1. The concept of a human act is of basic importance to the questions which are discussed in this book. It is not one of the aims of deontic logic to clarify this concept. The notion of an act is more like a tool which this logic has to use for other purposes of clarification. Considering however the complex and obscure nature of this notion we must try to throw light on some of its aspects before we can be reasonably sure that our use of it as a tool in deontic logic stands on a firm basis.
I find it surprising that the concept of a human act has as such been relatively little discussed in philosophic literature. The same is true of the related notions of activity and behaviour. Traditional philosophic discussion bearing on these concepts has concentrated on the problem of the so-called freedom of the will. In this discussion it is all too often taken for granted that it is clear what action is. In fact much of what has been said about the problem of freedom can be shown to be void of interest because based on some logically defective notion of acting.
In our discussion of acts we renounce every pretension of being systematic and shall try to confine ourselves to a necessary minimum of conceptual distinctions and observations. The freedom of the will we shall not discuss at all. But of the related topic of the ability to act (do) we shall have to say something.
2. The notion of a human act is related to the notion of an event i.e. a change in the world. What is the nature of this relationship?
It would not be right I think to call acts a kind or species of events. An act is not a change in the world. But many acts may quite appropriately be described as the bringing about or effecting (‘at will’) of a change. To act is in a sense to interfere with ‘the course of nature’.
An event we have said (Ch. II Sect. 6) is a transition from one state of affairs to another or from a state to a process or from a process to a state; or it is a transformation of processes. The Logic of Change which we sketched in the preceding chapter is primarily a logic of events of the first type. Events of the second and third type are also called the beginning (commencing starting) and the end (ceasing stopping) of processes.
The events which are effected through action can be of any of the several types just mentioned. The acts of opening a window or of killing a person effect changes in states of affairs. Starting to run or stopping to talk may be acts which effect a change from a state to a process and from a process to a state respectively. But when a walking man starts to run his action effects a transformation of processes.
The Logic of Action which we are going to outline in the next chapter will primarily be a logic of acts which effect changes among states of affairs. Other types of action will not be explicitly dealt with in our formal theory.
The examples of acts which we have here mentioned are examples of what I shall call generic acts or act-categories. There is an in principle unlimited number of cases of window-opening or of starting to run.
The several cases of generic acts I shall call individual acts or act-individuals. It is noteworthy that the word ‘act’ is used ambiguously in ordinary language to mean sometimes a generic and sometimes an individual act. It is e.g. correct to call murder an act; this is an act-category. It is also correct to call the murder of Caesar an act; this is an act-individual.
To the generic act of opening a window there answers the generic change of a window becoming open. To the individual act which was the murder of Caesar there answers the individual event of Caesar's death.
The logical difference between acts and events is a difference between ‘activity’ and ‘passivity’. An act requires an agent. An individual event is the taking place or happening of some generic event on a specified occasion. An individual act again is the doing of a generic act on a specified occasion by a specified agent.
3. When we say that an individual event happens on a certain occasion we may regard this occasion for the happening of the event as constituted by two successive occasions for the obtaining of certain states of affairs (see above Ch. II Sect. 6). Similarly when we say that an individual act is done on a certain occasion we may regard this occasion for the doing of the act as constituted by the two successive occasions for the corresponding individual event.
Not every occasion (or pair of successive occasions) is an occasion on which just any individual event may happen or act be done. Thus for example only on an occasion when a certain window is closed can this window open or become opened. Generally speaking only on an occasion on which the generic state of affairs described by p obtains can the generic change described by pT ~p or that described by pTp take place or become effected (‘at will’).
We shall say that an occasion constitutes an opportunity for the happening of a certain generic event or for the doing of an act of a certain category when the occasion has some generic feature which makes the happening of this event or the doing of this act (logically) possible on that occasion. For example: Only on an occasion when the window is closed is there an opportunity for opening it.
Any opportunity for the doing of an act of a certain category is also an opportunity for the happening of the corresponding generic event i.e. for the event effected through the act. The converse entailment however does not hold. Not every opportunity for the happening of a certain generic event is also an opportunity for the doing of a corresponding act. The occasion for the happening of the event has to satisfy additional conditions in order to constitute an opportunity for the doing of the corresponding act i.e. the act of bringing about this event. Which these additional conditions are will be discussed later (Sect. 7).
4. The notion of an agent is essential to the distinction between acts and events. We shall here make only a few brief comments on this notion.
