But in this way the last weak link between the 1wo parts is also snapped, and the signs 0, 1, = receive yet a third meaning 111 addition to their Boolean and
arithmetical
ones.
Gottlob-Frege-Posthumous-Writings
This opposition or conflict is such that we automatically reject one limb as false when we accept the other as true, and conversely.
The rejection of the one and the acceptance of the other are one and the same.
? Boole's logical Calculus and the Concept-script1 [1880/81]
In his writings, Leibniz threw out such a profusion of seeds of ideas that in IIus respect he is virtually in a class of his own. A number of these seeds wrrc developed and brought to fruition within his own lifetime and with his 11tllaboration, yet more were forgotten, then later rediscovered and drvcloped further. This justifies the expectation that a great deal in his work lhul is now to all appearance dead and buried will one day enjoy a rr,urrcction. As part of this, I count an idea which Leibniz clung to lluuughout his life with the utmost tenacity, the idea of a lingua l'haracterica, an idea which in his mind had the closest possible links with lhnl of a calculus ratiocinatur. That it made it possible to perform a type of ru111putation, it was precisely this fact that Leibniz saw as a principal ndvanlage of a script which compounded a concept out of its constituents rnl her than a word out of its sounds, and of all hopes he cherished in this muller, we can even today share this one with complete confidence. I will IJlrulc just the following from the relevant passages:
'Si daretur vel lingua quaedam exacta, vel genus scripturae vere philosophiae, . . . omnia quae ex datis ratione assequi, inveniri possent l. fUodam genere calculi, perinde ac resolvuntur problemata arithmetica aut jcomctrica. '*
? I>c Scientia universali seu calculo philosophico.
1 In 188 I, this article was submitted by Frege in turn to the Zeitschrift fiir Matlrematik und Physik, the Mathematischen Annalen and the Zeitschrift fiir l'lr/lu. wphie und philosophisehe Kritik, but was in every case rejected by the editors. lllluully remained unpublished.
l'rom the report of H. Scholz and F. Bachmann: Der wissenschaftliche Nachlass r? mt {,'o/1/ob Frege (Paris 1936) we learn that the lost original was 'a manuscript rrrJlnrcd for publication of 103 closely written sides of quarto'. Scholz and lt11chmann mention besides that Frege also submitted the manuscript to R. Avenarius for the Vierteljahrsschrift fii. r wissenschaftliche Philosophie. However it 11uuld be that what Frege submitted to Avenarius was the essay published in this v11lumc on pp. 47 IT. 'Boole's logical Formula-language and my Concept-Script', llnce Avenarius in his letter to Frege of 20/4/1882 cites the title of the manuscript rtlurncd by him as 'Boole's logical Formula-Language'.
The article can scarcely have been composed before 1880, the year in which the rtview by Schroder mentioned on p. 11 appeared. A great number of the reflections
? ? ? 10 Boole's logical Calculus and the Concept-script
Among the various sorties Leibniz made upon his goal, the beginnings of a symbolic logic come closest to what seems to be indicated by the phrase 'calculus ratiocinatur'. They are to be found in the essays:
Non inelegans specimen demonstrandi in abstractis and Addenda ad specimen calculi universalis
In these Leibniz stuck very close to language. Just as the words we use for the attributes of a thing follow one another, so he simply juxtaposes the letters corresponding to properties in order to express the formation of a concept. If, for instance, A means right-angled, B isosceles, C triangle, then Leibniz represents right-angled isosceles triangle by ABC. He uses a sign for identity, oo, and the sign + with the definition:
'A +Boo L significat A inesse ipsi L'.
This seems to coincide with the meaning recent logicians have given this sign, according to which A + B represents the class of individuals which belong to A or to B or to both. Since I am passing over less important details, the only other fact I will mention is that Leibniz allows the words 'non' and 'ens' to occur in his formulae. In this project he surely has the lingua characterica in mind, even though he made no express connection with the attempts he made to represent a content.
This way of setting up a formal logic seems to suggest itself naturally. At any rate recent German and English logicians have arrived at the same conception quite independently of Leibniz, though, as far as I know, in doing so they do not have a general characteristic in mind. However much Boolean logic* may stand out as a systematic working out of the fragmen- tary hints in Leibniz, it only goes beyond him in one point of fundamental importance-in the way it reduces hypothetical and disjunctive judgements to categorical. ? ? The Leibnizian 'ens' is left out.
In a short monograph,? ? ? I have now attempted a fresh approach to the Leibnizian idea of a lingua characterica. In so doing, I had to treat in part the same subject-matter as Boole, even if in a different way. This has
? Boole's main work is An Investigation ofthe Laws ofThought on which are founded the mathematical theories of logic and probabilities. London 1854.
** On one point indeed Boole has taken a retrograde step away from Leibniz, in adding to the Leibnizian meaning of A + B the condition that the classes A and B should have no common element. W. Stanley Jevons, E. Schroder and others have quite rightly not followed him in this.
*** Begriffschrift, eine der arithmetischen nachgebildete Formelsprache des reinen Denkens. Halle aS. 1879. (Note continued, pp. 11-12. )
put forward here by Frege are along the same lines as the contents of the essay published in 1882 Uber der Zweck der Begriffsschrift.
The last two sides of the lost original are preserved on a photostat, which was presented at the Paris Congress in 1935 as a specimen of the Frege NachlqfJ.
? ? BooZe's logical Calculus and the Concept-script 11
prompted many critics to draw comparisons between the two, of which the most detailed are those set out by E. Schroder in the Zeitschrift fiir Math. 11. Phys. Vol. XXV. He will not regard it as detracting from the gratitude which I here express for his thorough study of my monograph and its friendly review, if in what follows I attempt to supplement and correct that comparison.
A few of the signs I introduced there should be explained briefly here:
I deviate from usual practice in drawing a distinction between judgement und content of possible judgement. For me the relation in a hypothetical is uot one between judgements but between contents of possible judgement. 11ut ifl affirm that this relation holds, I then express a judgement.
( I) The content-stroke is horizontal, it is always prefixed to the expression of a content of possible judgement, serving to connect it to the judgement uml negation-strokes, and also to combine it with other contents of possible 111dgement by means of the conditional-stroke(? 2): e. g.
- ( 2 + 3 = 5).
(2) The judgement-stroke is placed vertically at the left hand end of the
content-stroke, it converts the content of possible judgement into a Judgement (? 2): e. g.
1--(2 +3 = 5).
(3) the negation-stroke is placed vertically under the content-stroke dividing it into two parts. The part to the right of the negation-stroke is the content-stroke of the original content, the part to the left that of its negation (~ 7): e. g.
. . . . . -(2 + 3 = 6).
(4) The conditional stroke connects two content-strokes by running from
the under side of the upper one to the left hand end of the lower. Like the negation-stroke it splits the upper stroke in two, of which the part to the right remains the content-stroke of the upper content, while the part to the left becomes the content-stroke of the whole, which means the negation of Ihe case that the upper content is false, and the lower true(? 5): e. g.
tx=7
X+ 3 = 10: means:thecasethatx+3= 10withoutx= 7doesnotoccur. Wecansay
in this case: ifx + 3 = 10, then x = 7. The roman letters such as the x here, confer generality on the content of the judgement by presenting it as true, whatever one may substitute for the roman letter (? 11, p. 21). If, for instance, we replace x by 2 we obtain
t,-2=7
L 2 + 3 = 10.
This is a true judgement, although the upper content, as well as the lower, is 1'11lsc. Here, the rendering 'if would jar with normal usage. Yet here too, the CIUIC is denied that 2 + 3 = 10 is true and 2 = 7 false.
