Berkeley'S External World: The Isaiah Berlin Virtual Library

The Isaiah Berlin Virtual Library
BERKELEY’S EXTERNAL WORLD
This is a lightly edited transcript of a text in Berlin’s papers. No attempt has

been made to bring it to a fully publishable form, but this version is posted here
for the convenience of scholars. A great deal more work is needed before the
transcript is faithful to the text in every detail.The bracketed numbers refer to
the pages of the original text, which is partly typed and partly in MS.
Lecture I: Introductory
1. Everyone knows what made Berkeley notorious. He said that
there were no material objects. He said the external world was in
some sense immaterial, that nothing existed save ideas – ideas and
their authors. His contemporaries thought him very ingenious and
a little mad. Dr Samuel Clarke thought there was no way of
refuting a man who thought life was a sort of coherent dream; only
it was such an absurd view. Descartes, who died fifty years before,
thought the same: only a beneficent deity guaranteed it wasn’t so.
Dr Johnson refuted it ‘thus’. Dr Whitehead still thinks it is not
quite true but cannot be refuted. The Russian Marxist refutes it in
three moves: children of minds = parents one = nonsense.
2. Berkeley would have been very shocked, and all his life
protested that he meant no such silly paradox. Of course he knew
a real gold coin or a real noise or a real fire as well as anyone else.
Fires singed and burnt you and the idea of fire didn’t; a real gold
coin buys something, the idea of one, even the painting of one,
doesn’t, etc. He was saying nothing paradoxical or odd, he was
saying only what everyone knew to be true. But he admitted that
the words in which he chose to say it might sound odd to some
people. Why then did he do it? Only because the ordinary way of
saying these things led to something he thought much too strange
and dark – to talk about material substance, or physical matter,
which according to the physicists and philosophers was colourless,
soundless, tasteless, odourless etc. Nobody had ever seen any, or
heard it, or touched it, or smelt it, or tasted it; nevertheless
apparently wise and respected persons not only claimed that it
1
B E R K E LE Y ’ S T H E O R Y O F T H E E XT E R N A L W O R L D
existed, but said that virtually everything consisted of it, and seemed
to know a good deal about its properties and habits – such as that
it was composed [2] of a very large number of very small spherical
bodies whirling inconceivably fast and possessing mysterious
powers, such as gravitation or impenetrability, which nobody had
actually come across (as they had across qualities like scarlet or
sharp, which ordinary things had), but which nevertheless the
learned Galileo and learned Dr Newton had proved were there,
somewhere there all right, and Descartes and Leibniz and Locke all
agreed.
The more an ordinary man thinks about such questions as
‘What is matter?’, ‘What is everything made of?’, the more he is
liable to get into a state of mental cramp. Thales said everything
was made of water, and Heraclitus of fire, and Pythagoras of
numbers, and Plato of imperfect examples of ideas, and Aristotle
of primal matter. But this was not the answer he really wanted.
‘What is coal?’ A black hard globular combustible etc. substance.
‘No, I mean what is it really? Not what does it look like?’
Appearance alters, coal remains, when someone spoke of
molecules of carbon, hydrogen etc. moving about and colliding,
composed of whirling atoms etc., he was in a muddle: what he saw
was a black hard shiny heap, perfectly still, not moving and
apparently continuous, occasionally with tongues of flame or a dull
glow etc. How could one and the same object be both continuous
and broken into globules, at rest and in motion, black and
colourless? Scientists offered no help, then or now; they merely
said that things could ‘really’ be one and look ‘the other’, and
provided some rules for passing from one to the other, plus
propositions about brains, eyes, optic nerves, effects of light, laws
of refraction etc. But all this language presupposed two levels – of
invisible physical entities and the world as we thought we knew it –
and didn’t explain the original paradox of the two worlds, and how
the one of physics came in at all; and how you got from one to the
other.
This has puzzled people ever since. It is this that Berkeley
thought he could answer; this which he regarded as a
philosophical, not a physical or grammatical, puzzle. His treatise is
a model of how philosophical puzzles are stated and should be
treated.
[3] So the puzzle is: What do scientists mean by atoms, electrons
etc., and generally what do people mean who say that an object
2
may be different from what it looks like? How do clouds come to
look like snowy mountain tops, or mountain tops like clouds? And
how do we ever know which are really mountain tops, and which
are really clouds? This doesn’t matter for an impressionist painter:
he paints what he sees, the picture does not tell you which is
which. Nor does the camera, and if you later say you did discover,
because one of the cloudy whiteish shapes was hard to the touch
whereas the other was filmy and you flew through it in an
aeroplane – when you say that, how do you know that what is hard
is a bit of a mountain, and what is filmy and penetrable is a cloud?
You say you mean by mountain something hard, and by a cloud
something you go through; but then how do you know the
mountain was really hard and did not merely seem so? People are
subject to strange errors and delusions, we are told.
The argument from illusion

Locke’s argument is that if you have a bowl of water which feels
hot to one hand and cold to the other, it follows that the water,
which cannot be both hot and cold at the same time, cannot be
either, but since it exists it must have some other properties which
give rise to or cause different sensations in the observer. Berkeley
points out that if this is supposed to prove the subjectivity of
secondary qualities, it does it equally in the case of so-called
primary ones. If I cannot decide by mere touch whether the water
is really hot or really cold, by the same criteria I cannot decide by
using my senses whether what I am seeing is one thing or two,
round or square, at motion or at rest, fast or slow, long or short,
etc. In other words, that primary qualities are no more exempt
from delusions than secondary. No doubt delusions create
difficulties. But once again he repeats that this perplexity, even if it
is real, this doubt, even if it is irresoluble, cannot demonstrate the
existence of matter, which is a meaningless word in any case; nor
yet of primary qualities, which, if they do not resemble those
provided by the senses, equally seem to mean nothing.
Since the Argument from Illusion has often been used to prove
that something non-sensible exists, it is worth dwelling on it. What
is the argument for or against? It is considered to be a powerful
argument against what is called naïve realism, against the view that
the external world is what we think it to be before we start
philosophising. We are said to believe that objects are what they
look, that they really do possess the colours they appear to have,
3
really give forth the sounds which appear to proceed from them,
etc. This of course is also Berkeley’s view; he thinks in some
curious sense that all these attributes are mental or ideas, but he
does not believe in any dualism. The main obvious difference lies
between monistic and dualistic views, that is, views which
distinguish physical reality from appearance as two kinds of being,
and those which do not, etc. By this classification Berkeley and the
naïve realists believe in one world, not two; that is, they reject the
theory of the Iron Curtain. In this sense Berkeley may be called a
naïve idealist. The Argument from Illusion, which Berkeley
recognises, draws attention to the fact that a physical object, say
the wall of my room, looks darker when no light appears to be
shining from anywhere than when it is sunlit or artificially lit. But
what can be meant by asking what is the colour of the wall? We feel
it to be self-contradictory to say that an object has two colours in
the same sense at the same time in the same place, yet it seems
equally queer to ask which of the many colours which, we say, the
wall ‘takes on’ as the sun sinks, is the true or the real colour.
Philonous points out to Hylas in the First Dialogue that clouds
which look red and purple are ‘really dark mist and vapour’. Dark
looking? If so, to whom? Or ‘really dark’ and invisible? A thing
which looks to have one uniform colour to me may present a
kaleidoscopic variety of colours to the eye of a fly. It may look
yellow to the sufferer from jaundice. Why do I persist in saying
that the colour is really, say, pink, although it ‘appears’ to be
variegated or motley to the fly or through the microscope, yellow
to the jaundiced observer, still odder to the man who takes the
drug mescal, etc. Well, which is the true, the real, colour or pattern
of colours?
Similar anomalies occur in the case of visual shape. The
microscope, the gnat’s eye, the distance at which the observer
stands make smooth surfaces seem ravaged by hills and valleys,
square towers seem round, round pennies seem oval, and so forth.
And this applies to all sensible qualities.1 Or what happens to a
headache? I have a violent headache; I take a drug, and the pain,
we say, decreases in intensity. Do I say that the headache is as
violent as before, but seems milder, or is it actually milder? What is
the test I use? What is the difference between the ‘objective’ yellow
of paper painted with yellow paint and the ‘subjective’ yellow of
1 cf. Gulliver and the Houyhnhms.

4
what the man with jaundice sees? Yellow is yellow and the rest is a
causal story. Or again, how do I measure the time that passes
quickly when I am absorbed, and very slowly when I am bored or
in pain? Or again, take Locke’s instance: if I put my hand into a
bowl of water, then one hand which previously felt colder than the
other now feels greater warmth than the other, being plunged into
what is surely the same water. How warm is the water? For it cannot
be both as hot and as cold as it feels to one and the other hand.
To all this common sense tends to reply that there are certain
objective standards. ‘Is the wallpaper really yellow?’ can be settled
only by a spectroscope and a vibrometer, which settles the
question of how many light waves are emitted. The shape of the
tower or the penny can similarly be determined by photographs
taken in a standard light from a standard distance. The length of
time is settled by a chronometer or a metronome, and so on. So
far so good. There is no doubt that, so far as this goes, it is
perfectly true and is a useful guide to our use of the adjective ‘real’
as opposed to ‘seeming’ in such cases. Is a given image a mirage or
is it not? What we mean by mirage is that a photographic camera
placed roughly where our eye is now would not have recorded any
palm trees waving round an oasis, or alternatively that if we
approach what looks the place where the palm trees are, they
dissolve. ‘An oar looks bent in water but is not really so’: that
means that if we feel along it with our finger we shall not
experience the feeling of an angle; if we pull out the oar, it will
look straight; nor is there any reason for supposing that water
bends or breaks things in the way that a blacksmith does; and so
on.
To all this Berkeley need say only that the findings of cameras,
metronomes and any other so-called objective criteria are just as
sensuous as what they are supposed to check and verify. All we are
really doing is to draw an ultimately conventional and arbitrary
distinction between, say, real yellow and apparent yellow, by calling
real yellow that which can be correlated in a certain prescribed
fashion with other visual data, namely the recordings of dials,
continuity, similarity to more than one observer, and so on. While
apparent yellows could be correlated with abnormal, i.e. rare,
events such as the experience called jaundice, or the projection of
yellow light, that they are likely to be intermittent, not last long, be
confined to one observer, the observer has something wrong, i.e.
differences with brain, eyes, to other observers, etc. To say the oar
5
looks but is not bent in the water is merely to say that, contrary to
the normal state of affairs, visual data which usually can be
correlated with a tactual datum called ‘feeling an angle’ in this case
cannot. And cases where they cannot can in their turn be classified
as examples of reflection of light in a certain way, and so become
normalised again. If the question arises whether an oar which
looks bent should, or should not, be called a bent oar, that is a
matter of convenience. By calling a thing bent we normally mean
to indicate a correlation between visual and other sorts of data. In
the case of the oar the correlation is an unusual one; for this
reason it is not convenient to use the same term, and we prefer to
speak of a systematic or necessary illusion; but if we know the
facts, the illusion no longer deludes, though force of habit may
make us forget a particular correlation. Do we still call it an
illusion? Yes, to call it something to indicate an unusual
correlation.
There is nothing mysterious about a mirror. It presents data
similar to ones we normally correlate with tactual ones, but for once
not so correlated. We know this in a general way, yet we
sometimes try to walk through or into mirrors. This is no more
mysterious than that we forget that some red berries are poisonous
though others just like them are not, and suffer the consequences.
In the second case the colour red is not enough to ensure
digestibility. In the first case visual data in general are not [enough]
to guarantee touchability or walkability through. The tower is
square and looks round. The pillar box is and looks round. The
moon is larger than a dinner plate or half a crown and yet
sometimes looks like one or the other. The earth seems immobile
yet is said to rotate. As Berkeley says in the Third Dialogue ‘a man
is not mistaken with regard to the ideas he actually possesses, but
in the inferences he culls from his present perceptions’. The
inference is that if he approached the tower it would still look
round, but if he approached the moon it would still look no bigger
than a plate, and that the earth would still look still even if you
stood on a convenient star and watched it through a telescope.