We may distinguish between empirical (natural) and super-empirical (super-natural) agents. An agent is empirical I shall say if the agent's existence is a contingent or empirical fact. Super-empirical agents have necessary existence. The difference between the two categories of agent can also be expressed by saying that an empirical agent is a ‘perishable’ a super-empirical agent an ‘eternal’ being.
The ideas of necessary existence and of super-empirical agents cannot be discussed within the limits of this work.
Agents who perform human action are empirical. But not all agents of human acts are human individuals.
We can distinguish between personal and impersonal agents. An impersonal agent is for example any so-called legal or juristic person (such as a corporation) a law court a legislative assembly or the state.
The action of impersonal agents is certainly ‘human action’ in some sense of the word. The question may be raised whether acts which we impute to juristic persons and other impersonal agents of human acts are ‘logical constructions’ i.e. could be defined (conceptually explicated) in terms of acts of some personal agents. This question however we shall not discuss.
Of what I here call personal agents we can further distinguish two kinds viz. individual and collective agents.
When an act is performed by one man we shall say that it is performed by him individually.
Sometimes the performance of an act requires the joint acting of several men. The table may be too heavy to be removed from the room by one person alone but two or more persons may do it by their joint efforts. We then say that the act of removing the table is performed by two or more men collectively.
That an act is performed by several agents collectively may also be described by saying that the agent who performs this act is a collectivity of men or a collective agent.
A collective agent must not be confused with an impersonal agent such as say a corporation or the state or some other juristic person. But the acts of a juristic person may entail the collective acting of some men.
Whenever several men perform an action collectively (‘by joint efforts’) each man does something individually. The question may be raised whether acts attributed to collective agents could not be regarded as ‘logical constructions’ of acts of some individual agents. This is another problem which will not be discussed here.
5. To every act (of the kind which we are here considering) there corresponds a change or an event in the world. The terms ‘change’ and ‘event’ must then be understood in the broad generalized sense which covers both changes (events) and not-changes (not-events). This correspondence between act and change is an intrinsic or logical tie. The act is as it were ‘defined’ as the act of effecting such and such a change. For example: the act of opening a certain window is logically the act of changing or transforming a world in which this window is closed to a world in which it is open.
By the result of an act we can understand either the change corresponding to this act or alternatively the end-state (see Ch. II Sect. 6) of this change. Thus by the result of the act of opening a certain window we can understand either the fact that the window is opening (changes from closed to open) or the fact that it is open.
On either way of understanding the notion of a result of action the tie between the act and its result is intrinsic. The act cannot be truly described as being an act of the category in question unless it effects a change or ends in a state of affairs of the kind in question which we call its result. An act cannot be truly called an act of opening the window unless it ends (results) in the window's being open—at least for a short time. Trying to open the window need not of course result in this.
When the world changes in a certain respect it may happen that it also by virtue of so-called causal or natural necessity becomes transformed in a certain other respect. We then say that the second transformation is a consequence of the first. If the first transformation is effected through action is the result of an act then the second is a consequence of action is a consequence of this act.
For example: a consequence of the act of opening a window may be that the temperature in a certain room sinks (is subsequently lower than it was before).
Whether a certain transformation will cause a certain other transformation to take place or not will usually depend upon the presence or absence of a number of other features of the world beside the states associated with the two transformations themselves. This is true also of human action. Whether the temperature in a room will sink or not as a consequence of opening a window will depend among other things upon the antecedent difference in outdoor and indoor temperature. Sometimes the temperature will not sink but rise.
Unlike the relation between an act and its result the relation between an act and its consequences is extrinsic (causal).
Someone may wish to object to our terms ‘result’ and ‘consequence’ (of action) on the ground that what is here called a consequence is quite commonly in ordinary language spoken of as the result of an act and vice versa. Thus for example we say that as a result of the window's being opened he caught a cold. The catching of the cold however was a consequence to use our terminology of the act whose result was that the window became open.
I am of course not anxious to correct the ordinary use of ‘result’ and ‘consequence’. What matters are not the terms but the conceptual distinction between such changes and states as have an intrinsic and such as have an extrinsic relation to a given act. This distinction is important to note and it is somewhat unfortunate that no clear terminological indication of it in ordinary parlance should exist.
