Some Sub-Atomic Particles of Logic1

R. M. HARE

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
I can best begin this contribution to the articulation of speech acts with
a consideration of the Fregian 'assertion-sign', or, to speak more accurately,
'judgement-stroke'. This correction itself warns us that there is not just
one kind of sign that has to be examined, but several, and that an
elucidation of this difficult subject has to begin with a careful distinction
between them. Only then can we see which of these signs are necessities,
or even possibilities, for logic—that is, what we are to say about
Wittgenstein's complete dismissal of Frege's sign in Tractatus 4.442 and
subsequently as 'logically quite meaningless'. I submit this discussion as
a penance for having failed to make the necessary distinctions clear in my
first book The Language of Morals, in spite of being at least partially aware
of them at the time.2 I later made some of them in print, especially that
between what I shall be calling signs of subscription and signs of mood.3
In spite of this, the distinction is still often neglected; in particular, both
Michael Dummett's and Donald Davidson's discussions would have been
a great deal clearer if they had been more attentive to it.4 I should perhaps
add that I have found Dummett's treatment of assertion rewarding, and
agree with most of it.
Let us then look at some of the different things that are done by what
I called in that book the 'neustic'. I am going to start with the easiest
customer. I do not think that anybody could, on reflection, deny that a
logical notation needs a sign of mood, if it is to handle sentences or speech
acts, or different kinds of things that are said, in different moods. This is
1
This paper was to have appeared in the Festschrift in honour of J. O. Urmson. It had to be
withdrawn because the Stanford University Press would not print it with the punctuation I think
correct. I could not agree to a style which obscures the relation of punctuation to logical form,
especially in a paper much of which is concerned with this relation. Naturally I am very grieved not
to be able to honour Jim Urmson, as an old friend and a distinguished philosopher, in that volume;
but, being himself a man of principle, he has been most understanding of my position, and I honour
him now in the happier environment that the editor of Mind has kindly afforded.
1 was asked to contribute to the Festschrift a paper about speech act theory, to which he has been
from the beginning a notable contributor, although it is far from being his only interest, any more
than mine. In particular, his paper 'Parenthetical Verbs', Mind, 1952, could be said to have anticipated
some of Austin's insights, and his 'On Grading', Mind, 1950, was of such importance that, as I can
testify from my own development, it opened up a whole new front for advances in moral philosophy.
He read it, I think for the first time, at a seminar in which he kindly invited me to join him at Christ
Church, Oxford in the late forties.
2
See appendix to the first paper of my Practical Inferences, London, Macmillan, 1971.
3
See my 'Meaning and Speech Acts', reprinted ibid., p. 74 from Philosophical Renew, 1970.
4
D. Davidson, Inquiries into Truth and Interpretation, Oxford, Oxford University Press, 1984,
p. 110; M. Dummett, Frege: Philosophy of Language, London, Duckworth, 1973, pp. 308 ff.

Mind, vol. xcviii, no. 389, January 1989 © Oxford University Prera 1989
24 R. M. Hare

evident, at any rate, if we need, in our logical notation, to have different

expressions for sentences whose meaning is different. For it is clear that,
for example, the Latin words 'i' and 'ibis' have different meanings, namely
'Go' and 'You will go'. I can see only two ways in which this conclusion
could be avoided. One would be by confining logic to sentences in the
indicative mood. This is sometimes advocated on the ground that logic
deals with what follows from what, and nothing follows from anything

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
unless both are propositions having truth values, and only sentences in
the indicative mood can express propositions having truth values. In reply
to this it is not necessary to say that imperatives like 'Go' can express
something that can follow or be followed from logically—that thorny
question can be left on one side. All that need be said is that logic is
concerned also, and perhaps more primitively, with inconsistency, and
that anybody who does not think that the two commands 'Go' and 'Do
not go' are inconsistent does not know English. We need therefore a logic
which can tell us how to separate the pairs of commands which are
inconsistent from those which are not.
The second way out would be to say that, although 'Go' and 'You will
go' mean different things, the difference is of no interest to logic, because
it makes no difference to the inferences and other relations that logic
studies. But this at any rate appears to be false. There is an inconsistency,
as I have said, between 'Go' and 'Do not go'; and there is an inconsistency
between 'You will go' and 'You will not go'; but there is none between
'You will go' (where this sentence expresses a statement) and 'Do not go'
(expressing a prohibition). At least, not an inconsistency of the same sort.
It may be that if a single person uttered both, he would be saying
something logically odd (although if he were an officer giving an order
that he was sure would be disobeyed, in order to get his subordinate into
trouble, he might say it). But at any rate if two different people said to
the same person 'You will go' and 'Do not go', there would be no logical
inconsistency between what they said. Similarly, the inference from 'You
are going to bring me five apples' to 'You are going to bring me at least
four apples' is valid; but that from 'Bring me five apples' to 'You are
going to bring me at least four apples' is invalid. Therefore a logic which
could not distinguish between the moods would be unable to distinguish
a valid from an invalid inference in this case. It is no answer to this
argument to say that the second of these inferences is invalid only because
all inferences containing imperative elements are invalid; for, even if this
were so, we should need, in order to tell whether inferences were valid,
to have some means of identifying the imperative elements in them which
would destroy their validity, or at least the indicative elements in those
whose validity was above suspicion.
It will not do, even, for logicians to put at the beginning of their books
a rubric to say that all the sentences in the formal part of the books are
Some Stib-Atomic Particles of Logic 25

