R. G. Collingwood, An Essay on Metaphysics, Oxford, at the Clarendon Press, 1940.

Collingwood was concerned by attempts to eliminate metaphysics, by which he meant critical analysis of the philosophical assumptions that scientists and everyone else use as they go about their business of systematic thinking. He argued that if we lose the capacity to subject these assumptions to analysis then systematic thinking and hence science and civilisation will become impossible.

‘The sciences, both natural and historical, are at present in a flourishing condition. By means of heroic efforts they have succeeded in disentangling themselves from the fallacies of method that vitiated much of their apparent progress in the nineteenth century. Their prospects of advance along the lines upon which they have now established themselves are incalcuable. Internally, they have nothing to fear. The only dangers that now beset them are external'.

Defective metaphysics, and tragically deluded anti-metaphysicians are on the march.
'This is my reason for offering to the public what might seem essentially an academic essay, suitable only for readers who are already, like myself, committed to an interest in metaphysics. The fate of European science and European civilisation is at stake. The gravity of the peril lies especially in the fact that so few recognize any peril to exist. When Rome was in danger, it was the cackling of the sacred geese that saved the Capitol. I am only a professorial goose, consecrated with a cap and gown and fed at a college table; but cackling is my job and cackle I will'.

Collingwood framed his discussion in terms of the logic of question and answer, whereby “Every statement that anybody ever makes is made in answer to a question.” He then offers a number of propositions.

1. A question is logically prior to its own answer.

2. Every question involves a presupposition (and usually a large number of them).

3. The logical efficacy of a supposition does not depend upon the truth of what is supposed, or even on its being thought true, but only on its being supposed.

4. A presupposition is either relative or absolute.

He used the example of checking the accuracy of his surveyors' tape to show the attitude of a person who realises that the tape will tend to stretch in the course of use, in contrast with a person who uncritically assumes that the length is an 'absolute presupposition'.

"The business of logical inquiries, like that on which we are now engaged, is to study high-grade or scientific thinking: their conclusions are not impaired by the fact that low-grade or unscientific thinking also exists"

"...to speak of verifying a presupposition involves supposing that it is a relative presupposition" because absolute presuppositions do not lend themselves to verification or criticism.

Somewhere he wrote that absolute presuppositions do not raise the question of verification otherwise they would be relative and not absolute presuppositions.

Proposition 5. Absolute presuppositions are not propositions.

The categories of truth and falsehood do not apply to absolute presuppositions.

Nor are absolute presuppositions ever actually propounded;

"To be propounded is not their business; their business is to be presupposed. The scientist's business is not to propound them but only to presuppose them. The metaphysician's business, as we shall see, is not to propound them but to propound the proposition that this or that one of them is presupposed".

V The Science of Absolute Presuppositions (APs)

Introspection is not enough to get over the problems of defective Metaphysics because improved methods are required, which he calls ANALYSIS.

Scientists are likely to violently resist an invitation to examine their absolute presuppositions; here, in Collingwood's scheme, we encounter the limits of criticism (or justification); "APs do not need justification". Instead they need to be scientifically described.
However they tend to provoke "numinous terror" and so are hard to handle in an age that claims (falsely) to have no superstitions, hence no methods to dissipate the terror and face up to the source of it.

In the event of the eradication of metaphysics "...the eradication of science and civilisation will be accomplished at the same time". For more on this see chapter XIII.

VII The Reform of Metaphysics

Metaphysics will be reformed by consciously making use of the historical method, not the 'scissors and paste' history that Collingwood deplores. The proper historical approach will widen the scope of the subject and bring many new problems into view. The scope will widen to accommodate the absolute presuppositions of all eras, giving an infinite number of constellations of presuppositions.

We can then compare different constellations, and at the most interesting stage of all, "...find out on what occasions and by what processes one of them has turned into another".


VIII What Anti-Metaphysics Is

Science and metaphysics are bound together and it makes no sense to suggest that knowledge will be advanced by the abolition of metaphysics. He is concerned to investigate the springs of anti-metaphysics, apart from defects in the programme and practice of people who explicitly concern themselves with metaphysics.