Perhaps this ‘defect’ of ordinary language is connected with the fact that the distinction between the result and the consequences of an act although logically sharp is at the same time in an important sense relative. What I mean by this can be explained as follows:
Consider again the act whose result we said was that a certain window is open (at a certain time and place). Could one not truthfully answer the question of what the person who opened the window did by saying that he let cool air into the room (thus lowering the temperature)? Is not the cooling of the room the result rather than the consequence? A consequence may be that someone in the room began to shiver and went out or subsequently caught a cold.
The answer is that we can certainly speak of an act of cooling the room but that this is a different act from that of opening the window. The act of cooling the room requires logically that the temperature should go down and may require causally that the window should be opened. The act of opening the window again requires logically that the window is opened and may lead causally to the fact that the indoor temperature sinks.
Thus one and the same change or state of affairs can be both the result and a consequence of an action. What makes it the one or the other depends upon the agent's intention in acting and upon other circumstances which we shall not discuss in this work.
The act of opening a window and that of cooling a room are logically distinct because of the nature of their results. But there is a sense in which the two acts may be said to ‘look’ exactly alike. The sense in which they look alike is that the activity involved in performing the two acts may be identical viz. certain muscular contractions and movements of the agent's limbs.
6. We shall distinguish between act and activity. To close a window or to kill a person is to perform an act. To smoke or to run or to read is to be engaged in activity.
The distinction is obviously important but philosophers have so far done very little to clarify it. Here I shall make only a few scattered observations on the topic.
As acts are related to events so are activities related to processes (cf. Ch. II Sect. 5). Events happen processes go on. Acts effect the happening of events activities keep processes going.
Activity is not internally related to changes and to states of affairs in the same manner in which acts are related to their results. Activity however may be externally or causally related to changes and states which are consequences of performing this activity. Running need not leave any ‘imprint’ on the world but smoking may leave smoke. As a consequence of drinking a person may get drunk. Getting drunk is an event and drunkenness a state.
The question may be raised whether activity is logically prior to acts or vice versa.
In some sense activity seems to be prior. Action may be said to presuppose or require activity. The bodily movements which are a prerequisite of most human acts may be regarded as activity in which the agent has to engage in order to perform those acts. The changes and states which we call results of action may be viewed as consequences of such prerequisite activities.
Yet in another way action seems prior. Human activity has a beginning and an end. The beginning and the ending of activity have sometimes at least the character of acts. To run is an activity but to start running or to stop running are acts of a kind. These acts however differ characteristically from acts which effect changes in states of affairs. The first of them implies a change or transition from a state to a process the second again from a process to a state (cf. Ch. II Sect. 6).
Beside the distinction between act and activity we have to note a distinction between acting and doing. To do something is to perform an act. To be doing something is to be engaged in some activity. That which we have called the result of an act is that which any agent who (successfully) performs this act on a certain occasion has done on that occasion. When an act fails of its intended result the agent has tried to do something which in fact he failed to accomplish (do). Trying is thus a ‘logically incomplete’ mode of acting. It is not immediately clear whether trying should be counted as falling in the category of act or in that of activity (cf. below Section 10).
7. We have previously (Ch. II Sect. 7) introduced the notion of an elementary change. The four types of elementary change we said are the four types of change and not-change which are possible with regard to a given (atomic) state of affairs and a pair of successive occasions. As schematic descriptions of the four types of change we introduced pTp pT ~p ~pTp and ~pT ~p.
We now introduce the notion of an elementary act. By an elementary act we shall understand an act the result (Section 5) of which is an elementary change. The correspondence between elementary act and elementary change is one to one.
We shall use the symbol d for acting. The schematic descriptions of the four types of elementary act shall be d(pTp) d(pT ~p) d( ~pTp) and d( ~pT ~p). It should be observed that d(pTp) etc. are schematic representations of sentences which describe acts just as pTp etc. are schematic representations of sentences which describe changes and p etc. are schematic representations of sentences which describe (generic) states of affairs.
We shall now consider the nature of the four types of elementary act in turn. For purposes of illustration let p represent the sentence ‘The door is open’.
Take first d( ~pTp). It describes the act of changing or transforming a ~p-world to a p-world. In terms of our illustration: it describes the act of opening the door. We could say that d( ~pTp) represents the sentence ‘the door is being opened’. A way of reading the schema d( ~pTp) would be ‘it is being done that p’. For purposes of convenience however we shall read it though this is somewhat inaccurate ‘p is done’ and call the act which the schema describes ‘the doing of p)’. (A more accurate but clumsier name would be ‘the doing so that p’.)
