John Maynard Keynes was born in Cambridge on 5 June 1883 and died in Tilton, Sussex on 21 April 1946. His father, John Neville Keynes , was a lecturer in logic and, later, university registrary, so the young Maynard imbibed Cantabrigian academia, dining with his father’s philosophy friends such W.E. Johnson . Maynard was educated at Eton and, in 1902, went up to King’s College, Cambridge. Brilliant at school, brilliant at university, he subsequently entered the India Office of the British Civil Service, then secured a fellowship at King’s, and in 1924 became Bursar (a position he held until his death), significantly enhancing the college’s wealth and intellectual reputation. From 1908 to 1915 he was university lecturer in economics. He was elected fellow of the British Academy in 1929 and, through his pioneering work in economics, became the eponymous creator and hero of Keynesianism. Much of his life was spent at the British Treasury: he was its Principal Representative at the 1919 Paris Peace Conference, virtually the de facto Chancellor of the Exchequer during World War II and key economic figure at the Bretton Woods Conference and Washington Negotiations. Arguably the greatest economist since Adam Smith, he secured worldwide fame. In 1942 he received a barony – becoming Lord Keynes of Tilton – a Cambridge Honorary Doctorate of Science, fellowship of the Royal Society and, but for his death, would have received the Order of Merit.
Keynes was a man of many worlds, a polymath, walking, seemingly effortlessly, between the bohemian Bloomsbury Group and the Civil Service establishment, between homosexuality and heterosexuality – he praised homosexual sodomy, yet married a Russian ballerina – and between the highest international negotiations and weekend Cambridge college meetings. Founder of the Cambridge Arts Theatre, he transformed the British Arts Council, collected nineteenth and twentieth-century French and English paintings and was a serious book collector, identifying (with Piero Sraffa) Hume as the author of the anonymously written Abstract of Hume’s Treatise. He was a fine essayist, yet also enjoyed a day’s pig farming. In philosophy he encouraged Braithwaite and Ramsey , funded Wittgenstein and drew Sraffa to Cambridge, who proved a catalyst for Wittgenstein’s later work. Bertrand Russell spoke of Keynes’s intellect being the sharpest and clearest that he (Russell) had ever known; and, for many, Keynes was the ablest Englishman of his generation. Although initially trained in mathematics and later securing, by far, his greatest influence as an economist – his collected works, mainly of economics, amount to thirty volumes – he wrote his fellowship dissertation in philosophy, philosophy and economics being then grouped within the moral sciences. This dissertation developed into his one philosophical work, A Treatise on Probability (1921) though miscellaneous philosophy notes exist in King’s College’s archive and some philosophical reflections occur in his essay on Ramsey (in Essays in Biography’, 1933) and his Two Memoirs (1949).
Keynes’s Treatise is a work of mathematical theorems, yet also of clear English prose, ranging over chance, statistics and induction, using examples from lotteries, astrology and other minds, dealing with, amongst others, Bayes, Bernoulli, Bertrand and Boole. In contrast with Moore , who hoped to make vague notions precise – in contrast with forced regimentations sometimes in Russell – Keynes explicitly valued the English of Hume. He sought to avoid wet clouds of doubt, as he termed them, in order to give his arguments a chance, acknowledging a pretence to more conviction than sometimes felt. Few ended up accepting Keynes’s probability results, though Russell acknowledged indebtedness and the Treatise’s original and clear presentation of the logical relation probability theory, with a valuable history and substantial historical bibliography, proved a stimulus to many. Ramsey, reacting against Keynes’s theory, gave birth to subjective probability. Later Rudolph Carnap defended a probability akin to Keynes’s, Carnap’s probability also being an objective logical concept. Paradoxically, Carnap measures the confirmation degrees using ideas from Ramsey’s work critical of Keynes.