1. Progressive Anti-Metaphysics. If the work of professional metaphysicians becomes out of date and ceases to contribute to ordinary, non-metaphysical thought, then people such as scientists may come to think of metaphysics as a hindrance rather than a help.

2. Reactionary Anti-Metaphysics. People engaged in some kind of ordinary thought may be threatened by advances in metaphysics and so demand its abolition as a defence mechanism.

3. Irrational Anti-Metaphysics. There may be a more complex case where there is a move to abolish science itself, aiming at the elimination of systematic and orderly thinking in every form. In a civilisation where for some centuries rational thinking has been valued, such a movement would need to conceal its ultimate aim and operate under a disguise. Such a disguise may well be afforded by any existing movements of types 1 and 2, and better still by a confused mixture of both.

Progressive Anti-Metaphysics.

It is absolutely necessary that the work of metaphysics should be done for the sake of science and all other activities, however 'practical' and removed from academic discourse they may be.
Work on metaphysics can never be completed; it needs to be done piece by piece when the need arises.

The work of metaphysics is to ascertain whether certain presuppositions recently brought to light are absolute or relative. If the professionals do not keep up, then amateurs, including scientists, will have to do it for themselves but they are likely to resent the inactivity of the professionals.

'This resentment makes him think of his own metaphysical work not as a contribution to metaphysics but as an exposure of metaphysics; not as piece of help given by himself to metaphysicians in the pursuit of their own proper business, but as an attack on them and their business too'.

An example of progressive anti-metaphysics was Newton's warning against metaphysical hypotheses in experimental philosophy; this was done to forestall criticism of his own metaphysics which was an integral part of his own physics.

Reactionary Anti-Metaphysics

If professional thinking in metaphysics gets ahead of ordinary thought then the 'amateurs' are likely to take a stand against metaphysics at large. This is really a defence of outdated metaphysics. It is not really opposed to metaphysics (though it appears to be so);it is really opposed to advances in science and associated advances in metaphysics).

Collingwood offered an account of the rise of this attitude in the nineteenth century. He described this century as an age of metaphysical reaction where progress consisted of working out the consequences of ideas from the previous century with fundamental criticism and thus fundamental progress ruled out. For example James Watt's steam engine was not really improved for almost a century after 1796; people simply made bigger and more powerful engines at the same level of 7% efficiency.

Similarly Locke's ideas on the parliamentary system were simply reproduced without critical revision, especially without attention to the presupposition of nationality which persisted as an absolute when it should have been exposed as a relative (dispensable) presupposition.

Hence 'The things that were done in the nineteenth century in the name of nationality, the things that are still done today, at what expense of life and property I shall not try to estimate, are done for the sake of an eighteenth-century "metaphysical" idea'.

IX Psychology as Anti-Metaphysics

Collingwood disputed the claim that psychology is the subject that really tells us how we think, and so 'owns' metaphysics.

X Psychology as the Science of Feeling

Collingwood suggested that psychology came into being to investigate feeling which was not covered by the existing sciences of bodies or of minds. The sciences of mind such as logic and ethics he labels normative or 'criteriological' because they are self-critical and very much concerned with locating and correcting error. In contrast feeling has no element of self-criticism and so a science of feeling has nothing to offer the criteriological sciences.

XI Psychology as the Pseudo-Science of Thought

Collingwood suggested that psychology set out to capture the whole of the business formerly done by the sciences of thought. The critical period was the eighteenth century, an age of fanatical revolt, "a Rousseauesque age of nostalgia for the forest, a Humian age of reason as the slave of the passions, a Wordsworthian age whose prophets told their disciples to close up the barren leaves of science and of art".

He postulated that the great destructive task of the century was to liquidate Europe's debt to Greece; this was manifest in religious innovations and in the attack on biology which was the last refuge of teleological natural science (the science of the organism). Under this impulse biology was 'materialised' by the doctrine that organisms are nothing but complexes of material particles which operate according to mechanical principles. On this, compare Barzun’s critique of the heritage of Darwin, Marx and Wagner.

In the theory of knowledge the revolt took the form of maintaining that thinking is merely the aggregation of feelings, and thus a special case of sensation and emotion. In his view, if the new wave merely swept away the tyranny of ancient error then all might have been well but its own programme was wrong.

