Other notes on actions, knowledge and language

Also from last year (see previous post).


People use to distinguish between theoretical reason and practical reason. In the first case, they talk about reason used to gain knowledge (or, at least, truths, since one is supposed to use theoretical reason in order to have a justified belief). The use of reason for making decisions and acting in various life circumstances belongs to the realm of practical reason.

Knowledge (scientific knowledge, at least) is technologically mediated. Even if it weren’t, it would still require the researcher to make a lot of decisions (why perform one experiment and not another, why test a particular hypothesis and not others, how to chose between theories which account for the same set of experimental data etc.).

The pragmatist view seems to be that theoretical reason is only a species of practical reason – reason used when making decisions in order to know something. I tend to agree to this. However, what I want to say is different. In my view, reason is not used only in making decisions in order to perform actions, but also in performing actions. In other words, our actions are performed in a “space of reasons” (see Sellars). I am not interested in decision making as a psychological process, so I am not interested to talk of making decisions with respect to psychological states. For me, in this context, “John decided to go to the movies” means “John went to the movies after considering the reasons to either go to the movies or to do something else”. “To consider the reasons to do A rather than B” means to argue for the belief that it is better to do A than B. Arguing for a belief can be done out loud or not, but is still an activity consisting in a series of actions.

Since one can know something only if one does something, it seems that one must be able to do at least some things without knowing anything. I can sit in a chair without leaving my room only if there is a chair in my room, but do I need to know that there is a chair in my room in order to sit in a chair without leaving my room?

What does it mean _to need something_ here? Of course, my true belief that there is a chair in my room would be enough for me to successfully sit in a chair without leaving my room (perhaps, together with the belief that there is a chair). One can believe something which happens to be true without having any knowledge about what one believes.

Now, what if one asks me: “What was your reasons for sitting in that chair instead of standing up?”. If I answer that I had the belief that there was a chair (and intended to sit down without leaving my room), it seems that I am describing the psychological process which resulted in my sitting in the chair, and not my reason for doing it. Does this change if I try to offer my reasons for that belief (e. g., “I had the visual impression of a chair” etc.)? Not really. Given my intention, my belief (for which I had reasons) made me sit in that chair. This, I suppose, is Jason Stanley’s idea behind his claim that I could have had a reason to sit in that chair only if I knew that there was a chair there.

However, what if say: “My intention was to sit down and there was a chair there”? Now I am not talking about either what I have believed or what I have known. I am talking about a fact. But how could a fact be part of the reason for an action?

Well, if I mention the aim of an action in the form of a fact, isn’t that fact a reason for my action? Suppose that doing A is a necessary condition for the occurrence of F (a fact). “The cup won’t get clean by itself, so I have to wash it.” Isn’t the fact that the cup is clean a reason for my washing the cup? Time is of no importance here. The black king not being able to move (within the game) is a reason for my last winning move even if it occurred after my move.

Now, when I state a fact I could say what I believe or what I know, of course, but even if I can successfully state (or assert) that F only if I know that F, there is a difference between having a reason and stating a reason. It might be that I have to know that there was a chair over there in order to state my reason for sitting in that chair, but I do not have to know that there was a chair there in order to have a reason for sitting in that chair. The fact that there was a chair there is a good enough reason. By looking at the chair (and at my rested feet after sitting in the chair) one could see my reasons.

Is this answer good enough?

It is not the case that whenever A is a necessary condition for B, B is a reason for A. By this rule, my sitting in the chair would be a reason for there being a chair in my room, by the way, so it seems that I’ve got something wrong here again.

George is a bachelor only if he is unmarried, but this does not mean that the reason for George being unmarried is that he is a bachelor. When we talk about “the reason for George being unmarried” we don’t want to talk about a fact. We do not give reasons (or ask for reasons) for facts. The fact that George is not married is the effect of George’s behaviour. Here we look at causes and not at reasons. Only actions (refraining from marriage is also considered an action here) require reasons. So even if F1 were a necessary condition for F2, that wouldn’t mean that F2 was a reason for F1.

If this is correct, then, facts can only be reasons for some actions, but they need no reasons to happen.

What about the fact that there was a chair in my room, then? It was a necessary condition for my sitting in a chair without leaving my room, but not the reason for what I did. The reason was the fact intended by me to happen – the fact that my legs were rested (or at least less tired) after sitting in that chair (perhaps taken together with the fact that my pants were clean afterwards, since I could have rested my legs by sitting on the floor as well).

But why did I believe that I needed to know that there was a chair there in order to have reasons to sit in that chair? The fact that there was I chair there made me sit in it and in order to state that fact I might have needed to know it. However, this seems irrelevant now.

