MIDWEST STUDIES IN PHILOSOPHY, I11 (1978)

RULES, RIGHTS, AND PROMISES.

I
H U M E had two theses about promises: one, that a promise is “naturally unintellig-
ible,” and the other that even if ( p e r impossibile) it-were “naturally intelligible” it
could not naturaZly give rise to any obligation.
I regard his discovery of natural unintelligibility as a great one, of wider application
than he gave it. But it is also quite difficult to understand.
His own exposition of these doctrines is tied up with his philosophical psychology
and metaphysics and ethics, as is made clear by the following quotation. “If anyone
dissent from this [sc. that promises have no force antecedent to human convention] he
must give a regular proof of these two propositions, viz that there is a peculiar act of
the mind, annext to promises; and that consequent to this act of the mind, there arises
an inclination to perform, distinct from a sense of duty.” Those who are familiar with
Hume’s ethics will understand this last bit, others will not. But my interest in the
subject is not exegetical. I believe that Hume did hit upon a problem of intelligibility,
as indeed is attested by the large literature on the topic of promises in present day
analytic philosophy, and that he was correct both in calling promises “naturally unin-
telligible” and in framing two theses - and consequently for us distinguishing two
problems - one, concerning what sort of beast a promise is, and the other, Concerning
how, given that there is such a thing, it can generate i n obligation.
Hume obviously believed that he was pointing to a contrast between promises and
words for perceived objects and events. These, then he thought to be “naturally intel-
ligible,” and the activity of meaning in which they are employed, itself also to be so. A
word evokes an image, or (presumably) an image or sensation evokes a word: the
image (or sensation) gives an example of what the word stands for. It will be the upshot
of the present paper that no language is in Hume’s sense naturally intelligible. Nor is
the contrast necessary for getting hold of what he was getting hold of. He saw that a
promise was not a phenomenon, and so that “promise” was not a word for which his
318
G. E. M. ANSCOMBE 319

picture of our understanding would work at all. I shall be arguing that no naturalistic
account of a rule, as of a promise, will work: it will follow that words and their relation
to their meanings aren’t “naturally intelligible” either. For the use of words involves
following rules; hence an account of language must make reference to rules. (Not
merely to regularities.) This, if I can put my finger on it rightly, will show just what is
wrong with empiricism. Here I want to approach this matter by explicating Hume’s
insight about the two problems concerning promises.
One may fail to notice that there are two problems, because a promise signijies the
creation or willing of an obligation. It might be thought that if you could show how
there can be a sign with that signification, you would be home and dry: the obligation
is generated by the giving of a sign which has that signification! Hume’s clarity of mind
perceived that this is not so.
We might indeed argue that it was not so, from the fact that a promise may not
generate any obligation to perform it - a promise to do something wicked, for exam-
ple. Showing this might however leave someone puzzled; or it may be seen as a matter
of defeasibility. So a direct attack is better.
I might say “Let there be a constraint upon me to do such and such.” This is a sign
signifying a will to be constrained. It is clear that we could understand this, and still go
on to ask “Will there be any such constraint?” Suppose I say “Let there be a legal
obligation on me to. . . .” we may ask whether one was in fact brought about - for
example whether my lawyer, to whom perhaps I gave this instruction, effected what I
asked. If - by my merely having pronounced those words in appropriate circum-
stances and before witnesses - there is automatically created the legal fact, then this
must be by a special nile of law.
Suppose that instead of “Let there be a constraint upon me to . . . ,” I had said “Let
there hereby be a constraint . . .” or “There is hereby a constraint. . . .” We may
still ask “And was there one? - with, of course, the extra question: was it thereby? It
just wouldn’t b e an acceptable answer to say “Of course, because she said it - it
doesn’t need anything but that (given the circumstances) to make it so, and you can see
that from the meaning of the words.”