You jump to these conclusions from force of habit or whatever
cause you like, and you are wrong. That is all there is to it. That is
Berkeley’s view. ‘We do not see what we feel,’ says Philonous;
‘neither is the same object perceived by the microscope which was
by the naked eye.’ But we do not like to multiply names
indefinitely for each differentiable experience, because it makes
6
things difficult to refer to: hence common names given to regular
clumps of co-existent or successive data. The datum yielded by the
microscope and that to the naked eye are of course different in a
describable way. I give them the same name because I can always
infer one from the other. The great mistake is treating what has the
same name as literally one and the same. The names stand for
strings of different data given to different senses or to the same
sense at different times under different circumstances. Identity is
thus resolved as by Hume into a regular succession, uniform
similarities, performance of analogous functions, and so forth. So
that the only thing distinguishing the coherent dream from reality
would be difference of vividness, coherence of the rest of
experience, i.e. validity as testimony for what happens next,
support by the testimony of the majority of observers, and so
forth. But then what about Descartes’ opinion? Perhaps everybody
is dreaming dreams which fit each other and coherent experience
is nevertheless in some sense unreal? And though there are no
criteria to distinguish it from reality, nevertheless it could be a
dream and we not know? But this would then be rendered
meaningless, for the meaning of the word ‘dream’ and the meaning
of the words ‘waking reality’ are drawn from a distinction in the
field of sensible experience, so that to say that everything is a
dream is to use the term ‘dream’ in a new sense, and when that
sense is examined it is seen to be the same sense as that in which
the term ‘reality’ is normally used. What we mean by a dream is
something which we shall wake from, or something at any rate
which other observers do not experience. What we mean by saying
the penny is really round is that it looks round, at least from above,
and feels round at most times, and looks elliptical from the side.
So far from its elliptical appearance casting doubt on its reality, it is
one of the principal ingredients of what we mean by a real penny.
It is if it ceased looking elliptical that we should begin to wonder
whether what we had before us was a real penny. It is if the oar
emerged from the water still looking bent that we should ask
ourselves whether something gravely perplexing had not occurred,
whether the laws of refraction of light had not failed us, and the
water had not acquired new and wholly unexplored physical
qualities. Whether, in short, we were not dreaming or having
hallucinations. The real is the normal, and our standards of
normality rest on a certain amount of experience of what
sensations go with what others under what conditions. If we do
7
not have enough such experiences we are apt to make mistaken
inferences, i.e. predict wrongly. And if someone says ‘How much
experience is enough?’, the answer is pragmatic, ‘Enough is
enough.’ When we no longer make annoying mistakes of wrong
correlation and wrong induction. All life and science is a like trial
and error. Trial and error as to what experiences go with what
others, and when and how often and how intensely – and can they
be repeated? Can the water be both hot and cold? Yes indeed, it is
cooler than one hand, and hotter than the other. What we mean by
its objective temperature is something to do with the position of
mercury on the thermometer dial. What we mean by its felt
warmth is whether it is warmer or colder or indiscriminably the
same as the felt warmth of the limb with which we test it. The
Serpentine will seem warm in spring to persons who bathe in it in
the winter, and icy cold to the inhabitants of South Sea islands. It
does not merely seem, it is both. Its so-called real warmth can be
gauged only from what all these different witnesses have to say.
When we have heard the testimony of the Eskimos and the South
Sea Islanders about the warmth of the average day of spring in
Oxford, and discount accordingly, we shall know roughly what to
expect. There is no mystery and no problem. A thing is real and
not illusory when it fits into our framework of beliefs and
expectations. The frameworks of different sorts of observer will
naturally be somewhat different, but if communication is to be
possible, they must be correlatable. The world describable in terms
of such regular correlation is ‘real’.
Objections
When a correlation seriously breaks down, the result is called
prima facie illusion, and demands a special new correlation to fit it
into the real framework. As to how the correlation works – and
what is meant by saying that observers exist and are in place, etc.,
these are difficulties that we shall have to deal with. But there is
one objection to even this which eminent philosophers such as
Professor Moore and Professor Price urge against this treatment of
the Argument from Illusion, and that is the case of double vision.
A thing may indeed be said to have all the properties it has
depending on the place from which it is observed, and the
conditions and the physiological state of the observer and so forth.
To say the penny is round and oval – and is bright is to say it looks
bright to X from Y in conditions Z, but dark to A in B in
8
conditions C. So a bright light is in this sense really both bright
and dim; depending on place and physiology etc. As a book is
intensely interesting or dull depending etc., so it is both equally
really. But if on pressing my eyeballs I see two glittering or dim
disks instead of one, am I not forced to say according to this view
that one penny occasionally displays the property of two-ness or
doubledness which is one of the attributes which constitute the
meaning of the word? But two-ness is not a property. Can I really
say [that to say] there are two pennies in my pocket is to say there
is any one entity there qualified by two-ness? We say that, in the
case of the subject–predicate Aristotelian logic, to say that a thing
may have two-ness is to entail that it may have three-ness or four-
ness; though that may entail that there is only one thing in the
Universe which is sometimes 77 or sometimes 210, and that robs
the term ‘one’ of any meaning. How does one differentiate one
from 1,000? If a thing may, as it loses redness and turns purple,
lose its three-ness and turn to two-ness, why should it not lose its
one-ness and revert to zero-ness? And if so, could it be a
something which is a nothing? Two-ness is no more a property
than oneness or being or existence is. At first this seems a
formidable objection, but it rests on a mistake. Of course two-ness
is not a property, but then what we mean by ‘one’ is not as simple
as it seems. If I say that Lewis and Short’s Latin Dictionary is one
book and not two, Lewis’s book and Short’s book, I do not mean
that it is not composed of many different constituents. The fact
that a whole is in some sense describable as one, although its parts
are not, seems too trivial to need elaboration. But what is relevant
is that we use the term ‘one’ in certain contexts to mean something
which in another sense is many. Now, applying this to our case, to
say of a penny that it is really one is to say many things about how
it would look to different observers in different lights, etc. But it is
also to say, so it turns out, that under certain circumstances my
visual field will contain data in it of a certain sort, which, when I
count as sense data are counted, I call seeing two circular patches:
and when this happens in the way it does, we normally speak of
there being only one real object. That is the rule for using the term
‘one’ in the case of physical objects; i.e. when two or more similar
data appear under certain more or less specified circumstances and
are not believed to be likely to appear except under those
circumstances, and only one datum of this sort appears under a
much larger variety of often experienced other circumstances, we
9
speak of ‘one’ object, whereas when two data appear under a great
many more kinds of circumstances, but sometimes only one
appears (e.g. if two trees in the distance appear to coalesce into
one blurred whole), we speak of two real objects in spite of the
one-ness of the rare, or blurred, associated datum. It is not a
question of attributes, it is a question of a rule for the use of words
like ‘one’ or ‘two’ in such a way as to supply information which
you check and find correct. We discover the rule by discovering
what in fact we would say.
Sometimes there are funny borderline cases when we aren’t sure
what to say.
For example:
(a) echo: doesn’t count: if we hear a note and an echo or a note

and harmonies, one or two?
(b) like melting headache: does it disappear and can ‘it ’ be
‘suppressed’? I have a pain in my tooth, sometimes it seems
like two pains, sometimes ‘like one’ – which? Like colour of the
sea, sometimes bluish gray, sometimes greenish blue, etc.
Well, which? It alters in colour? And what about different
observers at the same time? It is colourless? Then what do
poets talk about?
(c) Sometimes it becomes silly to ask ‘how many?’, e.g. how
many colours does a continuum of colours – from blue to
yellow – have? I can establish conventions: e.g. ‘Whatever I
discriminate after looking for 5 minutes’, or if A is like B,
and B like C, and A not like C, then A, B, C are different.
But these give different results or just vague [ones].
Conventional, depends on purpose or usage.
Physics cannot answer this – nor does it pretend to. Typical

philosophical puzzle. Now. There are broadly two methods
commonly employed for answering questions: (1) empirical, (2)
deductive.
1. Ordinary empirical investigation. In order to discover what
happens you use your eyes etc.
In order to find out how kangaroos behave in cold climates you
watch kangaroos. In order to discover whether people are happy
under dictatorships, you watch their behaviour, or ask them
questions. In order to discover what happens in the centre of the
earth, you observe other portions of the earth, and bring the
10
results of other sciences to bear on the formulation of general
propositions from which you infer conditions at the centre of the
earth. This needs a lot of logical apparatus, but the data are
empirical, i.e. gained through the senses.
2. The second or deductive method is that used in formal disciplines,
principally mathematics, but also other sciences where the rules
and concepts are defined in advance, e.g. games, or heraldry. What
happens there is that maxims, definitions and rules are laid down
in advance, so to speak, and the practitioner’s skill consists purely
in seeing what is and what is not permitted within the [4]
framework, or compatibly with the definitions of the rules, etc.
You do not, in the case of geometry, if you study conic sections,
try to catch them at it in order to see how the creatures behave.
You do not, in Latin grammar, having laid down the rule ‘Let
adjectives take the same case as the noun which they qualify’,
suddenly expect one fine morning to find various adjectives,
examples in your book of grammar, behaving eccentrically,
rebelling and refusing to agree with their nouns. When I say you
do not find this, I do not mean that it is unlikely that you will, or
even that it is a Law of Nature that you will not, but that it is
somehow nonsense to suggest that you will. Words, numbers,
symbols, chessmen behave as they do because you have laid down
the plan and defined numbers, chessmen etc. as whatever follows
the plan. If somebody one day makes his King in chess move
across three squares instead of one, or breaks a rule of bridge, you
do not say ‘Nowadays chess Kings are more enterprising or mobile
than they used to be, and get further.’ You say ‘I see you have
ceased playing chess; what you are playing, if you are playing, must
be some other game, say Royal-Chess.’
Now what about philosophy? Does it follow the first or
empirical method of noting what happens, and trying to formulate
general rules to cover this, or is it a game whose rules you yourself
impose? Too large a question, too many answers. Metaphysicians
think it is observation of data, but in a special non-sensuous
manner which gets the best of both worlds, i.e. you try to describe
the behaviour, i.e. the structure, qualities, relations of real, i.e. not
invented, entities, but they are guaranteed against ever altering in
their behaviour as much as if they were fictions obeying rules
invented by an author. Kant gave a somewhat different answer.
Certain modern logicians want to say that whatever is not logically
certain – i.e. follows from rules which you have invented or have
11
accepted from others, but which could easily be replaced by
another set of rules – whatever is not deducible from such rules, or
is such a rule, is empirical, i.e. subject to natural science and not
suitable for philosophic enquiry.
[5] (1) The problem may be seen to be composed of two or
several quite genuine problems confounded together, and when a
distinction is drawn, as Descartes recommends, a solution is
thereby at least partially provided. Example:
What is Life?
(1) Tale told by an idiot full of sound and fury, signifying

nothing.
(2) Dictionary definition in biology (chemical process within
organic entities involving metabolism etc.)
(3) Mrs Woolf: ‘life is a luminous halo, a semi-transparent
envelope surrounding us from the beginning of
consciousness to the end.’ 2
(4) Series of actual and hypothetical behavioural data which
differ in certain assignable ways from data defining deed or
inanimate entities.
(5) That which the Lord infused into Adam. See Genesis 1. 4.3
Which?
Mental Cramp.
Who answers question?
Mystery etc. – and empirical one about origins.
Suppose someone says ‘What is the University of Oxford?’
A. Name of a group of buildings of various shapes and ages in
S.E. Midlands.
Yes but Q. How can Oxford have defeated Cambridge? Names
don’t defeat; nor [do] groups of buildings.