But what I’ve actually said was that my true belief that there was a chair there was a sufficient condition for the success of my action. So perhaps my idea was that the sufficient conditions for the success of my actions are also reasons for them. So what about this rule – whenever A is a sufficient condition for doing B, A is a reason for doing B?

I transform a piece in a game of chess. There are two “moves” – the one making the transformation possible (M1) and the transformation (M2). If I regard M1 as necessary for M2, then M2 is a reason for M1. If I regard M1 as sufficient for M2, then M1 is a reason for M2. Here it seems that my intuitions with respect to reasons are stronger than my intuitions with respect to sufficient and necessary conditions.

Perhaps this is so because in judging necessity and sufficiency here I am only bound by the rules of the game. No intuition with respect to any sort of causal connections is present. When B can only be caused by A, it is easy to say that A is a necessary condition for B. When A causes B, but is not the only possible cause for B, it is also easy to say that A is a sufficient condition for B. (Remember Mackie: A is the cause of B iff A is a sufficient condition for X and X is a necessary condition for B.) I am not sure.

In any case, one could perhaps get clear by examining the formulation of the rules. Saying that it is enough to be in position P1 (obtained by performing M1) in order to transform your piece does not sound like a rule, since it does not specify all the cases in which a transformation is possible in the game. Saying that only if you are in P1 you can make a move like M2 sounds more like a rule of the game (things are getting even better if we think of the general form of a rule for transformations in chess). Also, P1 does not guarantee that the next move is going to be M2 (in the way in which M2 guarantees that the previous move was M1 – thus enabling us to say that M2 is a sufficient condition (and also a reason) for M1.

I have to note here that the true belief that there was a chair there was not said to be a sufficient condition for sitting in the chair but for the success of that action. But even this is not correct, since one might fail to sit in a chair in a lot o different ways, even if one believes that there is an empty chair in a particular reachable spot in the room.

(Necessary conditions appear in rules specifying interdictions, sufficient conditions in rules specifying obligations; if something was permitted by a rule, then this is sufficient for it being true that the conditions in the rule have been met; if an obligation (a “hypothetical imperative”) was issued, than it is necessary that the conditions specified by the rule have occurred [if the action was performed].)

Let us now think of the following rule:

(R) If one can do A only if B, it is reasonable for you to do A only if you know that B.

Now, this does not say that knowing that B is one of the reasons for doing A. “It is reasonable to do A” does not mean “There are reasons for doing A”, but “Doing A is not irrational” (somehow in the same way in which not being unhappy does not mean that you are happy).

Then, even if knowing that B is not a reason for doing A, R could still be justified. As someone pointed out, however, R seems too strong, since it contradicts our intuitions with respect to gettierized cases. In such cases, it seems reasonable for S (who has a justified true belief that B) to do A, but according to R it isn’t, since S does not know that B.

Suppose I am looking at the hologram of a chair which is furtively replaced at the last moment with an actual chair (which was hidden in my room) as I sit in it. If the only necessary condition for sitting in a chair were that there was a chair in my room, then it was reasonable for me do sit in that chair. If one still feels that my action was not reasonable, this is because another necessary condition was for a chair to be there, at the place were I saw the hologram.

I could analyse this in a different way, though. My empirical action of looking at a chair did not succeed, so not even my belief that there was a chair in my room could have been justified.

Isn’t there a problem with the fact that successful empirical actions guarantee the truth of the beliefs which are their necessary conditions?

In any case, one could of course ask me: “How could you look at a chair without knowing that there is a chair in your room? And if you couldn’t, how could the success of your action justify the belief that there is a chair in your room?”

This, of course, is the central problem here.

I could avoid it by saying that not all empirical actions are reasonable. When I look at an object in order to inspect it or contemplate it or for whatever purpose I already know that the object is present, but when I look for an object, it is not the same. I might get to (successfully) look at a chair in a room I never was before by looking for a chair. In order to do so I only need to believe that there could be a chair present.

One is not usually looking for a chair in the woods, for instance. Also, there are smaller chances for a chair to be found in a bathroom than in a dining room. Does this mean that is less reasonable to look for a chair in a bathroom? I don’t know.

It could also happen that a chair is in my line of sight and I start looking at it. At first, I am not paying attention to my visual input, but at some point I am starting to do so (several problems regarding the relation between empirical actions and sense data could be raised here; I have already tried to consider some of them elsewhere). If one said that my action of looking at that chair was irrational, I might agree.

