Home | Bibliography | Page 1 of 2 | Outline of Page Two

The Philosophy of Science of Arthur Eddington

Background: This page was written from the point of view of Wittgenstein's logic of language and may not be understood without first understanding that logic.

Outline of this page ...

Science Education - What is Needed

Children at school are taught to believe in the solar system of Copernicus -- but they are not taught how to question the application of that picture to reality. And this is what makes the picture so powerful: that we don't know how to question what its relationship to the truth is. Our education in science ought to give us the understanding we need to question whatever is presented to us as scientific truth, but it doesn't.

Where then ought it to begin? I think with the question: What is a scientific theory? distinguishing theory from fact, that a theory is a selection of facts plus imagination [cf. PI § 156]; that the theory is not itself one of the facts.

The Copernican solar system is an example of a scientific theory: the particular picture it presents is not itself an astronomical fact, but instead one of several possible ways to organize those facts; the organization is the imagination part of the theory.

Beyond that is the question of the presumptions that underlie scientific method, as e.g. Newton's "Rules for Reasoning in [Natural] Philosophy". In other words, the beginning of the study of natural science should coincide with the beginning of study of the philosophy of science and the question of sciences relation to metaphysics (when it confuses theories with facts) .... But nothing is less likely than that any subject at school be taught philosophically. And so we ourselves in later life have to think philosophically about what we are taught in science lectures when we are young.

How can a non-scientist talk about science?

How can I -- or anyone who is not a scientist -- talk about this? [cf. Wittgenstein about the philosophy of mathematics.]

How can I? Because I have no choice if I want to resolve the conceptual confusion that was created for me at school. And because the things that scientists say about science aren't necessarily to be trusted, because the things scientists say about science no more belong to the "account books" of science than the things mathematicians tell us about mathematics belong to the account books of mathematics.

What is said about the nature of science belongs to the philosophy of science, which, just as the philosophy of mathematics, is the view from outside. And scientists are not necessarily good philosophers of science, despite the famous conceptual clarifications that Einstein made.

In this paper I try to understand what Arthur Eddington said about physics in his Gifford Lectures, "The Nature of the Physical World", in 1927. [Note 1]

The Great Change in the Ideal of Scientific Explanation

One of the greatest changes in physics between the nineteenth century and the present day has been the change in our ideal of scientific explanation.

It was the boast of the Victorian physicist that he would not claim to understand a thing until he could make a model of it; and by a model he meant something constructed of levers, geared wheels ... The man who could make gravitation out of cog-wheels would have been a hero [then].

Nowadays [we] are dealing in physics with a symbolic world [i.e. mathematical equations showing the relations among pointer-readings have replaced the mechanical models made by the Victorian engineer]. (Eddington, Chapter X, p. 209)

We have dismissed all preconception as to the background of our pointer readings [i.e. numbers gotten by operations of measurement by e.g. scales, clocks, spectroscopes], and for the most part we can discover nothing as to its nature. (ibid. Chapter XII, p. 259)

The essential point is that, although we seem to have very definite conceptions of objects in the external world, those conceptions do not enter into exact science [i.e. physics] and are not in any way confirmed by it. Before exact science can begin to handle the problem they must be replaced by quantities representing the results of physical measurement. (ibid. p. 253)

... since we must cease to employ familiar concepts, [mathematical] symbols have become the only possible alternative. (ibid. p. 249)

Einstein's theory ... insists that each physical quantity should be defined as the result of certain operations of measurement and calculation. (ibid. p. 255)

The project of physics is quantification; what is not quantified does not exist for physics. (Galileo said that abstraction is the key to progress in science.) Therefore the "symbols" -- i.e. signs [the word 'sign' in Wittgenstein's jargon means e.g. marks on paper, spoken sounds, the purely physical face of language in contrast to language's meaning] -- of physics must be defined in terms of how they are obtained: the method of measurement is the meaning, so to speak. This looks very much like "The use of a sign -- and there are many types of uses, not just one -- is its meaning" rather than: "All signs are names and the meaning of a name is the object the name stands for".

Pointer-readings (i.e. the numbers given by the various techniques of measurement) and the relationships among them as stated in scientific equations (-- If we look from the general to the particular, the numbers indicated by the pointers replace the "symbols", e.g. v = d/t, in the equations --) would be the accounts books of physics. If, that is, Eddington gives a true account of what physicists do.

How Physics sees an Elephant sliding down a Hillside

Professor McTighe (ca. 1978) [Note 2] referred to this in a lecture:

"An elephant slides down a grassy hillside ..."

Eddington gives this as an example of a question in a physics examination. Then he says how physics handles it: the elephant is replaced by a mass, the hillside is replaced by a slope, the grass is replaced by a coefficient of friction.

And so we see that the poetry fades out of the problem ... (Eddington, p. 251-2)

McTighe made "the poetry" more vivid: he imagined that at the bottom of the hill was a pond and that the elephant landed in the water with a big splash.

McTighe's point, I believe, was that the grasp physics has of reality is not what everyone is looking for. And Eddington recognized this:

But the View of Physics is not what Everyone is Looking for

We have not reached the practice of the Laputans [the inhabitants of Jonathan Swift's "land of mathematicians"], who, "if they would, for example, praise the beauty of a woman, or any other animal, they describe it by rhombs, circles, parallelograms, ellipses, and other geometrical terms". [It is the rare physicist who holds] the belief that his wife is [merely] a rather elaborate differential equation. (Eddington, Chapter XV, p. 341)

The Sliding Elephant - Does "the poetry" belong to Reality?

But Eddington also seemed to believe "the poetry" to be a "mental illusion":

As scientists we realize that color is merely a question of the wave-lengths of ethereal vibrations ... (ibid.)

By 'merely' Eddington meant "really only": that color does not belong to "objective reality". This is wrong. What physics offers us is a possible way of looking at things (i.e. quantification); but it is no more the "objective" way than the Laputan way is.

... the color in the familiar world and its counterpart electromagnetic wave-length in the scientific world. (ibid. Chapter V, p. 94)

Eddington takes this correlation ("counterpart") for granted, overlooking the fact that, on his account, this correlation is an hypothesis and can never be anything else. This needs to be explained.

The relation between our concepts [i.e. the "grammars" or rules for using the words] 'sight' and 'eye'

In the context of the Philosophy of Mind, I question myself: "You call correlations between mind and body hypotheses, but do you call the relation between sight and the eye hypothetical?"

What it makes sense to say is the we see, and if we don't see we blame our eyes; we go to the eye doctor, wear eyeglasses, and so forth.

But it is not nonsense to say of a man without eyes that he sees, or it would not be (since it never happens). The presence of eyes is not our only criterion for applying the word 'see'. (This is a different question from the question of how someone without eyes could see. The correct answer to that would be: we don't know.)

Our concept 'sight' is not fixed according to causal criteria? The use of the word 'sight' is connected with the use of the word 'eye'. Strange as it sounds, beyond that language does not tell us where to go.

We believe in the causal relation between sight and the eye. Nothing counts against this correlation, but it does not belong to logic that nothing can count against it. (And there is no other necessity than logic -- i.e. grammar, which in Wittgenstein's jargon includes "rules of sense and nonsense", i.e. semantics, not only syntax -- in language.)

Why do we say that anyone sees? Because of what they say and do (i.e. how they behave). A child learns to use the word 'see' in particular surroundings, and the word 'eye' is also used in those surroundings.

Is 'color' a psychological-word?