going to be in the indicative mood; for we need to be able to satisfy

ourselves that the rubric has been observed. On the other hand, if the
rubric says that all the sentences in a book are to be understood as being
in the indicative, that amounts to the same as putting in front of each
sentence a rubric to the effect that it is to be understood as being in the
indicative; and that would be a sign of mood.
I propose to take it, then, that a complete notation needs signs of mood;

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
and for these signs I propose to use, as I have before,5 the term 'tropics',
from the Greek word for grammatical mood. There are many serious
problems about tropics; but I shall not have time to do more than mention
some at the end. In particular, there is the problem of how to distinguish,
for example, the meanings of the imperative and indicative verb-forms.
To this I hope to return in another paper.

The next sign I shall discuss is much more controversial, namely the sign
of assertion. Indeed, many people have, like Wittgenstein, denied the
necessity or possibility of such a sign. What I have said about mood-signs
or tropics may be taken as a warning against a confusion which has
sometimes been made. I mean by 'assertion' here not 'assertion as opposed
to commanding, etc.'; the sign of assertion in that sense would be a mood-
sign or tropic. Nor do I mean 'assertion as opposed to negation or denial'.
Following Frege, but without defending myself here, I shall take it that
negation of the ordinary sort, as opposed to what Searle and Vanderveken
call 'illocutionary denegation',6 can be dealt with as part of what is asserted;
that is, that we can, in this sense, assert either that the cat is on the mat,
or that the cat is not on the mat. I mean 'assertion as opposed to merely
supposing, entertaining or the like'. I shall be explaining what I mean by
this at considerable length. Because the word 'assertion' can be confusing,
I propose now to abandon it, and speak instead of a sign of subscription.
This has the further advantage of being readily applicable to other kinds
of sentences, speech acts, etc. than those expressed in the indicative mood.
As a shorter term for 'sign of subscription', I propose to use my old word
'neustic', from the Greek word meaning 'to nod assent'; I hope that this
paper will purge the word of its former disgraceful ambiguity.7
Are there, can there be, do there have to be, signs of subscription? I
feel inclined to reply that there are, and therefore obviously can be; that
there do not have to be, because signs of non-subscription would do just
as well if employed systematically; so would, alternatively, a ban on saying
anything that you do not subscribe to; but I think that it is a necessity to
5
See ref., n. 3.
6
See J. R. Searlc and D. Vanderveken, Foundations of Illocutionary Logic, Cambridge, Cambridge
University Press, 1985, p. 4, and my 'Some Alleged Differences between Imperatives and Indicatives',
Mind, 1967, s.f, repr. in my Practical Inferences, cited in n. 2.
1
See refs. in nn. 2 and 3.
26 R. M. Hare

have one or other of these devices or rulings—otherwise we should not

know what people were subscribing to and what they were not subscribing
to. Ordinary language is a bit more free and easy than that; but in principle
a language in which it is to be made clear what is being said has to have
some provision for indicating subscription or its absence.
I will start with an illustration of the trouble that can arise if we do
not have such a sign. Professor Davidson, who gave the John Locke