What am I trying to do here, then? I am trying to say that there is no circularity in the idea that one can know things by performing empirical actions, in spite of what the Action-Knowledge Principle might suggest. If I focus on reasonable empirical actions, then I am tempted to invoke the distinction between practical and theoretical reason here. It is practically reasonable to look for a chair if there might be a chair in the room. It might even be that the higher the probability of a chair being present, the more reasonable it is to look for it (if finding a chair is what you intend to do, so to speak).

Since you couldn’t know that there is a chair in the room without looking for it and you couldn’t reasonably look for it without employing practical reason, it seems to follow that non-accidental knowledge is not possible without practical reason (there are several respects in which I would agree to Stanley’s anti-intellectualism).

For now, I will stop with this train of thought here.


I am wondering, however, whether I should insist on the link between knowledge and our actions right now. Isn’t it more pressing to talk about the link between language and our world? Isn’t semantics more important than epistemology? Shouldn’t I avoid at all costs that my semantic intuitions are contaminated by epistemological considerations? Wasn’t this Russell’s mistake?

The link between language and world is established by performing communicative actions directly on real objects. Such actions are not necessarily (or at first) speech acts in the class of assertives. They have nothing to do with knowledge. When I offer you an apple, my reasons have nothing to do with knowledge in regular contexts. Of course, the apple has to exist (in some special case, I could offer you the future fruits of my apple tree, but let’s put that aside for now).

Suppose I say to you “Here, have this!”, while I extend my arm towards you, holding an apple in my hand. Now, I could not hold an apple in my hand if the apple was not present, but it could also be said that even if I could have spoken the same words even if the apple was not present, I could not have performed the same speech act on the apple if the apple was no present. Saying “Here, have this!” in order to offer you an apple right now without the apple being present (to perform the speech act on it) is, from this point of view, similar to performing the same gesture without an apple in my hand (perhaps performing it as if there was an invisible object in my hand etc.).

[If I pretend to have an invisible ball in my hand, when does the pretension end? If i suddenly grasp my cup of tea with the same hand, it is clear that the pretension did end, but if I pretend to put is on my table before grasping my cup of tea it seems to continue at least as long as I do not occupy the same space with another object. It might continue even afterwards, if at some point I pretend to pick the invisible ball from somewhere near my table, where it was knocked off by my use of that other object etc.]

If the object proves not to be an apple, but a theatrical prop, it is still okay, so perhaps in order to succeed with the performance of “Here, have this!” there is no need for an apple to be present. But an object must be present.

[Joke: I appear to pretend to offer you an (invisible) object, but then it turns out I was actually offering you a speck of dust on my finger.]

In short, the presence of the offered object is a necessary condition for the success of my communicative action of my offering an object to you. Does this mean that my speech act justifies (as a reason) the presence of the offered object? No, of course.

Not even my sitting in a chair justifies the presence of the chair, but only the belief that there is a chair present (since that was also a necessary condition for my sitting in the chair; I will not return to those things now). So do I need to have a justified belief that an object is present in order to offer it to you? If so, is my action of offering the object to you a reason (a justification) for that belief? It could be an additional reason if I was interested in knowing that the object was present, perhaps. but as such, it isn’t.

So why is it reasonable for me to offer you that apple? Because I think that you might accept my offer. And also because I believe that the apple exists, and I am justified to believe that because I am holding the apple (or I was looking at it etc.), but here no problem is raised with respect to what I know.

We act reasonably on our justified beliefs and if they prove to be wrong, our actions fail. But usually our beliefs are justified by basic empirical actions which guarantee the truth of the beliefs they justify. Those basic empirical actions, in their turn, if reasonable, are reasonable because they follow actions (looking for, trying to grasp etc.) which are reasonable themselves due to some beliefs (that an object could be seen, touched etc.) which are supported inductively by the success of our past actions. However, most of those past actions were not performed with the aim to know something.

In a sense, our actions already are in the world, since they supervene (for lack of a better concept) on events occurring in our world. In another sense, our actions have no ontology. Their entire structure – agent, addressee, overseer, object acted upon, object used, manner etc. – is non-ontological (in the same way in which a social institution or structure is non-ontological).

In any case, the semantic link is possible because:
– communicative and non-communicative actions can be performed on the same object (by the same agent, with the same addressee etc.;
– communicative and non-communicative actions both have conceptual structure (see grouping, tagging, predication etc.)
– communicative and non-communicative actions are all performed in the same space of reasons (the responsibility transfer space etc.) they can oppose each other and follow by reason from each other etc. the place they occupy in this space fixes their conceptual content (holism, inferentialism etc.)

I think this picture can be detailed further without addressing epistemological issues. As for the ontology of actions, persons, responsibility relations, I think it should be left aside altogether. The problem of the two vocabularies – naturalist and personal/intentional – shouldn’t be thought of as an ontological problem. I mean, even if its solution would show how to make the two vocabularies compatible (in a sense), it should be thought of as supporting some sort of ontological monism. Also, even if the solution is going to be anti-eliminativist with respect to the second vocabulary, that shouldn’t be thought of as supporting any kind of dualism.