Eddington, of course, means "the color we see", not the color of objects we point to. But by presenting the problem this way, he has made 'color' a psychological-word. So that, it is now only an hypothesis that color corresponds to "whatever" the spectroscope detects. ('Whatever' because Eddington has said that physics knows nothing of "the background of our pointer-readings"; color belongs to our "definite conceptions of objects in the external world".) Eddington's grammar for 'color' is not our held-in-common ordinary everyday grammar for that word.

Eddington is uttering nonsense. If 'color' were a psychological-word, then the grammar of 'color' would be like the grammar of 'pain'; color would be like Wittgenstein's "beetle in a box" (PI § 293; cf. ibid. § 304: "A nothing would serve just as well as a something about which nothing could be said"). The word 'color' would not be applied to objects. But, of course, that is exactly how we do apply that word.

'Objective' versus 'Subjective'

Here we have little hesitation in describing the waves as objective and the color as subjective. (Eddington, Chapter V, p. 94)

But the waves (or rather the scientific instruments by which they are detected; no one has ever seen a "light wave") are just as much an object of perception as the color is. (Eddington does not see that his Idealism requires him to also make 'pointer-reading' a psychological-word.) So how strange then to say:

The wave is the reality -- or the nearest we can get to a description of reality; the color is mere mind-spinning ... In this and other parallelisms we find the objective in the scientific world and the subjective in the familiar world. (ibid.)

Eddington uses the words 'objective' and 'subjective' equivocally. All perception is subjective -- i.e. it is the activity of a human subject. In that sense the color is subjective, as is the pointer-reading. (It is nonsense to call our perception of a weighing-machine -- or of anything else -- 'objective'.)

But what does Eddington mean by 'objective' -- that the wave is "really" to be found in the "object" (i.e. "reality") but the color is not? Eddington seems to have forgotten what he said about the language of physics -- that its signs are defined as pointer-readings; and what meaning has 'electromagnetic wave' as a pointer-reading? None. Those words are a picture of the very "thing in itself" of which Eddington said: "We have dismissed all preconception as to the background of our pointer readings, and for the most part we can discover nothing as to its nature." (ibid. Chapter XII, p. 259) Eddington's is the very foolishness Drury spoke of:

We have in immediate experience our one sole contact with reality .... [but then] Reality which lies before us at every moment is replaced by the abstract picture we have ourselves created. (M. O'C.Drury, The Danger of Words (1973) p. 109, 100)

The wave, not the color, is "mind-spinning" -- i.e. it belongs to a picture-theory. But not as if we were disputing facts with Eddington: it is not a question of fact which is real, which illusion, which subjective, which objective. We are making reminders about how we use our language.

Eddington's grammar for the word 'objective' would make the distinction between sense and nonsense impossible in the language of everyday. The statement 'The sky is blue' is an objective statement of fact in that language, even though the grammar of quantification-word is not the grammar of 'blue' in this statement. Of course 'quantifiable' = 'objective' is one way we use the word 'objective' -- but it is not the only way. Though not quantified, objective statements can be made about colors, ponderous elephants, and even spectroscopes and scales.

Eddington's "two tables"

Eddington and his two tables: "the table [as in table and chairs] we see" versus "the scientific table", i.e. electrons and so forth. This scientific table is "mind-spinning", telling a fairy tale "with verifications". (This is not like "Geometry with proofs" but like the Mock Turtle's "French, music, and washing -- extra", i.e. there is a category mistake here). The electrons belong to the fairy tale, as follows from what Eddington has said; the pointer-readings are the "verification" -- except that there is nothing here to verify (i.e. there is no object, no thing-in-itself named 'electron').

If the familiar table is an illusion, then the familiar pointer is of course an illusion too. -- But if everything we see, including the pointer-readings themselves, must be an illusion, then what is the word 'real' to mean?

[cf. Eddington's one table, for there is only one, not two.]

Is the word 'table' the name of an object?

What right does Eddington have to apply the word 'table' to his arrangement of atoms in space? How e.g. would he define the word 'table' in this case? Not ostensively; the word 'table' would not be a name-of-object. Or if Eddington said that it were, that would not fix the word's grammar; the question of how it was to be applied (i.e. verified) would still need to be answered. Otherwise there would be no criterion for applying the word correctly, in which case it would be meaningless.

There is a case where Eddington can say that the familiar table we see is an illusion: when the table we see is a holographic projection (i.e. hologram). We reach out to touch what we see, and our hand passes right through it. Here we can say that the solid table we see is an illusion, and that what is real is the light -- i.e. the light we see.

If we are using the word as it is normally used (and how else are we to use it?), then ... (cf. PI § 246)

Eddington does not use the word 'illusion' the way it is normally used. Does a scientific theory justify this (CV p. 44)? No, Eddington's use (or rather, non-use) of the word is metaphysical; it is the product of his Idealism, for the expression he actually uses is 'mental illusion'. Now what do we call, what is an example of, a 'mental illusion'? In normal English that combination of words is not used, unless it means the same as a 'delusion'.

Different criteria of verification give different meanings to the word 'illusion'

But I am still tempted to say that there is a sense in which what Eddington says could be correct. What would that sense be?

Is the micro-biologist talking nonsense when he says 'It is an illusion (or, it is false) that the soil that looks so lifeless to us is lifeless, because if we look at it under a microscope we find countless micro-organisms'? Is the astronomer talking nonsense when he says 'It is an illusion that the sun rises at one end and sets at the other end of a flat earth: it is rather that a round earth spins ...'? If these examples are not instances of nonsense, it is because a different standard of verification is applied in them than is applied in our life of every day. By that standard -- and only by that standard -- can we apply the word 'illusion' or 'false' here. But Eddington does not talk the way these other imagined scientists do:

my scientific table ... does not belong to the ... world which spontaneously appears around me when I open my eyes ... My scientific table is mostly emptiness. Sparsely scattered in that emptiness are numerous electric charges rushing about with great speed; but their combined bulk amounts to less than a [million millionth] of the bulk of the table itself. (Eddington, Introduction, p. xii)

... the process by which the external world of physics is transformed into the world of familiar acquaintance in human consciousness is outside the scope of physics. (ibid. p. xiv)

Is the nonsense in the supposed picture physics presents or is it in Eddington's Idealism?

"Really" belongs to metaphysics not physics

Would the physicist be talking nonsense if, like the other scientists, he said 'It is an illusion (or, false) that the table we see is solid: because according to pointer-readings the table is numerous electrical charges sparsely scattered in mostly empty space'? Because here too, a different standard of verification is given (i.e. pointer-readings) than the standard used in our life of every day. (Here you really can say that the meaning is the method of verification.) If we express ourselves this way, it becomes clear that the sentence 'The table is solid' is used differently by the physicist than we use this sentence in day to day life (same sign, different grammars). In the former case, the sentence is false; and in the latter case the sentence is true. And since we are using different frames of reference here, Eddington cannot say that the solidity of the familiar table is an illusion -- i.e. that the every day sentence 'The table is solid' is false -- without uttering nonsense, unless the table is not solid in the familiar way, e.g. because its wood is rotten (BB p. 45). The nonsense is produced when we say that 'The soil isn't really lifeless', 'The sun doesn't really rise and set', 'The table isn't really solid' -- because those statements would only make sense if 'the absolute (or, right) frame of reference' also made sense. That is, our dispute with Eddington is not one of facts, but one of grammar.

Reality and Illusion [Fact versus Theory] - Which is which?

