A Treatise on Probability and “My Early Beliefs”
Abstract
The purpose of this paper is to examine Keynes as a
philosopher in light of
The conclusion runs as
follows;
(1) In his youth,
Keynes firmly believed in human rationality, the relevant philosophical work
being his Treatise on Probability.
But he became increasingly skeptical about human rationality and ever more conscious
of the emotional aspect of human nature, in the midst of the chaos of post-WWI
Europe, as evidenced typically in his criticism of the market society and his
advocacy of the New Liberalism in The End
of Laissez-Faire (1926). As the years went by, Keynes’s skepticism about human
rationality grew ever deeper, and he came to stress the importance of custom
and tradition, re-evaluating the market society in face of the cruel realities
of the Soviet society.
(2) Keynes abandoned
the most essential part of his Treatise
on Probability, which lies in regarding probability as degrees of rational
belief between propositions, and the justification of induction based on it, in
the face of Ramsey’s criticism. This is clearly recognizable in his obituary of
Ramsey and My Early Beliefs. However, he did not abandon his theory of
probability completely, as can be seen in his critical review of Tinbergen
(1939). In Part V (The Foundations of Statistical Inference) of A Treatise on Probability Keynes had
criticized the mathematical use of statistical frequencies (Methods of Laplace)
and defended the inductive use of them (Methods of Lexis). Keynes’s criticism
of Tinbergen’s method is mainly based on this stance . We have the Keynes who
accepted Ramsey’s criticism on the one hand, and the Keynes who retained the
Methods of Lexis on the other. There seems to run a logical fissure between
Part III (Induction and Analogy) and Part V.
1. Introduction
Stimulated by the spate of
documents made public from the 1980s onwards, searching studies on “Keynes as a
philosopher” have been pursued, leading to a trend to re-examine and re-evaluate
Keynes’s performance from the point of view of the way the philosophy of young
Keynes developed in A Treatise on Probability (Keynes, 1921. Hereafter TP)
came to influence “Keynes as an economist”.
There is, however, a high wall which researchers
find themselves up against: namely the fact that Keynes did not write any philosophical
papers after the TP, into which he poured
his intellectual energies in his twenties. This meant for scholars painstaking research,
looking for philosophical traits in the writings of Keynes as economic
theoretician and economic policy expert. Under these circumstances, the only
evidence concerned is “My Early Beliefs”.
The purpose of this paper is to examine to
what degree “My
Early Beliefs”
could be said to reflect his philosophical and ethical path accurately. For
this purpose, first we will look into the theoretical substance of A
Treatise on Probability. We then go on to examine his philosophical path, with
the focus on Ramsey’s criticism of Keynes’s TP and Keynes’s (substantial)
acceptance. Thirdly, “My Early Beliefs” is examined. Finally some problems are raised on the relation between “Keynes as a philosopher” and “Keynes as an economist”.
2. The Stance of A
Treatise on Probability
To examine this paper we should
start by explaining the theoretical content of A Treatise on Probability
(hereafter the Treatise), for it is only in this book that “Keynes as a
philosopher” appears as the real protagonist.
There are three scholars who greatly
influenced the writing of the Treatise – Moore, Russell and W. E. Johnson. Firstly, it
was Moore’s
Principia Ethica (Moore, 1903) that spurred Keynes to start work on a
theory of probability. In MEB, Keynes said that in youth he accepted the “religion” of the Principia
Ethica, while rejecting the “moral”. However, in relation to the Treatise, Chapter 5 in Principia Ethica
matters, for his study started with his essay, “Ethics in Relation to Conduct” (1904), which is a
critical essay on the chapter, read at the Society in January1.
Secondly, he was greatly
influenced by Russell’s Principles of Mathematics (Russell, 1903) and Principia
Mathematica by Whitehead and Russell (1910). This emerges clearly from Part
II, “Fundamental
Theorems”.
Thirdly, there was the
part played by Johnson. In relation to Part II, Keynes started it independently
of Johnson, but on the way made many exchanges of views and ideas with him, the
result of which was incorporated in the Treatise2.
Keynes evoked the path he traversed under
those influences with thus :
… it was through trying to
prove the fundamental theorems of the subject [a theory of probability] on the
hypothesis that probability as a relation that I first worked my way
into the subject; and the rest of this treatise has arisen out of attempts to
solve the successive questions to which the ambition to treat probability as a
branch of formal logic first gave rise. (TP, 125)
Among other things, the Treatise
aims at exploring an epistemological and logical research in the world of
probability. We will examine the Treatise below, focusing on its essential
points of contention – the definition of probability; the question of whether
probability is objective or subjective; the sphere under consideration; the
argument for justification of induction.
2.1 Definition of
Probability
Probability is defined as a “degree
of rational belief between two sets of propositions”. The exact definition of
terms and their development are argued along the formal logic of Moore and
Russell. The fundamental view is very clear in the following passage.
The proposition (say, q) that we know in this case is not the same as the
proposition (say, p) in which we have a probable degree (say, α)
of rational belief. If the evidence upon which we base our belief is h,
then what we know, namely q, is that the proposition p bears
the probability-relation of degree αto the set of propositions h; … and
this knowledge of ours justifies us in a rational belief of degree α in the
proposition p. It will be convenient to call propositions such as p,
which do not contain assertions about probability-relations, ‘primary
propositions’; and propositions such as q, which assert the existence of
a probability-relation, ‘secondary propositions’ (TP, 11).