A. = a group of persons selected by other persons etc. who
have to be admitted etc. not buildings
Yes well Q. Oxford University has always been a home of lost
causes.
2 Virginia Woolf, ‘Modern Fiction’, in The Common Reader (London, 1925), p.
189; this article was previously published, anonymously, in The Times Literary
Supplement, 10 April 1919, but this passage (p. 189, col. 4) has been somewhat
altered.
3 sc. 2. 7.
12
[6] Does that mean that a group of members of Oxford
University, selected by a certain process – a team were engaged in
working for something which they could not gain? Absurd:
Men – people in Oxford always did that.
How many? When? Who?
Oh. Vague. Can’t specify. Some members of it at almost all
times are doing a Sisyphus.
But what is common? ‘Light’ – cause. ‘Foot’ – genetic. Working
class – similar. Etc.
‘Matthew Arnold wrote a poem about Oxford University’
doesn’t mean ‘wrote about buildings – or name: although he did
say something: not about selected teams: nor about a sufficient
number of persons there at most times’ – etc.
Context supplies meaning. Words vary with context –
conditions: words direct men. We do: we are right etc. What is
common and what is different. True. This a very late discovery: [it
is] because it wasn’t made that much intellectual nonsense
[occurred] and scientists had to put up with so much nonsense
from philosophers.
Descriptive theory: ‘Why does clock strike 12 when Sun is high
in sky?’ [back to 5] 2 questions. (1) Clock mechanism. (2) What
motive has clock? Why do I say ‘It is 12?’ Because I want to. Clocks
don’t want: hence question absurd. Loves and hates of atoms is
earlier view why this not grasped.4 (3) Problem can be shown to be
meaningless and due to a confusion of words or thoughts, so that
when it is analysed it is seen to be such that it can no longer be
taken to be a real question. Example: Can substances interact?
What is the purpose of the Universe? Is time unreal? Is space
crumpled? And if not, what surface has it? Is it smooth? Or
curved? Or crescent like a moon? Etc. Genius means the power of
rendering paradoxes as platitudes. The very fact that they are
platitudes makes them unnoticeable. To realise how much we owe
to Berkeley we must try to project ourselves into the atmosphere
of the seventeenth century, let alone the Middle Ages. Exactly this
sort of question we asked. Descartes liberated us in the matter of
geometry and to some extent in philosophical method; Leibniz on
space and time and certain mathematical techniques. But Descartes
and Leibniz both believed in the ontological argument, that is, they
believed that the existence of something could be deduced from a
4 Very hard to read, and obscure in import; also the position isn’t certain.
13
description of our idea of it. Leibniz believed that if your intellect
was powerful enough you could deduce [7] all the facts of Caesar’s
life. Joseph believed that causality could be deduced from the First
Law of Thought. Spinoza and Leibniz believed that atoms dimly
think, and so does Dr Whitehead.
It is not only philosophers who breed such problems. Children
who ask what happened before the beginning of the world, or
outside it, have to be answered somehow. So do people who want
to know what is meant by saying that the table is a solid extended
brown-coloured object at rest, of a definite size and shape, and at
the same time a whirl of dancing atoms and electrons or gamma
rays or beta particles which have no colour, no smell or taste or
hardness, and shape and size only in a very curious sense of the
words; and when they are told that it is really a collection of
packets of energy moving from one point to another in space,
without traversing the intermediate gap, they cease understanding
altogether. Or else if one asks the plain man whether he does not
think that everything in the world might suddenly become six
times bigger than it is, including our own body, he says yes, it
might; but then how would he notice that? And why should not
everything be becoming much bigger or much smaller, as it were,
all the time, uniformly, so that everything retained proportion vis-
à-vis everything else, and merely escape notice because the
measuring instruments were changing in size too? He has no ready
answer. It is by emancipating people from obsessive questions
such as these – e.g. that life is a dream from which there is no
waking – that philosophy does her work. Berkeley and Hume were
great liberators of this type: Berkeley with regard to matter or
substance, and Hume with regard to causal connection.
Berkeley saw (1) that puzzles arise not because we can’t define
our terms: we can (see Mrs Woolf etc.) but because definition not
= meaning; (2) that ‘meanings’ are not discoverable a priori; (3)
that empircism [is] not compatible with Dualism: Real v. nom[inal]
essences.
I propose to begin with Berkeley’s Principles of Human Knowledge,
which mainly deal with this topic of what the external world,
referred to by us, is, and therefore cannot avoid dealing with the
views of matter prevalent when he was [8] a young man.
What would a seventeenth-century scientist – and scientists
were philosophers in the seventeenth century – have said about a
material object? Matter is that which can be described in clear
14
terms, i.e. quantitative; possesses mathematical and logical
properties, e.g. measurability and causal necessity. These
discovered by non-sensuous intuition. Empirical Knowledge can
be clearer and darker: at one end is total confusion and darkness –
at the other perfect knowledge, utter clarity. All philosophers
believed this till the seventeenth century.
Locke rebelled against the centre of this doctrine, which was that
of real essences. Real essence: the search for the nature of that
permanent, irreducible, ultimate reality which lies behind or under
the appearance; for something which cannot be doubted;
something which when once looked at seems incorrigibly known
to be true. What Popper calls essentialism, something given to a
special mode of cognition, something which the vulgar see only
dimly. When you know it you know it, and if the facts of
observation do not seem to corroborate it, so much the worse for
the facts.
This metaphysical dogmatism is a continental tradition which
was never very popular in Great Britain. Descartes knew, he didn’t
believe or suppose or wonder; he knew that matter was colourless,
extended and had causal properties. Leibniz knew that the soul
could never stop thinking. Kant knew that what was right for one
man was right for ever and for everybody despite appearances.
Hegel knew that the political State was an organism with a right to
absolute obedience on that part of all its subjects. Filmer knew
Kings had divine right. And so on.
Locke could not be so sure. He believed in the absolute
correctness of what the scientists said, but could not be sure why
this was so. He was sure the world was full of tables and
mountains and planets, and sounds and thoughts and persons, but
he could not be sure how he knew this, so he attempted to
discover the sources and methods of information, and to describe
what the scientists or theologians were talking about in terms of
what he was sure he did know directly.
Locke’s account of substance and physical matter

Matter’s primary qualities, extension, figure, number, [9] solidity,
motion, rest, very different from the use of the same words by
Descartes. Account of Locke’s primary sources as wobbling
between empirically given[?] and c[omplex?] powers.
15
Locke’s confusion between (a) metaphysical substance, (b)
logical particulars, (c) reality versus appearance and (d) scientific
substratum:
Substance: what is permanent through change. That of which
the real essence can be apprehended. That whose qualities belong
solely to it. That whose behaviour can be discovered by
considering solely its inner structure or function or purpose. A
kind of island, a self-subsistent closed windowless something.
Something about which everything could in principle be known
and to knowledge of which no information about other things
could make a difference. Something which is what it is, and not
another thing, which could exist and go on by itself even if
everything else ceased to be. In fact what we mean by ceasing to
be is confused with that which cannot cease to be – i.e. the
permanent substance: it can change, because it is that in terms of
which change is change. A substance is something which is
completely independent, describable by a set of words not
applicable to anything else, because it cannot be qualified by
anything which is true of anything other than itself. Consequently
eternal. If it were not eternal the definition of it, which is not in
time, being an eternal truth, might not apply to [it] after it had
perished. But we know the definition to be true, and it cannot
therefore ever cease to apply.
One of the ways of thought which lead to the view that there
are substances, say, as held by Aristotle or Descartes or Spinoza is
the general principle, which crops up again and again in
philosophy, that movement implies rest, change implies
permanence, uncertainty implies certainty: therefore error implies
truth. Pre-relativity notion of Absolutes: Absolute Space: Absolute
Time: without which nothing for exact science to be about.
Develop. Connected with belief in non-sensuous intuition of
matters of fact; if such intuitions are incorrigible, that which they
are about must be for ever and ever what the intuition says it is.
Hence a rigid unchanging world as solid and firm as the
incorrigible intuitions which assert its existence. But if there are no
such intuitions then this [10] world turns out to be a myth or at
least a misleading way of describing the world. This obtains
support from Aristotelian subject–predicate logic, where it was
thought that whatever was a ‘true’ subject could not be a predicate.
If I say ‘The kangaroo is brown’, there is something denoted by
the word ‘Kangaroo’, and something else denoted by the word
16
‘brown’, and the Kangaroo may cease to be brown and still be a
kangaroo, but it cannot cease to be a kangaroo and be a kangaroo,
and will be a something of which the attributes are the attributes; the
attributes may come and go, but the something is permanent and
immovable, and has a nature of its own distinguishable from its
attributes. But if the subject–predicate logic is at fault, i.e. does not
always adequately describe our experience, this by itself won’t be
sufficient to demonstrate this thesis.
Examples: ‘Alexander is the son of Philip.’ Which is the subject?
If so could the attributes of the subject change without affecting
the attributes of any other subject? Or ‘Nine men came into the
room.’ Which of the nine men is the subject? i.e. of which single
substance is the proposition true in particular? A came in and B
came in and C came in etc. If I say ‘Nine men didn’t come in’ –
‘Group of nine’ is not doing something else.
(2) The next notion is that of particulars as opposed to
universals. This is a logical entity and simply refers to the subject
of predicates, which is what names denote. It is what has qualities
and relations. It is that which in some sense exists or is or occurs
even if only in my imagination or thought. It may be analysed as
things or experience or the flow of experience or events or point-
instances or here-nows or whatever you like. Cannot be described
because all words classify; but can be pointed to. If you strip it of
its qualities and relations it will be nothing, yet it is the only thing
that exists. Can be learnt only by ostension.5 It is not what words
stand for at all. Words always describe. If it were not so, all names
would denote characteristics [11] and all propositions would be
analytic for us. ‘This desk in brown.’ Someone might ask ‘Which
desk?’ Answer ‘This brown desk, of course.’ But ‘This brown desk
is brown’ is not very informative. So I say something is deskish in
form and brown. But what is something? Something made of wood.
But then if I say ‘Is this desk made of wood?’ – yes. What is made
of wood? Something brown and darkish; i.e. it is now brown and
darkish and wooden. What is it? In the end, if this peeling goes on,
nothing will be left of the desk.
Particulars are instances of qualities. Or what occurs in
experience (whether or not it is described). Words can refer only
5 Take proper names: ‘Charles is here.’ And you have to say ‘The Charles
who has long hair: the one with the long nose and a squint and all the rest of it.’
But that means ‘The man etc.’ who is ‘called Charles’, and this is an attribute.
17
to what anything has in common with other bits of experience, but
is not itself that which is common to various bits of experience.
Otherwise there would be no bits. Not all statements
grammatically about particulars are logically so; for instance, if I
say ‘Inattentive listeners to lectures are very reprehensible’, there may
be no such inattentive people in the room or the world. If so, yet
something seems to be said about somebody or something. Or if I
say ‘Nobody has ideas of genius nowadays’, it is difficult at first to
see what particular bits of experience or space or time that is
about, for it cannot be ‘nobody’ we say it about. If we mean by
nobody nobody, then there is nobody to say it about. Particulars are
what everything consists of, but can’t be precisely mentioned save
via the qualities that clothe it.
(3) Reality.
Quite different from preceding. If there are substances they must
all be real, but not all realities need be substances, certainly not all
particulars need be real, nor all realities particular. ‘Reality’ very
dangerous and confusing word which does not stand for a single
characteristic as words like ‘blue’ or ‘disagreeable’ or ‘expensive’ on
the whole do stand. Yet out of the proposition that it does so
stand metaphysics has derived its hold and its glory.
Examples: This is a real sheet of paper. That usually would be
taken to mean it is a material object and not an illusion or a mirror
image. This is a real omelette made of real eggs, i.e. nothing to do
with hallucinations; we mean not of egg-powder or some other
substitute. A real bearskin, i.e. not an artificial one, made of plastic.