Eddington takes two positions throughout his book: sometimes he says that the Rutherford-Bohr picture of the atom is a picture of reality, and sometimes he says that it is idle. E.g. he suggests that talk about electrons could be replaced with Jabberwocky -- so long as the numbers remain the same, no knowledge is lost. 'Electron' has no meaning beyond its metrical (i.e. pointer-reading) definition: 'slithy toves' could be substituted for 'electron' with no loss of meaning. (Eddington, Chapter XIII, p. 291) But then he says:

A rainbow described in the symbolism of physics is a band of ethereal vibrations arranged in systematic order of wave-length from about .000040 cm. to 000072 cm. From one point of view we are paltering with the truth whenever we admire the gorgeous bow of color, and should strive to reduce our minds to such a state that we receive the same impression from the rainbow as from a table of wave-lengths. (ibid. Chapter XV, p. 328-9)

But what this "impression" would be is undefined. If 'wave-length' means no more than the pointer-readings of a spectroscope, then any picture we invent of waves is idle: Eddington says again and again that we know nothing of the supposed reality behind our pointer-readings. What is Eddington asking us to do?

What Eddington suggests is nonsense. If anyone could see the way a spectroscope "sees", then we would not need spectroscopes, because that would be like looking through transparent window-glass and claiming that we were making use of a scientific instrument.

But although that is how the rainbow impresses itself on an impersonal spectroscope, we are not giving the whole truth and significance of experience -- the starting point of the problem -- if we suppress the factors wherein we ourselves differ from a spectroscope. (ibid.)

If it made sense to call "the rainbow we see" an illusion, then it would also make sense to call what the spectroscope "sees" an illusion. "Ethereal vibrations" -- how can a theoretical construct -- because "ether" is merely a postulate: no one has ever seen these vibrations -- what on earth would it mean to call that real and "the rainbow we see" an illusion?

There are broken-line drawings of the patterns of magnetic waves, but these, like the Rutherford-Bohr atom are only "models, pictures, maps" (DW p. 99-100). We apply the word 'model' both to the planetary systems and to the atom. But the planets of the model resemble the planets in the sky. But Eddington says that the protons and electrons of the atom model do not resemble the actual atom; indeed that there is no "actual atom", only pointer-readings. (If we apply the word 'model' to both, we should not allow that word to blind us to the difference between these cases.)

'Appearance' versus 'Reality'

How do we use the words 'real' and 'illusion'? They are used to contrast with one another, like the words 'true' and 'false'. When we call what we see an 'optical illusion', it is because we appeal to a method of verification. E.g. the case of the lines of equal length that are presented in contexts where they appear to be of different length; in this case we measure the lines with a ruler, and that is why we speak of 'appearance' and 'reality' here. Eddington says that the rainbow is appearance and "a table of wave-lengths" the reality; -- but these cases are utterly different. In the rainbow case it is not that we are deceived by appearances, but that Eddington is saying that everything we see (except for pointer-readings, which unaccountably have a privileged position) is appearance rather than reality. Eddington is not making a scientific remark, he is making a metaphysical remark (i.e. a remark which is either grammar of nonsense; it is nonsense).

Wittgenstein: how is 'solidity' defined?

We have been told by popular scientists that the floor on which we stand is not solid ... This is liable to perplex us, for in a way of course we know that the floor is solid ... Our perplexity was based on a misunderstanding; the picture of the thinly filled space had been wrongly applied. For this picture of the structure of matter was meant to explain the very phenomenon of solidity.

... in this example the word 'solidity' was used wrongly [i.e. without an antithesis] and it seemed that we had shown that nothing really was solid ... (BB p. 45-6)

But these ... are only hypotheses, models designed to explain, to sum up, what you observe. (PI § 156)

Which was precisely Drury's point. To turn the model into "the really real" is to confuse science with metaphysics.

Does 'real' mean 'quantifiable'? (Physics versus Art)

To equate 'real' to 'quantifiable' as Eddington does is to state a definition which is not in agreement with our usage of 'real'. 'What is not quantifiable is just as real as what is quantifiable' is a rule of grammar in our language of every day.

And Eddington's definition has nothing to recommend itself: why should what is "revealed" through the intermediacy of measuring instruments be regarded as real and what is found "immediately" be regarded as unreal? The "electromagnetic waves" are a "mental illusion" in a way that colors can never be: the waves are hypothetical constructs; -- now why call can a hypothetical construct the "really real"? Eddington says that the table (as in table and chairs) we see before us and feel under our hands is not real; what is real is the hypothetical atoms of which the table is "composed".

There is nothing necessary about esteeming quantification more than non-quantification. Calling the unquantifiable "unreal" is just making propaganda. (Why is it unquantifiable? For we ourselves made it unquantifiable; cf. Z § 259) Why shouldn't Beethoven's exploration of the countryside in his Sixth Symphony, or the work of any landscape artist be more esteemed than that of the natural scientist? (Of course this is not a question of having to make a choice.)

People nowadays think that scientists exist to instruct them, poets, musicians, etc. to give them pleasure. The idea that these have something to teach them -- that does not occur to them. (CV p. 36 [MS 162b 59v: 1939-1940 § 2])

I hope that nothing that I say in this paper will be taken as an "attack on science". The greater part of my life [-- Drury was a doctor of psychiatric medicine --] has been taken up with an attempt to think scientifically about the problems of mental pathology, their cause and their treatment. (DW p. 73-4)

... I want to fix a more precise meaning to that phrase, to determine its limits. (ibid. p. 77)

Drury's aim was to determine the limits of science, to recognize when scientists extend science beyond those limits and say more than they know (BB p. 45). This is also my aim.

Physics can contribute only a part to our picture of reality

If it cannot be measured, it does not exist for physics. That in itself indicates what an abstracted knowledge of reality physics offers us. Physics has a method of verification (of how do you know) -- but it is not the only one: measuring is not the only way we verify. This method may be wholly appropriate in physics, but hardly appropriate in our life, and hardly in building up a picture of reality. By its very nature, physics can only contribute to part of this picture, as can the other sciences. Even the totality of science leaves out much, if not most, of reality.

In science it is usual to make phenomena that allow of exact measurement into defining criteria for an expression [i.e. linguistic sign]; and then one is inclined to think that now the proper meaning has been found. Innumerable confusions have arisen in this way. (Z § 438)

Plato's science of ethics

"... that allow of exact measurement." This is how Plato seeks to define piety ('piety') in the Euthyphro, that is, by a method of measurement (6d-7d), for then piety will be knowledge, not something about the meaning of which men are at variance and may fall out with one another over. A method of measurement = a method of knowledge. (Cf. natural science = measurement; Plato wanted a natural -- or apparently, rather, a supernatural (because piety is not a thing with a mass that can be weighed, a length that can be measured, or a sum that can be calculated) but knowable -- science of ethics.)

But what is the relation of that notion to the essence or common nature of piety -- i.e. to what all pious acts have in common which differentiates pious acts from all other acts -- if there is such a thing? Rules can be derived from essences, and a method of measurement is a rule; both are objective standards: they exclude variance. (I don't see any other relation.)

Eddington's illusions

Eddington's proposed application of the word 'illusion' is nonsense -- but there is no one way to definitively show that it is.

Rather than a table or a rainbow, Eddington could have chosen a human being. If what we see when we see a human being is an illusion, then all our psychological language is nonsense when applied to other people (and so it is also nonsense when applied to ourselves, since 'identification' is nonsense etc.) [See Philosophy of Psychology (i.e. of the language of feeling and of mind)]

Or an amoeba. What would the micro-biologist think if he were told that he was wasting his life studying what in the end is only an illusion? At the atomic level of investigation, cancer doesn't exist. Eddington frequently alludes to the theory of evolution, but surely that is also an illusion; evolution -- that type of change -- has nothing to do with the laws of particle physics. And on Eddington's account, there is no science except physics.