If a negation-stroke stands to the left of the conditional stroke, as in
~2 = 7
2 + 3 = 5,
? ? 12 Boole's logical Calculus and the Concept-script
Above all, if we are not to go astray it is necessary that we should always bear in mind the purpose that governed Boole in his symbolic logic and the one that governed me in my Begriffsschrift.
If I understand him aright, Boole wanted to construct a technique for resolving logical problems systematically, similar to the technique of elimination and working out the unknown that algebra teaches. To this end, he represents judgements in the form of equations that he constructs out of letters and arithmetical signs such as +, 0 and 1. Logical laws then assume the form of algorithms, although these only coincide in part with those that hold in arithmetic for the same signs. In all this there is no concern about content whatsoever. In the main these means fulfil their purpose, at least as far as the range of problems that Boole has in mind are concerned. But one may think of logical problems lying outside this range. The use of arithmetical signs for logical purposes has the advantage that one is spared the necessity of learning a completely new algorithm. A large number of the transformations already familiar to our eyes still remain valid. ? Higher demands, which however Boole does not set himself, are naturally not met by this enterprise. Anyone demanding the closest possible agreement between the relations of the signs and the relations of the things themselves will always feel it to be back to front when logic, whose concern is correct thinking and which is also the foundation of arithmetic, borrows its signs
from arithmetic. To such a person it will seem more appropriate to develop for logic its own signs, derived from the nature of logic itself; we can then go on to use them throughout the other sciences wherever it is a question of preserving the formal validity of a chain of inference.
In contrast we may now set out the aim of my concept-script. Right from the start I had in mind the expression ofa content. What I am striving after is a lingua characterica in the first instance for mathematics, not a calculus restricted to pure logic. But the content is to be rendered more exactly than is done by verbal language. For that leaves a great deal to guesswork, even if only of the most elementary kind. There is only an imperfect correspondence between the way words are concatenated and the structure
the negation of the case that 2 + 3 = 5 without 2 = 7 is thereby converted into its affirmation we may render: 2 + 3 = 5 and it is not the case that 2 = 7 (? 7, p. 13). In
fTrr-2= l+l L2+3=5
-. - 2 = 1 + 1takes the place ofthe 2 = 7 ofthe preceding example. In rendering this, two denials come together and yield an affirmation:
2+3=5,and2=1+1(? 7. p. 12).
? I should mention that the deviations from arithmetic are for all that so fundamental ti}at solving logical equations is not at all like solving algebraic ones. And this greatly diminishes the value of the agreement in the algorithm.
? Boole's logical Calculus and the Concept-script 13
of the concepts. The words 'lifeboat' and 'deathbed' are similarly constructed though the logical relations of the constituents are different. So the latter isn't expressed at all, but is left to guesswork. Speech often only indicates by inessential marks or by imagery what a concept-script should spell out in full. At a more external level, the latter is distinguished from verbal language in being laid out for the eye rather than for the ear. Verbal script is of course also laid out for the eye, but since it simply reproduces verbal speech, it scarcely comes closer to a concept-script than speech: in fact it is at an even greater remove from it, since it consists in signs for signs, not of signs for the things themselves. A lingua characterica ought, as Leibniz says, peindre non pas les paroles, mais les pensees. The formula- languages of mathematics come much closer to this goal, indeed in part they nrrive at it. But that of geometry is still completely undeveloped and that of urithmetic itself is inadequate for its own domain; for at precisely the most important points, when new concepts are to be introduced, new foundations laid, it has to abandon the field to verbal language, since it only forms numbers out of numbers and can only express those judgements which treat 111' the equality of numbers which have been generated in different ways. But arithmetic in the broadest sense also forms concepts-and concepts of such richness and fineness in their internal structure that in perhaps no other science are they to be found combined with the same logical perfection. And there are other judgements which arithmetic makes, besides mere equations und inequalities. The reason for this inability to form concepts in a scientific manner lies? in the lack of one of the two components of which every highly developed language must consist. That is, we may distinguish the formal part which in verbal language comprises endings, prefixes, suffixes and auxiliary words, from the material part proper. The signs of arithmetic correspond to the latter. What we still lack is the logical cement that will bind these building stones firmly together. Up till now verbal language took over this role, and hence it was impossible to avoid using it in the proof itself, and not merely in parts that can be omitted without affecting the cogency of the patterns of inference, whose only purpose is to make it easier lo grasp connections. In contrast, Boole's symbolic logic only represents the formal part of language, and even that incompletely. The result is that
Boole's formula-language and the formula-language of arithmetic each solve only one part of the problem of a concept-script. What we have to do now, in order to produce a more adequate solution, is to supplement the signs of 111athematics with a formal element, since it would be inappropriate to leave the signs we already have unused. But on this score alone Boole's logic is nlready completely unsuited to the task of making this supplementation, since it employs the signs +, 0 and 1 in a sense which diverges from their nrithrnetical ones. It would lead to great inconvenience if the same signs were to occur in one formula with different meanings. This is not an
? In the case of the formula-language of arithmetic.
? 14 Boole's logical Calculus and the Concept-script
objection to Boole, since such an application of his formulae obviously lay completely outside his intentions. Thus, the problem arises of devising signs for logical relations that are suitable for incorporation into the formula- language of mathematics, and in this way of forming-at least for a certain domain-a complete concept-script. This is where my booklet comes in.
Despite all differences in our further aims, it is evident from what has been said already that the first problem for Boole and me was the same: the perspicuous representation of logical relations by means of written signs. This implies the possibility of comparing the two. If I now turn to this, it cannot be done in the sense of adjudicating between the two formula- languages, which is to be preferred. To raise such a question would mean referring back to their ultimate aims, which are more ambitious in my case than in Boole's. It would indeed be more than possible that each set of signs was the more appropriate for its own ends. Nevertheless, it seems to me worthwhile to work out the comparison in detail, since in that way many of the peculiarities of my concept-script are thrown into sharper focus.
To begin with, as far as the extent of these sign languages is concerned, one is entitled to expect that these would have to coincide, once everything had been sifted from my concept-script which lies beyond the confines of pure logic. Yet even then, not everything that I express can be translated into Boolean notation, whereas the converse transformation is possible. E. Schroder indeed thinks that my concept-script would only correspond to the second part of Boole's. This is refuted by the fact that I too can easily express such a judgement as 'Every square root of 4 is a fourth root of 16' by
~a4 =16 La2 = 4
although Boole would count it as one of the primary propositions. I represent all judgements in the first part of Boole in a similar way though in so doing I admittedly construe them quite differently. The real difference is that I avoid such a division into two parts, the first dedicated to the relation of concepts (primary propositions), the second to the relation of judgements (secondary propositions) and give a homogeneous presentation of the lot. In Boole the two parts run alongside one another, so that one is like the mirror image of the other, but for that very reason stands in no organic relation to it.
On the other hand, Boole had only an inadequate expression for particular judgements such as 'some 4th roots of 16 are square roots of 4'
~a2 =4
La4 = 16,
and for existential judgements such as 'there is at least one 4th root of 16'
IT&ra4 = 16,
So it transpires that even when we restrict ourselves to pure logic my
apparently no expression at all.
? Boole's logical Calculus and the Concept-script 15
concept-script commands a somewhat wider domain than Boole's formula- lunguage. This is a result of my having departed further from Aristotelian logic. For in Aristotle, as in Boole, the logically primitive activity is the formation of concepts by abstraction, and judgement and inference enter in through an immediate or indirect comparison of concepts via their extensions. ? The only difference is that Aristotle places in the foreground 1he case where the extension of one concept completely includes that of unother-i. e. that of subordination-whereas Boole reduces other cases to lhe case of equality of extensions. ? ? Here hypothetical and disjunctive 111dgements are for the moment left on one side. Boole now construes the hypothetical judgement 'If B, then A' as a case of subordination of con- cepts, by saying 'the class of time instants at which B is included in the class ,,rtime instants at which A'.
