identifies a way we could achieve some goal --- then it's unclear how I could be acting on it at all.

open-ended intentionsTo solve this dilemma we need to appeal to some ways in which intentions can be open-ended.: whatIt's a familiar idea that intentions can be open-ended with respect to what is intended.For instance, you can intend to visit the Weinachtsmarkt without intending to do so on any particular day., and whoIt's also true that intentions can be open-ended with respect to who will act on them.Consider a couple planning some tasks at the start of the weekend: they need to buy bread, to clean the bath, ... At this point, their intention is that one or both of them will do each of these things, but there is no further specification concerning who will act. Now you might say that you can't intend something without settling who will act. But this seems wrong given that (i) the couple's attitudes are practical, and (ii) generate requirements concerning agglommeration. (Even before it's determined who will do what, I know that I'm not going to be able to spend the afternoon in the pub.)[*skip] You might also say that open-ended intentions generate pressure to filling in details. This is true, but the details are not always filled by further intentions. At some point intentions give out and we just act. The point of appealing to the table-moving example was that here there is no need for the intention to specify the agents.I want to suggest that appeal to the open-endedness of intentions will help with the dilemma I had.The problem was, what attitude could I have to another's actions?

The attitude I can have is this: with respect to the whole plan, I intend that we implement it.
And with respect to its components, I intend that you or I or we do it.
So my intentions don't specify who will do what. But they don't need to, because this is already adqeuately specified by the fact that you're holding the baby and I'm nearest the clean nappies.
So in this case what determines who does what are the constraints, not the intentions.
This is why, despite appearances, I think the notion of parallel planning is coherent. Without irrationality or ignorance, it is possible for us each to plan all of our actions, yours and mine, and to act on these plans. In doing so we achieve coordination and manifest collective intentionality not by thinking about each other's plans but, more directly, by planning each other's actions.
Switch from intentions to motor representations.

shared intentional agency consists, at bottom, in interconnected planning agency of the participants.’

(Bratman 2011, p. 11)

parallel planning

You plan my actions as well as yours, and I do likewise.

The question was how to generalise from planning-like processes concerning actions involving just one agent to planning-like processes concerning actions involving two or more agents.
My suggestion is that we can generalise by appeal to something like parallel planning rather than by appeal to anything like interconnected planning.

Two (Or More?) Agents

\section{Two (Or More?) Agents}

\section{Two (Or More?) Agents}

Any evidence for parallel planning-like motor processes?

1. Each represents a single outcome motorically, and
2. in each agent this representation triggers planning-like processes
3. concerning all the agents' actions, with the result that
4. coordination of their actions is facilitated.
What do we need? (i) Evidence that a single outcome to which all the actions are directed is represented motorically. (ii) Evidence that this triggers planning-like processes, (iii) where these concerning all the agents' actions, and (iv) the existence of such representations facilitates coordination of the agents' actions.

Kourtis et al (2014, figure 1c)

I think we're a long way from having a large body of converging evidence, but there is some that points in this direction. One of the most relevant experiments is this one by \citet{kourtis:2014_attention}.
They contrasted a simple joint action involving two agents clinking glasses.

Kourtis et al (2014, figure 1c)

Kourtis et al (2014, figure 1c)

Kourtis et al (2014, figure 1b)

Here's the procedure.

joint

individual-bimanual

individual-unimanual

This is fine but what are we going to measure?

Kourtis et al (2014, figure 1c)

CNV

This is a signal of motor preparation for action which is time-locked to action onset. In previous research, Kourtis et al show (i) that the CNV occurs when joint action partners act, suggesting that when acting together we represent others' actions motorically as well as our own \cite{kourtis:2012_predictive}; and (ii) (roughly) a stronger CNV occurs in relation to actions of others one is engaged in joint action than in relation to actions of others one is merely observing \cite{kourtis:2010_favoritism}.
Kourtis et al hypothesised that in actions like clinking glasses, A single outcome represented is motorically, which triggers planning-like processes concerning all the agents' actions. This leads to the prediction that the CNV in joint action will resemble that occurring in bimanual action more than that occuring in unimanual action.

Kourtis et al (2014, figure 4a)

... and this is exactly what they found.

Conjecture

Sometimes when two or more actions involving multiple agents are, or need to be, coordinated:

1. Each represents a single outcome motorically, and
2. in each agent this representation triggers planning-like processes
3. concerning all the agents' actions, with the result that
4. coordination of their actions is facilitated.
There is more evidence for this conjecture than I have given here, but there is not a lot more converging evidence. This is a conjecture that we hope will be tested further rather than something we take to be established already.

prediction

Need to emphasise that this conjecture is a source of predictions. Here's one.