If the rainbow we see is an illusion, then the sentence 'I see a rainbow' must be false -- or rather nonsense, since it must always be false. If the color we see is an illusion, then it must always be false to say 'I see a blue chair'; or rather all our color words must be nonsense.

I.e. with the grammar of what other words is the grammar of the word 'illusion' indissolubly connected?

This is a reduction to absurd consequences; it is another way to approach Eddington's remarks. (Wittgenstein used this method in his "private language" investigation.)

Science and Truth

The scientific truth of one age is the scientific delusion of another; that is e.g. how Eddington saw the mechanicism of the 19th Century. The destiny of every scientific theory is to become part of the fossil record. The trouble with basing our lives on the conclusions -- i.e. on the theories and points of view -- of scientists is that science never finishes with anything. (But then neither does philosophy.)

The thing about progress -- it always appears to be much greater than it is. -- Nestroy (PI motto)

That is, a development may have the form of an improvement without essentially being an improvement. It may instead simply be a change, even a change for the worse.

Now, why am I quoting from a seventy year old book about a science that is constantly changing? Precisely because it is constantly changing. Eddington's views as a physicist are not subject to criticism by me; it is his philosophy of physics that I criticize. (E.g. whether Einstein posited the ether to the very end or whether most physicists posit it today, I do not know. It does not concern me, unless the physicist claims that the ether belongs to the "really real".)

Eddington was an astronomer of Trinity College, Cambridge, the college of Isaac Newton. He wrote: "Venus, so far as we know, would be well adapted for life similar to ours ... If transplanted to Venus we might perhaps continue to live without much derangement of habit --" (Eddington, Chapter VIII, p. 170). Seventy years later no astronomer would say that; -- but so what? But this is something we should remember whenever we hear astronomers speculating about "cosmology", because Eddington talked about Venus apropos of the possibility of life on other planets.

In Wittgenstein's logic of language, of course, we also look at our language at a specific point in time. But why do we look at it at all? Because certain concepts -- and that means the uses of certain words -- are perplexing us. It is the business of philosophy to resolve these perplexities.

Your questions refer to words; so I have to talk about words. (PI § 120)

Philosophical investigations -- conceptual investigations. (Z § 458)

And this description [of our concepts] gets its light, that is to say its purpose, from the philosophical problems. (PI § 109)

Scientific versus Everyday Language

The concepts of science (i.e. the ways scientists define words) are transitory: they are used only so long as they are useful to scientists; then they are dropped. (CV p. 44 [MS 127 36v: 27.2.1944 § 1]) They are invented to be used. But the concepts of our language of every day are for the most part inherited, i.e. passed on from one generation to the next. Obviously they too may change time (PI § 23); but, importantly, some concepts don't change, e.g. our applications of the words 'rainbow' and 'pain' and 'thunder'. There is no justification for abandoning concepts as old as human language is -- at least not in favor of the concepts belonging to a particular scientific theory.

Perhaps only human beings can perceive an elephant in the "poetic" way that human beings do and perhaps no two human beings perceive this phenomenon quite the same way; -- but, of course, for that conjecture to have a meaning, we need to imagine e.g. a god who can perceive these differences between species and between individual human beings. If a lion could talk it might employ concepts (i.e. language) that were incomprehensible to us. (ibid. p. 223) All we know is that we human beings agree [coincidentally, not contractually, agree] in the language we use (ibid. § 241); -- the rest is an "eternal question without an answer".

Of course, by saying that our normal, naive way of speaking shows concepts, not hypotheses (Z § 223), we are stating what our interest in language is. In Wittgenstein's philosophy the concern is with the grammar (i.e. the conventions) of the language of everyday. This language is not eternal; some of our concepts are different from those of the Greeks; who can say what the language of everyday may look like in another 2500 years. So we are not talking about "eternal truths", but only about language at a particular time in its history: those are the facts that concern us. And we should be very wary of drawing any conclusions about reality from our present concepts. (This is the temptation Rationalism succumbs to.)

Is Space Euclidean or non-Euclidean?

In our four-dimensional world ... distances obey a non-Euclidean geometry.... By inventing extra dimensions we can make the non-Euclidean geometry of the world depend on a Euclidean geometry of ten dimensions ... (Eddington, Chapter VII, p. 158-9)

Eddington rejects the view that "space is neither Euclidean nor non-Euclidean; it is all a matter of convention" (ibid. p. 159). -- But this is not one view; it is two views. Eddington has just said that "we can make", which means that space is neither one nor the other; the logical 'can' makes this a matter of convention. But it is not a matter of convention that a four-dimensional non-Euclidean geometry or a ten-dimensional Euclidean geometry is required to account for the evidence. Likewise it is a matter of convention, not of fact, whether our world is geocentric or heliocentric; but it is not a matter of convention that Ptolemy requires epicycles to account for the evidence and Kepler does not.

Looked at Eddington's way, it is not an hypothesis whether space is non-Euclidean or Euclidean. The hypothesis is whether space is four-dimensionally non-Euclidean or four-dimensionally Euclidean. (cf. TLP 6.342)

We can place the origin of the Cartesian coordinate system wherever we like. But once we have chosen a place for the origin, the location of the other points is not a matter of choice, although, of course, the scale we use is. This is the difference between convention (frames of reference) and evidence.

It is always difficult to make a fence of objections so thorough as to rule out all progress along a certain line of explanation. But even if it is still possible to wriggle on, there comes a time when one begins to perceive that the evasions are far-fetched. (Eddington, Chapter IX , p. 188-9)

'Evasions' means that: more and more qualifications are added to an hypothesis in order to save it from disconfirmation, and in this way the theory becomes more and more complicated. What Eddington's view amounts to is that we are told "something about the world by the fact that it can be described more simply with one system ... than with another" (TLP 6.342). And that Newton's Rules for Reasoning direct physicists to always choose the simplest system; of course, to say this is not to imply that the simpler system is the "true system" (The innocent move in the conjuring trick obliterates the distinction between facts and the theories that are imaginative arrangements of them).

Propositions (statements of fact), Hypotheses, and Tautologies

A proposition is coupled with reality with varying degrees of freedom. In the limit case there's no longer any connection: reality can do anything it likes without coming into conflict with the proposition. (cf. Wittgenstein's Philosophical Remarks [1929-1930], tr. Hargreaves and White (1975), p. 282)

Where there can be no conflict, the proposition has the status of a tautology. The other limit is a verification or falsification. A scientific hypothesis (theory) falls anywhere in between those two limits; it can be falsified -- i.e. require an amendment (qualification) -- by an anomaly, but it can never be verified, although more and more evidence may be found in favor of it -- i.e. more and more facts may be found that are consistent with it. (The "theory of evolution" has for some scientists become, according to Drury, not a fact, but a tautology: it has become anomaly-proof (cf. "bomb-proof", "water-proof"): any fact whatsoever can be fitted into it.)