A real owl, i.e. not stuffed. A real rainbow. A real mirage. A real
man. A real train, i.e. not a toy – not as opposed to hallucinations
or synthetic material. A real disaster. [12] A real event, i.e. not
fictional. But ‘He is really a character in a book’, i.e. not mistakenly
inserted into it by me. An image really in a mirror. Clearly this is a
matter of context, and ‘X is real’ means the word ‘X’ is being
applied in the normal and approved manner, and not something
about a single characteristic which all real entities have as opposed
to real or less real entities which lack it. Old view was ‘Is it real?’ =
‘Is there substance behind it?’
(4) Newtonian Physics and Atoms

What do scientists measure?
What is mathematics applied to?
18
(a) Identical through space and time.
(b) Susceptible to precise description.
(c) Independent of idiosyncrasies of particular observers of
their physical position, mental condition etc.
It is the owner of causal properties. It is certainly particular.
Everything which composes the universe is. That is merely saying
it is something and not nothing. It is not a substance because it is
empirical. If it is empirical, it is in some sense in space and time. It
used to be thought of as composed of indestructible minimal
particles, i.e. atoms, but this is only a particular physical
hypothesis, and physics continues to use terms not drawn from
ordinary life, although this hypothesis is no longer held in its
original form. It is certainly not the same as ‘real’, since the Greeks
and Romans knew perfectly well what they meant by the
difference between a ‘real’ manuscript and a forgery, or a ‘real’
palm tree and a mirage, or a ‘real’ mirage and somebody’s false
idea that he was seeing one when he wasn’t. They knew all this,
whatever theory of physics they may have held, or even if they
held none at all. So they meant by ‘real’ something distinct from
whatever physics may say.
Now Locke, when he talks of material substance, confuses [13]
these four notions hopelessly. I do not wish to go into his
arguments here, but:
Four arguments for four different conclusions:
(1) His view that matter is permanent through change, that all
changes are changes of it, but of an ‘it’ which remains fixed
and unaltered, is an argument in favour of substance.
(2) His argument that when we ask what a cherry is, then we
take away one by one its sweetness, its redness, its
moistness, its particular texture, its spherical shape, we have
nothing left but an unknown something or other, is the
search for bare particulars.
(3) His argument from illusions, that things are not always what
they appear to be; that square towers look round in the
distance, and the same water seems cold to a warm hand and
warm to a cold one, is an argument to prove that there must
be something materially real as against appearance.
(4) His argument in favour of primary qualities as the causes of
our perceptions, i.e. atoms, vibrations etc., is an argument in
favour of the existence of entities which physicists
19
investigate and describe, but which have life histories in time
and space unlike substances. Small wonder that his
conception of matter is so dark and so bewildering.
Berkeley rightly saw in the concept of matter one of the main

obstacles to progress, and attacked it from two sides at once – as
an abstract idea and as something unperceived and imperceptible.
The work in which this done most thoroughly is The Principles of
Human Knowledge. (It might be best to follow the plan of the
Principles, and to start at any rate by taking it as the text to consider
the development of the argument, as he presents it very coherently
there.) Worth mentioning the method he employs:
[14] What emerges from Berkeley’s treatment is that his method
is something like this.
(1) There are certain topics such as theology, morals etc. where
he repeats what his predecessors said, i.e. that it is discovery by
means of a special light with which we are endowed by God etc.
but
(2) With regard to the topics which interest us, namely ‘What is
a material object?’ ‘Does it exist outside our consciousness?’, ‘What
is a concept?’, ‘Has it a counterpart in reality independent of our
thought?’, ‘What is a colour?’, ‘Does it characterise something
within or without us?’ – with regard to these questions Berkeley
thinks the proper method is to examine how we use the words
which are thought to stand for these entities, and he claims that if
you consider very carefully what you mean when you notice such
words as ‘thing’, ‘real’, ‘red’ etc. you will realise that certain theories
about what they are will mean nothing to you, or are even self-
contradictory, i.e. contradictory to the meanings which you
originally claimed to attach to them. Now this is of course Euclid’s
method in Geometry, and so far it is deductive and does not
depend upon the facts of experience. But the original meanings
which the conclusions are said to contradict are not defined as a
rule in a conventional manner. Berkeley, for example, argues that
when you say you feel pain, and when you say you feel heat, you
may be referring to the same sensation. Whether you do or not is a
matter which you can settle only by, so to speak, looking within
yourself and considering whether you do or don’t, not by looking
up the original definitions and rules at the back of the book. The
[15] method may be deductive, but the data are supplied by
experience, experience of what we mean, i.e. what we want to say
20
by using this or that word. And this, in a sense, is an empirical
investigation of how people use such words as ‘colour’, ‘a physical
object’, ‘perceive’, ‘is’ etc, and whether anything can be discovered
from this about themselves, or the world. An investigation on just
this subject is what Berkeley begins with when he attacks the
notion of abstract ideas.
[16]
Lecture II (23 January 1947)
1. Berkeley begins The Principles of Human Knowledge by saying
that, before getting on to such large topics as ‘What are things?,’
‘What are minds?’, ‘What exists?’ etc., he must say something
about the part played by words. He thinks if any problem seems
unanswerable in principle something is wrong in the formulation.
It is only we who put difficulties in our own path in the matter.
We should believe that God has dealt more bountifully with the
sons of men, than to give them a strong desire for that
knowledge, which He had placed quite out of their reach. […]
Upon the whole, I am inclined to think that the far greater part,
if not all, of those difficulties which have hitherto amused
philosophers, and blocked up the way to knowledge, are
entirely owing to ourselves. That we have first raised a dust, and
then complain, we cannot see.6
Hence the need for looking for false principles of enquiry as the
source of our difficulties (Important. Pseudo-questions: if we
understand a question we must know what kind of answer it has –
otherwise meaningless), and this involves us in considering the
‘Nature and abuse of language’. One of the major abuses of this
sort is that which involves belief in the existence of abstract ideas.
Normally if we ask what a thing is – say a Kangaroo, or mate in
three moves – we can produce a specimen. Wherever it is a
thought, e.g. in the case of entities like ‘Oxford University’ or ‘life’
or Truth or Reality, we appeal to abstract ideas. i¸ppon men o¸rw,
i¸ppothta d¡oux o¸rw.7 We intuit ‘horseness’, an abstract idea. In
sections 7, 8 and 9 of the Principles Berkeley gives a very reasonable
6 Principles, introduction, paragraph 3.

7 ‘Hippon men horō, hippotēta d’oukh horō’ (‘I see a horse, but I don’t see
horseness’).
21
account of what Locke, for example, must sometimes have meant
by, say, the ‘Idea of Colour in Abstract’. That which is common to
red, blue and white in virtue of which we call them colours,
according to Locke, is itself neither red nor blue nor white nor any
other determinate colour. Now this cannot exist, but we can think
of it in some sense. Nor can colour exist without extension, yet the
mind can ‘frame to itself by abstraction’ the idea of one of these
without the other. The notion of ‘man’ or ‘humanity’ or ‘human
nature’ leaves out the ideas which it has of Peter, James and John –
those ‘circumstances and [17] differences which are determinate to
any particular existence’. Motion is ‘neither walking nor flying nor
creeping, nevertheless it is motion, though difficult to conceive’.
Berkeley observes of this: ‘Whether others have this wonderful
faculty of abstracting ideas they best can tell. For myself, I dare be
confident I have it not.’ He then says that he can imagine new
combinations of the original ideas. ‘I can imagine a man with two
heads or the upper part of a man joined to the body of a horse or a
nose or an eye by itself.’ But ‘I deny that I can abstract from one
another or conceive separately those qualities which it is
impossible should exist so separated or that I can frame a general
notion in abstracting from particulars in the manner aforesaid.’
Locke said that the capacity for abstraction is what divides men
from brutes. For ‘brutes abstract not’. Berkeley fears that, if this be
the criterion, then a ‘great many of those that pass for men must
be reckoned into their number’. The reason why brutes are
excluded is, it seems, because they do not appear to use words or
symbols. But it can follow only that what they lack is power of
abstraction if ‘the use of words is impossible without having abstract
general ideas’. What sort of thing is an abstract idea? Berkeley does
not deny (section 12) that we may in some sense have ‘general
ideas’. ‘An idea which considered in itself is particular becomes
general by being made to represent or stand for all other particular
ideas of the same sort.’ The geometer’s particular black line, an
inch long, becomes general by ‘being made a sign’, ‘so the same
line by being a sign is made [18] general’. A line is general because
it refers to ‘the various particular lines which it indifferently
denotes’. And again, ‘A word becomes general by being made the
sign not of an abstract general idea, but of several particular ideas,
any one of which it indifferently suggests to the mind.’ The word ‘any’
is going to give trouble later. He then quotes Locke, book 4,
chapter 7, paragraph 9 of the Essay on Human Understanding, where
22
Locke (very unfortunately) says that ‘A general idea of a triangle is
neither oblique nor rectangular, neither equilateral, equicrural nor
scalenon, but all and none of these at once.’ But this, so to speak,
jumbles all these together in its haste to communicate some
general notion. Berkeley mocks at this and says that if anyone
believes that they can perform such an operation he can think of
no argument against it, and invites the reader to look into his own
mind and see whether he can find something in it which is neither
oblique nor rectangular, neither equilateral, equicrural nor
scalenon, but all or none of these at once. Of course Berkeley is
right: it is to say the least difficult to frame such a notion. Not
difficult but obviously quite impossible. Something can certainly
not be qualityless. ‘Particulars can be bare’ is self-contradictory.
Locke says that it will be found a hard task for children – a hard
task for that tender age. ‘Is it not a hard thing to imagine that a
couple of children cannot prate together of their sugar plums and
their rattles and the rest of their little trinkets before they have first
tacked together numberless inconsistencies and so framed in their
minds abstract general ideas and annexed them to every common
name they make use of’?’ Of course there are universal notions,
which merely means that ‘a particular triangle doth equally stand
for and represent all rectilinear triangles whatsoever’. This is
considered quite different from abstract ideas (read quotation from
Sections 18 and 19; also marked bits in sections 23 and 24). Now
the difference between general ideas and abstract ideas is obviously
crucial. Suspend this. Consider what Berkeley wants to say first.
[19] Berkeley’s principal motive for discussing abstract general
ideas is to kill the view that there are entities in the external world
not perceived through the senses. If there are such, no telling what may
not be postulated. Of course Locke had routed innate ideas as
advocated by Descartes or Leibniz, and showed that we ‘did not
discover such within us, implanted from birth’, but Locke had not
gone far enough, and was still deceived by words. The war
between intuition and empirical knowledge must necessarily be
war to the death. You cannot have it both ways. If the mind is
capable of deriving information from sources other than the
stream of empirical impressions, it can construct an entire world
therefrom. The proposition that it can do this is derived from the
view that certain words have no obvious empirical counterpart.
While most sync[ategorematic] words have empirical counterparts
– ‘red’, ‘sweet’, ‘mountain’ etc. – there are words – like substance,
23
cause, infinite divisibility, square root, triangularity, humanity which
have none. But they mean something. Therefore they must have
some counterparts in the external world of which they are the
proper names; if these words are proper names, their owners, since
they could not be seen, smelt, heard, touched etc. can be made to
present themselves only by some other faculty: the named entities
were non-sensible, the faculty was a non-sensible mode of
acquaintance, the science that dealt with this was ontology or
metaphysics and subdivisions of it were physics, mathematics,
logic etc. Innate ideas might not exist; Locke might very well have
shown that infants at birth or untutored American Indians or
congenital idiots did not enjoy innate information; it comes only
later with adequate mental development. It is, as Leibniz believed,
potential. In mind, not before it, etc. But no matter how or
whence it comes, if it comes at all, non-empirical knowledge is
possible, that is knowledge which is neither of the data of the
senses nor yet knowledge of how words and other symbols are
used; awareness that is neither of particular colours or scents or
sounds [20] nor yet of the rules of language or the rules of chess or
the rules of mathematics, all of which might be represented as
man-made. And once that is allowed, the enemy is within the
gates. If non-sensible information was a principle allowed to be
possible, then how was one to deal with persons who maintained
that you see brown extended patches and touch hard extended
surfaces, but know in some other way that these are merely
appearances of ‘real things’ which you know are there in some
other way? That in hallucinations or dreams, or indeed in the
stream of images or daydreams which a hearty visualiser could
summon up at will, what is lacking is precisely the presence of this
hard, non-sensuous, substantial core? That and nothing else is what
is meant by distinguishing your data into appearance and reality,
into the physical world and the world of dreams and imagination.