Conjecture

Sometimes when two or more actions involving multiple agents are, or need to be, coordinated:

1. Each represents a single outcome motorically, and
2. in each agent this representation triggers planning-like processes
3. concerning all the agents' actions, with the result that
4. coordination of their actions is facilitated.
What about coordination? There is a little bit of direct evidence for this that I won't mention. But I do want to take you through why the interagential structure of motor representation might in theory result in the agents actions being coordinated.
Earlier we considered what is involved in performing an ordinary, individual action, where an agent fills a glass from a bottle, taking one in each hand and moving them in a carefully coordinated way. Compare this individual action with the same action performed by two agents as a joint action. One agent takes the glass while the other takes the bottle. The joint action is like the individual action in several respects.
First, the goal to which the joint action is directed is the same, namely to move the mug from here to there.
Second, there is a similar coordination problem---the agents’ two hands have to meet.
And, third, the evidence we have mentioned suggests that in joint action, motor representations and processes occur in each agent much like those that would occur if this agent were performing the whole action alone.

Suppose the agents' planning-like motor processes are similar enough that, in this context, they will reliably produce approximately the same plan-like structures of motor representations.
Then having a single planning-like motor process for the whole joint action in each agent means that
\begin{enumerate}
\item in each agent there is a plan-like structure of motor representations concerning each of the others’ actions,
\item each agent's plan-like structure concerning another's actions is approximately the same as any other agent's plan-like structure concerning those actions,
\item each agent's plan-like structure concerning her own actions is constrained by her plan-like structures concerning the other’s actions.
\end{enumerate}
So each agent’s plan-like structure of motor representations concerning her own actions is indirectly constrained by the other agents' plan-like structures concerning their own actions
by virtue of being directly constrained by her plan-like structures concerning their actions.
In this way it is possible to use ordinary planning-like motor processes to achieve coordination in joint action.
What enables the two or more agents' plan-like structures of motor representations to mesh is not that they represent each other's plans but that they processes motorically each other's actions and their own as parts of a single action.

So how does the joint action differ from the corresponding individual action?
There are at least two differences.
First, we now have two plan-like structures of motor representations because in each agent there is a planning-like motor process concerning the whole action.
These two structures of motor representations have to be identical or similar enough that the differences don’t matter for the coordination of the agents’ actions---let us abbreviate this by saying that they have to \emph{match}.
The need for matching planning-like structures is not specific to joint action;
it is also required where one agent observing another is able to predict her actions thanks to planning-like motor processes concerning the other’s actions (we mentioned evidence that this occurs above).
A second difference between the joint action and the individual action is this.
In joint action there are planning-like motor processes in each agent concerning some actions which she will not eventually perform.
There must therefore be something that prevents part but not all of the planning-like motor process leading all the way to action.
Exactly how this selective prevention works is an open question.
We expect bodily and environmental constraints are often relevant.
There may also be differences in how others’ actions are processed motorically \citep[compare][]{novembre:2012_distinguishing}.
\footnote{\citep[p.\ 2901]{novembre:2012_distinguishing}: 'in the context of a joint action—the motor control system is particularly sensitive to the identity of the agent (self or other) of a represented action and that (social) contextual information is one means for achieving this distinction'}
And inhibition could be involved too \citep[compare][]{sebanz:2006_twin_peaks}.
My proposal, then, is this. In both practical reasoning and motorically, sometimes agents are able to achieve coordination for joint action not by representing each others’ plans but by treating each other's actions and their own as if they were parts of a single action.
So perhaps joint action is not always only a matter of intention, knowledge or commitment: perhaps sometimes joint action constitutively involves motor representation.

collective goals

So far this has all been about coordination. But our question was about collective goals.

In virtue of what do actions involving multiple agents ever have collective goals?

Suppose Ayesha and Beatrice have a shared intention that they fill the glass. Then, on many accounts of shared intention,
the shared intention involves each of them intending that they, Ayesha and Beatrice, fill the glass; or each of them being in some other state which picks out this outcome.
The shared intention also provides for the coordination of their actions (so that, for example, Beatrice doesn't start pouring until Ayesha is holding the glass under the bottle). And coordination of this type would normally facilitate occurrences of the type of outcome intended. In this way, invoking a notion of shared intention provides one answer to our question about what it is for some actions to be collectively directed to an outcome.
What we've just seen is that a parallel answer can be given by appeal to i.a.s.m.r. ...