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
Lectures at Oxford many years ago, had been maintaining in them that
inverted commas (which are one way of indicating non-subscription,
though this is not their only function) could be eliminated from logical
notation by means of what he called, following the grammarians, a 'para-
tactic' construction; thus, instead of 'He said "Mary has gone home"',
we could write 'He said that: Mary has gone home', where 'that' is a
demonstrative having as referent the sentence that follows. There is some
etymological support for this device; for it is said that the conjunction
'that' does indeed derive from the demonstrative, and that the sentence
'He said that Mary plays tennis' is derived from a more primitive 'He
said that: Mary plays tennis'.8
I am not suggesting, nor was Davidson, that 'that'-clauses mean the
same as quoted expressions; there are obvious differences, and Davidson
explains them in later work.9 But that is not my point. The point I made
in subsequent conversation with him was that, if the sentence 'Mary has
gone home' is taken out of its insulating inverted commas, we need some
way of telling that it is not being asserted or subscribed to by the speaker.
In answer to this, he said, in effect, that it was perfectly easy to tell that
it was not being subscribed to. I said that I would one day show him that
it was not.
That evening I wrote him the following letter:
Dear Professor Davidson,
I have heard that some of our students, who disapprove of your government's
actions in Cambodia, are going to come and disrupt your next lecture, and I trust
that you will come prepared to shout them down.
The preceding paragraph contains 37 words.
My wife and I would be so delighted if you and your wife could come to lunch
with us on Friday week at our home.
The preceding paragraph contains 25 words.
I am sure that you will easily be able to tell which of the preceding paragraphs
express assertions.
Yours, etc.
I think it is in general true that if there could be a law requiring a
certain practice, there could equally be just a convention requiring it. Now
the law does require signs of subscription in certain cases, and, moreover,
8
Davidson, op. cit., pp. 105 ff. ' Ibid., 79 ff., 93 ff.
Some Sub-Atomic Particles of Logic 27

holds people to what they have so subscribed to. If anybody does not
agree, he should try getting up in court and swearing to speak the truth,
and then saying that he saw the prisoner do what he was accused of, and
then, afterwards, claiming that he was not, actually, asserting this but only
entertaining the idea, or supposing it, or joking, or pretending, or whatever.
For the same reason it was logically absurd, and not just morally
reprehensible, for Hippolytus to say 'My tongue swore, but my mind is

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
not sworn'.
Signatures, for example on cheques, provide another example. They
are not merely for identification of the drawer, whose name is now already
normally printed on the cheque form. They signify subscription (as the
etymology of that word indicates). If one has signed a cheque, one cannot
say one has not instructed one's bank to pay. If it pays, then one will
have no redress against it.
I said that if there could be a law requiring (and I might have added
'or permitting') a certain practice, there could equally be just a convention
requiring or permitting it. The law does not actually require the addition
of signs of subscription to all speech acts, and still less do our ordinary
conventions. Nor do they permit people to claim not to have subscribed
in the absence of such signs. Instead, both in law and by convention, we
have a rather elaborate, though actually pretty stringent, set of require-
ments, combining the use of signs of subscription with that of signs of
non-subscription. So we have always, in principle, a just cause of complaint
if somebody claims not to have subscribed to some utterance he has made,
when according to the conventions he has subscribed to it. Could we
simplify these requirements by insisting (as for purposes of logic we might
feel inclined to do) that all utterances subscribed to were prefaced by a
sign of subscription, and that none not so prefaced was subscribed to? To
understand this question we have to be clear, of course, that I am using
'subscribe' as a word, not for a mental act or state, but for the performance
of some kind of speech act or act of communication. Obviously my signing
of a cheque is not a mental act, and does not even need to be accompanied
by one in order to signify my subscription.
Let us ask, then, could we have a law requiring the universal use of
signs of subscription in front of subscribed-to utterances—a law which
made it impossible to hold someone to an utterance unless he had
subscribed to it by appending this sign? I think we could, though of
course it would be tedious. There is one extremely popular argument,
repeated by Davidson,10 which is designed to show that we could not. It
is said that, even if there were such a sign, there could be uses of it which
were non-subscriptive. Instances would be uses of it on the stage, on the
blackboard, inside quotation marks, and so on. So, it is claimed, the mere
presence of the sign could never guarantee subscription.
10
Ibid. 103.
28 R. M. Hare

I agree that the presence of the sign could never safely be relied on by
the audience as a sure indication that the utterer intended subscription
(as we saw, mental states do not come into the matter, and, moreover,
utterers can pretend to have mental states that they do not have; and
intention is, I suppose, a mental state). But that is not what is meant in
this context by 'guarantee'. What is meant is that the utterer, in using the
sign, and independently of any mental acts he may or may not be