And then would arise the appalling problem which faces anyone
who takes any such view as this of answering the question. ‘What
is the relation between the two worlds? Is it resemblance? Or
causal? Or special and unique? And how can you tell what if any
relation it is without seeing both ends of the relation at once,
seeing which is appearance and which reality in one act of
cognition? But if you see both, why should you bother with
appearances at all? Why should you bother with the senses at all?’
That is what Plato and Descartes, St Thomas and Spinoza all
24
thought. But if ordinary men certainly did not have the insight
which Plato and Descartes said they possessed, and did not see the
superior entities of the purely rational world, we are back with the
choice between one or two mutually exclusive sources of
knowledge. If the second, metaphysical, one exists, the first seems
unnecessary; if the first, i.e. the empirical, tells us all we do in fact
know, the second is either a shadow, an unnecessary copy of it, or
unintelligible, and this latter is in fact Berkeley’s position.
[21] Berkeley read the metaphysicians, and apart from his
theology and doctine of Selves and Relations, which is another and
later story, convinced himself that he attached no meaning to their
concepts, and did not posses the faculty of intellectual intuition,
superior to and different from the senses. How did they come to
make the vast mistake of thinking they had this power of almost
mystical vision when in fact they did not? Berkeley thought he
solved the mystery. It was due to a gigantic fallacy about how
words mean. The fallacy is one I mentioned last time. It is that of
unum nomen unum nominatum, i.e. that corresponding to every
syncategorematic word there is always some one entity for which it
stands – of which it is the name; this presupposes that language is
unambiguous, that words neither alter their senses nor stand for
related but different entities or are used in different ways to mean
differently in different contexts. He has in mind persons who
really think that in order to be aware of anything at all I must (as
used to be said) apprehend the universal in the particular; or
‘extract’ universals from particulars; a kind of process which leaves
particulars bare and unrelated like a kind of slag-heap of coke after
the valuable substance has been taken away from it, the valuable
stuff being the universals which I extract, strain off like a precious
juice from the grape; people who think that tend to argue that if it
weren’t so, I should not even be able to think of things as being
‘green’ or ‘grass’ because having seen green grass once, how could
I recognise it as green the second time I saw it? Unless I had in my
mind some notion or concept which I could apply like a
measuring-rod or a standard to see if it really is green and grassy?
And what is this but the ideas of green and grass, abstraction of a
universal, extracted from out of the first green grass I saw, retained
ever after in my mind’s eyes, which is a non-empirical intuition of
what it means to say that: [then not that?]. So, you see, a non-
empirical intuition is required even for empirical knowledge: but
25
this once granted may have objects of its own, number, substance
and all.
Again: words, we said, classify: to call a thing grass is to put it in
the same category with other things also called grass; what all
patches of grass have in common is grassiness. And I do not count
even[?] grassiness as I do this[?] grass – but unless ‘grassy’ meant
something to me, ‘grass’ would not. But it does: therefore ‘grassy’
does. So I must have a faculty for apprehending it which functions
whenever I identify anything as ‘such and such’. Of this I can say
only that in that event the brutes dismissed by Locke, who ex
hypothesi cannot perform such intellectual non-sensuous operations,
must be able to taste or smell abstractions, for unless we think
animals automata we certainly think that it is the fact that there is a
common quality between two experiences, say the look of milk
that makes the cat go for the milk or the dog recognise its master
the second time it sees it/him, or makes it avoid the fire after
singeing its claws once. Animals do ‘recognise’ the sound of a bell,
or the looks of things: by intuition? But this is irrelevant. This is
not a lecture on logic or the nature of universals and I cannot
therefore discuss the general problem involved, but this much
must be said: Berkeley must get rid of non-empirical analysing of
what it means to say that a given triangle becomes generalised by
being made to stand for any other triangles, or a given patch of
grass for other patches. The puzzling phrase is: ‘being made to
stand for’ or ‘standing for’. How does something ‘stand for’
something else? Because we make it do so. Words do not mean by
nature, nor [22] do geometrical designs on blackboards naturally
stand or fail to stand for anything. They are what they are, visual or
auditory experiences.
Now if I say of a given geometrical pattern, ‘Let it stand for all
triangles indifferently’, there are two difficulties at once.
1. In virtue of what do they do this? The triangles are
composed of thick black lines, let us say. Why then do they not
stand for a particular brand of black paint or draughtsman’s
charcoal? They are drawn on, say, paper. Why do they not stand
for a particular brand of paper as opposed to cloth, or something
else? And so forth. In other words somebody might say that it is in
virtue of only one property or a set of properties that the given
triangular patch in my visual field stands for triangles in general,
and this phrase ‘in virtue of’ is then what the defender of abstract
universals would say refers to the presence of, is the name of,
26
universals. That, according to logical realist or conceptualist, is
what this triangle and the triangles it stands for actually have in
common, and they actually have it, this single common property,
triangularity; the individual triangles are so many particular entities,
like headaches, or sensations of red, or loud sounds or bitter tastes.
I do not invent them, I come across them. And triangularity is
something else again, and they all have it in common, and I don’t
invent that either, I find it in the triangles, though what ‘in’ means
it is harder to say.
Berkeley gives countenance to such talk. In talking of triangles
Berkeley does not mean, or should not mean, what modern
mathematicians would mean by geometrical triangles, for these
obey rules which according to some theories are invented by us,
like the rules of chess, and if so have only that in common which
we put into them. We don’t learn about this by looking for
common qualities they have with other triangles any more than we
learn about Kings in chess by examining hundreds of chessmen.
He seems to mean the vaguely triangular shapes one might meet
with in nature, say pyramids or the flaps of envelopes; and the
difficulty is what is meant by saying that a given triangular shape
stands for, even[?] ‘in respect of’ its triangularity. Now the word
‘triangle’ doesn’t have anything in common with the triangles it
stands for, and therefore the [23] question ‘In respect of what
special characteristics of it do I use it to stand for triangles?’ does
not arise. I use any symbol I please: anything will do if I make it
serve me thus. But the difficulty lingers. The triangular shape
stands for whatever this, that or the other triangle have in
common. What does this ‘whatever’ stand for? What do entities
called triangles have in common? And if, as Berkeley thinks, there
is no one single ‘it’ which they all have in common, and which
‘triangle’ is the name of, what is meant by saying that it is proper to
use any one symbol for them all? Berkeley’s answer is that the
relation of ‘standing for’, ‘in virtue of’, etc. is really founded on
that of similarity between the various instances and this is very
important. First glimpse of light in the wood. (Note: What
Nominalism is; and explain difference between two types of
nominalism – where terms stand for similarity and where they are
purely arbitrary, i.e. collective noun: no general term can be purely
arbitrary: and no term if it is a token of a type.)
But if similarity is the cardinal relation, then the objection some
philosophers bring against it is something of this kind:
27
You say that you call two patches of colour ‘red’ or ‘scarlet’ or
even ‘this particular shade of scarlet’ because they resemble each
other or look similar. In other words you christen some particular
constituent in your experience with the sound or mark on paper
‘red’ as a name and say that anything which resembles it will be
called ‘red’ too. ‘Red’ means ‘like the specimen’. So far so good.
But now what about the word ‘resembles’ (let us use one word,
‘resemble’, and drop ‘similar’). How is that used? By analogy with
the way in which we explained how the word ‘red’ is used, we have
to say that the word ‘resembles’, like all words, is used to denote
that something resembles something else: in this case whenever
two or more relations of resemblance in their turn resemble each
other. But if so, the definition is either circular or leads to an
infinite vicious regress. Either the term ‘resembles’ is defined in
terms of itself, and you say ‘Two things resemble each other’ when
the complex consisting of these two things ‘resembles’ some other
complex originally christened as the model or specimen of [24] the
resemblance relation, and that the word ‘resembles’ is used
throughout the expression in the same sense; in other words if ‘X
is red and round’ = ‘X exactly resembles the original sensation we
agreed to call red’ then ‘X resembles Y’ = ‘The X–Y relation
resembles the relation we agreed to call resemblance.’ To get out
of this circle we might try and hold that the sense in which ‘this
red patch’ and ‘that red patch’ resemble each other is a different sense
of ‘resemble’ from that in which the resemblance of these two red
patches and the resemblance of those two blue patches resemble
each other. But then you have to say that the special sense in which
only the complex entity or pattern consisting of two red patches
and the complex consisting of two blue patches are said to resemble
each other is given by the archetype specimen in which these
resemblances resemble; namely, there has to be postulated a
supercomplex consisting of the two single complexes; and there
has to be postulated a new relation of resemblance between this
supercomplex and some other supercomplex also consisting of
two sets of two characteristics tied by the same network of
relations of resemblance. But of course the two supercomplexes,
in order to resemble, must be conceived as connected in a super-
supercomplex which has its own relation of resemblance, still
[mind?] out, sensibly observed, with another super-supercomplex,
and so on to infinity. And this seems not merely complicated but
definitely wrong.
28
a. Because the regress is infinite and proceeds backwards. Each
foundation crumbles as the foundation under it needs propping.
Each case of resemblance means something else, and we cannot
even say they resemble each other in resemblingness – because
that yields a new resemblance. tritoj a¹nqrwpoj.8
b. It is not true. I can say that I know what I mean when I say
that two sounds which I do not name, i.e. even classify, are exactly
like each other, and if I am told that I do not mean that unless I
mean by likeness this infinitely regressive series of resemblances of
resemblances of resemblances, then I must protest and say, ‘But I
do not mean that, for I am sure that the two sounds sound as one
when they occur together, but I neither know nor think about the
resemblances between combinations of combinations of a
increasingly elaborate kind in which this sound might possibly be
involved.’ I must, as Berkeley says, be allowed to know what I am
trying to say, and reject any attempt to represent me as stating
something which I am not stating.
1. This is the objection to similarity analysis urged by Russell
and others, and it is pretty formidable. Are we then in despair to
turn back to Plato or Aristotle and speak of ‘real universals’? But
relations between them and their instances, and between them
among themselves reproduce these difficulties in much worse
form. Perhaps the answer to it is somewhat along the following
lines: that awareness of resemblance is not a judgement asserting
that two things have been compared in terms of a criterion which
is itself outside them (which therefore needs accounting for,
justifying etc.) but is immediate (develop a little).
2. Less formidable is the objection that Berkeley when talking
of ideas means images, that this is his reason for being perplexed by
abstract ideas: for of course nobody ever thought that I could have
an abstract and general image of something. Of course I cannot
have an image of something which is not red or blue or green or
coloured. Of course I cannot have an image of a circle of no
particular size or position or thickness etc., but general ideas are
not images, they are thoughts. Not all, perhaps not much, thinking is
thinking in images. When I think of a myriagon, i.e. a figure with
ten thousand angles, I do not imagine it. Or, if I do, my image of a
figure with ten thousand sides and my image of a figure with nine
thousand nine hundred and ninety-nine sides are not very
8 ‘Tritos anthrōpos’ (‘the third man’).