What does “hereby” mean? Imagine that I write on the blackboard: “ I am hereby
writing on the blackboard.” What I wrote would be true. Again, “Let there hereby be a
constraint on me. . . .” or “There is hereby a constraint. . . .” might be an utterance
that was required in order to set off a paralyzing device. The difference between these
two is merely that we can see that the proposition about the writing is true, if w e see
the writing going on, and we probably don’t hear that the constraint is brought about
by hearing the utterance. For the utterance is not itself a constraining mechanism as
the writing is itself the verification of the other sentence.
Now someone will say “But that’s not whut ‘hereby’ means in, say, a contract!” No
doubt he’d be right. But let us just notice that $“hereby” did mean that, “1 hereby lay
myself under an obligation . . .,” (which would be an expression saying that an obli-
gation was created) would necessarily always leave the question to be answered: “And
did he? Did the obligation get created?” And this exemplifies Hume’s point: we have
(a)tried to imagine a natural phenomenon of creation, by the utterance of a sign, of the
sort of constraint we call obligation; (b) understood the sign to be one whose meaning
is that such an obligation is being created; and (c) seen that the question would still
arise whether the obligation had actually got created, and could not be proved to have
the answer “Yes” by reporting that the words had been uttered. (Just as, if a door opens
when 1 say to i t “I hereby open you,” that doesn’t mean that my saying those words
itself, in suitable circumstances, is enough to prove that the door is open.)
320 MIDWEST STUDIES IN PHILOSOPHY

The natural phenomenon might not have occurred, the mechanism might not have
worked. Note that this is a completely different point from the point that the mere
utterance of an expression: “I promise to do such-and-such” isn’t even necessarily a
-
promise a suitable setting is needed - and that even when it is, it doesn’t necessar-
i l y generate an obligation to perform it.
So we are down to understanding this special use of “hereby.” For “hereby” - not
as in “I am hereby writing on the blackboard” etc. - is what SO to q e a k at least
implicitly enters into promises. By the way, it was absurd of Hume to write as if there
had to be a special sign of promising. “I’ll help you today; will you help me tomor-
row?’ “Yesl” Here are promises given and received. The question is what it is for
them to be promises. They are or contain descriptions of possible future states of
affairs. They are made true by performance. But they are not mere predictions, not
even merely the sort of prediction which is an expression of intention. And one might
say: this “hereby” is something attached to that in them which goes beyond being
predictions and expressions of intention.
Hume’s own conclusion was “that promises have no force antecedent to human
conventions.” If this is found offensive, that will be by a misunderstanding. Cod
himself can make no promises to man except in a human language. The regularity of
the seasons, and the applicability of the rational probability calculus in matters of
chance, are not divine promises. A spouse who has always come home at a certain hour
has not eo ips0 broken any promise (only perhaps acted inconsiderately) by suddenly
and wilfully failing to do so. There is indeed such a thing as implicit contract, and there
might be one in this case. But that is two-sided. Mere fostering of expectations can’t be
making an implicit contract.
The rightness of Hume’s conclusions is independent both of his psychology and of
his theory of the foundation of morals in a peculiar “sentiment.”
I am not interested here in the conditions requisite for making ‘I will. . . .” into a
promise. What I am interested in is this “hereby” aspect of promising. We have seen
what the “hereby” is not. Because it is not that, the significance of a promise is that it
not only of itself (that is, without a mechanism) but by its signijicance purports to make
it the cuse that there is a new ob1igatio.n. The promise contains (perhaps on the face of
it just i s ) a future-tense description which the giver then makes come true - or he
breaks the promise. The obligation is a kind of necessity to make the description come
true. But what sort of necessity is that?