The full incongruity of the introduction here of the idea of time instants slunds out most clearly if you think of eternal truths such as those of mathematics. Schroder seems to want to avoid the artificiality this involves, siuce, in company with Hugh McColl, he explains expressions like A = 0, A t 8 = 1 etc. -whose sense, on the Boolean conception, is self-explanatory when taken in conjunction with the stipulations of the first part-all over U! (ain without referring back.
But in this way the last weak link between the 1wo parts is also snapped, and the signs 0, 1, = receive yet a third meaning 111 addition to their Boolean and arithmetical ones. According to Boole 0 111cans the extension of a concept under which nothing falls, as for example lhc extension of the concept 'whole number whose square is 2'. By 1, Boole uuderstands the extension of his universe ofdiscourse. These meanings hold for the first as much as the second part. I f one now ruptures this connection, then strictly speaking 0 has no longer an independent meaning in the Nccond part; combined with the identity sign it means a denial expressed as a
Judgement, while '= 1' designates an affirmation, which I express by the judgement-stroke. Then, besides this, the identity sign still has an iudependent meaning in formulae like A = B. But that even Boole on his npproach does not establish an organic connection between the two parts cun be seen from the fact that he does not use the equations of the first part us constituents of equations of the second part, and, if you hold strictly to Iheir meanings, cannot so use them. For in the first part A = B is a judge- ment. whereas if it were made a constituent of an equation of the second purl, as say in
( A = B )C = D
? This is not meant to imply that concept formation takes up a great deal uf space in their presentations. Rather, that their logics are essentially doctrines of inference, in which the formation of concepts is presupposed AN something that has already been completed.
** The relation of identity has greater content and so is less extensive nru. l general than that of subordination.
? ? ? 16 Boole's logical Calculus and the Concept-script
A = B would mean the class of time instants at which the content of the judgement 'A = B' was to be affirmed. On the other hand, if, to avoid this, you do not resort to the expedient of the idea of a class of time instants, then you get in return the double sense of the 0, the 1 etc. in the same formula, which would then, to say the least, become unperspicuous.
As opposed to this, I start out from judgements and their contents, and not from concepts. The precisely defined hypothetical relation between contents of possible judgement has a similar significance for the foundation of my concept-script to that which identity of extensions has for Boolean logic. I only allow the formation of concepts to proceed from judgements. If, that is, you imagine the 2 in the content of possible judgement
24 = 16
to be replaceable by something else, by (-2) or by 3 say, which may be
indicated by putting an x in place of the 2: x4 = 16,
the content of possible judgement is thus split into a constant and a variable part. The former, regarded in its own right but holding a place open for the latter, gives the concept '4th root of 16'. ?
We may now express
by the sentences '2 is a fourth root of 16' or 'the individual 2 falls under the concept "4th root of 16"' or 'belongs to the class of 4th roots of 16'. But we
? The reader may be surprised I don't put this in the form
2 = {116.
The reason is that {116 may not in the same context mean now this, now
that, 4th root of 16. Otherwise you could form the inference 2 = y/i6
- 2 = ~16 2=-2.
By using ~ on one side of an identity sign, you have laid it down that \li6 is to mean a particular 4th root of 16, just as the letter a too must be given the same meaning throughout a given context. I f it is in fact stipulated that ~is to mean the positive real root then -2 = ~is false and 2 =
~ isnottoberead:
'2 is a 4th root of 16', but
'2 is the positive real root of 16'.
We can see from this that the root sign cannot properly be used to help express that an individual falls under the concept of a root.
? Boole's logical Calculus and the Concept-script 17
may also just as well say '4 is a logarithm of 16 to the base 2'. Here the 4 is being treated as replaceable and so we get the concept 'logarithm of 16 to the base 2':
2X = 16.
lhc x indicates here the place to be occupied by the sign for the individual lulling under the concept. We may now also regard the 16 in x4 = 16 as replaceable in its turn, which we may represent, say, by x4 = y. In this way we arrive at the concept of a relation, namely the relation of a number to its ? lth power. And so instead of putting a judgement together out of an utdividual as subject* and an already previously formed concept as predicate, we do the opposite and arrive at a concept by splitting up the content of possible judgement. ** Of course, if the expression of the content of possible judgement is to be analysable in this way, it must already be 1tsclf articulated. We may infer from this that at least the properties and 1l'lations which are not further analysable must have their own simple dl'signations. But it doesn't follow from this that the ideas of these properties 1111d relations are formed apart from objects: on the contrary they arise . \lmultaneously with the first judgement in which they are ascribed to things. llcnce in the concept-script their designations never occur on their own, but always in combinations which express contents of possible judgement. I could compare this with the behaviour of the atom: we suppose an atom nl'vcr to be found on its own, but only combined with others, moving out of uuc combination only in order to enter immediately into another. *** A sign ll>r a property never appears without a thing to which it might belong being ut least indicated, a designation of a relation never without indication of the things which might stand in it.
In contrast with Boole, I now reduce his primary propositions to the ,\'t'condary ones. I construe the subordination of the concept 'square root of
* The cases where the subject is not an individual are completely different from these and are here left out of consideration.
? ? A great deal of tedious discussion about negative concepts such as 'not-triangle' will, as I hope, be rendered redundant by the conception of the relation of judgement and concept outlined here. In such a case one simply doesn't have anything complete, but only the predicate of a judgement which as yet lacks a subject. The difficulties arise when people treat such a fru~ment as something whole.
In this connection, I find it extraordinary that some linguists have recently viewed a 'Satzwort' (sentence-word), a word expressing a whole judgement, as the primitive form of speech and ascribe no independent existence to the roots, as mere abstractions. I note this from the KMtlngschen ge/ehrten Anzeigen 6 April 1881: A. H. Sayee, Introduction to the? Science ofLanguage 1880 by A. Fick.
? ? ? As I have since seen, Wundt makes a similar use of this image in his l. oglk.
? 18 BooZe's logical Calculus and the Concept-script
4' to the concept '4th root of 16' as meaning: if something is a square root of
4 it is a 4th root of 16:
trx4 = 16 Lx2 = 4.
I believe that in this way I have set up a simple and appropriate organic relation between Boole's two parts. Moreover, on this view we do justice to the distinction between concept and individual, which is completely obliterated in Boole. Taken strictly, his letters never mean individuals but always extensions of concepts. That is, we must distinguish between concept and thing, even when only one thing falls under a concept. The concept 'planet whose distance from the sun lies between that of Venus and that of Mars' is still something different from the individual object the Earth, even though it alone falls under the concept. Otherwise you couldn't form concepts with different contents whose extensions were all limited to this one thing, the Earth. In the case of a concept it is always possible to ask whether something, and if so what, falls under it, questions which are senseless in the case of an individual. We must likewise distinguish the case of one concept being subordinate to another from that of a thing falling under a concept although the same form of words is used for both. The examples given above
tx4 = 16 24
x = 4 and f--2 = 16
show the distinction in the concept-script. The generality in the judgement
'TX4 = 16 Lx2 = 4
'All square roots of 4 are 4th roots of 16' is expressed by means of the letter x, in that the judgement is put forward as holding no matter what one understands by x. I stipulated that roman letters used in the expression of judgements should always have this sense.