29
different. Yet this does not prevent me from realising that their
properties are very different. But because Berkeley talks loosely,
his opponents cannot score as heavily as they think. Berkeley’s
thesis is that knowledge may be analysed into ideas – = sensations
and images and any other empirical data – and words. What this
means is that any proposition claiming to assert a matter of fact –
not a law or a logical truth, but to give information about the
world – can be analysed into basic empirical data and compounds
thereof – Locke ‘simple’ ideas. I cannot imagine (i.e. I cannot see it
as) a myriagon. But I can both imagine and give rules for
constructing a pentagon: and I can imagine what it would be like
to have a powerful enough intellect to imagine a myriagon. The
image of one is not absurd. Whereas that of something without
sensible attributes is. That is Berkeley’s case. (Indicate that
nevertheless the basic currency is empirical = ostensive.)
3. A far more fatal difficulty, on any empirical analysis of
language, is about general terms such as ‘all’, ‘any’, ‘if’, propositions
about the past, and notions such as those of material things,
persons and scientific entities. These difficulties are not solved by a
non-nominalist view of universals. Let us consider one of these.
Generality[?].
[26] It seems clear that by his method of killing abstract ideas
Berkeley kills off too much. After all, it might be possible that,
although the data of experience could be defined in purely
sensible, ostensive ways – reds, blues, hard – and soft – touch
sensations, scents and flavours, sounds and inner images, muscular
sensations and emotions such as discomfort, rage, surprise, hate,
perplexity, sense of similarity or congruence etc. – although all
propositions of a factual kind might be reduced to this, the
relations between the data might be of a special or peculiar nature,
i.e. in some sense not derived from simple data of sense or
introspection or memory: that is, what propositions are about
would always be empirical, but the structure of what they were
about, whether it was hypothetical, [ ], or past or future etc. might
be due to something like thought or reason, e.g. coloured or
extended. All Ideas concrete.9
This is a very primitive way of suggesting some of Kant’s later
theories, and it may lead to conclusions repugnant to Berkeley’s
9Transcript corrected up to this point by Henry Hardy, hereafter by James

Chappel. HH will correct the rest in due course.
30
empiricism, but the reason for it is that his form of it involved
difficulties of a scarcely less formidable nature, (and although this
takes us away a little from the external world) it is just worth a
glance:
1. We have already dealt with difficulties about similarity. Now
consider difficulties about generality. How are we to analyse
propositions about ‘all x’ or ‘if x’: what ‘idea’ corresponds? ‘In the
higher [ ] all cats are grey’: I have idea of this cat or of a very large
number of cats: e.g. resemblances between the various cats I think
but even if number of cats in mind is finite and I have labelled it
all, ‘all cats’ now cover all possible cats [ ] infinite number possible
and I cannot collect infinity: all men, this and this and this and this
etc and etc’ cannot be diluted – into a finite collection: I cannot
point to a collection and say ‘that is what I mean by “all ” ’. I can
say ‘A, E, I O U Y’ are all the vowels of the English Alphabet: but
that means ‘nothing not there is a vowel etc.’ and nothing [ ], not X
nor B, no letters [ ] – and that is infinite. ‘Lecturers who talk too
rapidly are apt to be misunderstood’. Who are these lecturers? [ ]
We cannot tell: We cannot run them all: and say these are all the
lecturers there are. [ ] [There] could be others.
2. On B’s view geometry becomes odd, very odd. We have just
seen that the notion of infinity is unintelligible, because again we
can never say ‘etc.’. This perhaps is a special case of generality. But
worse is to come. If I say that a given visual surface is infinitely
divisible that means that it has an infinite number of parts. What
can this mean for Berkeley? Obviously the number of parts which
I can actually distinguish with the naked eye is finite and I could, I
[27] suppose, count it if I was careful enough. But only that which
I can discriminate can exist, for to say that there is something else
is all to him nonsensical and therefore meaningless. Therefore I
cannot infinitely divide and subdivide an empirical datum or an
empirical surface. Therefore surfaces do not consist of an infinite
number of parts, but only of so many. But as geometry, according
to Berkeley, (and algebra,) do purport either to describe or
symbolise the world, the notion of infinite divisibility is an absurd
and meaningless one. This odd conclusion he boldly embraces. Of
course someone might say ‘What about an eye differently
constructed from ours, say the eye of a fly (Berkeley speaks of a
cheese mite), to which an inch on a table surface may present
many more discriminable data than to our eye; and what about the
bit of the table seen magnified by the microscope’, again many
31
more data etc. To this Berkeley replies that the fly and the man
with the microscope literally do not see the same thing as the normal
observer with the naked eye. What the microscope reveals is
literally bigger and with more parts than data of naked eye. And we
say that what we see through the microscope is the same as what
we see without it in a very peculiar sense of ‘same’. It is not literally
‘same’. It is ‘same’ in a special sense of ‘same’. But that we shall
have to deal with in discussing identity of material objects. In the
meanwhile if five lines are not infinitely divisible, the doctrine of
infinitesimals on which the differential calculus of Leibniz and
Newton is based, as well as the doctrine of the incommensurables
which it is meant to solve, go by the board. (Give brief exposition
here of irrational numbers). E.g. if extension equals collection of
minima sensibilia, only those lines can be bisected which contain
an even number of them, i.e. a line containing an odd number
cannot be said to have two halves at all: v. odd. Berkeley literally
says that 1/10000 part of an inch doesn’t exist: and that explaining
what we mean (1” = 10000 miles, which is 1 mile = ?) contradicts
himself. [28] Now this, although it seems absurd, cannot be
brushed aside. There is a sense in which Berkeley is right, and
people do not see an infinite collection of points or magnitudes
infinitesimally small, and the application of or relation of a formal
system like arithmetic or algebra to the sensible world are in any
case a problem which the Greeks failed to solve, and the solution
of which by modern thinkers is not so clear as to be self-evident. But
the conclusion Berkeley reaches is that Newton and modern
mathematics and physics are simply wrong in using such concepts
as infinitesimals. That everything is made up of a finite collection,
either odd or even, of minima sensibilia not further sub-divisible,
but varying from time to time from place to place, from observer
to observer, from type of eye, brain etc. is kind of sensationalist
atomism, at least equally paradoxical and if worked out, absurd.
(How many? How to specify conditions?) It makes mathematics an
empirical science, and all proofs approximations, as Mill finally
said it was. And whatever else it is, we are clear that it is not that.
All this is done in order to get rid of arguments for the
existence of abstract ideas, under cover of which there is a liability
to smuggle in the mysterious and undesirable sub-stratum matter,
alleged by other philosophers to be the object of our knowledge
and the cause of our sensations. Words cannot show for non-
entities. Hence no abstract ideas. The world consists according to
32
Berkeley of sensations and selves, and symbols, and also God. But
what then misleads us into thinking of abstract ideas or words A:
World.
Let us briefly deal with symbols next.
There are 6 views in Berkeley of what these are:-
1. They are specific objects.
2. They are images.
3. They are names.
4. They are meanings.
5. They are signs.
6. They are notions.
On the first view anything may stand for anything. This
observed desk ‘stands for’ all other desks. ‘Stands for’ means ‘I
make it [29] stand for’. Thinking then means shuffling about
objects in this sense. Very much as I count on the fingers of a
hand where the fingers are the symbols. But (a) manipulation
might be too clumsy to make it useful frequently (b) we have no
way of knowing when arrangement ‘natural’ and when
‘conventional’ (c) Confusibility with representation. Several Terms
– ‘If’ etc. not very easy to represent with desks. – not desks e.g. or
if desks, or some desks. Easier than desks, but not too many[?] of
that and (c) still should.
2. Images. Useful stress by Berkeley on images for images’ sake
(artists and daydreamers) versus images as deliberate symbols.
3. Symbols are names. This won’t do, because names can be only
labels attached to specific entities, and not to classes defined in
terms of a property which if [it] has [a] name of its own becomes
Berkeley’s worst enemy, – an entity, the abstract universal. Fido
can be name of this dog, but ‘Dog’ name of what? Caninity? Which
is what?
4. Meanings – Symbols are meanings is even worse if what words
stand for are unqualified are ‘meanings’. What are meanings if
they’re not ‘meants’ – entities – non-sensible objects of the mind.
(golden[?] mountains) [ ] objective of false or nonsense
propositions.
5. Symbols are Signs. If this merely means that they in fact conjure
up whatever the words are thought to refer to, it will not do, since
we distinguish between precise meanings of words and their casual
causal associations.
6. Notions. This is a valuable and interesting idea indicating that
Berkeley realised the need for generalisation, but at the same time
33
wished to guard against the idea of having to have objective
counterparts of general terms i.e. of looking on all, including
general terms as names of things in the world; which would entail
that some entities – named by general names have no particular
but only a general nature i.e. there are universals as entities; this is
thought to be absurd if only because of the infinite regress which
Plato and Aristotle had pointed out (if universals are entities then
relations between them are entities too and require further
universals to relate them. Also Entities. Midas Touch. While found
in fallacy that meaning is transitive correspondence. ‘x’ means x
means if I want[?] ‘x’ [ ] etc. You will grasp the X whether it is so
or not: what ‘grasping that X’ is is another question, which
involves [the] whole of metaphysics, theory of knowledge and
logic. Fortunately irrelevant.
Things.
Having got rid of, as he thinks, abstract ideas, and determined
therefore that whatever ordinary men or physicists mean by
material objects cannot at any rate be abstract, Berkeley [30] is at
last in a position to spring his great new principle on the world.
‘The obvious though amazing truth’ as he calls it in the
Commonplace Book, of ‘Esse est percipi’ – ‘to be is to be perceived’.
How is this reached? ‘Consult, ransack your understanding’. How
Berkeley ransacked his understanding has been revealed to us by
his Commonplace Book, which was not meant for publication. He
began thinking about the nature of Time, and observes, ‘Time is a
train of ideas succeeding each other’ (Commonplace Book 158)
and then ‘duration not distinguished from existence’ i.e. whatever
exists last, whatever lasts exists. But duration is longer in pain than
in pleasure, so measuring time differs from moment to moment,
and from individual to individual. The same tō nûn (the now) not
common to all intelligences’, therefore ‘Time a sensation, therefore
only in ye minde’. ‘Tempus est percipi’ but ‘esse est tempus; therefore esse
est percipi’. That’s how the startling conclusion is reached. No doubt
there are a thousand contexts in which the word ‘Thing’ or the
word ‘Matter’ could be employed, but what we are after is what
people mean by material objects as distinct from mere ideas or
mere images or illusions and the like. Berkeley uses the term idea
roughly to mean what we should now mean by sense data, of ‘sense-
experiences’. He can not mean that ‘to be’ and ‘to be perceived’ are
interchangeable terms although he often talks as if he did. That is, he
does not mean the proposition is analytic or a tautology. And he
34
does not mean that because he thinks that there is something in
the word besides ideas or sensations i.e. percipient minds, these
minds clearly also exist, they have ‘esse’, but they are not necessarily
perceived. Consequently ‘esse est percipi’ is either not an analytic
proposition i.e. to say of a thing that it exists without being
perceived is not self-contradictory, but merely false in some way,
or else [ ] it is ambiguous and ‘esse’ means something quite different
in the case of sensations, and therefore physical objects on the one
hand, and of selves or spirits on the other. It becomes fully
analytic only in later e.g. Müller[?] or Ayerian phenomenalism.
At first whatever sense we take ‘esse est percipi’ in, whether
analytic or not, it seems an obvious paradox. The following
immediate and obvious objections spring to mind as they have to
those of many subsequent philosophers:
1. When I say that I perceive X – say a tree, do I mean that I
see it, or do I mean that I touch it? And is not ‘it’ common to sight
and touch and if so is there not something which is neither seen
nor touched, but is the same whether seen or touched.
2. What happens to X when I am not seeing it? Does it
disappear, clicking in and out as I open and close my eyelids?
When I say, upon returning to a garden after an absence ‘I see the
tree is still here’ do I really mean ‘Here it is again’? Not ‘still’, but
‘again’. Or if my fire has gone out in my absence, do I say, ‘Oh, I
should have remembered to keep my idea or image of the fire
vivid and warm in my mind – or before my eyes; because I was
thinking about and looked away at other things it went and
expired’?