Knowledge is recognizing that
there exists a rational belief of degree, α, between the evidence h and
the primary proposition p, which is the “secondary proposition” q. (It should be noted that p is
a certain rational knowledge (See Table１). 3
There is a clear distinction between “rational beliefs” and “irrational
beliefs”
An idea – “A rational belief of degree” – can be said
to be an intellectual inquiry which tries to grasp a probability in the system
of “rational” thought. Although at first glance the Treatise seems to
explore the sphere of “uncertainty”, what it aims at, in fact, is to grasp a
“rational belief of degree” in terms of a logical relationship, and to develop
it in an axiomatic way in the best possible axiomatic way4.
In this epistemology, there is a distinction
between “to know in a probabilistic way” and “not to be able to know in that
way” 5. The phenomenon of probability is a rational recognition in
the sense that it can be grasped through the process of knowledge (the
secondary proposition). In MEB it is repeatedly stated that in his youth Keynes
placed profound trust in the “rationality” of human nature, which perfectly
fits with the fact that the Treatise is a book which explores the
“rationality of human nature”.
2.2 The Question of Whether
Probability Is Objective or Subjective
The key to knowing whether a
“probability” dealt with in the Treatise
is objective or subjective is to pay attention to the primary proposition and
the secondary proposition.
…our knowledge of propositions seems to be obtained in
two ways: directly, as the result of contemplating the objects of acquaintance;
and indirectly, by argument, through perceiving the probability-relation
of the proposition, a bou t which we seek knowledge, to other propositions. (TP,
12)
It appears to be indirect knowledge that the Treatise
aims at analyzing6.
Turning to the topic, Keynes treats
probability as “objective”. This is his basic stance. Although he does not deny
the sub jectivity that probability involves7, he emphasizes that
probability is basically objective.
What we know and what probability we can attribute to
our rational belief is, therefore, subjective in the sense of being relative to
the individual. But given the body of premises which our subjective powers and
circumstances supply to us, and given the kinds of logical relations, upon
which arguments can be based and which we have the capacity to perceive, the
conclusions, which it is rational for us to draw, stand to these premises in
an objective and wholly logical relation. (TP, 19. my underlining)
Fundamental here is the
passage, “But given the body … relation”, in which a logical relation between a
premise and a (primary) proposition is regarded as “objective”8.
Keynes aims at constructing the philosophical foundation for judgment in the
sphere of “probability” through appl ication of (formal) logic9.
The “Indirect knowledge” the Treatise
deals with is obtained through argument. In cases where argument is composed of
a complicated set of arguments, a knowledge of formal logic is indispensable.
Thus recog nition in the form of a system of axiomatic theory is required there.
2.3 Sphere under
Consideration
The concept of “probability”
looms large in the Treatise. It is discussed in detail in Chapter 3,
“The Measurement of Probabilities”, the essence of which runs as follows ― In a
case where points lie on the same path, there exists a numerically calculable
sphere. Even if they diverge from the path, there exist non-numerical probabilities.
In this situation, there are cases where to some degree the difference is distinguishable
in terms of larger or smaller, and cases where it cannot be distinguished. And
yet both of them are also the object of theory of probability.
Some argue that the Treatise mainly treats the sphere
in which numerical measurements are impossible. In fact, however, the Treatise
does not state that probability is immeasurable, but rather treats the sphere
of rational knowledge, in which both measurable and immeasurable probabilities
are included.
It is true that the Treatise expresses
the view that even in the sphere in which measurability is impossible the concept
of probability workable, and criticizes mathematical statistics which excludes
such a possibility. It is worth emphasizing, however, that the sphere in which
measurability prevails occupies an importan t role in the Treatise.
The purpose of Part II, “Fundamental Theorems”, for
example, is stated as follows.
My object in it [Part II] is to show that, starting
from the philosophical ideas of P art I [Fundamental Ideas], we can deduce by
rigorous methods out of simple and precise definitions the usually accepted
results, such as the theorems of the addition and multiplication of
probabilities and of inverse probability. (TP, 125)
Par II, which is greatly
influenced by Principia Mathematica, is composed of a series of formal
logical proofs and reflective consideration on them. In the former, proofs of
various theorems are deductively made, provided that prob ability can be
measured numerically.
2.4 The Argument for
Justification of Induction
Another main objective of
the Treatise is the justific ation of
“induction” (Here we consider “universal induction” in Logic. “Inductive
theory” in Statistics is to be referred to in Section 4.)
Although induction has made a great contribution to
the development of human knowledge, logicians have not succeeded in justifying
it as yet. This was what Keynes set out to tackle.
What we wonder at this moment is how a
deductive and axiomatic analysis based on the method of analytic philosophy
which we saw above and an argument for justification of “induction” could be
consistently related. Here it is worth noti ng how Keyes emphasizes that, from the
point of view of logic, there is a fundamental relation between probability and
induction.