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
performing, gives his guarantee that he is uttering the words subscriptively.
We have to ask, then, whether it could be the case that by law anybody
who put this sign in front of an utterance was to be taken to have
subscribed to the utterance and could not wriggle out of it.
Such a law would obviously be very restrictive. It would, for example,
make play-acting impossible. But, I claim, one could have such a law. We
might compare the laws which there have been against blasphemy, foul
language, etc. If you get up on the stage and blaspheme, it is no use, if
such laws are in force, claiming subsequently that you were not subscribing
to the blasphemy but only acting the part of a blasphemous person. You
might as well try actually killing someone on the stage and then claiming
that the fact that you did it on the stage exempted you from prosecution
for murder.
In our actual practice, as opposed to such a restrictive law, we are much
more flexible. We have conventions which allow us to cancel subscription,
as well as those I have mentioned which allow us to indicate it, like the
signature. The proscenium arch which protects actors is an obvious
example of such a subscription-cancelling device. It is a convention that
things said behind it are not being subscribed to. That is why, as I have
been informed by Christopher Taylor, actors have a rule that if a real fire
breaks out back-stage, the person who discovers it has to shout, not 'Fire!'
(for fear that it might be thought to be part of the play) but some other
expression earmarked for this purpose.
That such protection is necessary is evident. None of us thinks that in
a performance of Cosi fan tutte the singers taking the parts of Ferrando
and Fiordiligi and the other pair have really got married; and, even if it
were a representation of a real lawyer and not Despina dressed up as one,
it would make no difference. But suppose that the performance takes place
in church, and is done by a clerk in holy orders who has read the banns
three times on previous Sundays, and after the whole liturgy has been
gone through according to the rubrics, the parish register is signed and
witnessed. Can we then say that they have not got married but were only
play-acting?
Even on the stage, as we have seen, people are sometimes held to things
which they say. The producer comes on and announces a change in
casting; the chorus-leader says 'Plaudite' (clap) to signify the end of a
Roman comedy; in the parabasis of a Greek comedy the chorus comes
Some Sub-Atomic Particles of Logic 29

forward and addresses the audience with views which are certainly
expressed in earnest, albeit by an agent rather than by their author. Greek
playwrights were sometimes successfully prosecuted for the political
indiscretions they put into the mouths of their actors. So the proscenium
arch and its ancient equivalents are not a complete protection.
I do not think we can get any further without introducing an important
distinction between two quite different kinds of non-subscription.11 The

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
first is that which, up to now, we have been considering almost exclusively.
This I shall call mimesis. It consists in the use of expressions which are,
by convention, signs for performing a certain speech act, but not
performing it in earnest. We have already had examples of this. I wish to
maintain that language would break down unless there were some way of
indicating when one was using mimesis and when one was seriously
subscribing.
Such non-serious, mimetic uses of language are to be distinguished
from a quite different class of cases in which subscription is withheld,
namely those in which expressions that, if isolated, would signify
subscription to something, sometimes do not do so because they are
embedded inside other expressions. Whether embedding always has this
effect is disputed. The most obvious examples are the embedded sentences
contained in conditional clauses, 'that'-clauses (in some contexts), disjuncts,
sentences in quotation marks, and the like. Such non-subscription is
obviously different from the mimetic kind. Whether these two possibilities
exhaust the kinds of non-subscription, I am not so sure. It is not easy to
decide into which category to put the writing of propositions on the
blackboard in a philosophical (as opposed to a historical) lecture. Are we
to say that the blackboard functions like the proscenium arch? Or are we
to say that there are implicit quotation marks round the sentence, just as,
in printed books, quoted passages which are indented are by convention
not given quotation marks, although strictly they should have them? On
the first alternative, we have a case of mimesis; on the second, one of
embedding (the sentence on the blackboard is embedded in the oral
remarks of the lecturer). I do not find it easy to decide whether the
blackboard case has one of these two explanations or whether it represents
some third kind of case different from both, which includes also what has
been called 'supposing' or 'entertaining'. The latter can be done to what
is expressed by complete isolated sentences, and yet is hardly non-serious
in the sense that play-acting is.
It might be argued that no sign of subscription could be of service in
logic or ordinary language, because such a sign would be useless as an
indication of non-mimesis (an actor could always, and would, put it in),
and unnecessary as a sign of non-embedding (because it is evident from
the form of sentences whether a given expression is embedded or not). I
11
I owe this distinction and the word 'mimesis' to Mrs Julie Jack.
30 R. M. Hare

have given my reasons for rejecting the first of these two arguments; we
could have, and in certain legal contexts do have, signs which give the
speaker's guarantee that he is being serious. As to the second reason, I
agree that //embedding and mimesis are the only two categories of non-
subscription (and I will for the present assume for the sake of argument
that they are the only two), then we could have a convention that all non-
mimetic, isolated, non-embedded utterances were to be taken as subscribed