It is easy to see that the act described by d( ~pTp) can (logically) be done only provided two conditions are satisfied. The first condition is that the state described by ~p prevails on the first of the two successive occasions which jointly constitute an occasion for doing the act. Only when the door is closed can it become opened. One cannot open an open door. The second condition is that the change described by ~pTp does not happen as we say ‘of itself’ i.e. independently of the action of an agent from the first to the second of the two occasions. If a door is so constructed that a spring or other mechanism pulls it open as soon as it has become closed then there is no such act as the act of opening this door. (But there may be an act of closing it and of keeping it closed. See below.)
Consider next d(pT ~p). It describes the act of changing a p-world to a ~p-world. If we call the act described by the schema d( ~pTp) ‘the doing of p’ we could call the act described by the schema d(pT ~p) ‘the destroying of p)’. If we apply the schema to our example it describes the act of closing the door.
The conditions for the doing of the act described by d(pT ~p) are as follows: The state described by p should prevail on the first of two successive occasions and not ‘of itself’ change to its opposite from this occasion to the next. For example: A door can become closed as a result of action only provided it is open and does not close ‘of itself’. Here the words ‘of itself’ mean that the change is due to some ‘natural’ cause such as e.g. the operation of a spring and is independent of the action of an agent.
Which act does d(pTp) describe? pTp means that the world does not change in the feature described by p on two successive occasions. Can this be a result of action? It certainly can. The world might change in a certain feature unless there is some agent to prevent it from changing. This is the sort of action that d(pTp) describes. We might call it ‘the preserving of p)’. This act can be done provided that the state described by p) initially obtains and would change into its contradictory state unless the change were prevented through action. For example: It could happen that a door which is open will close e.g. under the influence of a spring-mechanism unless somebody keeps it open.
d( ~pT ~p) finally describes the act of keeping the world unchanged in the feature described by ~p. We could call this act ‘the suppression of p’. This act can be done provided that the state described by ~p obtains but would change into the state described by p unless prevented through action. For example: It could happen that a door which is closed will open unless someone keeps it closed.
If by the result of an act we understand not a change but the end-state of a change then the correspondence between act and result is not a one-to-one correspondence. The same state may be the result of more than one act. Thus the state described by p can be the result both of the elementary act described by d(pTp) and of the elementary act described by d( ~pTp). The fact that a certain window is open can be a result either of an act of opening it or of an act of keeping it open.
Each one of the four types of elementary act can be performed only provided a certain state of affairs obtains. The types of act described by d( ~pTp) and d(pT ~p) moreover can be done only provided the change which is their result does not take place ‘of itself’ i.e. independently of action. The types of act described by d(pTp) and d( ~pT ~p) again can be done only provided that the changes described by pT ~p and ~pTp respectively would take place unless prevented by action. But this is the same as to say that the changes (not-changes) described by pTp and ~pT ~p i.e. the changes which are the results of the respective acts do not take place ‘of themselves’ i.e. independently of action.
The question may be raised what we shall say of the case when some other agent beside the agent in question effects the change which must not happen ‘of itself’ if we are to say truly that the agent in question has done it. Shall we say then that neither agent does the act? Or shall we say that both do it? If a person shoots at another at the very moment when the latter dies of a stroke the first person cannot be rightly said to have killed the second. The second died but was not killed. The first did not commit murder although he may have attempted to do so. Suppose however that two persons at the same time shoot at a third and that each shot individually would have killed him. Obviously we must say that the third man was killed i.e. that his death was a result of action. But by whom was he killed? If the assumption is that each shot individually would have killed him it is not correct to say that the two murderers killed him ‘jointly’ or ‘by joint efforts’ and that therefore the agent technically speaking was a collective agent. The right thing to say is in my opinion that he was killed by each one of the two murderers i.e. that his death was the result of an act of the one murderer and of an act of the other murderer. Both did it not ‘jointly’ but ‘individually’.
We must thus think of the changes the not-happening of which are conditions for the performance (performability) of an act that they are changes in nature i.e. such changes as happen independently of the interference of agents. This explains the meaning of the phrase ‘of itself’ which we have been using when speaking of those changes.
8. Beside acts we have also to study their ‘correlatives’ forbearances. What is it to forbear (to do) something?
Forbearing is not the same as not-doing simpliciter. That one forbears to produce through action the change described by ~pTp or the state of affairs described by p cannot be described by ~d( ~pTp). If for example a certain window is closed on a certain occasion one does not close it on that occasion—but neither does one forbear closing it then. Furthermore things which are beyond human capacity to do (e.g. to change the weather) one does not do—but neither does one forbear doing them.