The form in which we cast an hypothesis may make this sense of the word 'hypothesis' clearer. Wittgenstein once said:

A proposition can be verified [Then a proposition has the form of a simple statement of fact (assertion) e.g.: 'The book is on the table']; a hypothesis cannot, [An hypothesis has the form 'If p, then q' where p and q are propositions; it is] a law or rule for constructing propositions and looks to the future -- i.e. enables us to construct propositions which say what will occur and which can be verified or falsified. (Wittgenstein's Lectures, Cambridge, 1930-1932, ed. Desmond Lee (1980), p. 16)

Of course, part of the problem is that we use the word 'hypothesis' in many ways: there is no essential, or, general definition of the word 'hypothesis' (What Wittgenstein gives above is a selected sense of the word, i.e. jargon), but instead only similarities; or, in other words, an 'hypothesis' is a concept without an essence. Drury uses that word interchangeably with the word 'theory'. Whereas Isaac Newton used 'hypothesis' to mean 'unverifiable conjectures' (e.g. about the "really real" behind phenomena): "dreams and vain fictions of our own devising", about which he said hypotheses non fingo. So we have to be careful when we hear that word or utter it ourselves. The following text belongs to about the same period (I don't think Wittgenstein would have spoken of "the essence of an hypothesis" in his later years) but in the following remarks he seems to be talking about the specific type of hypothesis called a "scientific theory":

It is the essence of an hypothesis that its confirmation is never completed, but that doesn't mean that there is a verification of it which we may approach ever more nearly, without ever reaching it. No, an hypothesis simply has a different relationship to reality from that of verification. Hence the words 'true' and 'false' are inapplicable here, or else have a different meaning. (cf. Philosophical Remarks p. 285)

This is why I wrote above that Newton's rule that we are to regard an induced proposition "as very nearly true" is nonsense.

Well, if everything speaks for an hypothesis and nothing against it -- is it then certainly true? One may designate it as such [i.e. apply the word 'true' to it]. -- But does it certainly agree with reality, with the facts? -- With this question you are going round in a circle. (OC § 191)

The best case for an induction is that all known evidence supports it. But that "all known" is what makes an induction neither true nor false; unless 'true' means no more than 'there is no known counter-evidence'; -- and that more evidence has been found for Induction A than for Induction B, does not make A "more true" than B. (Here we talk about degrees of well-foundedness.) What is important is not to let the application of a word -- namely, 'true' -- cover up differences between cases. And so I would not apply the word 'true' to an induction at all.

Maybe to make the distinction between an induction (induced generality) and a proposition (simple statement of fact) explicit, we should always cast inductions in the form of hypotheses ('If p, then q' where p and q are propositions), as for example: "If all swans are white, then if A is a swan, A is white." (Of course, my example includes the induction "All swans are white" as well as the propositions 'A is a swan' and 'A is white'.) This form of expression would serve as a grammatical reminder.

Again: the question (for me) is whether the statement is certain, not whether we are certain about the statement.

On the other hand, maybe my notions about experimental verification are naive -- that science doesn't require anything like what I think it ought to require. -- But my question is: can I require, insist on my notion (i.e. is this sense or nonsense)?

What distinguishes a 'scientific' theory from all others?

Newton did not regard merely-consistent-with-the-facts speculations as science; those were his "hypotheses", our "tautologies". I would say that if speculations of that type were regarded by scientists as hypotheses, then science would not deserve any special respect. Hegel's suggestion about the "fossil record" ("The organic forms found in early geological strata never really lived. They are merely anticipations in stone of what was later to be clothed in living flesh and blood") is not, in my view and I believe by Drury's own definition, an example of a scientific hypothesis -- because it gives no explicit criterion for its falsification.

By definition, what distinguishes a 'scientific' hypothesis from all others is that it must be not only verifiable (i.e. clear which facts are consistent with it) -- but also, and more importantly, clearly falsifiable. (DW p. 18-19)

[But I think that Drury's criterion requires the distinction I made using the words 'hypothesis' and 'theory': "hypotheses", as in 'working hypotheses', are falsifiable, whereas "theories" are not. For instance, the relative complexity or simplicity of a theory is not a criterion of its truth or falsity: the Martian epicycles do not falsify the geocentric theory of the solar system. Again, a theory is facts plus imagination (Drury's "models, pictures, maps") -- or rather, plus an accommodating imagination ready to revise it picture of things in order to absorb anomalies (i.e. things that would not fit into the original picture), but an hypothesis is not. (These are of course all statements made to define the words 'theory' and 'hypothesis' in a particular way rather than another, to assign them a meaning, not to state what their meaning "really is".)]

An explanation does not replace the phenomenon it explains

We need to remind ourselves again and again that statements of fact are not science. Although scientists discover facts, it is their explanations (i.e. theories) of the facts that are what distinguishes science from simple fact gathering. The "picture of the [atomic] structure of matter was meant to explain the very phenomenon of solidity" (BB p. 45), the solidity that Eddington claims is "not real, but a mental illusion". The picture of light waves of different frequencies was meant to explain the phenomenon of color, not to explain it away. If we explain the rainbow we see as the refraction and reflection of sunlight in raindrops, we cannot then turn around and deny the reality of the rainbow we see, the very phenomenon we were trying to explain. That makes no sense.

A phenomenon is not a symptom of something else: it is the reality. (Philosophical Remarks p. 283)

Which means: There is no "really real" behind the rainbow that we see such that we could ever say that the rainbow we see is really an illusion. (It is not an example of what we call an "optical illusion", for example.)

What is hard for me to accept: that an hypothesis in science is never fully confirmed or disconfirmed, and that it can be modified indefinitely so as to never be disconfirmed. Because that is not what I want at all. I want to know (what metaphysics calls "the reality behind reality", which is not "the facts plus imagination" but things as they are in themselves, free of human perception and conception ("imagination") [But are there such "things"?, or is "the really real" really what it appears to be -- i.e. an undefined combination of words, a "picture" that can't even be drawn (i.e. not even logically possible), never mind compared with the world of experience]), and as such the geocentric and heliocentric models of the solar system are simply reminders that there is no absolute point of reference, that "the eye of God" is not our eye.

Pierre Janet and falsifiable theories (hypotheses)

Janet built up a most interesting psychology based on the twin concepts of "psychic energy" and "psychic tension". In the Introduction to one of his books he makes the profound remark that one great advantage of his theory is that time may prove him to be completely wrong. As a matter of fact I think Janet's hypothesis has not be substantiated, and if he were still alive he would perhaps have replied, "Well, what did I tell you".

I think that we must all be on the watch that in psychology and psychiatry we take care to formulate hypotheses which are capable of being refuted. (DW p. 18-9)

That is, or may be, the hallmark of a scientific hypothesis. But it is not what I want.

Knowledge - The longed-for ideal

I do not want the probability of inductions any more than I want mere ways of looking at things (frames of reference). I want the absolute perspective, something we can only speak of by analogy -- if we are not just uttering nonsense (and 'absolute objective certainty' is nonsense, undefined language). If the best we can hope for -- i.e. the only thing we can hope for -- is probability within some frame of reference or other, then is it worth having at all? Eddington says that we should not "confuse science with omniscience" (Eddington, Chapter X, p. 202).

I think that what philosophers -- at least those who are drawn to metaphysics -- really yearn to do is to see into the mind of God, to see as the eye of God sees. Augustine's Confessions has a prayer:

I shall know You, my Knower. I shall know You even as I am known. This is my hope, and in this hope I rejoice. (x, 1; this was also Plato's hope, to see, to know after death the reality projecting the shadows on the walls of the cave of his Republic)

Who sees God, sees all things. (St. Gregory the Great, pace Thomas Aquinas)

Now, no one expects any philosophy -- including natural philosophy (Newton) -- to offer a hope like that in this world, limited as our life is. What kind of hope does philosophy offer, 'philosophy' here as now distinguished from science? What does Socrates conclude in Plato's Apology [23b] if not that "the wisest of you is like Socrates, who has realized that he knows nothing of much importance". But we are better off for knowing our ignorance -- that is, not merely believing that we don't know, but understanding (or having reasons for hoping we understand) why we don't know. Socrates did not go beyond that -- maybe because there is no "beyond that" to go: You don't go beyond a philosopher; you just go someplace else. (Plato doesn't go beyond Socrates, but instead draws philosophy into the realm of endless speculation ("seeming") which is now called metaphysics: he calls the reality we know mere appearance (false reality), while the "reality" we only imagine he says is the real reality. Eddington also does this.)