“Shouldn’t you say that the difference between a name and a predicate is given by the different roles they play in our inferences? But now you seem to say that a name is a placeholder for the object acted upon in a speech act, while predication is an `evolved form` of tagging. The difference between a tag for a particular object and a general tag (a predicate) is not at all clear, though.”

Let’s start with sorting (a nonverbal action; perhaps one could perform it in order to communicate something, but usually one doesn’t do that). I have a few objects on the floor. I start moving them around and form a few heaps out of them. Now I pick up an object and put it into one of the heaps.

(Mind you, this is not a classification. Not only because I did not specify any criteria, but at the end of the sorting game I might be left with a few objects which I do not want to move into any of the existing heaps and make a ‘special heap’ out of them. Also, a heap is not a set, since it might have an order.)

When I pick up an object and put it into an existing heap, the difference between the object acted upon and the one I use to sort the first (i. e. the heap) is clear. (The difference is also clear in actual tagging of objects.)

“So you say that the difference between a name and a predicate is analogous to the difference between an object acted upon and an object used to act on the first?”

The name is a placeholder. Remember Frege’s idea of satisfaction? A predicate is satisfied by an object, not by a name. The name does not name (i. e. refer to) anything (although it can have other uses – we use a pet’s name to call it etc.). It stands in for the object.

[See, also, Kaplan – the referent of a demonstrative is part of the proposition expressed by the sentence containing the demonstrative]

“Are you sure this is what you want to say? After all, the name does not have any of the uses the name bearer has. Even a name token, when uttered, is nothing more than an ephemeral physical object. The name of a mountain, for instance, has no resemblance to its name bearer. You cannot climb it, admire it’s valleys a.s.o. So how could it stand in for the mountain?”

You should remember, though, that the name is used as a stand in for the only purpose of tagging (or predication; or sorting, for that matter). In this respect, I can use it even better than I could use the actual object. I cannot form a heap of mountains, but I could write their names on pieces of paper and form a heap out of that. For other purposes I use other stand-ins like drawings on paper (in a map), paintings, scale reproductions etc.

“But then you’ll have to say that any two names of the same object should be interchangeable, since they are stand-ins for the same object.”

Sure. Kripke has opened a path here, perhaps I could follow it.

“What about ‘This is George.’? Should it say that George is identical with himself?”

We never need to say that an object is identical with itself.

“Okay, but it has to say something, right? To me, it seems that the name is introduced as a tag for George.”

We need a context here. If I am presenting George to you, I am not introducing his name as a tag. I am saying that you could call him by using ‘George’, that you could address to him by saying ‘George’ and so on. The different uses of a name are known to you, including the placeholder use. What I am saying is that for this individual the word that fits the uses of a name is ‘George’.

“So you’re talking in metalanguage?”

Not necessarily. I am not saying that `George` is the name of this individual. I am only introducing the individual to you. Suppose you’ve already heard me talking about George and now you meet him. I could say ‘This is George’ to introduce him to you. This does not change if you haven’t heard me talking about him. In the second case I am still introducing George to you, together with his name.

“What if he is at some distance, so no proper introduction could be made, but you still say ‘This is George’?”

I am introducing him to you in a different sense. Suppose George was a pet tortoise. I could introduce it to you and also introduce its name to you.

“But in order to do that it would suffice to point to George and say its name. Why would you do more than that?”

Oh, well. Maybe I am in no mood for pointing. One does not need a reason to perform an action in a slightly different manner.

“How do we distinguish between performing the same action in a slightly different manner and performing a different action?”

By looking at the results, I suppose. Here, the result is that now you are familiar with George and also with his name. That is all.

“Couldn’t it be that the same result is achieved by performing completely different actions?”

If you think at mathematical operations, one could get 3 as a result of a lot of different calculations, but that is not what I mean by the result of an action. If one “wanted 3 as the result of one’s actions”, one could write it down on paper without any calculation.

“Or print it to look exactly the same, which is different from writing.”

Right, but now the actions are different because the technologies used are different. Our initial case, however, was not like that.

“What about ‘This is George’ said in order to baptize a small child?”

What about it?

“Isn’t the name introduced as a tag for that child?”

Of course not. The person baptizing the child is not trying to say anything about it. I suppose the name was already chosen by the child’s parents and the priest is only announcing it to the public. That is all. The parents did not need a special act of baptism. They settled on a name for their child and started using it directly.