3. What about my own body. Does that bob in and out of
existence as when I or somebody else choose to indulge in an act
of sensing or to cease doing so.
What about Exists? When I ask ‘does the material table exist’ is
it like ‘Does Queen Victoria exist’ [or] does the Queen of France
exist? Or Does Queen Semiramis exist? Or does Fairy Queen
(Mab) exist? Or does the Queen of Spades exist? Or does The
Queen Bee exist?: Does the [ ] Man exist? [32]
4. If the tree is described as ‘my own sensations’ how can they
also be somebody else’s. Is there one tree or two (2 observers)? My
tree and yours? If one, is there one headache (provided at least one
observer feels it)? Is there not something absurd in saying that I
can feel someone else’s headache by mistake in the sense in which
I can pick up their overcoat by mistake (‘How ghastly, I have
35
walked into Jones’ headache, and that always lasts a few days’).
This is impossible: [ ] impossible? Anyway: 2 headaches. One not
substitute for other. Surely visual data ditto.
5. ‘Exists’ There are two main questions which emerge from
this:
(1) Are we in some way compelled to assert that beside our
sensations there exists, as sensations exist, something else, whether
or not in specific relations with our sensations – but quite unlike it
and insensible? And more radically, what would make us believe
that this is true? Or false? And further still, does it make sense to
say that? Is it intelligible? A fortiori, if it is not sense it will not be
true, and if it is not true nothing other than psychological causes
can compel us to utter a set of words which turns out to be
meaningless. And that is Berkeley’s central thesis about material
substance.
(2) What is meant by saying that something – possibly sensible
but unsensed, continues to exist though not actually and
continuously sensed or perceived by us? And to this question,
which cannot be evaded on any theory of knowledge, Berkeley has
several conflicting views. The best way of treating this whole topic
is, I think, not by selecting Berkeley’s various replies subject by
subject, but by going through the Principles in Berkeley’s own
perfectly logical order and considering what he says point by point.
This is both an exercise [33] in philosophical method and
illuminating in itself. (There are 33 distinguishable points, each of
some interest and importance, which Berkeley makes, and in one
form or another they cover everything of interest that he had to
say. To these we must apply ourselves. They deal with arguments
or psychological causes which make people suppose that matter
exists. There are some inconsistencies in Berkeley’s answers,
nevertheless the argument in general is very coherent. The
remainder of these lectures will therefore be divided into 33
sections of very unequal importance.) I propose to treat the most
important arguments of subsequent philosophers, e.g. those of
Kant, Moore, Broad, Stout etc. under the relevant heads, but I
shall try to preserve Berkeley’s order as far as possible.
[34] What is matter? We don’t know but scientists use it. The
position is one rather like mysteries in Religion: we believe certain
sentences to embody true propositions but don’t know, or are not
sure, what they mean: all we know that they mean something and
that this is true.
36
Berkeley’s method consists in beginning by supposing that
something might be meant by matter, in order to see whether, if we
leave it so, that is, even if we simply leave the meaning of this
expression vague, and say that it is whatever helps philosophers
and scientists in the way in which they think it does, we can then
show that such a postulate, (it has any meaning) could not even then
perform the function claimed for it either because it is meaningless
or, if it has meaning, i.e. if what is described exists, because it does
not explain anything. Leaving aside for the moment the notion of
what is meant by attributing ‘causal activity’ to matter in the sense
in which Berkeley says that ideas are passive and only minds are
active, the first point that he makes is that some philosophers
think that matter must exist to be the possessor of primary qualities.
Locke had spoken of the difference between primary and
secondary qualities. The matter is in great confusion in Locke.
Descartes’ primary qualities are of course not sensible at all. In
Locke they oscillate between being sensible and being powers or
causes in things which make us aware of sensible qualities. The
primary qualities are: extension, figure, impenetrability, number,
motion and rest. In Descartes they are clearly not sensible, and
intuited by the powers of natural light or reason with which God
has endowed us. Locke, who rejected such non-sensible powers
but retained the distinction because it seemed sustained by science
and commonsense, had to find sensible meanings for these terms.
Consequently there emerges a curious view that whereas the
colour, taste, smell, [35] sound, touch, and hardness of a thing
somehow depends on the observer and are subjective, a shape, on
the other hand i.e. atomic structure, motion, number etc. are not
subjective although all seem equally discoverable only by the
senses.
Let us take Berkeley’s refutation of Rep. Perceptions from
fearful mare’s nest: Locke maintains for reasons which will appear
that my ideas are confident to me. I cannot have yours. But they
resemble entities which cannot be directly perceived. ‘Ideas’ are
sensible experiences: and the ideas of secondary qualities resemble
nothing and are mere by-products in me. Ideas of primary qualities
do resemble. Berkeley has no difficulty in disposing of this. If the
primary qualities are not sensible, then what can it mean to say that
they resemble ideas which are sensible? Can statue of J. Caesar
resemble non sensible original Caesar? If my so-called idea is
visible, whatever it may resemble it cannot be invisible:
37
resemblance, by definition entails common qualities. Even if
commonly that = resemblance. A colour patch cannot resemble
the invisible, a flute-y sound cannot resemble the inaudible. If
there are such things in nature as entities not firm to the senses,
then what is perceptible to the senses cannot be known to
‘represent’ them. If it were, you could not know this without being
able to comprise them i.e. being aware of both terms + the relation
of ‘being represented by’. And if it were neither firm to some nor
giveable, then it could not only not be known to you but could not
represent them. We are left by Locke with the absurdity of
supposing ourselves screened off from reality by a screen which ex
hypothesi we cannot know to be a screen or an iron curtain. If the
curtain is not absolute we would get glimpses of the [ ] entities
behind it. If it is as perfect as we are told [36] it is then ex hypothesi
how can we tell that there is anything behind it i.e. that it is a
curtain and not all there is, reality itself; and suppose, that without
evidence we suspect that there is such a reality, which is which? –
we know because the curtain is thought in some respects to
resemble that which it conceals. But we cannot tell whether this is
so or not because we have ex hypothesi no means of getting at the
originals to compare them with the curtain; but one thing is certain
and that is that if the originals do resemble the curtain the
difference in primary and secondary qualities breaks down because
the whole case for primary qualities is that they cannot be perceived
directly because, unlike secondary qualities, they are not sensible.
But if they are not sensible they cannot resemble the curtain which
is sensible. In which case gazing at the curtain will tell us nothing
about them, save only that they may be the unknowable and
indescribable causes of it and with that Berkeley deals later and less
exactly. In other words, when we say that something is say,
extended or figured, we either know what we mean by these words
or we do not. Either the word stands for something empirical, or it
does not. If it does, then it is drawn from experience and cannot
mean something behind the curtain – which is in principle incapable
of being directly experienced; if it does not, then no sensible
presentation of it is possible. On this point Locke was routed
completely, and except for later appearances among some
Materialist and Marxist writers the theory is never heard of again.
Berkeley goes on to spoil his argument slightly by saying
‘Extension, figure, motion, abstracted from all other qualities are
inconceivable’ because ‘it is not in my power to frame an idea of a
38
body extended and moving’ only. What he means of course is that
he cannot imagine a body whose only attributes are extension,
figure and motion. But then either [37] he means by these words –
extension and motion but no colour no shape sensible qualities, in
which case one would have thought that Wells’ Invisible Man was
something which one could define empirically, although not
perhaps have an image of, or else they are not empirically definable
in which case they do not refer to anything in [ ] experience even if
conjoined with colour and shape and it does not merely depend
upon my inability to frame an idea which suggests a merely
psychological deficiency and not meaninglessness of the terms.
So much for Rep. Perception. Now for [an] argument that is
[the] answer [to] a real need: Argument from Illusion. [38]
Phenomenalism
‘Sensible things are nothing but some sensible qualities and
combinations of them’.
Arguments that things cannot exist unperceived:
1. Heat is pain. Pain [is] not [the] effect of heat. Pain [is]
subjective, and isn’t ‘out there’, therefore heat [is] subjective too. Is
unheard thunder ‘terrifying’ unheard? If not, only noisy?
2. Principles Paragraph 53. Can you conceive sensible things to
exist outside the mind? Yes, trees and books.
Argument: All you do is to frame ideas and you omit to
conceive this. You try to conceive the unconceivable. You are
really trying to think of the unseen tree and the unseen books as
being seen unseen by you. When you think of the unseen tree,
what are you thinking of? The unseen tree – and that is the very
tree of which you say it exists outside the mind, and here you are
thinking about it. This is an appalling fallacy even for
Phenomenalism. It identifies my image or thought of X with X,
and destroys the difference between true and false propositions.
Supposing I say that there are three unicorns at present tethered in
the High Street to the building of the Examination Schools. There
must be some sense in which I must be allowed to maintain that
that proposition is not necessarily true because thought of by me.
According to Berkeley’s argument, if words mean anything there is
‘in my mind’ ‘the idea of’ three unicorns tethered, admittedly in an
‘ideal’ way, to the Examination Schools, they being also in my
mind. Therefore it is true that they are so tethered. They may lack
some property of tethered beasts, e.g. if I go out to look I shall for
39
example, not see them, fail to touch them, etc. But merely
imagining a seeing and touching of them should be enough if
objects are collections of ideas, and as the ideas of unicorns being
tethered etc. are [39] recurring, so do the objects. To say that
something is unseen is then to say that in a sense it is seen. To say
that something never happened is to say that it is happening now.
To say that Napoleon tweaked the ear of Marshal Ney is to say
that an image of Napoleon in my mind is even now tweaking the
ear of an image in my mind which is part of the reality of Marshal
Ney; and this is self-evidently absurd. Whether I think in terms of
images or not, and what they are doing to each other in my head,
has patently no relation to whether my proposition is true or false.
In other words the position is this: if I say that I understand
what a sentence means, I do not mean that I am having an image
of some sort; still less do I mean that when I say the sentence
expresses truth, I mean neither that I understand ‘I am having
images’, or that understanding cannot occur without images.
Berkeley’s argument is that to think of an unseen object is to ‘see it
in the mind’s eye’. This is in the first place untrue. In the second
place, even if it were true, it would make nonsense of negative and
hypothetical propositions, and general ones as well. (‘The Radcliffe
Camera is on fire’ would entail my having an image similar to it,
and an image of the fire etc. but what image would ‘The Radcliffe
Camera is not on fire’ entail, or even, ‘if I do not know whether the
Radcliffe Camera is on fire or is not on fire’ – does that mean that
I have an alternation of tongues of flame followed by nothing,
followed by more tongues of flame, followed by nothing, etc.?).
Alternatively, if I do have a procession of images parading in my
mind, if I am day-dreaming or musing, I am not necessarily saying
or asserting anything. If we translate Berkeley’s language and
suppose ideas not to be images but something more transparent
and similar to what Locke said – to be psychological entities of
some sort – even then the theory will not do. Supposing I say ‘Let
us assume there is a Number X which nobody happens to have
thought [40] of and that it is the product of two primes’. Then
surely it is absurd to say ‘Ah, but now this Number X has been
thought of and therefore loses the property – that which “nobody
happens to have thought of”, and is therefore not what we were
talking about. Therefore an unthought of number is a
contradiction in terms.’ Or if I say ‘All integers nobody has ever
thought of are either odd or even’ I don’t thereby contradict
40
myself and say in one breath, ‘All the unthought of integers,
hereby, now, by this very act of thinking thought of, therefore no
longer unthought of, are odd or even’. The fallacy consists in
supposing that thinking is a relation between symbols on the one
hand and objects on the other, so that corresponding to every
combination of symbols there must be something existent in the
Universe to which the symbols correspond, and that this is what I
am talking about. If that were so, all false propositions, all
unfulfilled hypotheticals, would correspond to be correct symbols
for the real entities; this means either (a) there are no false or
unfulfilled propositions, or [(b)] there are unreal or unfulfilled
entities: and all propositions are true, either about real or unreal
entities. Therefore, all false propositions are true. Absurd.