Is there a mind of God (i.e. is the combination of words 'mind of God' nonsense)? I really don't know. On the one hand, it suggests that we are asking for an "absolute frame of reference" (which is undefined language, i.e. nonsense), but on the other hand the introduction of a picture requiring "a god who sees what we cannot see" perhaps makes this a question without an answer. "The picture is there, but what is its application?" Of course it has none, but I think we might partly identify 'the mind of God' with 'absolute frame of reference', although the former concept includes much more than that. (This is a conceptual puzzle/muddle, not a question for religion.)

But there is another purpose to our study of philosophy. Not only to make us aware in particular instances of what we know (understand) versus what we don't know, but to make us more and more aware that there is something even more fundamental that we are never going to understand. (Kierkegaard, quoted by Drury) This is called "the riddle of existence" or "life itself".

And what does that have to do with the philosophy of science? It arises out of the question of what the relation of science to knowledge is. Because if we base our actions, our moral judgments, our understanding, on the facts, then we don't want to be mistaken about the facts. We want to be sure that what we have are facts, and not theories that we have mistaken for facts (through our wretched education). But that requires examining each particular statement to see if it is a statement of fact; it requires asking: how do you know? [The distinction Drury made between fact and theory.]

What does 'random' mean in science?

When a scientist characterizes an event as "accidental", he tells us nothing about the event itself (but only about the grid through which he examines it -- TLP 6.342). 'Accidental' is not a concept of science. (Now that is hardly a report of what scientists actually do; it is stipulative philosophy of science, a rewriting of the Rules for Reasoning.)

I want to say that: when scientists say that something is "accidental" or "random" or "by chance", they say no more than that they cannot (i.e. do not) imagine a way to explain it. They say nothing about the event, but only about their attitude toward the event; or that their way of looking at things, or that their investigative instrument (e.g. the microscope), yields no explanation in this case. That is, characterizing an event as "random" or "accidental" is in no way to explain (account for) the event. (I also want to say: science is deterministic; i.e. that is its program: strict causality.) But that seems to be a mistaken notion.

"Primary and Secondary Laws of Physics"

Some things never happen in the physical world because they are impossible; others because they are too improbable. The laws which forbid the first are the primary laws; the laws which forbid the second are the secondary laws. It has been the conviction of nearly all physicists that at the root of everything there is a complete scheme of primary law [that] fixes the history of every [particle or] constituent of the world ... (Eddington, Chapter IV, p. 76)

The "secondary laws" are statistical laws: they "relate to the behavior of crowds, and depend on the fact that although the behavior of each individual may be extremely uncertain average results can be predicted with confidence" (ibid. Chapter XI, p. 244).

The water in a teapot placed over a fire - What can/will it do?

So that, to the question: if we place a teapot on a fire, will the water boil?

Primary law replies, "It may boil; it may freeze; it may do pretty well anything. The details given are insufficient to exclude any result as impossible." Secondary law replies plainly, "It will boil because it is too improbable that it should do anything else." (ibid. Chapter IV, p. 76)

To answer the question, primary law would need to know "the positions, motions, etc., of some quadrillions of particles and elements of energy" (ibid.). But Eddington says that the present attitude of physics is that even then primary law could not answer; there is a limit to primary law: the behavior of an individual particle is not determined by strict causality, but can only be predicted as the probability that it will do this, that, or the other.

Eddington says that as a consequence of quantum mechanics, physics "is no longer pledged to a scheme of deterministic law" (ibid. Chapter XIV, p. 294). [I don't think it need be the case that 'strict causality' = 'universal determinism', but it may mean 'determinism in the particular system'.] Einstein, on the other hand, I believe, never accepted the quantum theory. That is what he meant by saying that God doesn't throw dice, because that is the picture of reality quantum physics offers (in contrast to the strict causality of Newton): events are predictable only by their relative probability, just as the outcome of throwing dice is. [Note 2a]

Forecasting Next Year's Weather

Eddington considers determinism in the context of predicting what the weather will be a year from now: there are so many factors involved, including the factor of whether a single human being discards a lighted match in a Canadian forest; -- it is absurd to suppose that next year's weather is already predetermined, he concludes (ibid. p. 310). But what Eddington does not see -- sorry, or rather, the view he does not accept -- is that this is not a question about the nature of things, but about the attitude of science to things -- and that this is all it is ever a question of. Or again: there are no laws of nature (which would concern "nature in itself", i.e. metaphysics' "really real", the "reality behind reality"), but there are laws of physics (and these are to I-don't-know-what extent determined by the Rules for Reasoning -- i.e. by the practices of physicists, by their attitudes and projects in science).

Strict Causality versus Statistical Knowledge

May it not have been precisely the program of seeking strict causality that led to science's fruitfulness, and may it not be that by abandoning strict causality science will become barren? The fruitfulness I am talking about here is intellectual (philosophical) -- i.e. "fruitful" from the point of view of understanding the universe for its own sake, rather than for the sake of being able to do something with your knowledge. We do classify the results of statistical investigations as knowledge, but statistical knowledge is not the same as the knowledge of causes. On the other hand, Eddington argues:

It is quite true that the quantum laws for individuals are not incompatible with causality; they merely ignore it. But if we take advantage of this indifference to reintroduce determinism as the basis of world structure it is because our philosophy predisposes us that way, not because we know of any experimental evidence in its favor. (Eddington, Chapter XIV, p. 303)

Eddington treats this as a matter of experimental fact: we look for strict causality, and if we do not find it we say that it does not exist. Is he correct -- i.e. can the question of determinism be treated this way? If an hypothesis can be formed about this, then yes. But can it be? Pascal experimentally demonstrated that there is void space (i.e. he managed to create a vacuum using an inverted bell jar, a basin and mercury, if I remember correctly). Einstein, on the other hand, posited the ether; he did not demonstrate that it exists. It can always be said: if light requires the ether, then we do not understand light; because there is no experimental evidence in favor of the ether's existence. Could Einstein reply: the requirements of light are just such evidence? It seems to me that this is what Eddington is doing ("The investigative instruments we use do not allow us to determine strict causality, and therefore strict causality does not exist"), but I may be mistaken.

"The only thing we cannot look for is nonsense" (Limits)

At what point do we stop -- (logically) must we stop -- looking for something -- i.e. is it possible to "prove a negative", i.e. a statement that such-and-such does not exist? How do you decide that what you are looking for is not to be found -- not as a practical (i.e. what we do in our life of every day), but as a logical question: the only thing you cannot look for is nonsense. Eddington would have to show that the case of quantum theory is like the case of "absolute space frame"; Einstein showed, not that no such thing exists, but that this combination of words is undefined.

Indeterminism is an Attitude of physicists, not a Law of physics

"You are seeking a frame of space which you call the right frame. In what does its rightness consist?

"You are standing with a label in your hand before a row of packages all precisely similar. You are worried because there is nothing to help you to decide which of the packages it should be attached to. Look at the label and see what is written on it. Nothing.

"Right" as applied to frames of space is a blank label. It implies that there is something distinguishing a right frame from a wrong frame; but when we ask what is this distinguishing property, the only answer we receive is "Rightness", which does not make the meaning clearer or convince us that there is a meaning. (Eddington, Chapter II, p. 21)

What Eddington means, of course, is that the combination of words 'absolute space frame' is undefined; it has no meaning because we have not given it one.