Also, a name can be used as a tag, but only in some special circumstances (“You’re an Einstein, aren’t you?”).

“So you’re basically saying that demonstrative indexicals are placeholders for objects and names are like some specialized demonstrative indexicals?”

Basically, yeah, this is my point.

The larger picture is this. If the problem is how could we conceive ourselves as talking about the world we are in, the answer is that what makes this possible is the fact that we can act together in the world. When two people act on the same object, they can also perform communicative actions on that object. There is, of course, a long way from primitive communicative actions to ‘George is a bachelor’, but I think one is better equipped to travel it by using concepts such as ‘tagging (an object directly)’, ‘using a placeholder / or a stand-in for an object / or an object handle’, ‘this-predication’ a.s.o.

Now, if you want, you can think of the following case. Let us say there is some practice X, such that there is only one practitioner of X – the Xer. Being an Xer is in this case a tag which can be used to refer to our Xer. ‘Xer’ could then be used as a nickname for that person, instead of her actual name. Some names could have a similar natural history, but others don’t.

Actually, even if most names had such a natural history, my view wouldn’t change much. I would concede that names are unlike demonstratives, but that would not change what I am saying about `basic reference` consisting in our ability to perform communicative actions directly on objects from our environment.


“Throw this away!” – by saying this I make a request to you that you throw some object away. The request is made with respect to that object directly (in a way similar to tagging; one could say, if one wanted, that I ‘tag’ the object in case with my request to be thrown away (or disposed off in some other manner) and the addressee of this communicative action of mine is you.

“I promise to give you this at the end of the day.” – here it can be said directly that I have promised that object to you.


So, what does it mean that ‘this’ is some sort of an “object handle”?

I suppose there are actually some choices here:

– ‘This’ is a verbal demonstration (similar to a demonstrative act) – I demonstrate the object to you (“This!”). Then I say: “I promise to give it to you at the end of the day.” I could have said: “I promise to give at the end of the day”. “It” is just a syntactical feature.

– ‘This’ is a syntactical feature. We already act on the object in case (look at it etc.). I say: “I promise to give this to you etc.” but I could have said “I promise to give it to you etc.” instead.

– The content of my utterance is the actual object tagged (in a pretense-like fashion) with my promise to give it to you at the end of the day. ‘This’ refers to that object (but how?). – Actually, I think this is the view I wanted to avoid.
– The content of my utterance is the occurrence of the word ‘this’ (as a placeholder of the actual object) tagged with the promise to give “it” to you at the end of the day. This seems absurd, although I seemed to support this view by something I have said previously.
– In order to actually use the object in case in a non-communicative physical action I usually have to grasp it. Even if you are already paying attention to it, I need a linguistic tool in order to grasp it and use it in a communicative action. ‘This’ is such a tool. – So what is the difference between ‘this’ and ‘it’ then (if ‘it’ is just a syntactical feature and not a ‘grasping tool’)? – One could reply that ‘it’ is a placeholder, not for an object but for another word – a name or some other referring word. So in “I promise to give it to you etc.”, ‘it’ appears to be a syntactical feature (which one could dispose of), but is actually a placeholder for the grasping tool (‘this’).

But in this case, it would not be possible to make that promise by saying ‘I promise to give at the end of the day.’, since in the absence of a ‘grasping tool’ one could not use the object in case to promise it to another person. However, primitive tagging did not need a linguistic grasping tool (or handle).

Also, one could (non-communicatively) act on a physical object without grasping it (or handling it in any other way) – see: breaking a rock into smaller pieces by smashing it with a bigger rock, covering something with leaves, painting something (without touching it) etc.



Maybe focusing on logical matters isn’t always helpful. I want to conceive the relations between actions as logical (or at least logico-semantical) and thus impose a logical model on them. Wittgenstein would perhaps tell me that this is misleading. The view according to which something can be thought of either in the logical space or in the causal space is rooted in Tractatus, but one does not need to hold to that. A different model might be better. Looking at simple cases of relations between actions (without assuming that they can be either logical, or causal) might provide me with some better insights.

[It is interesting to see that philosophers were relying on habits learned in doing scientific research when doing philosophy. Such habits were a lot more particular than what the empiricist / rationalist divide could suggest. Instead of simply using reason as the source of knowledge, Plato was using specific kinds of mathematical reasoning, one might say.]

So, how could this new approach look like? Let us think of a simple request: “Bring me that rock!”. The addressee of my speech act will perform the action of bringing the rock to me. The request is neither a necessary, nor a sufficient condition for the second action, but they are related nevertheless. Also, the request in itself does not cause the second action, since its addressee can choose to comply with it or not. The second action seems also related to the promise “I’ll bring that rock to you”.