It is also clear that forbearing cannot be defined as the doing of not-changes. d( ~pT ~p) does not mean that an agent forbears to produce the state of affairs described by p. It means that he (‘actively’) prevents this state from coming into existence—e.g. keeps open a door which otherwise would close.
It seems that forbearing cannot be defined in terms of action and change (and truth-functional notions) alone. But we can define it in terms of action change and ability. We propose the following definition:
An agent on a given occasion forbears the doing of a certain thing if and only if he can do this thing but does in fact not do it.
The notion of forbearing thus defined is the logically weakest member of a series of progressively stronger notions of forbearing. On our definition forbearing to do something which one can do does not presuppose awareness of the opportunity. In a stronger sense of ‘forbear’ an agent forbears only such action as he knows he can perform on the occasion in question. In a still stronger sense an agent forbears only such action as he knows he can perform but decides (chooses prefers) to leave undone on the occasion in question. If in addition he feels an inclination or temptation to do the action which he chooses not to do then he is in a still stronger sense forbearing it. Of this strongest sense of ‘forbear’ we also use such words as ‘abstain’ or ‘forsake’.
We shall introduce a special symbol f for forbearing.
f( ~pTp) shall mean that one forbears to change through action a ~p-world to a p-world. We shall call this kind of action forbearing to do. The forbearance described by f( ~pTp) is possible only in a ~p-world which does not on the occasion in question ‘of itself’ change into a p-world. For example: To forbear to close a door is possible only provided this door is open and does not close ‘of itself’.
f(pT ~p) means that one forbears to destroy (annihilate undo) the state described by p. This forbearance is possible only in a p-world which does not on the occasion in question ‘of itself’ change into a ~p-world.
f(pTp) means that one forbears to preserve the state of affairs described by p. This is possible only in a p-world which will on the occasion in question change into a ~p-world unless the change is prevented through action.
f( ~pT ~p) finally means that one forbears to suppress the state of affairs described by p. This is possible only in a ~p-world which will on the occasion in question change into a p-world unless the change is prevented through action.
The modes of conduct which we have just been discussing we shall call the four types of elementary forbearances. It should be clear in which sense we can talk of ‘corresponding’ elementary acts and forbearances. To the elementary act described by d(pTp) corresponds the elementary forbearance described by f(pTp) and so forth.
Forbearing just as much as acting has results and consequences.
Primarily the results of forbearing are that certain changes do not take place. Thus the forbearance described by f( ~pTp) results in that the change described by ~pTp does not occur. And similarly for the other elementary types of forbearance.
There is a prima facie objection to this way of arguing which has to be answered. The result of my forbearing to open a certain window say is that I do not open it. But what if somebody else opens it? Cannot in spite of my forbearance the change from ‘window closed’ to ‘window open’ take place as the result of some other agent's interference with the state of the world? The answer seems to me to be this:
At the very moment when another agent opens a window which I have up to this moment forborne to open the opportunity for (continued) forbearing gets lost. What I may forbear to do when the window is being opened by another person is to keep the window closed or to prevent the other person from opening it. But I can no longer forbear to open the window. Thus my forbearance to do this will necessarily be ‘reflected’ in the fact that the window remains closed.
Using our generalized notion of change which includes not-changes we can also say that the results of forbearing are that certain changes take place. There is a certain convenience in this mode of expression.
Thus instead of saying that the forbearance described by f( ~pTp) results in the fact that the change described by ~pTp does not take place we can say that it results in the fact that the change described by ~pT ~p takes place. For if the state of affairs described by ~p obtains on a certain occasion and if the world does not change in this feature then—by the laws of the Logic of Change—the world remains unchanged in this feature.
By similar argument we can say that the forbearances described by f(pT ~p) f(pTp) and f( ~pT ~p) respectively result in the changes described by pTp pT ~p and ~pTp respectively.
Instead of calling certain changes the results of forbearance we can also call certain states of affairs the results of forbearance. These states are the end-states of the resulting changes. Unlike the correspondence between forbearances and changes as their results the correspondence between forbearances and states is not one-to-one but one-to-two. Thus e.g. the state of affairs described by p can be the result either of a forbearance to prevent it from coming into being or of a forbearance to destroy it. It can also as will be remembered be the result either of an act of doing or of an act of preserving it. Finally this state can exist without being the result of either an act or a forbearance.