I have described probabilit y as comprising that part
of logic which deals with arguments which are rational but not conclusive. By
far the most important types of such arguments are those which are based on the
methods of [pure] induction and analogy. (TP, 241. [ ] is mine)
…the validity of every
induction, strictly interpreted, depends, not on a matter of fact, but on the
existence of a relation of probability. An inductive argument affirms, not that
a certain matter of fact is so, b ut
that relativ e t o certain evidence there is a probability in its favour.
…… The clear apprehension of this truth profoundly modifies our attitude
towards the solution of the inductive problem. (TP, 245)
Here “probability” is
defined as a rational belief of degree between propositions. An “inductive
argument” is defined under the influence of this definition. It is argued,
moreover, that induction should be a problem of formal logic (the existence of
probabilistic relations) rather than a ‘certain matter of fact’.
That is to say, a probability at each point
of time is a relation between each premise and each conclusion, and is
independent of a probability at the next point in time. Keynes argues that induction
is not a problem of experience but one of formal logic. This might seem a surprising
argument to most of us today.
However, we should understand that only by adopting
such a stance is the theory of probability developed in Parts I and II directly
connected with Part III, “Induction and Analogy”. The problem of induction thus
becomes the main subject of Part III, “probability” as so defined being framed.
He states that an inductive argument has very
general implications10. In fact, he argues that the first two
chapters (Chs. 19 and 20) dealing with induction show the following feature.
In the
enunciation, given in the two preceding chapters, of the principles of analogy
and pure induction there has been no reference to experience of causality or
law. So far, the argument has been perfectly formal and mi ght relate to a set
of propositions of any type. (TP, 269)
Induction works through “analogy”
and “pure induction”. Keynes propends for the former. His explanation of Hume’s
egg analogy – which is a criticism of Hume’s inducti on – goes a long way toward
clarifying these two elements and Keynes’s evaluation of them11.
The conditions in which pure induction, which
is a mere repetition of instances, strengthens an argument, are dealt with in
details in Ch. 20, “The Value of Multiplication of Instances, or Pure
Induction”. This chapter is characterised as follows.
The chief value of the chapter, in my judgment, is
negative, and consists in showing that a line of advance, which might have
seemed promising, turns out to be a blind alley, and that we are thrown back on
known analogy. Pure induction will not give us any very substantial assistance
in getting to the bottom of the general inductive problem. (TP, 260-261)
The relation between pure
induction and analogy is explained as follows12.
When our control of the experiments is fairly
complete, and the conditions in which they take place are well known, there is
not much room for assistance from pure induction. … But where our control is
incomplete, and we do not know accurately in what ways the instances differ
from one another, then an increase in the mere number of the instances helps
the argument. For unless we know for certain that the instances are perfectly
uniform, each new instance may possibly add to the negative analogy. (TP,
243)
Analogy is, as it were, a
spatial extension of knowledge, while pure induction is a temporal extension of
knowledge. Having made a generalization due to pure induction based on ‘his’ own
“probability”, Keynes points out that in order f or it to hold a “prior
probability” needs to be known. He states that “analogy” provides it13.
The prior
probability, which must always be found, before the method of pure induction
can be usefully employed to support a substantial argument, is derived, I
think, in most ordinary cases – with what justification it remains to discuss –
from considerations of analogy. (TP, 265)
Keynes argues that one can
obtain a “prior probability” by considering analogy in the spatial extension of
knowledge. Thus induction turns out, according to him, to depend on analogy as
a means to obtain a prior probability. And he states that
it is a scientific method to
heighten the “known analogy”14.
To sum up, Keynes clearly states that (i) what matter s most about the
argument on probability is induction; (ii) induction is composed of “analogy”
and “pure inductio n”, the former of which is the more important; (iii) the
prior probability is obtained through analogy.
The fundamental argument of analogy runs as
follows.
If some one
thing is true about both of two objects, if, that is to say, they both satisfy
the same propositional function, then to this extent there is an analogy
between them. Every generalization g (φ,f), therefore, asserts that one analogy
is always accompanied by another, namely, that between all objects having the
analogyφ there is also the analogy f. The set of propositional
functions, which are satisfied by both of the t wo objects, constitute the positive
analogy. The analogies, which would be disclosed by complete knowledge, may
be termed the total positive analogy; those which are relative to
partial knowledge, the known positive analogy.
As the
positive anal ogy measures the resemblances, so the negative analogy measures
the differences between the two objects. The set of functions, such that each
is satisfied by one and not by the other of the objects, constitutes the negative
analogy. We have, as before, the distinction between the total negative
analogy and the known negative analogy. (TP, 248)
Let us explain the above with a simple example.
Suppose that we have objects, a dog and a cat. If “one thing” is “four legs”,
then both dog and a cat meet propositional function φ “if A, then four legs” (that
is, a dog and a cat has analogy φ). If “another thing” is “pelage”, then both dog
and cat meet propositional function f “if A, then pelage” (that is, a
dog and a cat has analogy f).
Then a set of propositional functions g (φ, f)
[a dog and a cat have four legs and have pelage] becomes a positive analogy. Distinction
between perfect knowledge and imperfect knowledge denote, respectively,
certai nty and a rational belief of degree in Keynes’s sen se. In the case where
knowledge is perfect, “total” is used, while otherwise “known” is
used.