Similar things can be said about an observation and its report. Making the observation is a necessary condition for reporting it correctly, but not for the report per se. In addition, it is not a sufficient condition for making the report and cannot cause this second action by itself, since one can choose not to report the observation.

However, the report “answers” to the observation, just as the action of bringing that rock “answers” to either the promise or the request. What does this mean? Roughly speaking, it means that one action is expected as the next move in a particular language game after the other has been performed.

But why do we expect it to be performed? Saying that we observe the rules of the game does not help here. If I ignore a request, this does not mean that I am not playing the game of requests anymore.

Even if I talk about responsibility relations, this does not solve the entire thing. If I promise to do A, I am, indeed, responsible for doing A (to the person I made my promise to), but I still could fail to keep my promise and claim that I was playing the game of promises.

These three cases (the request, the promise and the observation) are similar, but not in all respects. In the case of a request, the person accepting the request (by saying so) is responsible to perform the requested action, but that is all. Still, the action “answers” to the request to perform it. Both have the same conceptual content (perhaps introduced by the request), but that is not enough. A request and a promise can also have the same conceptual content, without it being the case that one “answers” to the other (except when the request is the request to promise to bring that rock). One can perform the action without publicly accepting the request (and is disputable that some sort of unconscious, inward acceptance is necessary) and the acceptance of a request is not a promise (or is it?). In any case, if I bring that rock to you two times, the two actions might have the same conceptual content without being related in the sense that one “answers” the other. Also, an order and a request can have the same conceptual content without being related in that way.

So, how are the request to perform A and the performance of A related?

It could be said that by issuing the request I intend to perform A with your help. Or at least that I intend that A should be performed. The request and the performance of A can both be made with the intention to get the same result (here is a difference between a request and a promise, of course). But this does not explain why the performance of A answers to the request to do A (and not the other way around).

Suppose I ask you to give me a ball you are holding in your hands and you do not understand it. This is also some kind of request, one could say. Now suppose I take the ball from your hands. Perhaps in this way you could understand what my request was. A child could learn about requests this way. The point is that the child should see what I was trying to get from it as a result of my request. (It is easier to learn interdictions and encouragements, since you are already doing something.)

By calling a child one issues the request to come to her to the child. If the child does not understand, another person has to take the child by the hand and bring him to the person calling it. Of course, the child has to pay attention to what is going on, in order to learn something from such an incident.

The same thing has to be repeated over and over again. Communication is achieved when the conceptual content is the same in the case of the request and its execution and the request is most of the time followed by its execution (when it isn’t, some excuse can be found).

This is the game. (Compared to this, “waving your hand to say goodbye” is a more complicated game.)

So is the priority of the request to its execution only due to the way we learn things? Is it an accident? If we taught children requests but uttering them after the child did something, would we say that the request “answers” its execution?

It doesn’t seem so. In the second case, we would be teaching the child reports (of its actions), not requests.

Let us say that the request to do A is a means to get the result of A via A being performed by someone else.

One could execute requests without knowing how to perform actions together with another person, cooperatively (“Tap your feet to the ground, so that your side of the seesaw is up in the air and my side goes down!” – I say this to you and you do it every time, but it is not like we operate the seesaw together.)

“Clip your fingernails”, “Wash your teeth” – of course, at first the parent does all these things for the child. After the child learns how to do them, the parent asks the child to do them in her stead. At some point, the parent just reminds the child to do such things for himself (or herself).


It matters that the request is made first and the action is performed afterwards. By making the request one is making herself responsible to see to it that the request is executed. By doing the action (in answering to a request) one is not undertaking such a responsibility.

Doing A precludes the request to do A. (It is the same with the promise to do A and the performance of A, since one cannot promise to do what one has already done.)

Even in the case of an observation, it could be said: “If you report that you have observed two books on the table, this makes it useless for you to observe if there are, indeed, two books on the table.

‘But I might want to check on my report, or to improve it a.s.o.’ – Okay, but if it’s that kind of report, then it does not “answer” to an observation with the same conceptual content. (And perhaps it shouldn’t be called a ‘report’.)


‘So aren’t you, after all, still trying to apply a logical model on things?’ – Well, I am not sure that my concept of “preclusion” is logical.

Doing A does not contradict the request to do A (and does not oppose it either). It invalidates a necessary condition for issuing the request (perhaps) and in this sense it makes the request absurd (so to speak). It is not that the point of the request cannot be achieved anymore (or that more actions are needed in order to get to the end of the series). The point of the request is already achieved, so the request has no point anymore. In any case, I am not able to continue talking about this right now.