It should now be clear what has to be understood by the consequences of forbearances—and also clear that forbearance can have consequences. The consequences of a certain forbearance are the consequences of the state or change which is the result of this forbearance. Thus e.g. if the state described by p is the result of a forbearance to prevent it from coming into being then everything which is a consequence of the change described by ~pTp is a consequence of this forbearance. There is no difference ‘in principle’ between the consequences of acts and of forbearances. (This is a logical observation of some importance to a certain type of ethical theory.)
We can exhibit the correspondences between the elementary acts forbearances and changes together with the conditions of acting and forbearing and the results of acts and forbearances in a table. (See p. 49.)
In ordinary language it seems the words ‘act’ and ‘action’ are used pretty much as synonyms. The philosopher is free to give to the two words different meanings for the purpose of marking some conceptual distinction which he thinks important. Here I shall employ the term ‘action’ as a common name for acts and forbearances. Acts and forbearances we could say are two modes of action.1
9. From the discussion of acts (and forbearances) we now move to a discussion of abilities or the notion of ‘can do’.
We have distinguished between generic and individual acts (Section 2) between the result and the consequences of an act (Section 5) and between act and activity (Section 6). These distinctions are relevant to the present discussion.
When do we say of an agent that he can do a certain thing—for example can open a window or can get up from his bed or can tell a lie? This is a very complicated question. What is said about it here will be confined to a necessary minimum for our theory of norms.
To be able to do some act we shall say is to know how to do it. Sometimes we can also say that it is to master a certain technique. The mere fact that by some haphazard movements of my hands and fingers I succeed in opening a door with a complicated lock-mechanism does not entitle me to say that I can open a door with this type of lock. But if normally i.e. on most occasions when I set myself to the task I succeed in opening the door without much trial and error then I may be said to be able to do this sort of thing. I then know how to do it. I also master a certain technique.
|Condition of action||Act or forbearance||Result of action|
p is but vanishes
p is preserved
|unless preserved||d(pTp)||pTpp remains|
one lets p vanish
p is and remains
p is destroyed
|Same f(pT ~p)
one lets p remain
p is not and does not
p is produced
one lets p remain
p remains absent
p is not but happens
unless suppressed d( ~pT ~p)
p is suppressed
p remains absent
|Same||f( ~pT ~p)
one lets p happen
Ability to do a certain act must be distinguished from ability to perform a certain activity such as to walk to run or to speak. Of the ability to perform a certain activity we do not normally use the phrase ‘know how’. A child who has learnt to walk or to speak is not ordinarily said to know how to walk or to speak. But ability to perform an activity can sometimes quite naturally be characterized as mastership of a technique; for example when a child has learnt to handle knife and fork in eating.
The ‘can do’ which we are here discussing is the ‘can do’ of acts and not the ‘can do’ of activities.
One can make a distinction between ability and skill and relate it to a distinction between knowing how and having the mastership of a technique. The man who is able to do a certain thing knows how to do it. Only if the activity which is involved in doing the thing is of a complicated kind does this ability amount to mastership of a technique. When it does this we call such ability a skill.
We can also make a distinction between ability and capacity. Capacity often has the character of ‘second order’ ability. It is within a man's capacity to do a certain thing we may say when he can acquire the ability or skill needed for doing this thing although he does not yet possess it.
On the view of ability which we are here adopting a criterion for judging truly that a man can do a certain act is that normally on most occasions for doing it he should succeed. But is this not like saying that he can do something only if on most occasions he can do this? Are we not moving in a circle here?
I do not think that we have a circle here but a noteworthy shift in the meaning of certain words. That I ‘can do’ something has a different meaning when it refers to an act-individual and when it refers to an act-category. That on some occasion a certain state of affairs say that a door is open comes (came) into being as a consequence of some activity on my part say some movements of my hands and fingers is a necessary and sufficient condition for saying that I can (could) do this thing or produce this state on that occasion. The sole criterion of the ‘can do’ is here the success of certain efforts. Of this ‘can do’ no ‘know how’ and no reasonable assurance of success before the attempt is required. These are requirements of that ‘can do’ which refers to act-categories and which alone amounts to ability. It is moreover only when these requirements are satisfied that consequences of activity assume the character of results of action.