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
to by the utterer. And over a large part of discourse we do have such
a convention. But we equally could, and in parts of discourse do, have
a convention that, on the contrary, subscription should be taken as not
being given unless a sign of subscription is appended. I have given
examples of this. I can see advantages for logic in having the latter
convention made explicit and universal in formalized languages, as Frege
seems to have wished. Certainly it is not impossible to have such a
convention.
This whole subject will, I hope, become clearer when I have discussed
the third of my sub-atomic particles, which I am going to call the sign of
completeness. For then, I hope, we shall understand better what embedding
is. But before I do that, I must add two notes to my defence of the sign
of subscription. The first is that we have in language a class of performative
verbs which are (i) neutral as between the imperative, indicative, and
other moods, and (2) apparently devoted to some sort of subscription,
which can be of varying degrees of what I shall call 'insistence'. The verb
'I insist' is indeed one of these. It can be followed by an indirect command
(in the Kennedy's-Latin-Primer sense of 'command' which I used in The
Language of Morals), as in 'I insist that you go'; but it can also be followed
by an indirect statement, as in 'I insist that he has gone'. Another verb
in this class, but at the opposite end of the same spectrum, is 'suggest';
we can say 'I suggest that you go' or 'I suggest that he has gone'. In
between come some more neutral expressions like 'I tell you' and 'I say'.
We can say 'I say that you are to go' or 'I say that he has gone'; and we
can say 'I tell you to go' or 'I tell you that he has gone'.
I cannot see any need to say that there is a difference here between
senses or uses of these verbs. 'Insist' does not have different senses in the
above examples; in both it expresses very firm and unyielding subscription,
and the difference in the moods of the subordinate clauses looks after the
difference in meaning between the two whole speech acts. The same is
true of the other verbs. This is some support for the view that there is
an expressible operation which can be called subscription. Other specialized
verbs of this sort are 'I swear', which can be constative as in 'I swear he
has gone' or commissive as in 'I swear to speak the truth' (though it has
no use in indirect commands); and 'I advise you', as in 'I advise you to
go' and 'I advise you that the goods you ordered are ready for dispatch'.
But the last case may be a genuine ambiguity in the word 'advise'.
Some Sub-Atomic Particles of Logic 31
The second note is to forestall a possible objection. It may be said that
the conventions that I have appealed to, and the laws, are not linguistic
conventions or laws. On this view, it has nothing to do with the meanings
of words that I cannot get away with swearing in court to speak the truth,
and then stating something, and then saying that I was joking; it depends
on an extra-linguistic convention or rule, or on that (to my mind) damnably
vague thing 'the context'. And it is not a rule of language that gives the

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
proscenium arch and the blackboard the protective power that they have,
but rather the rules of the institutions called 'the theatre' and 'giving
philosophy lectures'. The effect of this objection is to make me wonder
what is the distinction between linguistic and non-linguistic rules. Linguis-
tic rules cannot be limited to rules which determine what one can or
cannot concatenate with what in what order. There are such rules, and
they are the formation-rules of a language, together with the rules (if they
are different) whose observance saves us from logical errors like self-
contradiction. But these are certainly not the only linguistic rules; there
has to be added at least the class of rules which determine what we can
and cannot say in what situations—I prefer this word to the word 'context',
because the latter means, properly, the surrounding words, and the
fashionable philosophical use of it is incorrect as well as deplorably vague.
The commonest kind of rule of this class is the kind that Austin called,
in his paper on 'Truth','descriptive conventions',12 such as that which
forbids me to call the paper on which this is printed 'purple'. There are
also Austin's 'demonstrative conventions', which regulate the use of
referring expressions like 'this'. All these involve for their observance an
attention, not merely to the words one is uttering, but to the situation in
which one is uttering them.
I do not see why we should not add to these two classes of linguistic
rules a further class, the subscriptive conventions, e.g. that which forbids
me to say in court under oath that I saw the prisoner kill the policeman,
and then claim that I had been joking; or to draw, sign, and hand over a
cheque, and then say I was only play-acting, when I am not on the stage.
These are certainly linguistic rules, in that they determine what I commit
myself to by what I say. That there are extra-linguistic penalties for
breaches of them does not distinguish them from other linguistic rules; if
I say the prisoner killed the policeman when he only struck him, I am
breaking a descriptive rule but may still be subject to the penalties of
perjury.

Having dealt all too sketchily with the tropic and neustic, I now come to
the third and last particle on my list, which I shall call the sign of
12
J. L. Austin, Philosophical Papers, Oxford, Oxford University Press, 1961, pp. 89 ff., repr. from
Proceedings of the Aristotelian Society, 1950.
32 R. M. Hare

completeness or clistic, from the Greek word for 'to close'. The commonest
clistic in ordinary language is the full stop; but this has other uses too,
which make a concise account of it hard to give. A well-designed clistic
would, it seems to me, take one of two forms; and the difference between
them arouses in me a suspicion that we have here not one possible sub-
atomic particle but two. The first is the sign of concatenation, which is
familiar among logicians and linguists. In his original notation, Frege's