We may say: the necessity is one of making the description come true - or being
guilty of something. Of what? Of breaking a promise. And what is that? A description
which someone gives and which because he has given it he must make come true or be
guilty. Of what? Not just to go on running round in the circle let’s try again and say: of
an injustice, a wrong against the one to whom the sign, the description, was given. But
.
what wrong was that? The wrong of breaking a promise. . . We are back in the circle
after all. A wrong is the infringement of a right. How does telling someone one will do
-
something give him a right against one? Well, it does if it’s a promise! Let’s have a
sign for its being that, say “I promise,” put in front of the prediction. Or, because we
know that too well, let’s invent one; “I blip.” It’s not the prediction by itself that it’s an
offence not to make come true, it’s the “blipping” of it, or its being a blip. And what is
the meaning of its being a blip? That it’s an offence not to make the attached descrip
tion come true. But what offence? The offence of going contrary to a blip. - It seems
clear that we just haven’t explained what blipping is at all. Even if we could somehow
get out of the circle, we have the problem: how on earth can it be the meaning of a sign
G . E. M. ANSCOMBE 321
that by giving it one purports to create a necessity of doing something - a necessity
whose source is the sign itself, and whose nature depends on the sign.
That is Hume’s first problem, translated into philosophically neutral terms. In the
next section I shall point to a solution of it, which fully justifies Hume in his own
solution.

I1
What we have to attend to is the use of modals. Through this, we shall find that not
only promises, but also rules and rights, are essences created and not merely captured
or expressed by the grammar of our languages.
Mod& come in mutually definable related pairs, a s : necessary, possible; must, need
not; ought, need not, etc.; together with modal inflections of other words.
When it is said that something must be, or can’t but be, sometimes this is true only if
it actually is then; sometimes only if it is later; sometimes neither matters. For the first
“It must be in this drawer!” is an example; for the second: “So and so can’t but win!”
For the third “You have to move your King”; or “You can’t wear that!” This is a minute
selection fiom the extremely multifarious use of modals, only a still smaller range of
which has much interested philosophy in the past. Aristotle indeed made a little noted
observation that one sense of “necessary” is: “that without which some good will not
be attained or some evil avoided”; and in our time there have been developments of
deontic logic, which shows a consciousness of kinds of modality beyond what used to
be attended to.
I want to arouse interest in one range of uses, which constitutes what I’ll call “stop-
ping modals.” These are of course negative; corresponding positive ones, or the posi-
tive form into which any negative modal can be put, we may call “forcing modals.” The
negative gets priority; it is I think more frequent than the positive, which restricts
one’s action to one thing. (Just as “thou shalt nots” tend to leave you freer than “thou
shalts.”) If I say “You can’t wear that!” and it’s not, for example, that you are too fat to
get it on, that’s what I call a stopping modal.
Think of the game played with very small children where several players pile their
hands on top of one another. Then, if one of them doesn’t pull his hand out from the
bottom, you say “You have to put your hand on top”; it’ he pulls it out too soon you say
“No, you can’t pull it out yet, so and so has to pull his out first.” “You have to,” and
I‘
you can’t” are at first, words used by one who is making you d o something (or
preventing you), and they quickly become themselves instruments of getting and
preventing action.
After all, once this transformation has taken place, the following is true: in such a
case you are told you “can’t” do something you plainly con, as comes out in the fact
that you sometimes do. At the beginning, the adults will physically stop the child from
doing what they say he “can’t” do. But gradually the child learns. With one set of
circumstances this business is part of the build-up of the concept of a rule; with
another, of a piece of etiquette; with another of a promise; in another, of an act of
sacrilege or impiety; with another of a right. It is part of human intelligence to be able
to learn the responses to stopping modals without which they wouldn’t exist as linguis-
tic instruments and without which these things: rules, etiquette, rights, infringements,
promises, pieties and impieties would not exist either.
A stopping modal is very often accompanied by what sounds like a reason. “YOU
322 MIDWEST STUDIES IN PHILOSOPHY

can’t move that, the shelf will fall down” for example shows the “can’t” to be of the
type Aristotle remarked.
Similarly if I say “You can’t sit there, it’s N’s place” -and it’s clear that this is not just
a personal decision of mine on a particular occasion. Now this form: “You can’t. . . ;
.
it’s N’s. . ,”though it has other applications as well, as also the form par excellence in
which a right is ascribed to N. The applications under this heading may be very
.
various. It may be “You can’t stop N from. . , it’s N’s to. . .”; or somthing may be
N’s only here and now (as we’d say of the parking-space that he’s got into) or it may be
quite generally something that only N can do or have, or N and people like N in some
respect; or others only if these do not seek to do or have the thing.