I shall call the ‘can do’ which refers to individual acts the can do of success and that which refers to generic acts the can do of ability. The first ‘can do’ is always relative to an occasion for acting. The second is independent of occasions for acting. By this I mean that it makes no sense to say that we can do—in this sense of ‘can do’—the thing on one occasion but not on another—unless that other occasion belongs to a stage in our life-history which is either before we have learnt to do this thing or after we have forgotten how to do it.
Before we have acquired the ability success and failure on the individual occasion for acting is the only sense in which we can or cannot do a certain thing. When the ability (or skill) has become acquired however we can also do things which we sometimes fail to accomplish in spite of efforts. We may fail because of some unforeseen obstacle or because another agent interferes and prevents us from completing the act. When this happens we describe what we did on the occasion by saying that we tried to do something but failed.
Yet as already observed we cannot be said to have the ability unless on most occasions when we set ourselves to do the act we succeed in doing it. In this way success can be said to be the measure and criterion of ability and yet the meaning of the ‘can do’ of ability be different from the meaning of the ‘can do’ of success.
10. It would be a mistake to think that whenever an agent has successfully accomplished an act he has also tried to accomplish it. A similar remark can be made of activity. Normally when I shut a door or walk or read I cannot be said to try successfully to shut the door or to move my legs or to read out the words. To construe every act as a result or consequence of trying to act would be a distortion.
Although doing does not entail trying to do it would seem that ability to do entails capacity for trying to do. If I can do I also can try.
It would also be a mistake to think that although one cannot do any given thing one can at least try to do it. One cannot for example jump to the moon. But can one not try to jump to the moon? It is not clear what sort of behaviour we should describe as trying to jump to the moon’. Not until we have at least some idea of how to do a thing can we try to do that thing. To ‘have an idea’ of how to do a thing again presupposes that we are not convinced that it is humanly impossible to do that thing. Since we are convinced that it is humanly impossible to jump to the moon in the ordinary sense of ‘jump’ we can rightly say that we are unable even to try to perform this feat. To say ‘I try although I know that I shall fail’ is to state a contradiction in terms.
There are also many things which I know are humanly possible and which I may learn or otherwise acquire the ability to do but which at present I cannot even try to do because of my ignorance.
The question may be raised: Is trying act or activity? In the course of trying to do something one may perform various acts. But basically trying seems to me to belong to the category of activity. Trying to do something may as we say ‘result’ in the act's being successfully performed. But performing the act is not tied to trying to perform it in the same way as the resulting change is tied to the doing of the act. One is inclined to call the successful performance a consequence rather than a result in our sense of the terms. It is contingent whether trying leads to result but it is necessary that acting result in change.
When an agent tries to do something which he can do but fails to accomplish the act has he then forborne to do the thing in question? We are free to answer Yes or No depending upon how we wish to mould the notion of forbearing. Here we shall understand ‘forbear’ in such a way that unsuccessful trying to do something which it is within the agent's ability to accomplish counts as forbearing. ‘Forbearing to do’ in other words will not be understood so as to entail ‘forbearing to try’.
On this ruling doing and forbearing are two jointly exhaustive modes of action. If an agent can do a certain thing then on any given occasion when there is an opportunity for him to do this thing he will either do it or forbear doing it.
We could however also mould the notions in a way which would make unsuccessful trying a ‘middle’ between doing and forbearing. That which is here called ‘forbearing’ could then be called ‘failing to do’ or ‘leaving undone’. Perhaps it would be more in accordance with ordinary usage to call the two jointly exhaustive and mutually exclusive modes of action ‘doing’ and ‘failing to do’ rather than ‘doing’ and ‘forbearing’. But usage seems to be vacillating. The important thing is not whether we should choose this or that terminology but that it should be clear how the terminology actually chosen is to be understood.
11. The notion of forbearing we have thus defined (see Section 8) that ability to do and ability to forbear doing the same thing are reciprocal abilities.
It may appear more plausible to say that what an agent can do he can also forbear doing than to say that what an agent can forbear doing he can also do. It is inviting to think that it is somehow ‘easier’ to forbear than to do and that there are many more things which we can forbear than things we can do.
The appearance of asymmetry between the abilities is partly I think due to confusions and to neglect to observe certain conceptual distinctions.
First the idea may be guilty of a confusion between not-acting and forbearing. On an occasion which is not an opportunity (see Section 3) for doing a certain act an agent necessarily does not do this act. Of course we could so define ‘forbear’ that an agent is said to forbear those acts also which he has not even an opportunity of doing. But this would be an odd use. We should then have to say such things as that an agent who is in a room where the window is open ‘forbears’ to open the window in that room.