Pos
itive analogy is counterposed by negative
analogy (in short, the objects are not similar). In the case of a dog and a
cat, if “one thing” is “dislike the cold”, then propositional function φ “if A,
it dislikes the cold” meets cat, but not dog. If “another thing” is “bark”,
then the propositional function f “if A, it barks” does not meet cat,
but meets dog. In this case, a set of proposition al functions g(φ, f) [a
dog and a cat dislike the cold, and bark] is called negative analogy.
According to Keynes, if knowledge is perfect,
there is no room for pure induction, for total analogy is there. On the other
hand, if knowledge is pa rtial, there is room f or pure induction. For by means
of pure induction negative analogy will be found, so positive analogy can move
closer to total positive analogy than in the case without it.
In the above argument Keynes takes the case
in which knowledge is imperfect, that is, the “probability” situation. A
“positive analogy” in that situation m eans for “one thing” to take the form of
“probability” propositional function between two objects. How can one perceive
the probability (the degree of similarity)?
If analogy is perfect, pure induction is not
needed. In this situation analogy is obtained as “direct knowledge”. Though direct
knowledge is, in a sense, individual introspection, it is stated to possess an
objectivity which everyone shar es in common. Keynes states this direct
knowledge in the sphere in which probability works.
… we can
justify the method of perfect analogy, and other inductive methods in so far as
they can be made to approximate to this, by means of the assumption that the
objects in the field, over which our generalizations extend, do no t have an
infinite number of independent qualities; … the use of inductive methods can be
justified if they are applied to what we have reason to suppose a finite
system. (TP, 285)
In the above passage “the
method of perfect analogy” means not needing pure induction. When knowledg e is
imperfect, analogy can attain a higher level with the aid of pure induction.
Here Keynes makes the point that insofar as the number of “the features of
things” (say, “four legs”, “pelage”) are finite, the inductive method can be
justified by “the method of perfect analogy” (which is obtained through individual
introspection) or “other inductive methods” (the method of coming close to
perfect analogy by finding negative analogies with the aid of pure induction).
There are some points in this argument which
I feel are doubtful, or at any rate odd.
Firstly, with “probability” and “inductive
method” defined à la Keynes, the fundamental relation between the two is
stressed, as a result of which the inductive method takes on an aspect
departing from what is usually supposed. The inductive method here is taken to work
in the world in which analogy plays an essential role, experience being
neglected.
Secondly, there is a peculiarity in thinking
that analogy can be grasped in the form of “probable” propositional function.
The world of formal logic is placed in the center of thought rather than the
world of experience.
3. After the Treatise
― Influences from Ramsey’s Criticism
F. Ramsey came up with
severe criticism of the Treatise in his paper, “Truth and Probability” (Ramsey,
1926) 15. And Keynes accepted it to a considerable degree. It was very
rare for him to admit so much publicly. This has attracted a great deal of attention,
for, as already noted, he published neither a paper nor a book on philosophy,
and criticism came from Ramsey who occupied an important place in the philosophy
developed in Cambridge.
In this section we will see the main points
of Ramsey’s criticism and Keynes’s response to them.
3.1 Ramsey’s Criticism
Ramsey’s criticism of the Treatise
rests mainly on three points.
The first is that there is no probability relation
between propositions - a sheer negation of the definition of probability by
Keynes.
If anyone
were to ask me what probability one gave to the other, I should not try to answer
by contemplating the propositions and trying to discern a logical relation
between them. I should, rather, try to imagine that one of them was all that I
knew, and to guess what degree of confidence I should then have in the other. (Ramsey,
1990 [1926], 59)
Here Ramsey speaks of a
subjective probability which an individual attaches to another proposition
rather than a probability between propositions. He insists that probability
should be a problem of individual judgment16.
The second point is that even the statement of its
main principles lacks consistency. This exposes an ambiguity between objectivity
and subjectivity of probability argued in the Treatise17
The third point is a criticism of Keynes’s attempt to
incorporate the sphere of the inductive method into that of the deductive
method.
The logical
relation which justifies the inference is that the sense of import of the
conclusion is contained in that of the premises.
But in the
case of an inductive argument this does not happen in the least; it is
impossible to represent it as resembling a deductive argument and merely weaker
in d egrees; it is absurd to say that the sense of the conclusion is partially
contained in that of the premises. (Ramsey, 1990 [1926], 82)
Positing a probability between proposition A and
proposition B does not mean that proposition B is deductively (and partially)
derived from proposition A. It is strange, says Ramsey, to argue as if this
were true.
These three criticisms clearly make sense.
3.2 Keynes’s Response
It was in an obituary to Ramsey, “Ramsey as a
Philosopher” (Keynes, 1931 [1933])
that Keynes responded to Ramsey’s criticism. This is
very important evidence
for an understanding of the stance of “Keynes as a
philosopher” as of October
1931. High regard for Ramsey’s attention to “human
logic” and self-criticism w ith
regard to the Treatise as argued based on
“formal logic” are expressed in a mixed
way.
Thus he [Ramsey] was led to consider ‘human logi c’ as
distinguished from ‘formal logic’. Formal logic is concerned with nothing but
the rules of consistent thought. But in addition to this we have certain
‘useful mental habits’ for handling the material with which we are supplied by
our perceptions and by our memory and perhaps in other ways, and so arriving at
or towards truth; and the analysis of such habits is also a sort of logic. The
application of these ideas to the logic of probability is very fruitful. .… So
far I yield to Ramsey – I think he is right. (1933, [JMK] 338-339).