Let my try to look at some other very simple cases. Such cases could be:

– offering, giving and receiving an object;
– pushing somebody out of the way and saying “pardon” (or something similar);
– trading an object for another one;
– getting someone’s attention and showing something to that person (perhaps also saying what it is you’re showing);
– disproving what someone says by doing something (“there are no apples left on the table” – pointing to an apple on the table);
– alerting somebody, that person’s reaction;
– helping someone to carry a thing from one place to the other;
– disputing a found object with another person;
– playing simple games: passing a ball, pushing someone in a cradle etc.
– asking for permission to do something;



‘And what do you hope to achieve with this?’

Well, I hope to understand the relations between our verbal and non-verbal actions (communicative and non-communicative).

‘Wouldn’t it be enough for you to see that there are some habits, some established regularities here?’

Communication and semantic content cannot be just a matter of regularities.

‘Why not? When you notice a regularity in nature, it is as if you are told something. The cat has noticed that when she meows, you pay attention, so she is meowing to tell you to pay attention to her. Weren’t you ready to say that actions have semantic content? If they do, then even the actions performed by a small child who does not know to talk yet do have some semantic content.’

They do, in a sense. But there is no robust semantic content there.

‘Right, but the robustness comes from the network of entailed responsibilities. This, at least, is what you seem to claim. So why not leave it at that?’

Maybe I will, but I need to make sure that my conceptual proposal is good enough. Maybe what I need to say is that a network of entailed responsibilities is not necessarily a “logical space” (and it clearly isn’t a causal space). I do not know. Maybe I need more concepts in order to talk in a satisfactory way about the way in which “language meshes with our life”.



‘In any case, I do not see how you could try and talk about actions outside a “causal space”. If an action has an object, the object has to be somehow affected by it. Even declarations must affect a social reality in some way. So actions must be in a natural / causal space. If you want to say that they can also be thought of as figuring in some logical space, although the two cannot coincide, that is your choice, but you cannot put them only in a logical space. You need a causal space for most of your so-called responsibility entailments as well.’

We’ve been through that before. Causal relations can be conceived in therms of responsibility entailments as well.

‘Okay, but in order to establish such responsibility entailments properly you need to notice the causal relations. You cannot say that painting an object red makes you responsible for it being green. Or that throwing a glass on the floor makes you responsible for it floating towards the ceiling (when in fact it makes you responsible for breaking the glass).’

Sure. But your first example has nothing to do with causal relations. I am responsible for the object being red when I paint it red due to the semantic content of “painting” (this seems circular, since I say that the semantic content is provided by the network of responsibility entailments, but it isn’t, due to semantic holism).

Your second example is more to the point. I do not want to say that whatever things we conceive as happening do actually happen. We are part of our environment and act in our environment. We do not create our environment by our actions but only modify it. I agree to all this.

I could also agree that our network of responsibilities also grows from our interactions with our environment (and with each other). However, the persons, their intentions, actions, responsibilities and the entire ‘logical space’ and the semantic content which it pervades do not have an ontology. It is pointless to locate them in our environment, but is also pointless to say that they are not in our environment.

‘But then I return to my question – how can an action, which does not have an ontology, affect an object acted upon, which does have an ontology?’

When we talk of actions affecting objects we are actually talking of events affecting objects. This is all. But we have the ability to call some events our actions and take responsibility for them.

Think of productive actions (in Aristotle’s sense of poiesis). A bird can make a nest, but the bird is not the author of the nest. It does not perform productive actions. We can regard ourselves as the authors of all our actions and talk about this. However, even if ‘authorship’ is something which appeared naturally, our vocabulary of authorship is kept for different purposes than those of describing what happens in the nature.

I do not know what else to say at this point.



“One more time, why are you so keen on talking about actions, after all?”

Well, I think this is central. Remember what Stegmueller said about philosophical paradigms? The linguistic paradigm in philosophy led me to this.

“Right, but, why not talk directly about responsibility, promises, other commitments, human relations, social institutions, existential matters and so on?”

I cannot ignore the past philosophical debates. It is due to such debates that the paradigms were established. Several metaphilosophical arguments got me convinced that any proper philosophical activity is conceptual engineering.

“Right, but why not directly engineer concepts which could be a lot more relevant to the everyday life of other people?”

I must confess I am interested in that too, but since my formation is that of a theoretical philosopher,…

“So it’s just a biographical accident that you,…”

I hope it isn’t. I think I need a more general and abstract conceptual framework in order to talk about the most mundane things. For instance, I am really interested in thinking about what my responsibilities are as a parent. The general framework says that my responsibilities issue from what I do (or did in the past), so this is why I think my responsibilities as a parent are related to my past action of bringing a child into this world. This seems reasonable to me precisely because I am using that larger conceptual framework.