'When it is a case of overthrowing tyranny one should not be squeamish about the choice of weapons. But the tyranicide's dagger is not the best instrument for governing the people it has liberated. Epistemological materialism, in attacking the criteriological science of logic and offering to replace it by psychology, deliberately proceeded on the assumption that thought did not possess that power of self-criticism which had in the past been rightly regarded as distinguishing it from feeling'.

XIII The Propaganda of Irrationalism

This chapter depicts with chilling accuracy the process that occurs in many courses where the critical faculties of students are systematically destroyed. He first asks us to picture a civilisation where respect for truth is a powerful belief and systematic thinking is prized in intellectual and practical pursuits. Each feature of this civilisation would have characteristics derived from that prevailing habit of mind.

'Religion would be predominantly a worship of truth...Philosophy would be predominantly an exposition not merely of the nature of thought, action & etc. but of scientific thought and orderly (principled, thought-out) action, with special attention to method and to the problem of establishing standards by which on reflection truth can be distinguished from falsehood. Politics would be predominantly the attempt to build up a common life by the methods of reason (free discussion, public criticism). Education would be predominantly a method for inducing habits of orderly and systematic thinking'. And so on.

'And suppose that now within this same civilisation a movement grew up hostile to these fundamental principles...an epidemic disease: a kind of epidemic withering of belief in the importance of truth and in the obligation to think and act in a systematic and methodical way. Such an irrationalist epidemic infecting religion would turn it from a worship of truth to a worship of emotion and a cultivation of certain emotional states...Infecting politics it would substitute for the ideal of orderly thinking in that field the ideal of tangled, immediate, emotional thinking; for the idea of a political thinker as a political leader the idea of a leader focussing and personifying the mass emotions of his community'.

This movement of thought would need to proceed by stealth because the healthy tissues of thought would strongly resist any open attack on the springs of rationality and scientific thinking.

'Let a sufficient number of men whose intellectual respectability is vouched for by their academic position pay sufficient lip-service to the ideals of scientific method, and they will be allowed to teach by example whatever kind of anti-science they like, even if this involves a hardly disguised breach with all the accepted canons of scientific method.'

'The ease with which this can be done will be much greater if it is done in an academic society where scientific specialisation is so taken for granted that no one dare criticise the work of a man in another faculty. In that case all that is necessary to ensure immunity for the irrationalist agents is that they should put forward their propaganda under the pretence that it is itself a special science, which therefore other scientists will understand that they must not criticise'.

Collingwood was concerned about the impact of psychology at that time (pre-1940). Lately sociology has taken over the function he describes.

XIV Positivist Metaphysics

Early positivists such as John Stuart Mill were too interested in the validity of scientific thought to join psychology in its attacks on logic. But they were too easily satisfied as to the criteria of scientific validity. They perceived two stages in scientific inquiry, first ascertain facts, then classify them.

Their account of sensation in the ascertaining of facts failed to take account of the theoretical perspective required to make the process effective – Collingwood, perhaps unhelpfully, called this the “historical nature of facts”.

“In the second place it was rash of the positivists to maintain that every notion is a class of observable facts. This amounted to saying, what in fact positivists have always tried more or less consistently to say, that scientific thought has no presuppositions.”

And so the positivists became proponents of a certain type of metaphysics while they maintained either that metaphysics did not exist, or was impossible, or should be swept away.

XV A Positivist Misinterpretation of Plato

This chapter is a digression on some Oxford Platonists who distorted Plato's thought to make it mean the opposite of his intention and to agree with John Stuart Mill. He commented, as an aside, on the lengths to which commentators will go to make their author talk good sense (as they see it) "which after all is nothing compared with the length to which a hostile reader will go in order to satisfy the desire to make out that his author is talking nonsense, especially if his hostility is not quite openly acknowledged, even to himself".

XVI Suicide of Positivist Metaphysics

Mr Ayer is selected as an example of an anti-metaphysician who has no understanding of the nature and function of presuppositions, with the result that he attacks psuedo-metaphysics instead of the real thing. He mistakes suppositions for propositions and inquires after their truth or verifiability instead of looking at the problem situation to find the function that they serve.