Let us now look at the case where the content of such a general affirmative judgement occurs as part of a compound judgement, say as the antecedent of an hypothetical judgement; e. g. :
If every square root of 4 is a 4th root of m, then m must be 16. The expression
does not correspond to the sentence, and is even false, which is why the judgement stroke has been left off the left-hand end of the uppermost horizontal stroke; for we may substitute numbers for x and m which falsify this content. Thus if we take m to be 17, then the consequent m= 16 would
? ? ? Boole's logical Calculus and the Concept-script 19 hecome 17 = 16 and so would be false. Of course that doesn't yet
necessarily make the whole
false; for if the antecedent
17=16 1(x4 = 17
x2 =4
x4 = 17 l x2 = 4
were also false, the whole would be true despite the falsity of the conse- lJUent. *
But we may take a value for x, 3 say, which satisfies the condition x' = 17
for
lx2 = 4 134 = 17
2 3=4
is true, since not only 34 = 17 but also 32 = 4 is false. Thus ifwe give mthe value 17 and x the value 3, the antecedent
is satisfied but the consequent
x'=m
1x2 = 4 m= 16
m= 16
x=m ~4
2 x=4
is false. Thus
is not true for all values of x and m, which is what would be asserted by prefacing the formula with the judgement-stroke. But the sentence
'If every square root of 4 is a 4th root of m, then m must be 16'
Nays something different. You could also express its content as follows:
'If, whatever you understand by x it holds that x4 = m must be true ifx2 = 4, then m= 16'.
We can see: the generality to be expressed by means of the x must not KOvern the whole
*Cf. Footnote above p. lOfT. and Begriffschrift ? 5, where the latter atdmittedly contains the mistake pointed out by Schroder in his review. This however had no effect on what followed.
? ? 20 Boole's logical Calculus and the Concept-script but must be restricted to
x4 =m Lx2 = 4
I designate this by supplying the content-stroke with a concavity in which I put a gothic letter which also replaces the x:
-6-r-a' = m La2 = 4.
I thus restrict the scope of the generality designated by the gothic letter to the content, into whose content stroke the concavity has been introduced (? 11 of the BegrifJsschrift). * So our judgement is given the following expression:
m= 16 a4 =m a2 = 4.
By means of this notation, I am now also able to express particular and existential judgements. I render the sentence 'Some 4th roots of 16 are square roots of 4' thus:
For
~a2 =4
La4 = 16.
n a2=4 -vra4 = 16
means the content of possible judgement
'If anything is a 4th root of 16, it is not a square root of 4', or 'No 4th roots of 16 are square roots of 4'.
We now designate this content as a false generalization by prefacing it with a negation-stroke, and present the result as an assertion by means of the judgement-stroke. Analogously
~a2 = 4means: 'There is at least one square root of 4'.
This is the negation of the generalization of the negation of the equation a2 = 4.
We can now also easily show the link between particular and existential judgements. For in
tr0,-ra2 = 4 La4 = 16
* In discussing my monograph, E. Schroder made the proposal that the designation of generality be introduced into Boolean logic by the use of gothic letters. However that is inadequate, since the scope over which the generality is supposed to extend is still left open. The drawback that a second negation sign is needed is connected with this.
? BooZe's logical Calculus and the Concept-script 21 we may insert two negation-strokes in immediate succession, which thus
nmcel each other out
IT&mra2 = 4 La4 = 16
nnd think of this as concatenated as indicated here: I~"T1Ta2 = 4
tust as you may analyse
La4 = 16
tr&ra2=4 into l~a2 =4.
Thus the only distinction between
IN lhat
IT(j;,(T a2 = 4
La4 = 16 and
'[a2 = 4 a4 = 16
t11kcs the place of a2 = 4. But according to ? 7 of the Begriffsschrift,* this mcnns that a is a square root of 4 and a 4th root of 16. So just as you can lrnnslate
~a2 =4 as
'There is at least one square root of 4', you may express
IT6-rrrra2 =4 or t-r&Tra2 =4 La4 = 16 La4 = 16
by Ihe sentence
'There is at least one number which is both a square root of 4 and a 4th ruul of 16'. But this is equivalent to the expression:
'At least one 4th root of 16 is a square root of 4'.
I have now shown that the difference in extent of the domains governed by 8oolean logic and by my concept-script, extraneous as it might at first llaht appear, is in fact as closely as possible bound up with their original uunatruction. A few examples may now serve to illustrate how the uonatruction of the concept-script enables it when combined with the signs orarithmetic to achieve the more far-reaching goals it set itself.
(I) There are at least two different square roots of 4. The sentence 'a- b follows from a2 = 4 and b2 = 4' is denied in its generality.
? Sec also above, footnote on p. IOff.
,
ba=b a2 = 4 b2 = 4
? ? ? 22 BooZe's logical Calculus and the Concept-script (2) there is at most one number whose double is 4
1f;a:! [2b =4
(3) 4 is a positive whole number (including 0). That is, 4 belongs to the series beginning with 0, in which the immediate successor of any member is obtained by adding 1. *
(4) 12 is a multiple of 4; that is, 12 follows 0 in the arithmetical progression with difference 4. ** Two numbers with opposed signs are here not counted as multiples of one another, and 0 only counted as a multiple of itself.
(5) 12 is a common multiple of 4 and 6; that is, 12 is a multiple of 4, and*** 12 is a multiple of 6.
1~(0, +4 ~12,)
~(07 +6 = 1211)
(6) 4 is a common (aliquot) factor of 12 and 20.
y
1~ (0, +4 ~ 12,) p(07 +4 = 2011)
(7) The multiple of a multiple of a number is a multiple of that number.
T~(07+a= c11)
[5(0,+a~ b1)
~(Oy+b) = Cp)
(8) A and B are congruent modulo M. Here A, B and M need not be
whole numbers; nothing further than their addibility is presupposed. This
*Cf. Begriffss. :hrift ? 29, where in the gloss on formula (99) for;f(x10z6 ) one should re~d~/(x1,z6)?
. . Begrilfsschrift ? 26.
. . . See footnote above, p. lOfT.
? BooZe's logical Calculus and the Concept-script 23 way of speaking would be inconvenient for ordinary use, but it is only meant
to have application when it is necessary to go back to the concept. ~(A,+ M~B1)
t{ J ( B 7 + M = A p )
(9) 13 is prime. Here 1 is counted as a prime number, 0 not. More
explicitly: 13 is a positive non-zero whole number(~(07 +1 = 13p)),and"'
whatever positive number b may be which is greater than 1(p(17 + 1 = bp)) und different from 13 (--. - b = 13), 13 cannot be a multiple of b ( --,~(07 + b = 1 3 p ) ) -
y
{i(07+b= 13p)
y
( IO) A and B are positive non-zero whole numbers that are eo-prime. IIere 1is treated as prime to every number.
p(07 +b = Ap) p(07 +b=Bp) ; (1y +1=bp) p(07 -+- 1= Ap)
7i (1y
b = 13
+ 1 = bp) 5(07 + 1 = 13p)
. . . . . . . __p(0
( 11) A is a positive non-zero rational number; that is, there is"'"' at least
+ 1 =Bp)
one positive non-zero whole number which is a multiple ofA.
~;(o, +A~ n,) X(0,-+- 1 = np)
{J
"' See footnote above, p. lOfT.
"'"' See above p. 14 and Begr(ffsschrift ? 12.
7
? ? ? ? ? 24 Boole's logical Calculus and the Concept-script
(12) A is the least common multiple of B and r. More explicitly: every common multiple of B and r is greater than or equal to A, and* A is a common multiple of B and r.
~(Or+B =Ap)
~(Or+F = A11) A . :;;:: a
~(Or+B = a11)
~(Oy +F = a11)
In neither this nor the preceding examples is the concept of a product
presupposed.
(13) The real function <P(x) is continuous at x = A; that is, given any positive non-zero number n, there is** a positive non-zero g such that any number b lying between +g and -g satisfies the inequality - n. :;;::<P(A+b)- <P(A)::;;n
n o b -n. :;;:: <P(A + b ) _ <P(A) :? : n -g. S:b. S:g
g>O n>0
I have assumed here that the signs <, >. :;; mark the expressions they stand between as real numbers.