That is, if this is not a logical matter, but a matter of just giving up at some point, this shows that indeterminism is an attitude of physicists, not a law of physics (let alone of nature).

But there is a Second Possibility - namely that It is a Law of Physics

The conditions of our exploration of the secrets of Nature are such that the more we bring to light the secret of position [of an electron] the more the secret of velocity is hidden ... When we encounter unexpected obstacles in finding out something which we wish to know, there are two possible courses to take. It may be that the right course is to treat the obstacle as a spur to further efforts; but there is a second possibility -- that we have been trying to find something that does not exist.

When the concealment is found to be perfectly systematic, then we must banish the corresponding entity from the physical world. There is really no option. (Eddington, Chapter X, p. 221)

That is the difference between science (which is falsifiable) and metaphysics (which cannot be proved false) -- and that is why the latter is idle except as a reminder of possibilities; but isn't it also why there is a limit to the former's claim to produce knowledge? Well, but isn't knowledge that something doesn't exist "true knowledge"? We used to say "All swans are white" which entails that "No swan is black". That some swans are white was irrefutably true, but that no swan is black was not. With respect to negatives we can (This "can" is logical possibility) only talk about our present state of knowledge -- that is, it is not that negative knowledge is not "true knowledge" but that it is different in kind from positive knowledge. Negative knowledge is actually deduced from positive knowledge, which is why it is so important "to say no more than we know" (BB p. 45) -- i.e. to call a deduction a deduction.

Justification by experience comes to an end. If it did not it would not be justification. (PI § 485)

I.e. that is the way we use the word 'justification'. And the use of a word belongs to logic: so that, giving up at some point is a matter of logic, because there is a point at which 'to look further' becomes nonsense. Or again: objective doubt, like objective certainty, requires grounds. So that, if physicists, as Eddington says, have run out of grounds for doubting that strict causality does not exist, then it does not exist. Though this does not preclude new grounds for doubt from arising in the future -- that verification comes to an end does not imply that further evidence is impossible (There are indeed black swans), but as of now, according to Eddington, strict causality is not to be found in nature (or should I say physics). And that is the nature of science. So then, Eddington is correct in his assertion -- if what he is asserting is an hypothesis.

Is 'Every event has a cause' a rule of grammar or a method or both?

If 'Every event has a cause' -- or, "strictly speaking": 'Every effect has a cause', but in everyday life we use these expressions interchangeably [although this needs to be treated differently in another context] -- is a rule of grammar, then 'Some events do not have causes' is nonsense. I think the former rule belongs to our language of every day. The latter rule belongs e.g. to the quantum theory; it is a new rule, the making of which is justified by a theory. Certainly if we insist 'But there must be a reason why such-and-such occurred', we are only making a grammatical remark (reminder).

But does that rule of our language of everyday imply "determinism"? That notion maybe does not belong to our language of every day. It is a metaphysical notion -- i.e. a picture, e.g. the picture of "underlying reality" as a single mechanism, say an old-fashioned mechanical clock, in which no part is idle: all parts are connected, and the movement of each is predetermined by strict causality. But our metaphysical notions arise out of the pictures that our language suggests to us.

But the rule 'Every event has a cause' does not entail that all events are inter-related, nor that any given series of causes does not have a first member.

When do we seek causes in everyday life?

When do we ask about causes? Normally when something happens that we do not see an immediate explanation for. E.g. "Why did the daffodils not come up?" -- (We dig in the soil) "The bulbs have disappeared." -- "Why did the bulbs disappear?" -- "Because mice ate them." -- "Why did mice eat the bulbs?" -- "Because that is what mice do." And that's the end of the matter. Or anyway, though children can go on uttering the word 'Why?' indefinitely, the initial question has been answered: an event has occurred and a cause has been identified (for the first event: "The bulbs have disappeared", and for the second event "Because mice ate them"); -- and when a cause has been identified, that is the logical (i.e. grammatical) end of the matter, the point at which we may choose to stop and not investigate further. Again:

Justification ... comes to an end. If it did not it would not be justification. (PI § 485)

"Because that is what mice do" (or, in other words, "Because bulbs are food to mice") is not an answer to the question "Why did the daffodils not come up?". It does not belong to an "endless series of causes" of the one event. In the language of every day, events are discrete.

Suppose we had answered "Why did mice eat the bulbs?" with "Because we did not protect the bulbs with wire cages". "And why didn't we protect the bulbs?" -- "Because we had never had any trouble before this." -- "And why had we never had any trouble before this?" This is probably the point at which we answer: "Who the devil knows!" or "God only knows!" At some point we lose interest, not because we must, but because that is the way we live. Scientists, however, do not lose interest; they go from question to question. (But note: a scientific investigation needs a particular type of event. If e.g. "Why didn't I protect the bulbs?" were answered "Because I forgot to do it", this is not necessarily an event that can be investigated scientifically.)

'Every event has a cause' is a rule, or, method embodied in a rule of grammar. 'Things don't just happen for no reason' could be restated 'Always look for a reason!'

Do you think there must always be a reason why things happen?

But why does anyone want to insist "But there must be a reason. No one know what it is, but --"? Is this 'must' logical (i.e. grammatical)? Or is this like the case of "Things don't just vanish into thin air"?

We say that this 'must' is logical, but we still want to insist that 'Every event has a cause' "says something about reality". (This is not the same as the case of the induced generality: "Usually when we look for a cause, we find one.") "Things don't just happen for no reason; ex nihilo, nihil." Why does anyone say that -- is it because "No, it just happened, but not for any reason" makes us feel as if we were facing the phantasm of an irrational, unpredictable world? "There's an explanation for everything, though we may never discover what it is." Is this a matter of instinct (of the urge to myth-making which is characteristic of mankind)?

What a curious attitude scientists have --: "We still don't know that; but it is knowable and it is only a [question] of time [till we know it"!] As if that went without saying. -- (CV p. 40 [MS 124 49: 16.6.1941])

Now in what way is this "curious"? Philosophers do not have that attitude; neither do theologians. Is it a question of method -- i.e. of having learned how to ask the type of questions that can be answered (because philosophers, theologians and scientists do not ask the same kind of questions, nor are they looking for the same kinds of answers)? -- Is not the attitude of scientists, within the context of science, justified by experience?

It is possible that the notions 'strict causality', 'determinism', 'predeterminism', arose out of our language of every day, out of the grammatical rule 'Every event has a cause'. -- But they are not identical to out common rule because that is not the whole of our rule, which continues: 'not necessarily a unique cause, and not necessarily a mechanical cause (like billiard balls), and not necessarily an endless string of causes'. The notion 'determinism' may arise out of attending to some of the circumstances in which we apply 'Every event has a cause', while ignoring others.

'There is an explanation, though we may never know it' = 'The world is essentially intelligible'. -- This is tautological -- i.e. it is not falsifiable. But determinism can be made hypothetical if we make the rule: 'If we do not know the reason because we have no technique for determining a reason, then there is no reason.' -- Isn't this the rule that the quantum theory employs?

On the other hand, if it makes sense to say that we believe in the uniformity of nature, then why shouldn't it make sense to say that we believe in the principle "Nothing happens without a reason"? And why shouldn't an instinct, an instinctual picture of things, be embodied in grammar?

Experiences, thoughts, -- life can force this concept on us.