Keynes’s work on probability was influenced by Johnson, by way of some theorems, concepts and notation. Curious as it might seem, another influence was Moore’s Principia ethica, the bible of the Bloomsbury Group and Cambridge Apostles (the well-known ‘secret’ society). While accepting Moore’s ideal – the ideals of passionate contemplation of, and communion with, beloved persons and objects of beauty – Keynes rejected his scepticism of the possibility of rational belief in any action being even probably right, given that moral status depended on unknown consequences. Moore’s demand for empirical evidence regarding the probable future was seen as a mistaken demand for causal certainty. Keynes argued that we could have degrees of rational belief concerning the future as common sense suggests – and his Treatise set out to show this. While his spirit was high in the Moorean abstract heavenly ideal, his pen was scribbling out practical examples of probable judgements – examples as diverse as the reinsurance rates for the Waratah, a vessel that sank in South African waters, the probable loss to a lady beauty contestant who missed the final viewing, and the legal tale of Cyllene, a racehorse, who should have serviced one of the plaintiff’s brood mares: now what was the probability of a resultant horse – and, if such result there be, of its being a winner?
Logical entailment sets the scene: we can perceive that certain conclusions logically follow from sets of premises. So too, argues Keynes, we can perceive unanalysable relationships between propositions weaker than entailment, namely partial entailments or probability relationships. In this way Keynes broadens logical relations to include probability, establishing logical connections as the proper matter of probability theory. The limiting cases of these partial entailments are one and zero, one, when conclusions logically follow from the propositions taken as premises, and zero, when they contradict the premises.
According to Moore, the intrinsically valuable was open to direct inspection: he and his disciples would, for example, try to see whether a violent short love affair was better than a longer tepid one. All this was considered as scientific and rational as investigating the material of our senses or seeing the validity of modus ponendo ponens. Keynes proclaims such a direct apprehension of some probability relationships. These relationships objectively hold, but probability judgements are subjective in that they depend upon which sets of propositions the judgers know, sets upon which to base their probable conclusions. Given the known propositions as premises, judgers might err over what they conclude probable, and to what degree probable; their beliefs might hence be irrational and akin to those based on invalid arguments with true premises. A rational belief in p as having a certain probability needs to be grounded in knowledge of a set of propositions h and some secondary proposition q asserting a probability relation between p and h. Although Keynes equivocates a little between perfect logical insight and rational human insight, his prominent claim is that if the probability relation is of degree a, then the rational belief in p is of degree a.
Thus it was that Keynes fitted probability into Russell’s and Moore’s empiricism and realism, showing how degrees of belief could be rational and independent of the believer’s psychology. Keynes stresses, however, that probability relations between propositions with different subject-matter are rarely commensurable; indeed, probability relations are often incapable even in principle of any numerical measurement at all. While it is true that under-writers can be persuaded to give numerical values for virtually any risk – values backed with hard cash – the values often contain some caprice. A problem appears (and remains) on stage – of how the unmeasurable probability relations help to identify degrees of rational belief, if such rationality there be. Also arise fundamental questions of the nature of the correspondence even when some measure can be given both to the probability relation and to, given h, the partial belief in p. Further, there is the matter of the relationship between that partial belief that p and the conditional belief that if h, then p, David Lewis showing, decades after Keynes, that, on reasonable assumptions, a conditional probability typically lacks the same value as the probability of the corresponding conditional proposition’s being true.
Measurement of probabilities is possible only when we have sets of possibilities that are mutually exclusive, equi-possible and collectively exhaustive. Keynes, although mindful of previous paradoxes, rehabilitates a Principle of Indifference as the criterion for identifying equipossibles: the alternatives must fall under a single determinable; they must have the same degree of determinateness, with the evidence being symmetrically related to all in every relevant respect. Of course, there is now the difficulty of identifying what counts as ‘same degree’ and ‘relevance’; and, but for the rickety rescue raft of Keynes’s direct judgement, we should end up explaining degrees of rational belief in terms of partial entailments which are explained in terms of equi-possibles which are probably to be explained by degrees of rational belief.
The main contemporary rival to Keynes’s a priori logical theory was the empirical frequency theory, according to which an event’s probability is the proportion of instances of relevantly similar events to some class of which the event is a member. This theory has problems even with the favoured cases of statistical hypotheses, though – as Keynes notes Venn, a leading proponent of the frequency theory, does at least in the main restrict his theory to the favoured statistical hypotheses. Keynes himself seeks comprehensiveness; he thus falls prey to the criticism that even if his theory explains how some evidential probabilities can justify degrees of rational belief, it cannot account for those probabilities that seem to be empirical facts, such as the 50 per cent probability that a radium atom will disintegrate within so many years.