(14) The real function <P(x) of a real variable xis continuous throughout the interval from A to B.
- n. :;;::<P(c+b)-<P(c). :;;::n -g. S:b. S:g A. S:c+b. S:B
g>0
. .
? Boole's logical Calculus and the Concept-script1 [1880/81]
In his writings, Leibniz threw out such a profusion of seeds of ideas that in IIus respect he is virtually in a class of his own. A number of these seeds wrrc developed and brought to fruition within his own lifetime and with his 11tllaboration, yet more were forgotten, then later rediscovered and drvcloped further. This justifies the expectation that a great deal in his work lhul is now to all appearance dead and buried will one day enjoy a rr,urrcction. As part of this, I count an idea which Leibniz clung to lluuughout his life with the utmost tenacity, the idea of a lingua l'haracterica, an idea which in his mind had the closest possible links with lhnl of a calculus ratiocinatur. That it made it possible to perform a type of ru111putation, it was precisely this fact that Leibniz saw as a principal ndvanlage of a script which compounded a concept out of its constituents rnl her than a word out of its sounds, and of all hopes he cherished in this muller, we can even today share this one with complete confidence. I will IJlrulc just the following from the relevant passages:
'Si daretur vel lingua quaedam exacta, vel genus scripturae vere philosophiae, . . . omnia quae ex datis ratione assequi, inveniri possent l. fUodam genere calculi, perinde ac resolvuntur problemata arithmetica aut jcomctrica. '*
? I>c Scientia universali seu calculo philosophico.
1 In 188 I, this article was submitted by Frege in turn to the Zeitschrift fiir Matlrematik und Physik, the Mathematischen Annalen and the Zeitschrift fiir l'lr/lu. wphie und philosophisehe Kritik, but was in every case rejected by the editors. lllluully remained unpublished.
l'rom the report of H. Scholz and F. Bachmann: Der wissenschaftliche Nachlass r? mt {,'o/1/ob Frege (Paris 1936) we learn that the lost original was 'a manuscript rrrJlnrcd for publication of 103 closely written sides of quarto'. Scholz and lt11chmann mention besides that Frege also submitted the manuscript to R. Avenarius for the Vierteljahrsschrift fii. r wissenschaftliche Philosophie. However it 11uuld be that what Frege submitted to Avenarius was the essay published in this v11lumc on pp. 47 IT. 'Boole's logical Formula-language and my Concept-Script', llnce Avenarius in his letter to Frege of 20/4/1882 cites the title of the manuscript rtlurncd by him as 'Boole's logical Formula-Language'.
The article can scarcely have been composed before 1880, the year in which the rtview by Schroder mentioned on p. 11 appeared. A great number of the reflections
? ? ? 10 Boole's logical Calculus and the Concept-script
Among the various sorties Leibniz made upon his goal, the beginnings of a symbolic logic come closest to what seems to be indicated by the phrase 'calculus ratiocinatur'. They are to be found in the essays:
Non inelegans specimen demonstrandi in abstractis and Addenda ad specimen calculi universalis
In these Leibniz stuck very close to language. Just as the words we use for the attributes of a thing follow one another, so he simply juxtaposes the letters corresponding to properties in order to express the formation of a concept. If, for instance, A means right-angled, B isosceles, C triangle, then Leibniz represents right-angled isosceles triangle by ABC. He uses a sign for identity, oo, and the sign + with the definition:
'A +Boo L significat A inesse ipsi L'.
This seems to coincide with the meaning recent logicians have given this sign, according to which A + B represents the class of individuals which belong to A or to B or to both. Since I am passing over less important details, the only other fact I will mention is that Leibniz allows the words 'non' and 'ens' to occur in his formulae. In this project he surely has the lingua characterica in mind, even though he made no express connection with the attempts he made to represent a content.
This way of setting up a formal logic seems to suggest itself naturally. At any rate recent German and English logicians have arrived at the same conception quite independently of Leibniz, though, as far as I know, in doing so they do not have a general characteristic in mind. However much Boolean logic* may stand out as a systematic working out of the fragmen- tary hints in Leibniz, it only goes beyond him in one point of fundamental importance-in the way it reduces hypothetical and disjunctive judgements to categorical. ? ? The Leibnizian 'ens' is left out.
In a short monograph,? ? ? I have now attempted a fresh approach to the Leibnizian idea of a lingua characterica. In so doing, I had to treat in part the same subject-matter as Boole, even if in a different way. This has
? Boole's main work is An Investigation ofthe Laws ofThought on which are founded the mathematical theories of logic and probabilities. London 1854.
** On one point indeed Boole has taken a retrograde step away from Leibniz, in adding to the Leibnizian meaning of A + B the condition that the classes A and B should have no common element. W. Stanley Jevons, E. Schroder and others have quite rightly not followed him in this.
*** Begriffschrift, eine der arithmetischen nachgebildete Formelsprache des reinen Denkens. Halle aS. 1879. (Note continued, pp. 11-12. )
put forward here by Frege are along the same lines as the contents of the essay published in 1882 Uber der Zweck der Begriffsschrift.
The last two sides of the lost original are preserved on a photostat, which was presented at the Paris Congress in 1935 as a specimen of the Frege NachlqfJ.
? ? BooZe's logical Calculus and the Concept-script 11
prompted many critics to draw comparisons between the two, of which the most detailed are those set out by E. Schroder in the Zeitschrift fiir Math. 11. Phys. Vol. XXV. He will not regard it as detracting from the gratitude which I here express for his thorough study of my monograph and its friendly review, if in what follows I attempt to supplement and correct that comparison.
A few of the signs I introduced there should be explained briefly here:
I deviate from usual practice in drawing a distinction between judgement und content of possible judgement. For me the relation in a hypothetical is uot one between judgements but between contents of possible judgement. 11ut ifl affirm that this relation holds, I then express a judgement.
( I) The content-stroke is horizontal, it is always prefixed to the expression of a content of possible judgement, serving to connect it to the judgement uml negation-strokes, and also to combine it with other contents of possible 111dgement by means of the conditional-stroke(? 2): e. g.
- ( 2 + 3 = 5).
(2) The judgement-stroke is placed vertically at the left hand end of the
content-stroke, it converts the content of possible judgement into a Judgement (? 2): e. g.
1--(2 +3 = 5).
(3) the negation-stroke is placed vertically under the content-stroke dividing it into two parts. The part to the right of the negation-stroke is the content-stroke of the original content, the part to the left that of its negation (~ 7): e. g.
. . . . . -(2 + 3 = 6).
(4) The conditional stroke connects two content-strokes by running from
the under side of the upper one to the left hand end of the lower. Like the negation-stroke it splits the upper stroke in two, of which the part to the right remains the content-stroke of the upper content, while the part to the left becomes the content-stroke of the whole, which means the negation of Ihe case that the upper content is false, and the lower true(? 5): e. g.
tx=7
X+ 3 = 10: means:thecasethatx+3= 10withoutx= 7doesnotoccur. Wecansay
in this case: ifx + 3 = 10, then x = 7. The roman letters such as the x here, confer generality on the content of the judgement by presenting it as true, whatever one may substitute for the roman letter (? 11, p. 21). If, for instance, we replace x by 2 we obtain
t,-2=7
L 2 + 3 = 10.
This is a true judgement, although the upper content, as well as the lower, is 1'11lsc. Here, the rendering 'if would jar with normal usage. Yet here too, the CIUIC is denied that 2 + 3 = 10 is true and 2 = 7 false.
If a negation-stroke stands to the left of the conditional stroke, as in
~2 = 7
2 + 3 = 5,
? ? 12 Boole's logical Calculus and the Concept-script
Above all, if we are not to go astray it is necessary that we should always bear in mind the purpose that governed Boole in his symbolic logic and the one that governed me in my Begriffsschrift.