Here we find endorsement of a
theory of probability which takes into due account Ramsey’s “human logic” rather
than a theory of probability as an objective relation between propositions. The
statement, “So far I yield to Ramsey – I think he is right.” carries a lot of
weight, given that the Treatise was a product of reflection prolonged
over a considerable span of time by Keynes as a philosopher.
4. On “My Early Beliefs”
Let us turn to “My Early
Beliefs”, in which Keynes’s own path of thought is stated (Following his will
[1941], the two memoirs were posthumously made public. This is one of the two).
It is summarized as follows18.
(i) Around 1903 ― “passionate contemplation and
communion” (MEB, 436. Moore’s “Religion”) was rated very highly. It was
considered to be rational and scientific. In contrast, Keynes rejected Moore’s
“Morals” following the Benthamite calculus and the general rules of correct
behavior” (MEB, 436). Keynes stood by individualism, which places trust in the
rationality of human nature.
(ii) Around 1914 ― Trust in the rationality of human
nature became weaker, while
human feelings began to receive more consideration.
(iii) 1938 ― Misgivings about the rationality of human
nature became ever deeper while trust in conventions became ever higher. He came
round to the idea that application of rationa lism and excessive individualism
should be restrained, and that Moore’s “religion” was narrow.
The most conspicuous term used
throughout “My Early Beliefs” is the “rationality of human nature”. When he speaks
of his path of thought, the basic tone as the years went by, was that trust in
it again ever and weaker. His regret that in his youth he had placed too much trust
in the rationality of human nature is repeatedly expressed19.
I think that profound trust in the
rationality of human nature lies at the root of the Treatise. There
Keynes tried to construct a magnificent theory of knowledge by incorporating the
framework of “rationality” into the sphere between “impossibility” and
“certainty”, and applying strict formal logic to it. Moreover, he tried to
justify an inductive method in the form of formal logic by approaching it from
the point of view of his concept of probability. It is no wonder that young
Keynes’s trust in the rationality of human nature stated in “My Early Beliefs”
ran through the Treatise as well.
Our
apprehension of good was exactly the same as our apprehension of green, and we
purported to handle it with the same logical and analytical technique which was
appropriate to the latter. … Russell’s Principles of Mathematics came
out in the same yea r as Principia Ethica; and the former, in spirit,
furnished a method for handling the material provided by the latter. (MEB,
438-439)
The circumstances w hich limited
Keynes’s scope when he embarked on his study of probability were what I
described above. The stance which discusses “green” and “good” with the same
logical and analytical method, and which uses the method of Principles of
Mathematics in order to handle material provided by Principia Ethica, is no more or less than the stance
taken with the Treatise.
However, as 1914 was drawing near, “trust in
the rationality of human nature” became weaker, Keynes says. Might it not be WWI
that was closely related to this change in philosophical view? 20
WWI
brought the Bloomsbury Group up against the burning issue of conscientious objection, which saw Keynes in a
delicate position vis-à-vis the Group. Moreover, he suffered a great
disillusionment with international politics during the negotiation process in
the Versailles Peace Conference. This was Keynes’s experience of a phase that
saw Western civilization , whic h had seemed to enjoy permanent progress, plunged
into the turmoil of chaos, confusion and revolution.
It was in the a rticles on social philosophy
such as “Am I a Liberal?” (Keynes, 1925) and The End of Laissez-Faire
(Keynes, 1926) that Keynes’s view of human society found expression. Here Keynes
criticizes that social philosophy based on the assumption that society is
compose d of rational individuals as mere fiction which neglects the fact that
the real world is composed of ignorant and weak individuals, arguing that attainment
of the public good would be difficult leaving everything to the private action
of individuals, and would be possible on ly by organizing social units.
The philosophy which underlies the Treatise
appears to be “ethics of rational self-interest” and belong to the current of
individualistic philosophy. It clearly aimed at examining “rational beliefs”, disregarding
irrationality. According as he shifted his interest from philosophy to
economics, the major events which attacked his interest from WWI onward seem to
have led him tow ard diffidence vis-à-vis the rationality of human nature and
respect of human feeling.
A “New Liberalism” is proposed in The End
of Laissez-Faire. It is rooted in misgivings about leaving the economic
society to ignorant and weak individuals. Ke ynes c onsiders institutions, in
size, between individuals and the state to be ideal. In this regard, he
positively evaluates the form of institutions which have been formed as a
result of historical development. That said, he mentions “risk, uncertainty,
ignorance” as one of the fields in which the state needs to be positively
involved in some way or another.
As the years went by, his trust in the
rationality of human nature seems to have become even weaker. As of 1938, Keynes's
position was:
The
attribution of rationality to human nature, instead of enriching it, now seems
to me have impoverished it. It ignored certain powerful and valuable springs of
feeling. Some of the spontaneous, irrational outbursts of human nature can have
a sort of value from which our schematism was cut off. (MEB, 448-449)
The error in approaching human
nature in the framework of schematism, and recognition that there exist
valuable feelings, even if irrational, in human nature finds clear expression
here.