In these forms of statement the second part: “It’s N’s . . .” has a peculiar role. It
appears to be a reason. And it is a “reason” in the sense of a logos, a thought. But if we
ask what the thought is, and for what it is a reason, we’ll find that we can’t explain them
separately. We can’t explain the “You can’t” on its own; in any independent sense it is
simply not true that he can’t (unless “they” physically stop him). But neither does “It’s
.”
N’s . . have its peculiar sense independent of the relation to “You can’t.’’ Of course,
once these linguistic practices exist, we can detach the two parts from one another and
“It’s N’s” can appear as an independent reason, for example, a reason why one will not
do something.
Let me now restrict the word “reason” (in the context of action) to something inde-
pendent which someone puts forward as his reason for what he does. And let me adopt
the word “logos” (I might also use “theme”) for the second half of “You can‘t. . . .”
“because. . . .” where the two halves are not independent. I shall say that there are
various logos-types, and that the name of the general logos-type is an abstraction from
many particular cases; a label which tells you the formal character of the stopping
modal. For example, one logos-type is a right, and another, very general one, is: a rule.
Thus if you say “You can’t move your king, he’d be in check” “He’d be in check,”
gives the special logos falling under the general logos type: a rule of a game.
Consider the learner in chess or some other game. Of course: “You have to move
your king, he’s in check” is equivalent to “The rules of the game require that, in this
position, you move your king.” But a learner may not yet have this idea: The rules of
the game require. . . . Accepting it when told “You have to move your king, he’s in
check,” is part of learning that very concept: “the rules of the game require.” Requir-
ing is putting some sort of necessity on you, and what can that be? All these things
hang together at some early stage: Learning a game, learning the very idea of such a
game, acquiring the concept of “you have to” which appears in the others’ speech,
grasping the idea of a rule. Nor is there a distinct meaning for “being a rule of the
game” (unless the general idea has been learned from other games) which can be used
to explain the “you have to” that comes into that learning.
Now for the parallel between rules and promises. This is obscured by the fact that a
promise is essentially a sign and the necessitation arises from the giving of the sign.
But the problem of necessitation is nevertheless similar. I may point to a sign which
states a rule. Like a promise, it contains some sort of description. What we conceive to
be the necessity of acting so because of a rule is indeed not generated by the rule’s
being uttered. But the problem is similar: just as we ask what a promise is more than a
mere expression of intention, we may ask what a rule is beyond a mere regularity. In
explaining this one will say for example that the rule is given in a formula for acting,
whose meaning is that one “must” act thus in accordance with it. But even if a formula
-
can have such a meaning, why “must” one? Because that is the meaning of the rule.
G . E. M. ANSCOMBE 323
But what is a rule? And, as with promises, even if we could somehow get out of this
circle, we’d still have the problem: how can a formula have such a meaning?
The problem does not seem so acute, because the rule may merely be the rule of a
practice which you are at liberty to engage in or not. There is after all no necessitation,
other than one of following these rules i j you wish to engage in a practice which is
partly defined by them. So it doesn’t strike one that even this conditional necessitation
is problematic.
When it comes to rules of logic, it is otherwise. Let u s not speak of variant logics; that
is a mere distraction. For even in a variant logic, there will always be the question
whether a rule has been followed. According to the rule, you can’t do that; perhaps you
must do this. You can’t have this and that, you must allow this transition. Now how can
a rule tell you that you can’t do something? A rock barrier may be a natural sign that
you can’t go this way; or a person can tell you you can’t do something. But a rule?
These “musts” and “can’ts” are the most basic expression of such-and-such’s being a
rule; just as they are the most basic expression in learning the rules of a game, and as
they are too in being taught rights and manners. But they aren’t, in Hume’s phrase
“naturully intelligible.” The mark of this is the relation of interdependence between
the “you can’t’’ and the “reason” where this is what I have called the theme or logos of
the “you can’t.” These musts and can’ts are understood by those of normal intelligence
as they aIe trained in the practices of reason.