If I understand him aright, Boole wanted to construct a technique for resolving logical problems systematically, similar to the technique of elimination and working out the unknown that algebra teaches. To this end, he represents judgements in the form of equations that he constructs out of letters and arithmetical signs such as +, 0 and 1. Logical laws then assume the form of algorithms, although these only coincide in part with those that hold in arithmetic for the same signs. In all this there is no concern about content whatsoever. In the main these means fulfil their purpose, at least as far as the range of problems that Boole has in mind are concerned. But one may think of logical problems lying outside this range. The use of arithmetical signs for logical purposes has the advantage that one is spared the necessity of learning a completely new algorithm. A large number of the transformations already familiar to our eyes still remain valid. ? Higher demands, which however Boole does not set himself, are naturally not met by this enterprise. Anyone demanding the closest possible agreement between the relations of the signs and the relations of the things themselves will always feel it to be back to front when logic, whose concern is correct thinking and which is also the foundation of arithmetic, borrows its signs
from arithmetic. To such a person it will seem more appropriate to develop for logic its own signs, derived from the nature of logic itself; we can then go on to use them throughout the other sciences wherever it is a question of preserving the formal validity of a chain of inference.
In contrast we may now set out the aim of my concept-script. Right from the start I had in mind the expression ofa content. What I am striving after is a lingua characterica in the first instance for mathematics, not a calculus restricted to pure logic. But the content is to be rendered more exactly than is done by verbal language. For that leaves a great deal to guesswork, even if only of the most elementary kind. There is only an imperfect correspondence between the way words are concatenated and the structure
the negation of the case that 2 + 3 = 5 without 2 = 7 is thereby converted into its affirmation we may render: 2 + 3 = 5 and it is not the case that 2 = 7 (? 7, p. 13). In
fTrr-2= l+l L2+3=5
-. - 2 = 1 + 1takes the place ofthe 2 = 7 ofthe preceding example. In rendering this, two denials come together and yield an affirmation:
2+3=5,and2=1+1(? 7. p. 12).
? I should mention that the deviations from arithmetic are for all that so fundamental ti}at solving logical equations is not at all like solving algebraic ones. And this greatly diminishes the value of the agreement in the algorithm.
? Boole's logical Calculus and the Concept-script 13
of the concepts. The words 'lifeboat' and 'deathbed' are similarly constructed though the logical relations of the constituents are different. So the latter isn't expressed at all, but is left to guesswork. Speech often only indicates by inessential marks or by imagery what a concept-script should spell out in full. At a more external level, the latter is distinguished from verbal language in being laid out for the eye rather than for the ear. Verbal script is of course also laid out for the eye, but since it simply reproduces verbal speech, it scarcely comes closer to a concept-script than speech: in fact it is at an even greater remove from it, since it consists in signs for signs, not of signs for the things themselves. A lingua characterica ought, as Leibniz says, peindre non pas les paroles, mais les pensees. The formula- languages of mathematics come much closer to this goal, indeed in part they nrrive at it. But that of geometry is still completely undeveloped and that of urithmetic itself is inadequate for its own domain; for at precisely the most important points, when new concepts are to be introduced, new foundations laid, it has to abandon the field to verbal language, since it only forms numbers out of numbers and can only express those judgements which treat 111' the equality of numbers which have been generated in different ways. But arithmetic in the broadest sense also forms concepts-and concepts of such richness and fineness in their internal structure that in perhaps no other science are they to be found combined with the same logical perfection. And there are other judgements which arithmetic makes, besides mere equations und inequalities. The reason for this inability to form concepts in a scientific manner lies? in the lack of one of the two components of which every highly developed language must consist. That is, we may distinguish the formal part which in verbal language comprises endings, prefixes, suffixes and auxiliary words, from the material part proper. The signs of arithmetic correspond to the latter. What we still lack is the logical cement that will bind these building stones firmly together. Up till now verbal language took over this role, and hence it was impossible to avoid using it in the proof itself, and not merely in parts that can be omitted without affecting the cogency of the patterns of inference, whose only purpose is to make it easier lo grasp connections. In contrast, Boole's symbolic logic only represents the formal part of language, and even that incompletely. The result is that
Boole's formula-language and the formula-language of arithmetic each solve only one part of the problem of a concept-script. What we have to do now, in order to produce a more adequate solution, is to supplement the signs of 111athematics with a formal element, since it would be inappropriate to leave the signs we already have unused. But on this score alone Boole's logic is nlready completely unsuited to the task of making this supplementation, since it employs the signs +, 0 and 1 in a sense which diverges from their nrithrnetical ones. It would lead to great inconvenience if the same signs were to occur in one formula with different meanings. This is not an
? In the case of the formula-language of arithmetic.
? 14 Boole's logical Calculus and the Concept-script
objection to Boole, since such an application of his formulae obviously lay completely outside his intentions. Thus, the problem arises of devising signs for logical relations that are suitable for incorporation into the formula- language of mathematics, and in this way of forming-at least for a certain domain-a complete concept-script. This is where my booklet comes in.
Despite all differences in our further aims, it is evident from what has been said already that the first problem for Boole and me was the same: the perspicuous representation of logical relations by means of written signs. This implies the possibility of comparing the two. If I now turn to this, it cannot be done in the sense of adjudicating between the two formula- languages, which is to be preferred. To raise such a question would mean referring back to their ultimate aims, which are more ambitious in my case than in Boole's. It would indeed be more than possible that each set of signs was the more appropriate for its own ends. Nevertheless, it seems to me worthwhile to work out the comparison in detail, since in that way many of the peculiarities of my concept-script are thrown into sharper focus.
To begin with, as far as the extent of these sign languages is concerned, one is entitled to expect that these would have to coincide, once everything had been sifted from my concept-script which lies beyond the confines of pure logic. Yet even then, not everything that I express can be translated into Boolean notation, whereas the converse transformation is possible. E. Schroder indeed thinks that my concept-script would only correspond to the second part of Boole's. This is refuted by the fact that I too can easily express such a judgement as 'Every square root of 4 is a fourth root of 16' by
~a4 =16 La2 = 4
although Boole would count it as one of the primary propositions. I represent all judgements in the first part of Boole in a similar way though in so doing I admittedly construe them quite differently. The real difference is that I avoid such a division into two parts, the first dedicated to the relation of concepts (primary propositions), the second to the relation of judgements (secondary propositions) and give a homogeneous presentation of the lot. In Boole the two parts run alongside one another, so that one is like the mirror image of the other, but for that very reason stands in no organic relation to it.
On the other hand, Boole had only an inadequate expression for particular judgements such as 'some 4th roots of 16 are square roots of 4'
~a2 =4
La4 = 16,
and for existential judgements such as 'there is at least one 4th root of 16'
IT&ra4 = 16,
So it transpires that even when we restrict ourselves to pure logic my
apparently no expression at all.
? Boole's logical Calculus and the Concept-script 15
concept-script commands a somewhat wider domain than Boole's formula- lunguage. This is a result of my having departed further from Aristotelian logic. For in Aristotle, as in Boole, the logically primitive activity is the formation of concepts by abstraction, and judgement and inference enter in through an immediate or indirect comparison of concepts via their extensions. ? The only difference is that Aristotle places in the foreground 1he case where the extension of one concept completely includes that of unother-i. e. that of subordination-whereas Boole reduces other cases to lhe case of equality of extensions. ? ? Here hypothetical and disjunctive 111dgements are for the moment left on one side. Boole now construes the hypothetical judgement 'If B, then A' as a case of subordination of con- cepts, by saying 'the class of time instants at which B is included in the class ,,rtime instants at which A'.