“So you want to think of what responsibilities one is reasonable to undertake as a parent, regardless of the social norms?”

Not necessarily. Think of my analogy. Bringing someone into this world is similar to bringing someone with you to a party. Now, you have to make sure that the person you have brought with you does not spoil the fun for the other people at that party, so you have to educate your child along similar lines. This is what I would say. However, there might still be some sort of social convention in place when we say that we have to make sure that our companion at a party does not spoil the fun for the other participants. That idea in itself needs justification and perhaps at the end of the justification process lies something like “doing otherwise would meet the disapproval of most people”. I don’t know.

“Doesn’t your talk about responsibility make ethical matters central to any philosophical endeavor, according to you?”

I hope not. Responsibility is not an ethical concept in my view. It is related to authorship, as you might have noticed, which makes it equally relevant to aesthetics, for instance.

“Don’t you think that Aristotle’s distinction between praxis and poiesis is important here? If you do not think that both are just different types of actions,…”

I know what you want to say. My concept of responsibility is too general. It covers regular actions and productive actions as well. Only responsibility for productive actions is related to authorship, while responsibility for regular (practical) actions is not. The two are different, so I cannot be interested in something so general that eludes the borders between different kind of value judgments. You want me to look at the differences between such cases and control my ‘thirst for generalizations’.

In this case, however, I think Aristotle’s distinction is not reasonable. The nature of the result of an action (a completely new object / a modified object / etc.) does not affect the nature of the action.

I would risk to say that ethical matters come into play when other persons are not only the addressees (or the overseers) of one’s actions but take the position of an object acted upon (or of an instrument used to act).

For instance, if I am offering my opinion to you, you are the addressee of my speech act (and only that). I could opine that you are a worthless individual and your life must be ended as soon as possible by the state, but if you remain the addressee of my speech act (things can be confusing), there is no need for an ethical evaluation of my action. When I verbally aggress you by telling you that you are worthless and should be put out of your misery etc., the addressee of my action is perhaps myself (it could be someone else in a special situations) and you are the object of my action.

“Okay, you do not need to continue with this sort of examples. I still am a bit reluctant to accept your approach. In short, you talk as if people are (or should be) computers. If they are unable to compute what the rules dictate with respect to their responsibilities (or expected behaviour) in particular circumstances, that is due to the fact that, in contrast to mathematical formulas, social rules accept more solutions. However, even in such cases (practical) reason can be conducive to unique solutions. Actual circumstances, particular context elements which are beyond any conceptualization, feelings, idiosyncrasies and biographies do not matter.”

Of course they do. But I am not trying to predict how a person is going to act in some particular circumstances. I cannot do this even for myself. I am interested in the rules of the game, that is all. And because we talk of a game, of course a ‘computer’ (artificial or human) should be able to play it, but that is only incidental to the whole thing.

“You are covering yourself with a lot of intellectual dust here. Your view is no longer clear to me. Perhaps we’ll talk about this some other time.”




“So you want to say (using Heidegger’s distinction) that your hand can sometimes be ‘present at hand’?”

Why do you say that?

“Well, you seem to think that any action should include not only something acted upon, but also some sort of instrument. When the instrument used to act is not an artifact, it must be your body (or a part of it). But then, if your hand can sometimes be an instrument without you noticing it (that is, it is in the ‘ready to hand’ mode), it must follow from this that at least sometimes your hand could be an instrument in the ‘present at hand’ mode.”

Okay. I am not sure that I want to say that every action must involve the use of an instrument, but I can agree that I sometimes use my hand as an instrument to perform certain actions. This is especially obvious when the hand replaces an actual artifact (I could use my hand instead of a special tool to flatten a piece of dough etc.).

So this might mean that at least in some cases, when I cannot control my hand the way I want to, due to various factors (coldness, arthritis etc.), it might get into the ‘present at hand’ mode for myself. The case might sound weird due to the choice of words to express Heidegger’s distinction, but this shouldn’t count against the possibility of such cases.


In the case of a rhetic act (to use Austin’s distinctions), for instance, It could be said I that use my mouth and vocal cords to produce sounds (or my hand to press keyboard keys, in order to produce some letters etc.). In any case, if a speech act is in question, then perhaps my instruments are the words uttered. In such cases, when we talk about the illocutionary force of the speech act in question, we might oscillate between thinking of what the agent is trying to do and what she actually does. Thing is, we sometimes use the same word for both.

‘To promise’, for instance, might mean:
(a) to propose to make a commitment to an addressee (which could accept your proposition or not) – what you do;
(b) to get the addressee to accept you commitment – what you are trying to do;