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
formulae were well and truly concatenated; after the vertical stroke which
was his sign of subscription, there was, attached to it, the horizontal
stroke, which he originally referred to as the content-stroke, but later
abandoned the term. To this content-stroke the rest of the formula was
literally attached: it branched out like a procumbent tree to signify
the articulation of the proposition subscribed to, with conjunctions,
disjunctions, etc.; and on the right-hand side, at the ends of the branches,
he put the elementary propositions which were being combined in this
complex logical formula.
This notation was abandoned, no doubt because it was typographically
so expensive and inconvenient; but its passing seems to me to have led to
the neglect of an aspect of sentence-formation which ought not to be
forgotten. This is the necessity for somehow holding together the constituent
parts of a sentence, not merely as a lot of live piglets might be held
together by an insensitive farmer in a sack, but in an articulate or structured
way, each constituent having its place. It is interesting that linguists, in
their efforts to represent deep structure (which is hardly different from
what used to be called logical form) have had recourse to a very similar
and typographically just as expensive device, the so-called 'tree', which
also relies on lines joining up the constituent elements in an ordered
structure. Both these devices illustrate graphically, and more clearly than
most logical symbolisms, the important notion that sentences are organized
wholes and that it has to be clear, not only what belongs to the sentence
and what does not, but where in the sentence it belongs. Linguists of a
slightly earlier vintage used as a sign of concatenation, in dealing with
surface structure, plus-signs between the morphemes, and logicians have
done much the same with concatenation-signs shaped like saucers or
inverted saucers linking the symbols. Neither of these devices is so good,
because they do not bring out the features of order and articulation
(including branching) that I have just stressed.
Because it brings out this feature, what I shall call the Frege-style clistic
has great advantages; but there is one disadvantage. It does not make so
clear the necessity for indicating where the formula begins or ends. Perhaps
we could say that it must begin with the vertical stroke or sign of
subscription; but how do we know where we have to stop in any other
direction? Could not bits be added on at top or bottom, left or right? Are
we to say that they could not, because the rule is that only elements
Some Sub-Atomic Particles of Logic 33
joined to the vertical stroke by a continuous line belong to the sentence?
In other words, are we to say that what confines the sentence in this
notation is the white paper on which sentences are printed; a sentence is
an island with a very indented coastline, entirely surrounded by white
paper? Unfortunately this would not do; for we have done nothing to
prevent people adding bits to Frege's island or tree by putting in more
branches. We would do better to insist that the sentence was written on

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
one or on a series of horizontal lines, and that other lines were drawn
enclosing the sentence or other expression that was 'all that was being
said'.
This, in fact, is what Her Majesty's Customs do, and what banks do.
In customs declarations one has to make a list of the articles to be declared,
and then draw a line all round the list (or at least at the bottom, printed
lines sufficing for the other boundaries). This is to prevent anybody
coming along afterwards and adding to the list; and banks used to insist
on a horizontal line after the amount in words, to prevent it being added
to (now they insist instead on ending with the pence in figures or the
word 'only', which has the same effect). And the Morse Code gives us
sequences for 'Message begins' and 'Message ends'.
I will now illustrate the logical utility of such devices. Suppose that we
are listening- to a radio commentary on a cricket match, and the
commentator says 'He's caught it. He hasn't, he's missed it'. In writing
this down, we would put a full stop after the first sentence. Nobody thinks
that the commentator has made any self-<;ontradictory statements, though
of course he has corrected an earlier statement by uttering another
statement which is the contradictory of it. He has thus contradicted
himself only in a weak sense. But if he had said 'He has caught it and he
hasn't', he would have contradicted himself in the strong sense in which
only utterers of self-contradictory sentences can do this. The difference
between the single conjoint sentence 'He's caught it and he hasn't' and
the pair of sentences 'He's caught it' and 'He hasn't', is therefore important
for logic.13 The practice of putting the two premisses of a syllogism on
two lines without any sign of conjunction between them needs to be re-
examined for this reason. Aristotle is usually more careful.14
In one place where I gave an earlier version of this paper, they had a
seminar room with a blackboard that one could shut up with wooden
doors, so that the room could be used for parties; and I wrote on the
blackboard 'We are now in China if two and two make five', and closed
one of the doors to cover up the 'if-clause. I then got the audience to
agree that what I had written on the board was false, but afterwards I
13
It seems that Dummett thinks otherwise (see op. cit., p. 336), though I im not sure that this
is what he means.
14
Sec discussion in J. Lukasiewicz, AnuoiWs Syllogistic, Oxford, Oxford University Press, 1051,
p. 2.
34 R- M. Hare

opened the door and showed them that it was not. If I had written a full
stop after the words 'We are now in China' on the exposed part of the
board, I should not have been able to perform this trick.
Full stops are not always given this kind of function; a writer is, for
example, held to have contradicted himself in the strongest sense if, in
the course of a paragraph or even of a book, he makes statements which
contradict one another; the full stops in between do not help him, because

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
in books it is a convention that a writer is taken as subscribing to the
conjunction of the statements printed therein. Commonplace-books and
'albums' like Philosophical Investigations are free from this convention, as
are philosophical dialogues. I sometimes think, however, that Wittgen-
stein's thought would be a lot clearer if he had, as well as using full stops,
put in names for the different characters in the dialogue, or, alternatively,
put in, in front of the propositions to which he himself wished to subscribe,
the sign of subscription about which he is so contemptuous.