Where induction is concerned – reasoning from all observed cases of F being G to everything F being G – Keynes stresses that the conclusions are not that things simply are so or must be so, but that, relative to evidence, there are probabilities in their favour. Such probability conclusions can be justified. That the probable fails to occur need indicate no faulty reasoning; and Keynes rejects the ridicule heaped upon ruder races’ reasonings, pointing out that their probability assessments might often have been justified given the known evidence. Of course, it can be sensible to seek further evidence. Suppose all observed instances of F have been G: observing further instances of F as also G can be inductively relevant analogically, that is as a means of eliminating other features that might be necessary for something F to be G. Further instances help positively to support everything F being G only if certain conditions hold, for example that the hypothesis possesses a certain antecedent probability prior to observation. This leads Keynes to postulate a finite a priori probability of some limitation of independent variety. Postulate we might indeed, but arguably we have no good reason to do so. Keynes reminds us, though, that experiences have increased the probability of the limitation holding, and such a posteriori support, he proposes, need not collapse because of circularity. Keynes registers the doubt-raising complexities of how we group characteristics and the assumptions of underlying causal generators of those characteristics, complexities given radical boost by Nelson Goodman’s ‘grue’ discussions in the 1950s.
Ramsey, as perceptive an observer of logical relations as any, confessed an inability to spot the objective probability relations between propositions – and also questioned Keynes’s justification of induction. Keynes seemingly accepted Ramsey’s demolition of his probability theory, though debate continues concerning the extent of Keynes’s acceptance. There is also debate over the extent to which Keynes’s approach to probability affected his economic theorizing concerning risk, uncertainty and expectations. What undoubtedly continued to concern him was the justification, one way or another, of the rationality of degrees of belief and of resultant actions; he remained dissatisfied with justifications that rested solely on their evolutionary success.
In contrast to Keynes’s Treatise and his writings in economics, Moore’s ideal – a timeless unworldly ecstasy – possessed, as Keynes later quipped, less place for action than the life of St Francis of Assisi, who at least made collections for the birds. True, Moore had written of social duties, but he and his disciples largely ignored them, feeling that they interrupted engagement with their private ideal. Keynes, however, lived in both the social world of duty and the private ideal; he was not, so to speak, merely the investment adviser to the Bloomsbury Group and his college, but to the world. His interventionist advice and cautious advocacy of social reform, his promotion of the arts and education – all being designed to increase the quantity of valuable things in the life of the nation – resulted from ethical springs,
As Keynes later wrote in ‘My Early Beliefs’, he and the other Apostles were, at the turn of the twentieth century, water-spiders gracefully skimming a stream, in little contact with the eddies and currents beneath. The Apostles’ values transcended mere Benthamite utilitarian pleasures – though Keynes and others were not immune to fleshing them out with the carnal. Keynes emphasized the need for individual judgement rather than rule following and held to this for himself and the educated elite. He became increasingly aware that he attributed an unreal rationality to human nature, one as unreal as the meliorists’ ungrounded optimism. As his economic work developed, he heeded human nature’s under-currents, supplementing his love of reason with support for spontaneity, irrationality and life’s changing patterns. His economics began to deal with such animal spirits, as he termed them, recognizing how a few individuals’ behaviour could generate waves of group optimism and pessimism.
Throughout his life Keynes was wayward and considered something of a rebel: he dismissed religion as hocus-pocus and challenged customary morality – yet he became so important a national figure that his funeral service was held in Westminster Abbey. He allowed his thinking to evolve and, when criticized for his mind’s changes, responded, ‘When the facts change, I change my mind. What do you do, sir?’ A high value was placed on reason throughout his life; and while he certainly looked to consequences in assessing right and wrong, he insisted, on rational grounds, that we should be very chary of the moral risk in making sacrifices now for uncertain distant advantages. His assessments of what should be done urged a reasoned disregard for much of the long run – for, in his well-known aphorism, in the long run we are all dead.