From Collective Goals to Joint actions

\section{From Collective Goals to Joint actions}

\section{From Collective Goals to Joint actions}
Sometimes there are two or more actions involving multiple agents where the actions have a collective goal while nevertheless not comprising a joint action. Here is an example (rope block case).

collective goal = joint action ? no !

collective goal grounded by i.a.s.m.r. + X = joint action? ...

Even if this isn't right maybe i.a.s.m.r. plus some further ingredient will give us a notion of joint action. Let me finish by exploring this possibility.

Simple Account

(Intentional) joint action occurs when there is an act-type, φ, such that each of several agents intends that they, these agents, φ (and their intentions are appropriately related to their actions).

*example: Ayesha and Beatrice must each intend that they ...
\textbf{Simple Account}
(Intentional) joint action occurs when there is an act-type, φ, such that each of several agents intends that they, these agents, φ.
Counterexample to the Simple Account.
Contrast two friends walking together in the ordinary way, which is a paradigm case of collective agency, with a situation where two gangsters walk together but each is forcing the other. It works like this: Gangster 1 pulls a gun on Gangster 2 and says: “let’s walk” But Gangster 2 does the same thing to Gangster 1 simultaneously. This is walking together in the Tarrantino sense, and clearly not a case of joint action. At least it’s not joint action unless the central event of of Reservoir Dogs is also a case of joint action.
Since in this case there is something which all the agents involved intend, it seems that our being involved in a joint action can't be a matter only of there being something such that we each intend that we do it together.
Bratman’s brilliant idea for avoiding this sort of problem is to suggest that we don’t just each intend the action but rather we each intend to act by way of the other's intentions.
We can put this by saying that our intentions must interlock: mine specify yours and yours mind.
Now this appeal to interlocking intentions enables Bratman to avoid counterexamples like the Tarantino walkers; if I intend that we walk by way of your intention that we walk, I suppose can't rationally also point a gun at you and coerce you to walk.

‘each agent does not just intend that the group perform the […] joint action.

‘Rather, each agent intends as well that the group perform this joint action in accordance with subplans (of the intentions in favor of the joint action) that mesh’

(Bratman 1992: 332)

each agent does not just intend that the group perform the […] joint action. Rather, each agent intends as well that the group perform this joint action in accordance with subplans (of the intentions in favor of the joint action) that mesh' \citep[p.\ 332]{Bratman:1992mi}.
Our plans are \emph{interconnected} just if facts about your plans feature in mine and conversely.
‘shared intentional [i.e.\ collective] agency consists, at bottom, in interconnected planning agency of the participants’ \citep{Bratman:2011fk}.
In making this idea more precise, Bratman proposes sufficient conditions for us to have a shared intention that we J ... ... the idea is then that an intentional joint action is an action that is appropriately related to a shared intention.

We have a shared intention that we J if

‘1. (a) I intend that we J and (b) you intend that we J

‘2. I intend that we J in accordance with and because of la, lb, and meshing subplans of la and lb; you intend [likewise] …

‘3. 1 and 2 are common knowledge between us’

(Bratman 1993: View 4)

\begin{minipage}{\columnwidth}
\emph{Bratman’s claim}. For you and I to have a collective/shared intention that we J it is sufficient that:
\begin{enumerate}[label=({\arabic*}),itemsep=0pt,topsep=0pt]
\item (a) I intend that we J and (b) you intend that we J;
\item I intend that we J in accordance with and because of la, lb, and meshing subplans of la and lb; you intend that we J in accordance with and because of la, lb, and meshing subplans of la and lb;
\item `1 and 2 are common knowledge between us' \citep[View 4]{Bratman:1993je}
\end{enumerate}
\end{minipage}

the simple account revised

I want to suggest there might be another possibility here. Let's go back to the simple account ...

(Intentional) joint action occurs when there is an act-type, φ, such that each of several agents intends that they, these agents, φ (and their intentions are appropriately related to their actions).

φ comprises two or more actions involving multiple agents where the actions have a collective goal in virtue of an interagential structure of motor representation.

I think this gets around the Tarrantino walkers counterexample. (Because if you have the iasmr you don't need a gun!)
But it leaves us with another problem ...
People can't intend phi under this description unless they can identify the interagential structure of motor representation we have been characterising.
So we have to give up on this idea as it stands. But maybe we can modify it.

interagential structure of motor representation grounds collective goal

> experience of act type

> concept of act type

φ is an act-type we know through experiences arising from interagential structure of motor representations grounding collective goals.