I shall not have room to say any more about the individual particles I
have mentioned. My guess is that, like their counterparts in physics, they
are a great deal more numerous and various than one might at first suspect;
I am therefore not claiming to have given a complete list. Nor shall I
have room to answer any of the many questions that arise about the
particles which I have listed. I will, however, raise some of them.
We need to ask what is left of the sentence if we subtract the tropic,
neustic, and clistic. In earlier writings I called this the phrastic.15 But if
the clistic is like Frege's and serves to articulate the sentence into clauses,
then we cannot take it away without destroying the articulation of the
sentence and leaving behind a mere collection of unrelated bits. It might
be better to have some other way of articulating the sentence (perhaps by
putting 'hooks and eyes' on all the words, determining the 'part of speech'
to which each belonged). The effect of this would be to make words fit
into each other only in certain arrangements; each word would in fact
carry with it a 'sentence-frame' into which other words would fit in
certain, but only in certain, places. For example a subject-term would fit
predicates but not (without some intermediary link such as 'is identical
with') other subject terms. This suggestion is not original. It would free
the term 'clistic' for a sign of enclosure, limiting the boundaries of a
sentence.
A phrastic, then, would consist of an articulate combination of words,
such that by adding to it a tropic we could give it a mood, and by adding
a neustic we could subscribe to what was said in it. It would be required
that the phrastic was complete in one sense of that expression; it would
have to be such that it 'made sense' after a tropic and neustic were added
13
Sec my The Language of Morals, Oxford, Oxford University Press, 1952, p. 18.
Some Sub-Atomic Particles of Logic 35
to it. But what of the clistic? At what stage should we add this? If we add
it at the very end after the neustic has already been added, then shall we
be engaging in the somewhat fraudulent manoeuvre of subscribing to what
is said in a sentence, but leaving it open to ourselves to add bits to it
afterwards if it suits us? Surely HM Customs would not allow this?
On the other hand, if we insist on leaving the neustic to the end, after
the clistic has been added, we shall be certifying the sentence as 'all that

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
is being said' when all has not been said; we still have to add the sign of
subscription. But perhaps the problem is unreal. After all, I can draw up
a customs declaration or cheque and put in marks to signify that that is
all there is I have to declare or pay, and only after that sign it.
There is the problem, which has been discussed in the literature, of
whether only whole sentences have tropics or moods, or whether subordi-
nate clauses too must, or can, have them. I think that the discussions
would have been clearer if neustics had been distinguished from tropics.16
This removes the main temptation to saying that subordinate clauses
cannot have tropics, namely that nobody is subscribing to them. That is
just a confusion. It seems fairly obvious that at least some subordinate
clauses have tropics, though not neustics; for example commands in oratio
obliqua.
I mentioned earlier the problem of how to distinguish between the
meanings of the different tropics or mood-signs. There is also that of
whether neustics can be of different kinds, or at least strengths. The use
of the words 'insist' and 'suggest', mentioned above, seems to indicate
that this is so. The idea has been carried further with the suggestion that
we can have negations or denegations of neustics.17 This operation must
of course be distinguished from ordinary internal negation (see above). It
is not even quite the same as what has been called 'external negation', as
in 'It is not the case that you ought'; for explicitly withholding subscription
is not the same as subscribing to the statement that something is not the
case.
Lastly, to what do truth-values attach—to whole subscribed-to utter-
ances or speech acts, complete with neustics, or to these minus their
neustics, or to tropics-cum-phrastics, or just to phrastics? I feel inclined
to say, but without confidence, that different sets of values apply to
different combinations. If a complete sentence with neustic, clistic, tropic,
and phrastic is uttered, the speaker is open to an accusation of speaking
falsely if the tropic is indicative and the phrastic specified something that
is not actually the case. If we remove the neustic, then nobody is open to
such an accusation, but the remaining clistic-cum-tropic-cum-phrastic
expresses something which can be true or false (if, again, the tropic is
indicative). The further removal of the clistic and the tropic leaves the
16 17
See ref. in n. 3. See n. 6 above.
36 R. M. Hare
phrastic, which expresses something which cannot indeed be true or false,
but can be or not be the case. If it is the case and the tropic is imperative,
then the command is satisfied; if it is indicative, the proposition is true.
A host of problems arises here which it would be interesting to explore;
but space has run out, and I shall have to return to them.18