In this transformation, disappointment with Soviet
socialism must have played no small part. Looking back to the severe criticism
of the “pseudo-morality” of capitalism in the 1920s, there is, clearly, a
change in his view of capitalism in the 1930s and 40s. This is emerges in all
evidence if we compare Chapter 24 of the General Theory (Keynes, 1936),
where the pursuit of profit is admitted and individualism affirmed, with The
End of Laissez-Faire.
5. Keynes as “a Philosopher” and as “an Economis t”
5.1 The General Theory
In the latter half of the
1920s Cambridge saw a major change in Wittgenstein’s philosophy from “the
former period” to “the later period”, touched off by criticism, again, from
Ramsey. Keynes was there at the time, and deeply involved in discussion.
Given these situations, the view has been put
forward that there is a distinction to be made between the earlier period and
the later period of Keynes as philosopher as well22. Accepting
Ramsey’s criticism and abandoning the Treatise, argues this view, Keynes
envisaged an alternative philosophy, e.g. “collective pragmatism”23.
In contrast, there is a view that the philosophy
developed in t he Treatise runs through up until the General Theory24. For example, stating that “Once this
post-1931 output is taken into account, a contrary pict ure emerges in which
literal interpretations of the review are revealed to be superficial” (O'Donnell,
1989, 140), O’Donell argues that up until the General Theory the
essential points of the Treatise were maintained. Similarly, Carabelli remarks
that in spite of the lip service which Keynes paid to Ramsey’s criticism in his
essay, Keynes did not change his view on probability in substance (cf. Carabelli,
1988, 97).
My view on this issue runs as follows. On the
one hand, Keynes abandoned the fundamental philosophy underlying the
Treatise (among others, Parts I, II and III). In this respect, Ramsey’s
criticism had a significant impact.
However, there
is no trace of Keynes’s having tried to work out an philosophy alternative to that
of the Treatise. Rather there is a possibility that Keynes’s change of
tack was brought about by events during and after WWI. Therefore, I am somewhat
skeptical about the possibility of coming to significant results by examining Keynes’s
economics writings to determine to what degree he changed in philosophical
terms after the Treatise. It does not seem that whether Keynes
philosophically changed or not had a significant influence on “Keynes the
economist” 25 26. Rather, in consi dering Keynes as an economist the
aspect of his social philosophy seems to be more important.
Incidentally, I understand the features of
the General Theory as follows.
(i) Theoretical System Which Is Operational – The General Theory
is purely theoretic with little economic policy analysis, which is rare in his
writings. And yet he is consistent, in the sense that a model is constructed in
such a way as to incorporate the viewpoint of the policy planner.
(ii) Monetary Economics of Underemployment Equilibrium – The main feature
of the General Theory is that it presents the monetary economics of
underemployment equilibrium. Keynes proposed it as a system of simultaneous
equations based on a causal analysis.
(iii) Model Determining the Volume of Employment – The model which the General
Theory proposes determines the volume of employment.
(iv) Two Contrasting Phases - The
view of the market society presented in the General Theory is as a
system showing two contrasting phases -
On the one hand, “stability, certainty, simplicity”, on the other,
“instability, uncertainty, complexity”.
After the publication of the
General Theory, Keynes mentions two
points of departure from the traditional theory in “The General Theory of
Employment” (Keynes, 1937). One is emphasis on the fact that the future is
uncertain, while the other consists in a theory of supply and demand for the
output as a whole. These two points correspond to the above-mentioned “two
contrasting phases”.
5.2 Keynes’s View of th e
Nature of Economics
After publication of the General
Theory Keynes famously advanced severe criticism of Tinbergen (1939) in
Keynes (1939b). Two points are emphasized there. One is that economics should
be a moral scienc e, the other that economics should belong to the sphere of
logic.
Here we shall dwell on the latter. Keynes
thinks that economics can make progress through improvement of models. However,
actual figures should not be put into a variable function, for if they were,
the model would lose generality and value as a way of thinking. The objective
of statistical studies is to test the relevance and effectiveness of a model
rather than to put figures for variables for the sake of forecasting27.
This view fully reflects the stance taken in the Treatise, where we find lucid criticism of theory
of probability as developed in the form of frequency theory. Part V, “The
Foundation s of Statistical Inference” is directly related. There, as corresponding
to the “universal induction” (in logic) developed in Part III, “Induct ion and
Analogy”, statistical inference (in statistics) is advocated.
On this occasion, two methods of using statistical
frequency in order to determine posterior probability are compare d. One is
“mathematical use” (or the method of Laplace), the other “inductive use” (or
the method of Lexis). Severely critical of the former, Keynes upholds the
latter (sta tistical inference). The inductive method in the Treatise criticizes the frequency theory
in which probability is dealt with depending on the multiplication of cases,
while the Treatise emphasizes the
importance of “analogy” and “the method of Lexis”.
Among other things, Keynes held that it is very
dangerous to use variables in probability, as is clearly seen in the following.
In the logic of implication, which deals not
with probability but with truth, what is true of a variable must be equally
true of all instances of the variable. In probability, on the other hand, we
must be on our guard wherever a variable occurs. … If x stands for
anything of which φ(x) is tr ue, as soon as we substitute in probability
any particular value, whose meaning we know, for x, the value of the
probability may be affected; for knowledge, which was irrelevant before, may
now become relevant. (TP, 62-6 3)
The view developed in the Treatise seems to lie behind the claim
that economics is a sphere of logic. The method of Tinbergen was merely an
object of criticism made by Keynes long before.