Compare individual action. We know what it is to grasp by virtue of experiences created by motor representations. ...

conclusion

In conclusion, my question was about in virtue of what ... collective goals. Many researchers have answered this question by appeal to one or another kind of shared intention. What we have shown is that in some cases the right answer to the question involves motor representation rather than intention. And, more tenatively, this suggests that there may be a role for motor representation in characterising joint action. Perhaps understanding joint action and shared intention requires appeal not just to intention, commitment and the rest, but also to motor representation.

Sharing a Smile

\section{Sharing a Smile}

\section{Sharing a Smile}
[THIRD POINT: smiling is a goal-directed action, the goal of which is to smile that smile]
My topic is sharing a smile. But first think about ordinary, individual actions like genuine smiles.
What distinguishes a genuine smile from a muscle spasm or the exhalation of wind?
I want to suggest that it's this: the smile is a goal-directed action where the goal is to simile that smile.
But why think of the smile as goal-directed? Because smiling the smile requires considerable motor coordination: it’s not a matter of simple muscle contractions but more like the production of a phonetic gesture where context affects what is needed to realise the smile.
Further, like grasping an object or articulating a particular phoneme, it is an action that can be realised by different bodily movements in different contexts.
This is why I put slides of two quite different but both genuine smiles.
[Objection:]
Now you might say that the smile can't be goal-directed because is isn't explicable by appeal to belief, desire and intention
This is because the genuine smile is spontaneous and not something that can be produced at will (although it could probably be inhibited, at least to some extent); after all, this is what distinguishes the genuine from the polite smile.
\footnote{
From web source: The Duchenne smile involves both voluntary and involuntary contraction from two muscles: the zygomatic major (raising the corners of the mouth) and the orbicularis oculi (raising the cheeks and producing crow's feet around the eyes). The zygomatic major can be voluntarily contracted but many people can't voluntarily contract the orbicularis oculi muscle.
}
So now we might be tempted by the view that a smile is merely caused by an emotion in the way that gasses can cause you to burp.
Maybe there are smiles like this, but some genuine smiles are sustained.
And what sustains them is a process of controll
How could this be if such smiles are not consequences of beliefs, desires and intentions?
I think a reasonably natural view here is to think that part of what makes an event a smile, a goal-directed action and not just a muscle spasm caused by excess wind, is the way that motor control is involved. Specifically, the genuine smile will involve a motor representation of the outcome, the smile, and this motor representation will lead to movements by way of planning-like motor processes.
But you don't have to buy this to agree with me.
All you have to accept is that actions like some smilings can be goal-directed and controlled even in the absence of relevant beliefs, desires and intentions.
I think smiles fall into the category of actions like graspings, reachings and gesturings which are goal-directed but do not necessarily involve intention.