The full incongruity of the introduction here of the idea of time instants slunds out most clearly if you think of eternal truths such as those of mathematics. Schroder seems to want to avoid the artificiality this involves, siuce, in company with Hugh McColl, he explains expressions like A = 0, A t 8 = 1 etc. -whose sense, on the Boolean conception, is self-explanatory when taken in conjunction with the stipulations of the first part-all over U! (ain without referring back.
But in this way the last weak link between the 1wo parts is also snapped, and the signs 0, 1, = receive yet a third meaning 111 addition to their Boolean and arithmetical ones. According to Boole 0 111cans the extension of a concept under which nothing falls, as for example lhc extension of the concept 'whole number whose square is 2'. By 1, Boole uuderstands the extension of his universe ofdiscourse. These meanings hold for the first as much as the second part. I f one now ruptures this connection, then strictly speaking 0 has no longer an independent meaning in the Nccond part; combined with the identity sign it means a denial expressed as a
Judgement, while '= 1' designates an affirmation, which I express by the judgement-stroke. Then, besides this, the identity sign still has an iudependent meaning in formulae like A = B. But that even Boole on his npproach does not establish an organic connection between the two parts cun be seen from the fact that he does not use the equations of the first part us constituents of equations of the second part, and, if you hold strictly to Iheir meanings, cannot so use them. For in the first part A = B is a judge- ment. whereas if it were made a constituent of an equation of the second purl, as say in
( A = B )C = D
? This is not meant to imply that concept formation takes up a great deal uf space in their presentations. Rather, that their logics are essentially doctrines of inference, in which the formation of concepts is presupposed AN something that has already been completed.
** The relation of identity has greater content and so is less extensive nru. l general than that of subordination.
? ? ? 16 Boole's logical Calculus and the Concept-script
A = B would mean the class of time instants at which the content of the judgement 'A = B' was to be affirmed. On the other hand, if, to avoid this, you do not resort to the expedient of the idea of a class of time instants, then you get in return the double sense of the 0, the 1 etc. in the same formula, which would then, to say the least, become unperspicuous.
As opposed to this, I start out from judgements and their contents, and not from concepts. The precisely defined hypothetical relation between contents of possible judgement has a similar significance for the foundation of my concept-script to that which identity of extensions has for Boolean logic. I only allow the formation of concepts to proceed from judgements. If, that is, you imagine the 2 in the content of possible judgement
24 = 16
to be replaceable by something else, by (-2) or by 3 say, which may be
indicated by putting an x in place of the 2: x4 = 16,
the content of possible judgement is thus split into a constant and a variable part. The former, regarded in its own right but holding a place open for the latter, gives the concept '4th root of 16'. ?
We may now express
by the sentences '2 is a fourth root of 16' or 'the individual 2 falls under the concept "4th root of 16"' or 'belongs to the class of 4th roots of 16'. But we
? The reader may be surprised I don't put this in the form
2 = {116.
The reason is that {116 may not in the same context mean now this, now
that, 4th root of 16. Otherwise you could form the inference 2 = y/i6
- 2 = ~16 2=-2.
By using ~ on one side of an identity sign, you have laid it down that \li6 is to mean a particular 4th root of 16, just as the letter a too must be given the same meaning throughout a given context. I f it is in fact stipulated that ~is to mean the positive real root then -2 = ~is false and 2 =
~ isnottoberead:
'2 is a 4th root of 16', but
'2 is the positive real root of 16'.
We can see from this that the root sign cannot properly be used to help express that an individual falls under the concept of a root.
? Boole's logical Calculus and the Concept-script 17
may also just as well say '4 is a logarithm of 16 to the base 2'. Here the 4 is being treated as replaceable and so we get the concept 'logarithm of 16 to the base 2':
2X = 16.
lhc x indicates here the place to be occupied by the sign for the individual lulling under the concept. We may now also regard the 16 in x4 = 16 as replaceable in its turn, which we may represent, say, by x4 = y. In this way we arrive at the concept of a relation, namely the relation of a number to its ? lth power. And so instead of putting a judgement together out of an utdividual as subject* and an already previously formed concept as predicate, we do the opposite and arrive at a concept by splitting up the content of possible judgement. ** Of course, if the expression of the content of possible judgement is to be analysable in this way, it must already be 1tsclf articulated. We may infer from this that at least the properties and 1l'lations which are not further analysable must have their own simple dl'signations. But it doesn't follow from this that the ideas of these properties 1111d relations are formed apart from objects: on the contrary they arise . \lmultaneously with the first judgement in which they are ascribed to things. llcnce in the concept-script their designations never occur on their own, but always in combinations which express contents of possible judgement. I could compare this with the behaviour of the atom: we suppose an atom nl'vcr to be found on its own, but only combined with others, moving out of uuc combination only in order to enter immediately into another. *** A sign ll>r a property never appears without a thing to which it might belong being ut least indicated, a designation of a relation never without indication of the things which might stand in it.
In contrast with Boole, I now reduce his primary propositions to the ,\'t'condary ones. I construe the subordination of the concept 'square root of
* The cases where the subject is not an individual are completely different from these and are here left out of consideration.
? ? A great deal of tedious discussion about negative concepts such as 'not-triangle' will, as I hope, be rendered redundant by the conception of the relation of judgement and concept outlined here. In such a case one simply doesn't have anything complete, but only the predicate of a judgement which as yet lacks a subject. The difficulties arise when people treat such a fru~ment as something whole.
In this connection, I find it extraordinary that some linguists have recently viewed a 'Satzwort' (sentence-word), a word expressing a whole judgement, as the primitive form of speech and ascribe no independent existence to the roots, as mere abstractions. I note this from the KMtlngschen ge/ehrten Anzeigen 6 April 1881: A. H. Sayee, Introduction to the? Science ofLanguage 1880 by A. Fick.
? ? ? As I have since seen, Wundt makes a similar use of this image in his l. oglk.
? 18 BooZe's logical Calculus and the Concept-script
4' to the concept '4th root of 16' as meaning: if something is a square root of
4 it is a 4th root of 16:
trx4 = 16 Lx2 = 4.
I believe that in this way I have set up a simple and appropriate organic relation between Boole's two parts. Moreover, on this view we do justice to the distinction between concept and individual, which is completely obliterated in Boole. Taken strictly, his letters never mean individuals but always extensions of concepts. That is, we must distinguish between concept and thing, even when only one thing falls under a concept. The concept 'planet whose distance from the sun lies between that of Venus and that of Mars' is still something different from the individual object the Earth, even though it alone falls under the concept. Otherwise you couldn't form concepts with different contents whose extensions were all limited to this one thing, the Earth. In the case of a concept it is always possible to ask whether something, and if so what, falls under it, questions which are senseless in the case of an individual. We must likewise distinguish the case of one concept being subordinate to another from that of a thing falling under a concept although the same form of words is used for both. The examples given above
tx4 = 16 24
x = 4 and f--2 = 16
show the distinction in the concept-script. The generality in the judgement
'TX4 = 16 Lx2 = 4
'All square roots of 4 are 4th roots of 16' is expressed by means of the letter x, in that the judgement is put forward as holding no matter what one understands by x. I stipulated that roman letters used in the expression of judgements should always have this sense.