Secondly the impression that there are more things one can forbear than things one can do may be due to a confusion between doing and trying to do. Consider a man on the bank of a river which is as a matter of fact too wide for him to cross swimming. He cannot cross it but if he can swim and is not certain whether he will reach the other shore he may try to cross it. If he can try he can also forbear trying. He forbears trying by not plunging into the water and setting out for the other shore. This is exactly the same ‘negative behaviour’ by which our man would manifest his forbearance to swim across the river could he perform this feat. But this does not mean that forbearing to swim across the river and forbearing to try to swim across the river are one and the same thing. They are different just because they are forbearances relative to different modes of action.
In a sense therefore forbearing is precisely as ‘difficult’ as doing. But in another sense forbearing can quite rightly be said to be normally easier than doing. That an agent can or is able to do a certain thing shall mean we have said (Section 9) that he knows how to do this has learnt it sometimes that he has acquired mastership of a technique. However in order to be able to forbear an agent need not normally learn anything in addition to learning to do the thing in question. We could express this insight in several ways. ‘Can do’ we might say is prior to ‘can forbear’ although the two ‘cans’ are reciprocal. There is no special ‘know how’ of forbearing.
Consider again the eight elementary acts and forbearances which answer to a given state of affairs.
An agent's abilities with regard to corresponding acts and forbearances we have found are reciprocal. That which an agent can effect as a result of his action he can also forbear to effect as a result of his action and conversely.
An agent's abilities with regard to acts and/or forbearances of different types but relative to the same state of affairs are not reciprocal. They are on the contrary logically independent of one another. The fact that an agent can through his action destroy a state of affairs which exists and does not vanish ‘of itself’ is no guarantee that he can produce this same state of affairs if it does not exist and does not come into being independently of his action. There are plenty of examples of this. To take a drastic but convincing one: Men can kill each other but they cannot raise the dead. Generally speaking: that a man can or cannot do the act described by d( ~pTp) is logically independent of the proposition that he can or cannot do the act described by d(pT ~p).
The same seems to be true ‘in principle’ of the pairs of acts described by d( ~pTp) and d(pTp) and by d(pT ~p) and d( ~pT ~p) respectively. That I can suppress something which happens unless it is suppressed does not entail that I can destroy it if it exists. Nor does the converse entailment hold. And that I can prevent from vanishing something which exists does not entail that I can produce it if it does not exist or vice versa.
12. There are two types of act which are of great importance to deontic logic and which relate to one agent's ability to interfere with the ability of another agent to perform a certain act. These are the types of act which we call hindering or preventing and compelling or forcing.
These two types of act are obviously interdefinable. Therefore we can here limit the discussion to one of them. To compel an agent to do something is the same as to prevent him from forbearing this thing. And to hinder an agent from doing something is the same as to force him to forbear it.
To hinder an agent from doing something is to act in such a manner that it becomes impossible for that agent to do that thing. To hinder or prevent is to ‘make impossible’. The result of the act of hindering an agent from doing a certain thing on a certain occasion is to change the world in such a way that the agent cannot do that thing on that occasion. But this result be it observed can be effected only on condition that the agent can do this thing. One cannot prevent people from doing that which they in any case cannot do. An act of preventing thus results in the fact that an agent in some sense cannot do that which he in some sense can do.
This may look like a paradox though it certainly is not one. But it is an interesting illustration of the two senses of ‘can do’ which we distinguished and discussed in Section 9. The sense in which one must be able to do something in order to become prevented from doing this thing is that sense of ‘can do’ which refers to act-categories. The sense again in which one is not able to do that which one has become prevented from doing is the sense of ‘can do’ which refers to act-individuals. Preventing from doing does not annihilate ability to perform the generic act. Preventing on the contrary presupposes this ability and destroys the successful exercise of it only on an individual occasion.
This of course is not to say that abilities could not become annihilated or destroyed as a result or consequence of action. By injuring a person I may temporarily or even permanently make him unable to perform a certain generic act which before he could do. This however is not what we ordinarily call ‘preventing’. We call it ‘disabling’.
- 1. Sir David Ross, too, makes a distinction between act and action. The distinction for which he uses the two terms is quite different from the distinction for which they are used here. See The Right and the Good (1930), p. 7.