So far, then, I've suggested that smiling is a goal-directed action, the goal of which is to smile that smile.
Now imagine a situation where a single individual encounters and event (a clown’s falling) which causes amusement which causes her to smile
Note that the smile also modulates the emotion; if, for example, she supressed the smile, the quality of her amusement would change.
How could we gain insight into the fine-grained dynamics of others’ emotions?
How could we ever appreciate the unfolding of another’s grief, or the way their engagement leads to an explosion of ecstasy at the climax of a concert?
Part of the answer is obvious: by being there, with them.
[Not that this is the only possibility --- in some cases we might be told.]
But how exactly does being there, in the same situation help?
Merely being in the same situation is surely not enough.
It’s not enough that we each experience amusement, grief or ecstasy.
After all, individuals are different. Different individuals’ feelings don’t unfold in the same way just because they are in the same situation.
It’s just here that collective intentionality is relevant.
\textbf{What is involved in sharing a smile?}
Minimally, I think there have to be two kinds of connection between us for us to share a smile.
First, the way your smile unfolds is shaped by how mine unfolds and conversely.
I also suppose that our smiles can be minutely coordinated with each other.
But it’s not just that our smiles are interdependent in this way ...
It’s also that each of our smiles is shaping the way our amusement unfolds.
So the way your amusement unfolds is being controlled by, and controlling, the way mine unfolds.
In sharing a smile, we are emotionally locked together.
[*todo: remove motor stuff for this talk! Also: don't lose sight of idea that control is a way of knowing.]
[*todo: need slide with control arrows highlighted (my emotion controls yours).]
[*Structure: (i) I know because my emotion controls yours; (ii) But if my emotion controls yours, how can yours be amusement at the clown's falling? because control is partial, and reciprocal; (iii) But the mere fact of control isn't enough for knowledge; rather, control must show up in experience somehow. After all, for all I have said so far, we might, in sharing a smile, be unaware that our emotions are locked together. (iv) There must be an experience that is distinctive of sharing a smile. (iv) Note that I don’t want to say that someone who is sharing a smile needs to understand the situation in the way I’m describing it. All I'm claiming is that the fact of reciprocal control somehow affects our awareness. (v) It may affect in our awareness insofar as we are sensitive to contingencies between our own actions' and others' actions, and between our actions and the causes of them. (vi) So my position is this: the reciprocal control justifies each agent in making judgements about how the others' amusement is unfolding, and this justification is at least indirectly available to the agents by virtue of their having experiences characteristic of sharing a smile. ]
Our being emotionally locked together means that to a significant extent I am feeling what you are feeling, that the way my amusement is unfolding matches they way your amusement is unfolding. So if you know how your own amusement is unfolding and you know that we are emotionally locked together, you can know much about how my amusement is unfolding. So joint expressions of emotion like sharing a smile have the potential to enable us to know not just that others are amused but how their amusement is unfolding.
But the fact of reciprocal control (which means our emotions are locked together) together doesn’t all by itself mean that we can know how each other’s emotions are unfolding. After all, for all I have said so far, we might, in sharing a smile, be unaware that our emotions are locked together. Now you might think this sounds implausible because its hard to imagine sharing a smile without an experience that is distinctive of sharing a smile. And it might be natural to describe this experience as an experience of sharing. But even if that is correct, it’s necessary to say exactly why someone who is sharing a smile is in a position to know things about how the other’s emotion is unfolding.
I don’t want to say that interaction only helps if you know that your emotions are locked together. That is, I don’t want to say that someone who is sharing a smile needs to understand the situation in the way I’m describing it. But minimally the fact of reciprocal control must somehow feature in our awareness.
[*The idea in outline: \begin{enumerate} \item the ways our amusements unfold is locked together \item this is in part because a single motor plan has two functions, production of your smile and prediction of my smile \item the single motor process means that we might experience being locked together in some way (not that our emotions are locked together but that our actions are, in something like (but not exactly) we experience actions when seeing ourselves in a mirror or on CCTV (check Johannes’ discussion of this)). \end{enumerate} ]
Here I want to offer a wild conjecture. In joint expressions of emotion there is a single motor plan with two functions, production and prediction. The motor plan both produces your own smile and enables you to predict the way the other’s smile will unfold. [*missing step about monitoring and experience. (The Haggard idea: motor planning can give rise to experiences concerning one's own actions \citep{Haggard:2005sc}.)] Because your plan has this dual function, your experience of the other’s (my?) smile is special. From your point of view, it is almost as if the other is smiling your smile.% \footnote{ Joel caricatured this idea seeing me eating fruit: ‘it’s almost as if I’m eating that fruit.’ } This means that sharing a smile has characteristic phenomenology.
This odd phenomenological effect means that in sharing a smile we can each think of the situation almost as if there were a single smile. And almost as if there were a single state of amusement. (In thinking of the situation like this it is important that we have a subject-neutral conception of the amusement and an agent-neutral conception of the smile.% \footnote{ Tom Smith asked about this. I clarified that I wasn’t suggesting there was a state of amusement which is ours, nor that the subjects are thinking of the situation in this way. That’s the point of the appeal to subject-neutral amusement. It’s a partial model of the situation. } [*Here I think I’m shifting back from the perspective of the participants in sharing a smile to the perspective of the theorist. Probably what I should say is, first, that a theorist can think of the situation in this way and use this to argue, second, that there is a simple, partial conception of the situation that doesn’t require understanding reciprocal control and interlocking emotions but is sufficient for each smiler to have knowledge of the way the other’s emotion unfolds.] So my suggestion is that in sharing a smile you experience my smile almost as if it were yours (or: you experience me almost as if I were smiling your smile), and so you might also experience our situation almost as if it involved a single state of amusement.
It's more like we each plan a single smile.
But---to reply to the objection---these plans have a dual function. Your plan both produces your own smile and enables you to simulate---to experience---my smile. And likewise for my plan. The interdependence of our smilings means that we could each think of the situation as if it were one in which a single state of amusement were responsible for our actions.