Department of Philosophy R. M. HARE

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015
University of Florida
Gainesville
Florida 32611
USA

" Two papers on related topics have appeared in Mind since this paper was written and committed
to the Urmson Festschrift. Neither makes necessary any alteration to my tent, but some comments
on both may be in order.
Michael Pendlebury's 'Against the Power of Force' (Mind, 1986, p. 361), contains some good
insights and arguments, with which I broadly agree. They would, however, be clearer if he had at
the start distinguished as I have between signs of mood and signs of subscription—a distinction of
which he shows himself aware later in his paper. I cannot agree with him that it is helpful to try to
explain the semantic significance of different moods in terms of different satisfaction-conditions. The
latter are conditions for the satisfaction of phrastics, and leave the meaning of the mood-sign
untouched. Much darkness has been shed by looking for surrogates for truth-conditions in the case
of prescriptions. T o understand the meaning of the imperative mood-sign is, rather, to understand
what difference the use of an imperative makes to the communicative situation, and in particular what
requirements are thereby incurred by the speaker and others. For 1 very good account of these see
W. Alston, 'Sentence Meaning and Illocutionary Act Potential', Ph. Exch., 1977, §§ iv f.
Jennifer Hornsby, in her 'A Note on Non-Indicatives' in the same volume, p. 92, demolishes
Davidson's view so easily that it is surprising to find her remaining so deferential to him. Surprising,
also, that after what has been written about it, she continues to believe in 'the clean separation of
locutionaxy and illocutionary matters' (n. 12, p. 97; see p. 24 above and my 'Austin's Distinction
between Locutionary and Illocutionary Acts', in my Practical Inferences, Basingstoke and London,
Macmillan, 1971, p. 100). A possible explanation is that she seems to confuse Austin's phatic with
his phonetic act (How to Do Things mth Words, Oxford, Oxford University Press, 1962, p. 92). dearly
the phonetic act is distinct from the rhetic, but the boundary between phatic and rhetic is very shaky.
I think she is wrong to suppose that there are different senses of 'say' when used of saying
something imperative and saying something indicative. Imperatives also have an oratio obliqua form
(the 'indirect command' of the old grammarians), which in English is expressed with 'tell to', the
analogue of 'tell that' (see p. 30 above); or, if 'say' is used, in the form 'She said that he was to go'.
The indicativity of the 'that'-clause here is superficial, but although what is reported as having been
said ('Go') was imperative, 'say' still means the same.
Her main suggestion revives an old dispute in a new form: the dispute as to whether what is
embedded in a fully articulated imperative should be a complete indicative sentence, or rather a
phrastic (sentence radical). I mentioned the point briefly in The Language of Morals, p. 21. I prefer
the phrastic solution for two reasons. The first is that there is no need for an unsubscribed-to
indicative tropic attached to the embedded phrastic of an imperative sentence. We do indeed, need
to know the satisfaction- or being-the-case-conditions of the phrastic in order to understand the entire
sentence (see pp. 35 f. above); ind one simple way of conveying this understanding is to say what
would be the truth-conditions of the corresponding indicative. (Another way would be simply to cuff
those who did not obey one's commands correctly^ and in this way a purely imperative language
could be taught, similar to that envisaged in Philosophical Investigations, § 2, which Wittgenstein says
in § 6 could be 'the whole language' of its users.) It does not follow from this that the complete
indicative has somehow to appear inside the imperative. The thought that it does may be due to the
prejudice that /m/A-conditions are basic to all kinds of meaning, whereas they are basic, if at all, only
to the meanings of indicatives. This prejudice has done harm in moral philosophy (see my 'A Reductio
Some Sub-Atomic Particles of Logic 37
ad Absurdum of Descriptivism', in Philosophy in Britain Today, ed. S. Shanker, Beckenham, Croom
Helm, 1986, pp. 124 f, repr. in my Essays m Ethical Theory, Oiford, Oiford University Press, 1989).
The second reason is that, in Homsby's formulation, if the so-called indicative that is embedded
is made unambiguously indicative (e.g. by putting 'You are going to' instead of'You will') the whole
sentence then expresses a clearly false statement, because one cannot say an indicative imperatively.
I do not find convincing her efforts to get out of this difficulty; the way out is to purge the embedded
phrastic of its indicative tropic.

Downloaded from http://mind.oxfordjournals.org/ at University of Bath Library & Learning Centre on June 9, 2015