Thirty
years ago [1908] I used to be occupied in examining the slippery problem of
passing from statistical description to inductive generalization in the case of
simple correlation; and today in the era of multiple correlation I do not find
that in this respect practice is much improved. In case Mr Loveday or others
may nurse inductive hopes, it is worth pointing out that Professor Tinbergen
makes the least possible preparation for the inductive transition. (JMK.14,
315-316)
Keynes (1939b) shows unequivocally that an
important feature (the method of Lexis) of the Treatise was still there
in 1939. Now, we have the Keynes who accepted Ramsey’s criticism on the one
hand, and the Keynes upheld the method of Lexis on the other. Are these two
aspects to be regarded as inconsistent, or are they actually compatible?
I would suggest they might be
seen as compatible on the following grounds. Even if
“a theory of probability as
a rational belief of degree between propositions” and the justification of a
“method of induction” based on it are abandoned, it is conceivable that Keynes
maintained a critical stance on the frequency theory as a mathematical practice,
while emphasizing the utility of the method of Lexis. By doing so, there seems
to be a cleavage in logical interrelation between Part III and Part V.
5.3 Two Problems
Finally, I would suggest two problems observed
in researches into the Treatise.
The first problem is doubt about a certain
view of the Treatise and the General Theory – inter preting the Treatise
as a book arguing “uncertainty”, and seeing the kernel of the General Theory
in “uncertainty”.
In my understanding, the Treatise does not argue “uncertainty”, but sees probability as “a
rational belief of degree”, so that it belongs to a category of “rationality”.
It is made clear in a reference to the Treatise in the General Theory
(n.1, 148) that “very uncertain” is different from “very improbable”.
“Uncertainty” is put outside the framework of rational beliefs. Additionally,
to see the kernel of the General Theory
in “uncertainty” fails to take into account the important point of view that
Keynes’s theoretical system is characterised by the above-mentioned “two
contrasting phases”.
The second problem is that when we try to
grasp the relation between the Treatise
and the General Theory, we are faced with the fact that Keynes did not
write any philosophical paper after the Treatise. In terms of documents,
although it is possible to argue Keynes’s economics from a social-philosophical
point of view, it is far more difficult to argue it from a philosophical and
logical point of view. It seems that the change which occurred in Keynes’s
stance toward economics could emerge more significantly with investigation into
his economic theory and social philosophy28.
1) Cf. TP,
341-343. Incidentally, it is worth noting that Keynes carried out interesting
research on the index problem at the same time. Cf. Ke
ynes (1909).
2) Cf. TP, 126 ,166-171. For W.E. Johnson, see Keynes
(1931a).
3) See also TP, 11.
4) See also TP, 356.
5) Cf. TP, 35.
6) Cf. TP, 13-14.
7) For such an ambiguous statement, see, e.g. TP,
56, 312.
8) A similar drift is seen at TP, 56, 76
as well.
9) Cf. TP, 3.
10) Cf. TP, 242.
11) Cf. TP, 242.
12) See also TP, 267-268.
13) See also TP, 265.
14) Cf. TP, 267.
15) Immediately
after the Treatise came out, Ramsey published
a review of it (Ramsey, 1922).
16) What is
called “modern functionalist views of the mind” (Ramsey, 1996,
p. xviii).
17) Cf. Ramsey (1996, 85-86).
18) Cf. Hirai (2000, Ch.7).
19) Cf. MEB, 447. See also Braithwaite (1975, 318).
20) Ramsey’s
criticism, which appeared after 1922, and Keynes’s self- criticism vis-à-vis the
Treatise do not seem to have a direct
relation with his diffidence about the rationality of human nature. Ramsey’s
criticism is of the concept of “probability” as defined by Keynes. And Ramsey’s
alternative itself als o has (a significant role for) rationality.
21) Concerning
this point, Bell (1995) takes a critical attitude to what he sees as Keynes ’s
shift towards conservativism. For recent evaluations of “My Early Beliefs”, see
Fitzgibbons (1988), Skidelsky (1991), Carabelli (1988), O'Donnell (1989), Bell (1995),
Rosenbaum (1998).
22) These stances are expressed by Davis (1995),
Bateman(1996) and Ito h (1999) among others.
23) Davis (1995)
and Itoh (1999) consider this is the philosophy underlying the General
Theory.
24) The representatives of this view are Carabelli
(1988) and O'Donnell (1989).
25) Despite t he high level of mathematical
knowledge shown in the Treatise, it is
worth noting that Keynes hardly made any use of it in
his economics thereafter. As a key to understanding this point, criticism of
“the method of Laplace” and defense of “the method of Lexis” developed in Part
V might be considered. It is also worth trying to make sense of the fact that
Keynes made ample use of a relatively simple statistical method (for example, A
Treatise on Money (Keynes, 1930)) while criticizing Tinbergen’s method.
26) For a case of a scholar taking this view,
see Bateman (1996).
27) Cf. Keynes’s letter to Harrod dated July
4, 1938 (JMK.14, 296-297).
28) My answer to this point is Hirai (2003
Ch.15; 2007 Ch.13).