So perhaps it is similar to the concept 'object'. (ibid. p. 86)

This instinct can be characterized as "religious" or "superstition" -- that does not matter. What is interesting is that this instinct exists, and that it is connected to our search for explanations (and that means to our searches in philosophy, science, our myth-making); these have the same foundation in instinct (or anyway I find it natural to make this connection). It is also interesting that while we apply the word 'instinct' to the behavior of most animals, we do not say e.g. of a dog that it is superstitious, nor that it makes myths, nor that it seeks explanations. This instinct is unique to human beings. And it is interesting that it exists: human beings want to know the truth, the how and why. The Apostle Paul's "God made all men to seek Him, that by groping for Him, they might find Him" can be regarded as a recognition of this instinct.

Greek gods ... not to be left alone

The Greek gods are a crude but telling expression of this absolute conviction that since man is somebody, and not merely something, the ultimate explanation for what happens to him should rest with somebody, and not merely with something.... Greek mythology expressed the firm decision of man not to be left alone, the only person in a world of deaf and dumb things. (Etienne Gilson, God and Philosophy (New Haven: 1941), p. 22-23)

Again, however this is characterized, the instinct exists. Although whether it exists in everyone, I do not know:

The examiner [for Drury's oral examination in physiology] then went on to say: "Do you think there must be a significance, an explanation?" As I see it there are two sorts of people: one man sees a bird sitting on a telegraph wire and says to himself: "Why is that bird sitting just there?", the other man replies "Damn it all, the bird has to sit somewhere"." (DW p. xi)

Do you find that reply satisfying? (That is what an explanation attempts to do: to put your mind at rest. (cf. ibid. p. 109)) [But there is also a question for logic here ("Must there be a reason?" versus "Can there be a reason?")]

The type of explanation or account people look for varies among individuals

However, it might also be remarked that the type of explanation or account people look for varies among individuals. What Wittgenstein wanted was descriptions ("Philosophy only states what everyone admits" (PI § 599)), but other people want scientific theories or some other type of myth (e.g. religious, metaphysical). The type of understanding people seek may also have its foundation in instinct, or, in what amounts to the same thing, in each individual's character (Heraclitus: "A man's character is his fate").

I am drawing attention to similarities here, rather than differences; making comparisons, rather than distinctions.

Is the belief in "Nothing happens without a reason" the same as the belief in "Reality is like a mechanical clock" (described above)? Not necessarily. Einstein's call for the return of the "spirit of Newton" wasn't a call for a return to belief in "determinism", but to "reality is intelligible". Is that a distinction without a difference? I don't know.

Are all events causally inter-related? If I drop a leaf, when that leaf strikes the ground will this have repercussions on the other side of the universe?

Is all knowledge subject to future rebuttal?

Science indeed produces experimental knowledge -- i.e. knowledge subject to future rebuttal. But is there any other kind? The difference between science and metaphysics is that in metaphysics statements need only stand to reason (i.e. be self-consistent), whereas in science statements must stand not only to reason but also to the evidence. The problem of speaking of 'metaphysical knowledge' is not that metaphysics is nonsense; there is, after all, nonsense in everything -- that is not a distinguishing feature of metaphysics. The problem with metaphysics is that it consists entirely of statements derived from its axioms (a rule of grammar derived from other rules of grammar is still a rule of grammar (OC § 1)). This gives it the status of a tautology, and tautologies tell us nothing about reality. ("If we know that either it is raining or it is not raining', we know nothing about the actual weather" (TLP 4.461).) Axiomatic geometry and pure mathematics are like metaphysics rather than like science. So that, if we want knowledge, we must accept science and its statements of falsifiable probability -- where the methods of science are applicable and strictly applied, that is.

Philosophers constantly see the method of science before their eyes, and are irresistibly tempted to ask and answer questions in the way science does. This tendency is the real source of metaphysics, and leads the philosopher into complete darkness. I want to say here that it can never be our job to reduce anything to anything, or to explain anything. Philosophy really is "purely descriptive". (BB p. 18)

When we describe the logic of our actual language -- i.e. when we report the way we use our language everyday -- we describe facts. By 'facts' we mean statements-of-fact (not artifacts); what we do in logic in this instance is to make statements of fact. What do we call a 'statement of fact'? A statement of what is known. What do we mean by 'known'? That we have adequate, or, compelling, or, sufficient grounds for saying that a statement is true (or false). In the case of where we report language usage, we do no more than make simple statements of fact that no one will deny the truth of (PI § 415; cf. Z § 211), but in other cases what compelling grounds look will vary from one statement of fact to another statement of fact.

But again -- I keep saying this because I am more than a little dense (i.e. I seem unable to be penetrated by a distinction which I have made countless times already) -- I am not willing to call science, which is essentially hypothetical, 'knowledge'; -- why?

That which is possible (or not "too improbable") is the domain of natural science; that which is actual is the domain of natural history. (Eddington, Chapter XII, p. 266)

This is an important distinction. I said above that I was not willing to call generalities and predictions 'knowledge'. But what is a statement of probability if not a prediction about what will happen, or a generality based on what has happened. By my lights science should not be called 'knowledge' at all -- because (again and again and again): facts alone are not science; the explanation of facts -- theory -- is science.

Is all knowledge subject to future rebuttal? I don't know. We call many different types of things 'knowledge'. As with the grammar of the word 'game', there is no essence of 'to know'. And that means: we must investigate the individual case (PG i § 24, p. 60).


[Arthur Eddington's Philosophy of Science - Page 2 | The philosophy of science of James Jeans | The philosophy of science of M. O'C. Drury]


Note 1: Arthur Eddington [1882-1944], The Nature of the Physical World [Gifford Lectures, University of Edinburgh, January-March 1927] (Ann Arbor: 1958), Chapter X, p. 209. [BACK]

Note 2: Thomas McTighe was [perhaps still is] a professor at Georgetown University who taught courses in the Philosophy of History and the History of Medieval Philosophy which I attended. His remark about Eddington comes from the first course.

When on the first assignment for the latter course I wrote that I felt it was dishonest for me to write about Plotinus, because the text seemed to me to be "plain nonsense" (cf. PI § 464), Professor McTighe patiently wrote as a comment on my essay: "Since you have to be here, try to make the best not the worst of it." (This is the best advice I have ever been given, and I have never forgotten it.) A very modest man, McTighe did not call himself a philosopher, but only an historian of philosophy. [BACK]

Note 2a:

May the spirit of Newton's method give us the power to restore unison between physical reality and the profoundest characteristic of Newton's teaching -- strict causality. (Message on the Newton Centenary 1927; quoted in The Nature of the Physical World, Chapter XIV, p. 294)

You believe in the dice-playing god, and I in the perfect rule of law in a world of something objectively existing which I try to catch in a widely speculative way ... The great initial success of quantum theory cannot convert me to believe in the fundamental game of dice ... I am absolutely convinced that one will eventually arrive at a theory in which the objects connected by laws are not probabilities, but the conceived facts, as one took for granted only a short time ago. (Einstein's correspondence with Max Born, in the latter's Natural Philosophy of Cause and Chance (1949), quoted by S. Korner in "Some Types of Philosophical Thinking" in British Philosophy in Mid-Century (1957), p. 123)

If a, then b rather than if a, then with x% probability b. "The great accomplishment of Newton -- strict causality" (Einstein). If the tools you have are inadequate or inappropriate for the work you want to do, then you need new tools, in this case, new conceptual tools or concepts. If strict causality should be the goal, then physicists nowadays are merely treading water with statistical or quantum physics, waiting for a new conceptualizer to appear.


Site copyright © September 1998. Send Internet mail to Robert [Wesley] Angelo. Last revised: 3 April 2006 : 2006-04-03

The URL of this Web page:

Back to top of page

Wittgenstein's Logic of Language - Introduction and Table of Contents | Bibliography | Site Search | Site Map