Ethics and theology are singled out as specially pressing cases of the metaphysics that is to be eliminated but the positivists' notions of ethics and theology bear little relation to the subjects as practiced by the greats. They possibly derive from foolish parents or nurses and are coloured by feelings of resentment against the humiliations of childhood in "what is called a virtuous and religious home".

The positivists see metaphysics as threatening to science, and they see criticisms of positivist science as part of an attempt by metaphysics to destroy science. But in so far as metaphysics is concerned with the absolute presuppositions of science, and attempts to find out what these are at a given time, then the positivist attack can only damage science "in proportion to its own degree of success...Any attack on metaphysics is an attack on the foundations of science; any attack on the foundations of science is an attack on science itself".

“This is the real danger of the 'logical positivist' attack on metaphysics, and may conceivably be the real motive behind it. It is a mass of inconsistencies and confusions; but one thing which it may very well serve to express is an aspiration towards the destruction of science and its supersession by the most recklessly unscientific kind of thought”.



He is concerned with a question about the history of thought in the fourth century AD, and especially with the Christian belief in God at that time. He suggests that civilisation and science require a special metaphysical framework and suitable institutions to perpetuate this framework. If the special absolute presuppositions are lost there is no way that we could be sure of recreating them. If they persist then science could be fairly swiftly rebuilt even if all records of its achievements were lost.

'The guardianship of the European "scientific frame of mind" is vested in the religious institutions of European civilisation'.

He suggests that the invention of monotheism was decisive for science because it made possible the idea of 'one world' which hangs together under the same set of principles; hence the disciplines can relate to each other instead of describing different worlds that are not penetrable by rival specialists.

“In a polymorphic science there is no sense in calling one science nearly or distantly akin to another. They are all just different. If anybody after a training in one science began to study another, his previous training would be valueless; he would have to start again at the beginning”.

Thales initiated a reform movement towards a monotheistic science and Aristotle's Metaphysics is the high-water mark of this reformation. However his own metaphysics fell into error, especially his belief that merely by using our senses we learn that a natural world exists; he mistook a presupposition for a fact. According to Collingwood this error was corrected by Christianity, by the Patrists who formulated the metaphysics of the 'Catholic Faith' which provided the main or fundamental presuppositions of natural science ever since. This did not happen in time to save Rome which was weakened internally by metaphysical error.

In speaking of the Holy Trinity in respect of the natural sciences, they meant, by belief in the Father, that there is one world of nature. By believing in the Son they meant that this world is a multiplicity of natural realms. “By believing in the Holy Ghost they meant absolutely presupposing that the world of nature, throughout its entire fabric, is a world not merely of things but of events or movements”.

Collingwood claims that the presuppositions that make up the 'Catholic Faith' have provided the metaphysical background for Western thought, including scientific thought, from the fourth century to the present day. It appears that one could subscribe to the metaphysics without having to swallow the theology as well, and so a task for modern scholars is to pick the various strands of metaphysics and theology apart.


XXII Kant's Problem and the Problem of Today

Kant's view of metaphysics was:
(1) It is concerned with God, freedom and immortality.
(2) It is an inquiry that cannot be avoided by people of integrity because we are bound to make assumptions about these things; the question is whether our assumptions are good or bad.
(3) The people of his time who attacked metaphysics were not friends of science because they really attacked the foundations of all knowledge.
(4) Metaphysicians were to blame for this because they misunderstood both their problems and their methods.
(5) Better metaphysics could not be found until errors were cleared away.

XXIII Metaphysics and Critical Philosophy

Kant wrote the Critique of Pure Reason to review the problem situation in metaphysics “very much as a man might approach any difficult problem in any kind of science by reviewing the more important attempts to solve it; asking how far they had succeeded and why they had not succeeded more completely; considering the state in which they had left the problem; and in general making quite sure he knew what the problem was that he wanted to solve, and upon what conditions he would be able to claim in the future that he had solved it”.

238 Looking at maths and physics, Kant decided that they had progressed when they got past the point of making measurements and observations and got to the stage of "arguing out the consequences of assumptions" (in maths) and "asking questions and demanding answers to them" (in science).

A similarly systematic approach is needed in metaphysics.