References
(JMK
indicates The Collected Writings of John Maynard Keynes, London: Macmillan.
The numbers following indicate volumes.)
Bateman, B. 1996. Keynes's
Uncertain Revolution, Ann Arbor: University of Michigan Press.
Bateman, B. and Davis, J. eds. 1991, Keynes and Philosophy: Essays on the Origin of
Keynes’s Thought, Aldershot: Edward Elgar.
Bell, Q. 1995. Elders and Betters, London: Murray.
Braithwaite, R.B. 1975. Keynes as a Philosopher in Essays on John
Maynard Keynes , edited by M. Keynes, Cambridge: Cambridge University Press.
Braithwaite, R.B. December 1972. Editorial
Foreword to TP (JMK.8, xv-xxii).
Carabelli, A.1988. On Keynes's Method,
London: Macmillan.
Coates, J.
1996. The Claims of Common Sense, Cambridge: Cambridge University Press.
Davis, J.
1995. Keynes's Philosophical Development, Cambridge: Cambridge
University Press.
Fitzgibbons, A. 1988. Keynes’s Vision, Oxford: Clarendon
Press.
Hirai, T. 2007, Keynes’s Theoretical Development
– from the Tract to the General Theory, Routledge.
Keynes, J.M. 1909. The Method of Index Numbers with Special Reference to
the Measurement of General Exchange Value, in JMK.11, 49-173.
Keynes, J.M. 1921. A Treatise on Probability,
London: Macmillan (JMK.8).
Keynes, J.M. 8
and 15 August, 1925. Am I a Liberal?, The Nation and Athenaeum in JMK.9,
295-306.
Keynes, J.M. 1926. The End of Laissez-Faire,
Hogarth Press in JMK.9, 272-294.
Keynes, J.M. 1930. A Treatise on Money, I ＆II, London: Macmillan
(JMK.5 & 6).
Keynes, J.M. 1931. Essays in Persuasion,
London: Macmillan (JMK.9).
Keynes, J.M. 1931a. W.E. Johnson, The Times, in
JMK.10, 349-350.
Keynes, J.M. 1931b. Ramsey as a Philosopher, The New Statesman and
Nation in JMK.10.
Keynes, J.M. 1933. Essays in Biography,
London: Macmillan (JMK.10).
Keynes, J.M.1936. The General Theory of
Employment, Interest and Money, London: Macmillan (JMK.7).
Keynes, J.M.
February 1937. The General Theory of Employment, Quarterly Journal of
Economics, in JMK.14, 109-123.
Keynes, J.M.
1939a. My Early Beliefs (in Two Memoirs, Rupert Hart-Davis, 1949) in JMK.10,
433-450.
Keynes, J.M. 1941 Will, 14 February, King's College.
Keynes, J.M.
September 1939b. Tinbergen's Method, Economic Journal in JMK.14,
318-320 (Related material is reproduced in JMK.14, 285-318).
Moore, G. 1903. Principia Ethica,
Cambridge: Cambridge University Press.
O'Donnell, R.
1989. Keynes: Philosophy, Economics and Politics, London: Macmillan.
Ramsey, F. January 1922. Mr Keynes on Probability, The
Cambridge Magazine.
Ramsey, F. 1926. Truth and Probability in Ramsey (1990,
75-133).
Ramsey, F. 1931. The Foundations of
Mathematics, London: Kegan Paul.
Ramsey, F. 1990. Philosophical Papers, Cambridge: Cambridge
University Press, edited by D.H. Mellor.
Rosenbaum, S.P. 1998. Aspects of Bloomsbury,
London: Macmillan.
Rosenbaum, S.P. 1995. The Bloomsbury Group, Toronto: University
of Toronto Press.
Russell, B. 1903. Principles of Mathematics, Cambridge: Cambridge
University Press.
Shionoya, Y. 1991, Sidgwick, Moore and Keynes: a
Philosophical Analysis of
Keynes's 'My Early Beliefs' (in
Bateman and Davis, eds.).
Skidelsky, R. 1992. John Maynard Keynes,
London: Macmillan.
Tinbergen, J. 1939. A Method and Its Application to Investment
Activity, League of Nations.
Whitehead,
A.N. and Russell, B., 1910. Principia Mathematica, Cambridge: Cambridge
University Press.
K. Itoh, 1999. Keynes’s Philosophy,
Iwanami Shoten (in Japanese).
T. Hirai, 2000.
Keynes, Schumpeter, Hayek - Searching for the Vision of the Market
Society, Minerva Shobo (in Japanese).
T. Hirai, 2003. Looking at Keynes’s
Economics from Multiple Points of View, University of Tokyo Press (in
Japanese).
***
1. Introduction
2. The Stance of A
Treatise on Probability
2.1 Definition of
Probability
2.2 Question of Whether
Probability Is Objective or Subjective
2.3 Sphere under
Consideration
2.4 Argument for
Justification of Induction
3. After the Treatise
― Influences from Ramsey’s Criticism
3.1 Ramsey’s Critici sm
3.2 Keynes’s Response
4. On “My Early Beliefs”
5. Keynes as “a Philosopher” and as “an Economist”
5.1 The General Theory
5.2 Keynes’s View of the
Nature of Economics
5.3 Two Problems