Let us now look at the case where the content of such a general affirmative judgement occurs as part of a compound judgement, say as the antecedent of an hypothetical judgement; e. g. :
If every square root of 4 is a 4th root of m, then m must be 16. The expression
does not correspond to the sentence, and is even false, which is why the judgement stroke has been left off the left-hand end of the uppermost horizontal stroke; for we may substitute numbers for x and m which falsify this content. Thus if we take m to be 17, then the consequent m= 16 would
? ? ? Boole's logical Calculus and the Concept-script 19 hecome 17 = 16 and so would be false. Of course that doesn't yet
necessarily make the whole
false; for if the antecedent
17=16 1(x4 = 17
x2 =4
x4 = 17 l x2 = 4
were also false, the whole would be true despite the falsity of the conse- lJUent. *
But we may take a value for x, 3 say, which satisfies the condition x' = 17
for
lx2 = 4 134 = 17
2 3=4
is true, since not only 34 = 17 but also 32 = 4 is false. Thus ifwe give mthe value 17 and x the value 3, the antecedent
is satisfied but the consequent
x'=m
1x2 = 4 m= 16
m= 16
x=m ~4
2 x=4
is false. Thus
is not true for all values of x and m, which is what would be asserted by prefacing the formula with the judgement-stroke. But the sentence
'If every square root of 4 is a 4th root of m, then m must be 16'
Nays something different. You could also express its content as follows:
'If, whatever you understand by x it holds that x4 = m must be true ifx2 = 4, then m= 16'.
We can see: the generality to be expressed by means of the x must not KOvern the whole
*Cf. Footnote above p. lOfT. and Begriffschrift ? 5, where the latter atdmittedly contains the mistake pointed out by Schroder in his review. This however had no effect on what followed.
? ? 20 Boole's logical Calculus and the Concept-script but must be restricted to
x4 =m Lx2 = 4
I designate this by supplying the content-stroke with a concavity in which I put a gothic letter which also replaces the x:
-6-r-a' = m La2 = 4.
I thus restrict the scope of the generality designated by the gothic letter to the content, into whose content stroke the concavity has been introduced (? 11 of the BegrifJsschrift). * So our judgement is given the following expression:
m= 16 a4 =m a2 = 4.
By means of this notation, I am now also able to express particular and existential judgements. I render the sentence 'Some 4th roots of 16 are square roots of 4' thus:
For
~a2 =4
La4 = 16.
n a2=4 -vra4 = 16
means the content of possible judgement
'If anything is a 4th root of 16, it is not a square root of 4', or 'No 4th roots of 16 are square roots of 4'.
We now designate this content as a false generalization by prefacing it with a negation-stroke, and present the result as an assertion by means of the judgement-stroke. Analogously
~a2 = 4means: 'There is at least one square root of 4'.
This is the negation of the generalization of the negation of the equation a2 = 4.
We can now also easily show the link between particular and existential judgements. For in
tr0,-ra2 = 4 La4 = 16
* In discussing my monograph, E. Schroder made the proposal that the designation of generality be introduced into Boolean logic by the use of gothic letters. However that is inadequate, since the scope over which the generality is supposed to extend is still left open. The drawback that a second negation sign is needed is connected with this.
? BooZe's logical Calculus and the Concept-script 21 we may insert two negation-strokes in immediate succession, which thus
nmcel each other out
IT&mra2 = 4 La4 = 16
nnd think of this as concatenated as indicated here: I~"T1Ta2 = 4
tust as you may analyse
La4 = 16
tr&ra2=4 into l~a2 =4.
Thus the only distinction between
IN lhat
IT(j;,(T a2 = 4
La4 = 16 and
'[a2 = 4 a4 = 16
t11kcs the place of a2 = 4. But according to ? 7 of the Begriffsschrift,* this mcnns that a is a square root of 4 and a 4th root of 16. So just as you can lrnnslate
~a2 =4 as
'There is at least one square root of 4', you may express
IT6-rrrra2 =4 or t-r&Tra2 =4 La4 = 16 La4 = 16
by Ihe sentence
'There is at least one number which is both a square root of 4 and a 4th ruul of 16'. But this is equivalent to the expression:
'At least one 4th root of 16 is a square root of 4'.
I have now shown that the difference in extent of the domains governed by 8oolean logic and by my concept-script, extraneous as it might at first llaht appear, is in fact as closely as possible bound up with their original uunatruction. A few examples may now serve to illustrate how the uonatruction of the concept-script enables it when combined with the signs orarithmetic to achieve the more far-reaching goals it set itself.
(I) There are at least two different square roots of 4. The sentence 'a- b follows from a2 = 4 and b2 = 4' is denied in its generality.
? Sec also above, footnote on p. IOff.
,
ba=b a2 = 4 b2 = 4
? ? ? 22 BooZe's logical Calculus and the Concept-script (2) there is at most one number whose double is 4
1f;a:! [2b =4
(3) 4 is a positive whole number (including 0). That is, 4 belongs to the series beginning with 0, in which the immediate successor of any member is obtained by adding 1. *
(4) 12 is a multiple of 4; that is, 12 follows 0 in the arithmetical progression with difference 4. ** Two numbers with opposed signs are here not counted as multiples of one another, and 0 only counted as a multiple of itself.
(5) 12 is a common multiple of 4 and 6; that is, 12 is a multiple of 4, and*** 12 is a multiple of 6.
1~(0, +4 ~12,)
~(07 +6 = 1211)
(6) 4 is a common (aliquot) factor of 12 and 20.
y
1~ (0, +4 ~ 12,) p(07 +4 = 2011)
(7) The multiple of a multiple of a number is a multiple of that number.
T~(07+a= c11)
[5(0,+a~ b1)
~(Oy+b) = Cp)
(8) A and B are congruent modulo M. Here A, B and M need not be
whole numbers; nothing further than their addibility is presupposed. This
*Cf. Begriffss. :hrift ? 29, where in the gloss on formula (99) for;f(x10z6 ) one should re~d~/(x1,z6)?
. . Begrilfsschrift ? 26.
. . . See footnote above, p. lOfT.
? BooZe's logical Calculus and the Concept-script 23 way of speaking would be inconvenient for ordinary use, but it is only meant
to have application when it is necessary to go back to the concept. ~(A,+ M~B1)
t{ J ( B 7 + M = A p )
(9) 13 is prime. Here 1 is counted as a prime number, 0 not. More
explicitly: 13 is a positive non-zero whole number(~(07 +1 = 13p)),and"'
whatever positive number b may be which is greater than 1(p(17 + 1 = bp)) und different from 13 (--. - b = 13), 13 cannot be a multiple of b ( --,~(07 + b = 1 3 p ) ) -
y
{i(07+b= 13p)
y
( IO) A and B are positive non-zero whole numbers that are eo-prime. IIere 1is treated as prime to every number.
p(07 +b = Ap) p(07 +b=Bp) ; (1y +1=bp) p(07 -+- 1= Ap)
7i (1y
b = 13
+ 1 = bp) 5(07 + 1 = 13p)
. . . . . . . __p(0
( 11) A is a positive non-zero rational number; that is, there is"'"' at least
+ 1 =Bp)
one positive non-zero whole number which is a multiple ofA.
~;(o, +A~ n,) X(0,-+- 1 = np)
{J
"' See footnote above, p. lOfT.
"'"' See above p. 14 and Begr(ffsschrift ? 12.
7
? ? ? ? ? 24 Boole's logical Calculus and the Concept-script
(12) A is the least common multiple of B and r. More explicitly: every common multiple of B and r is greater than or equal to A, and* A is a common multiple of B and r.
~(Or+B =Ap)
~(Or+F = A11) A . :;;:: a
~(Or+B = a11)
~(Oy +F = a11)
In neither this nor the preceding examples is the concept of a product
presupposed.
(13) The real function <P(x) is continuous at x = A; that is, given any positive non-zero number n, there is** a positive non-zero g such that any number b lying between +g and -g satisfies the inequality - n. :;;::<P(A+b)- <P(A)::;;n
n o b -n. :;;:: <P(A + b ) _ <P(A) :? : n -g. S:b. S:g
g>O n>0
I have assumed here that the signs <, >. :;; mark the expressions they stand between as real numbers.
(14) The real function <P(x) of a real variable xis continuous throughout the interval from A to B.
- n. :;;::<P(c+b)-<P(c). :;;::n -g. S:b. S:g A. S:c+b. S:B
g>0
. .