Popper described Kant's philosophy as an attempt to account for the phenomenal success of Newtonian physics which had apparently found 'the Truth'. Kant's quest was to explain how such final Truth could be found.

XXIV Metaphysics as Transcendental Analytics

Kant's transcendental analytics was a historical study of the absolute presuppositions used by natural scientists from Galileo to his own time. This limited frame was a characteristic of his time, even of the keenest students of history such as Voltaire and Hume. In the following four chapters Collingwood sets out to show how Kant's Transcendental Analytics can be read as a history of the absolute presuppositions of natural science from Galileo to Kant himself.

XXV Axioms of Intuition

Collingwood thought that Kant was quite wrong in his analysis due to a lack of historical perspective. Kant did not realise the limitation of treating the metaphysics of his own time as the ultimate or the only possible metaphysics (possibly due to his belief that Newton had come up with the ultimate theory in physics).

Kant grouped his subject matter under four heads of which the first is 'axioms of intuition'. These are a set of presuppositions based on the idea that natural science can be regarded as applied mathematics.
The impact of Newton is clearly apparent in this assumption, though Collingwood points out that this is a view shared by Pythagoras and Plato - i.e. all science is mathematical.

This principle is at odds with the Aristotelian metaphysics which sees the world as a world of 'qualities' rather than numbers.

'The God of Plato "always geometrizes". The natural science of the Renaissance represents a Pythagorian-Platonist reaction against the Aristotelianism of the Middle Ages; and Galileo's pronouncement that the book of nature is written by God in the language of mathematics was a deliberate echo of Plato's and a declaration of war on the Aristotelians'.

Collingwood considered that this revival of Platonism in a world that had gone Aristotelian requires explanation. The revival was not complete, with Aristotelian principles holding out among biologists and naturalists in spite of eighteenth century materialism (and reductionism). Also the revival is not pure, because genuine Platonism did not permit precision in the world of real (changeable) things.

Christianity changed this by making it a matter of faith that the world of nature should be regarded as a realm of precision, and so amenable to mathematical description. Galileo's principle still holds (1940) with scientists believing that their field is more scientific the more it approximates to applied mathematics. Though there is a symbol of revolt with the idea that laws of nature are 'statistical' and not absolute.

“The biological controversy between ‘mechanists’ and ‘vitalists’ is in effect a struggle between Platonists and Aristotelians, friends of applied mathematics and friends of quality, for the last citadel of Aristotelian natural science. To the mere spectator, there seems to be evidence that the "mechanists" are winning”.



XXIX Three Senses of the Word 'Cause'

I The free and deliberate act of a conscious and responsible agent.
II A state of affairs which we can manipulate to make things happen (or prevent things from happening).
III A one-one linkage of causal priority between events.
Type I is the historical sense of causation where people and their decisions are the focus of interest. Type II applies in the world of practical affairs and technology, of engineering and medicine.

Type III applies in the theoretical sciences of nature, and it is here that various forms of confusion arise, unlike the case with I and II which Collingwood suggests are straightforward and easy to understand. It may be that things have gone downhill since he wrote because the notion of free and deliberate acts by responsible agents has been eroded by various forms of historical and psychological determinism.

As he suggested in chapter VII the scientific notion of causation (and hence the whole structure of science) is under strain due to incompatible elements that go to make up the definition of type III causation.

XXX Causation in History

XXXI Causation in Practical Natural Science

Type II causation has elements of I and III about it because it involved deliberate choices (type I) and it also makes use of type III regularities.

XXXII Causation in Theoretical Natural Sciences

This type of causation is independent of human agents and their perceptions or wishes. A necessary cause of type III produces its effect "no matter what else exists or does not exist...the relation between cause and effect is a one-one relation. There can be no relativity of causes, and no diversity of effects due to fulfillment or non-fulfillment of conditions".

This he described as a tight sense of causation in contrast with the loose causation of type II. For example if he sets fire to a time fuse and in five minutes there occurs an explosion, lighting the fuse is a loose cause of the explosion because intermediate steps are required (the burning of the fuse).
There is another form of tightness involved; the one-one relation must be simultaneous, so there is no difference in time between cause and effect.

He put the question; what does it really mean to talk about causes producing effects in the tight sense?

The rationalist and empiricist answers.

(i) The rationalist says "necessitation means implication". A cause is a ground and the effect stands in a logical relation to it. But this breaks down as a programme because the causal propositions advanced by modern science cannot be obtained by pure thought from first principles. Hence the need for observation and experiment.

(ii) The empiricist says "necessitation means observed uniformity of conjunction", as described by Hume.

(iii) Earl Russell (in an account which Collingwood greatly admires) says "necessary is a predicate of a propositional function meaning that it is true for all possible values of its argument or arguments". Collingwood calls this a "functional" theory and rejects it on two grounds, first it implies the rationalist or implicational theory that he has rejected, second he is not happy with Russell's use of 'necessary'. This may become clearer when we pursue Collingwood's thoughts on anthropomorphism in physics.

Collingwood suggests that when we talk about causation in a tight sense we mean more than logical implication and more than uniformity of conjunction; this something "is in the nature of compulsion" (which is certainly implied in talk of natural laws, especially in so far as these are different from man-made laws that can be broken).

Compulsion is something that we learn about in our social life and we tend to apply this notion in a metaphorical way to our relations with natural things, and to the relations that they have among themselves.

He suggested that there is an anthropomorphic tradition in natural science that arose from animistic theories of nature. The animistic concept of causation persisted when mechanism replaced animism in modern physics, so Newton used a whole vocabulary which ascribed to 'causes' a kind of power that properly applies to one human inducing another to act in a certain kind of way.
Modern physics has developed the Newtonian doctrine to eliminate causes altogether; so of the two Newtonian classes of events:

(a) those that happen according to law and (b) those that happen as the result of causes, class (a) has swallowed (b) entirely.

XXXIII Causation in Kantian Philosophy

Collingwood claimed that Kant over-reached himself by sweeping into one bag the Baconian tradition (type II causes), the Cartesian identification of type III causes with grounds, the Leibnitian law of sufficient reason and the Humian concept as causes as events prior in time to the effect.

“and, neglecting the one thing in Newton which modern physics has found most valuable, namely the doctrine that what happens according to a law happens without a cause, devised a doctrine which was soon accepted as orthodox. The central points are three.”

(a) That every event has a cause. (b) That the cause of an event is a previous event. (c) That (a) and (b) are known to us a priori.

So according to Collingwood, Kant ended up with a metaphysical theory of causation that was incoherent. His two suppositions do not hang together, or do so only under "a pressure which must produce a somewhat violent strain in the resulting structure".

In so far as these suppositions were widely accepted, the fabric of science based upon them was in a dangerously unstable condition. The hatred of metaphysics among modern scientists may reflect a sense of insecurity due to this situation.
Collingwood poses two metaphysical dilemmas, the anthropocentric dilemma (brought to light in the late 19th century) and the anthropomorphic dilemma which lies deeper and is less understood.

1. Two solutions to the anthropocentric dilemma

(a) The natural scientist is trying to build a practical or experimental science of the Baconian type, using type II causation. [This sounds like instrumentalism; if it predicts OK it is OK]. Success spells out in bending nature to our own purposes.

(b) The scientist wants to construct a theoretical science of nature, gaining knowledge about the world in itself, regardless of what man has done or may hope to do by manipulating it.

334 Since Kant 1(b) has been replaced by 1(a).
2. Two solutions to the anthropomorphic dilemma.
Do we presuppose that the world is animated by something like a human mind or psyche?
(a) The first answer is that nature only becomes intelligible to man by use of analogies drawn from the conscious life of man.
(b) The second answer is that such analogies confuse the issue and a well-devised vocabulary would avoid them. Nature does not devise, adapt or invent things to bring about her ends, only humans do this and the scientist should seek explanations in terms of physical and chemical processes.

2(b) was the orthodox view at Kant's time. Natural science was opposed to anthropomorphism (the 'a' response) and pinned its faith to type III causation as its best weapon. But this contained an element of anthrophmorphism that could only be detected by metaphysical analysis; because this was barred, anthropomorphism persisted at the core of natural science. This is a dangerous situation for science especially while the anti-metaphysicians remain in the ascendant and so block the work that is required for metaphysical house-cleaning.

So while the sciences continue to make spectacular progress they are threatened by defective metaphysics.