2. Berkeley is obviously not over-convinced of the validity of
this theory of the existence of unperceived objects, and therefore
has other expedients whereby to rescue them. When I am not
looking at the object, someone else may be. The problem is this:
must we really say that bodies, so to speak, flick in and out as we
look and cease looking? Every time we fall asleep or close our eyes
or even blink: do tables really cease to exist for as long as we fail to
attend to them? We certainly do not believe this, and should be
hard put to it to give an adequate account of even the most
primitive propositions of physics or indeed commonsense if our
sole evidence were the actual experience [41] of the observer – the
odd browns and blues and squares and hards, the odd chair-like,
table-like and book-like data in the broken succession in which
with little regular order between them we come across them – to
this Berkeley says:
a) That the data of various senses do occur in a certain
predictable connection (we dealt with this in the section on the
Argument from Illusion).
b) He brings in other minds. I see the tree in the quad, and I
presently cease seeing the tree in the quad, but someone else may
continue to look at it, and this is what I mean when I say that the
tree continues to exist in the quad. This argument comes in
Paragraph 90 of the Principles, at the beginning of the Second
Dialogue. This bristles with difficulties. It means that after I cease
perceiving the tree its only chance of continuing to be is that
somebody else will continue to keep it simmering, as it were, until
I take it up again, but in what sense is your idea or image or
perception of the tree identical with mine? Supposing we both
41
look at the tree: Then are there two trees or one? According to
Berkeley surely there is some sense in which I can say that there
are two ideas of a tree occurring at this moment, and if ‘sensible
things are nothing but combinations of sensible qualities’, there are
two lots of such combinations and hence two trees, and it does not
help my tree even if your tree goes on or takes up where my tree
perishes. Moreover is the tree to be the sum of everything seen,
thought, imagined etc. in a certain way at one and the same time? I
mean this: I look into the quad and say to myself ‘I see a tall
spreading chestnut tree with green leaves’. You look in the same
direction and wearing blue spectacles, as you do, say, ‘I see a tall
blue chestnut [42] tree’. Someone else in China, thinking about the
quad, imagines the entire quad covered by plane trees. Someone in
Timbuctoo imagines all the trees razed to the ground and the quad
a desert, and so on. Now, is the quad all these conflicting
constituents at once?
But far worse is to come. If the situation could be imagined
whereby for one split second nobody happened to think or have
an image of the tree, it really would cease to exist, and cease to
exist in precisely the same sense though from a different cause as if
someone destroyed it by fire? Only that in the second case there
would be ashes, and in the first none? It then becomes very
mysterious how the tree is resurrected in its full pristine glory
merely literally by the flicking of an eyelash when I open my eyes
and gaze at it again. Even worse is to come: what is to happen to
the unobserved and unimagined portions of the world, the centre
of the earth, the most distant stars, and this may bring it home
more closely, the back of my head or my heart and lungs which I
do not see, and nobody normally has in most cases ever seen. Do
they not exist at all? Do my eyes not exist when I see by means of
them, and unless I imagine them at the same time or see an image
in a mirror which is in some sense a constituent of them? Unless
someone is kind enough to continue to look at the back of my
head, and get someone else to take over when he gets bored by the
process, my head has no back, my eye has no existence etc. during
vastly the greater part of their alleged existence. But this
contradicts the premise from which we start, namely: that
appearance of the so-called outside object varies as the conditions
of my body do. In this case my body exists only intermittently and
then depends on other acts of sensing for its existence. If people
permanently decided to [43] stop looking at or thinking of me I
42
should perish, as surely as from a bullet or poison. It was a wonder
that this easy method of extinguishing one’s enemies by the easy
and painless method of forgetting them has not been applied more
frequently.
Berkeley again must obscurely feel that this explanation will not
really do because it is really at this point that he introduces the
notion of God to help him. Although other people may cease to
think about the tree, God continues to do so permanently and
perpetually. But God is so defined by Berkeley as by all Theists
that it is in the first place a little difficult to see how he can without
being at all similar to finite creatures nevertheless entertain ideas as
they do, or why, because he sees patches of colours, smells smells,
hear[s] sounds, etc. it should follow that they hear precisely what
he hears and we shall get into the same difficulties as we multiply
observers. Is God required to sense everything everywhere that is
given to any finite creature? There is the penny which I say I see. I
see an oval brown patch and expect to touch something hard, cold
etc. My friend Jones who is bent over it sees a round brown patch,
and expects to touch something cold, hard etc. My jaundiced
friend Smith sees a yellow patch. My friend Robinson who is
unfortunately in a state of Delirium Tremens says that he sees a
small pink rat where I claim to see the penny. A large number of
other observers are thinking of different thoughts and seeing very
different sights at this moment. What occurs in the mind of God
to correspond to all this? This may to the pious Christian be a
ludicrous question, since God is not finite or in space or time, He
cannot be said to see or hear or smell in the empirical sense etc.
But Berkeley unfortunately assigns all these functions to God, and
we cannot therefore avoid asking so crude a question [44] if we
accept his argument. Does God entertain the round penny and the
oval penny and the yellow penny and the brown penny and the
pink rat and everything else whatever simultaneously? Or if
‘simultaneously’ cannot apply to God who is not in time, in what
sense does He entertain them at all[?]. Or if He does entertain
them, how can they not conflict? Moreover, what is an Atheist to
say if we analyse the proposition ‘objects exist unperceived’ as
objects are perceived by other minds. Are there no unperceived
objects for Robinson Crusoe before he meets Man Friday? And if
we analyse this as continuous being in the mind of God, then
suppose Robinson Crusoe is an atheist, does it mean nothing to
him, literally nothing, to say that the mango tree continues to be
43
when his back is turned? Or that his own heart is beating when he
happens not to listen to its palpitations? Indeed it is all too clear
that Berkeley’s God is not due to theological bias on his part, but
is desperately necessary to prop his system as he expounds it. (It is
really a very poetical view of the creation including man in the
same creation as God as Hamlet is to Shakespeare. Shakespeare
conceives Hamlet, who in some sense exists only in his mind. And
then the reader or the audience become acquainted with Hamlet
too, and if Shakespeare’s mind is eternal, then when the audience
goes home or the reader closes his book, Hamlet goes on in the
mind of Shakespeare. This is an eccentric view.) Professor Broad
in his book, ‘Reality, Physics and Perception’ puts this with some
force by saying that the argument takes the form of saying ‘my
theory would hold water only if God exists, therefore God exists
for otherwise my theory would be false, but that is unthinkable’.
This argument is like saying ‘What I have just said is nonsense
unless someday someone will invent a language in which [45] it will
make sense. Therefore some day a language will be invented’.
From this we can conclude, as Professor Broad does, that pure
phenomenalism i.e. one which analyses statements about material
objects in terms of the actual data of actual observers, must be
absurd, because we do in fact in our notion of physical object
imply more than this. We imply, by giving specific names, that,
even on a very loose Humean analysis of identity, the table is at the
very least a series of data bound by such relations as succession
and similarity and possibly performance of certain functions, and
through these possess a certain degree of vividness, coherence and
so forth which presupposes not merely actual but also memory
data, forgotten data and the data of hypothetical observers who
would observe what we ex hypothesi cannot observe. If I call
something a door, I mean to imply that it has an outside as well as
an inside and that from certain angles I can see only one side of it
and infer the existence of the other. And if I am told that some
other observer can see the other side, then apart from the
difficulty about the existence of other percipients and positions of
their bodies in space I may ask ‘Does he see the same thing as I
do? Is not each man confined to the circle of his own ideas?’
But this brings up the difficulties about the word ‘same’, for the
word ‘same’ is ambiguous. I may mean by the ‘same’ something
like ‘absolutely indistinguishable’ as when I say ‘This is the same
mistake as you made yesterday’. Or as I say ‘the same constitution
44
governs England and the Isle of Wight’. Or do I mean by the
‘Same’ something forming part of the succession of experiences
given to one and same individual? Is this circular? I am not sure: if I
ask you ‘Do you see the tree I see?’ the question is ambiguous. No
doubt there are in some sense familiar to us two experiences of
seeing. No doubt they are similar. We describe what we see in
similar language. If we so define our use of ‘one’ and ‘similar’ as to
say that one observer, one percept, two [46] observers, two
percepts, then of course nobody ever sees or feels what anybody
else feels. But if I define ‘one’ in terms of more than one person,
then it does not make sense to say ‘I see what you see’. Someone
may object that there is some sense in which I cannot feel your
toothache. Your toothache is yours, private to you. I can feel a
toothache similar to yours, but then there are two toothaches. But
now if it makes sense to say we see one and the same chair, there
is in theory no sense why we could not be said to feel the same
toothache, if the physical conditions of your toothache turned out
to be the same as mine. This may sound very eccentric and
physiologically so it is. Normally I trace the causes of my
toothache to my body and yours to your body. But if some
mysterious and empirically describable but, so far as we can tell,
unlikely causal connection between your body and mine began to
occur, or if you were amused I laughed (and you did too – or not),
so that every time ‘you’ stood in a cold wind ‘I’ sneezed, it might be
possible that I could feel a cold, and realising that it was the
behaviour of your body rather than mine that had brought it on,
say ‘What a fearful nuisance, I have got Jones’ cold in J’s body,
some distance from this one’, or ‘I have caught Jones’s sense of
humour’ as I say, ‘By mistake I have taken away Jones’s umbrella’.
Jones’s relation to what I call his umbrella now in my hand is not
then logically so very different from his relation to the pain or
amusement now suffered by me; only it presupposes very
abnormal physiological circumstances. Evidence of this is perhaps
provided by the phenomena of hypnotism and telepathy.
If this is so, a physical object is analysable in some way into a
succession of experiences not all of which are actually ‘mine’, and
therefore are actually other people’s; but some of them may be
hallucinatory i.e. unreliable and incompatible with the notion of an
unfailingly veracious God; Hence we must look for some other
explanation of continuity, and Berkeley finally [47] provides it by
talking about hypothetical data. The tree is what I should see if I
45
were looking. The most vivid illustration of that he gives in asking
what is meant by speaking of the motion of the earth (in Paragraph
60 or thereabouts of the Principles): the earth is said to move yet
nobody sees it do so. But if, says Berkeley, we were in a suitable
place we would see it move ‘among the choir of planets’. In other
words, one of the things that I mean by saying that things exist
unobserved is not what one does see, but what any observer might
see if he were appropriately placed. Now this at once makes things
much easier of course. The table is no longer what I see + what
you see + what I imagine + what God sees, but what I would see if
I wee looking, and, when I am not looking, still what I would see if
I were. The fact that I am not looking does not destroy it. Here
explain Ayer’s Argument. [Ambiguity of identity: ‘how do I know
lowering my eyelids does not destroy [a] table as fire does?’ (a) Fire
destroys. True but its denial intelligible. If same true of eyelids, not
only intelligible but probable. But in case of fire not.] This is the
theory accepted by modern phenomenalism, but [it] is nevertheless
far from easy. What does it mean to say that I have an unseen
heart, or that the particles of which my skin is composed are too
small to be observable? It means that if anyone chose to cut me
open, they would see such and such a shape, feel such and such
data etc, i.e. my heart. If I used a powerful enough microscope, I
would see such and such cellular data. The world thus consists of
so many actual experiences flanked by ‘what would be the case if
any observer were observing so and so from a point so and so at a
time so and so.’ This is presumably what Mill meant by saying the
physical objects were the permanent possibility of experience. The
tree in the quad is both what I see when I look and what I don’t
see, but would see if I had looked, when in fact I didn’t. This is
certainly the most important view Berkeley advances of the
meaning of the concept of a physical object. It gets rid of the
primitive theory of images of the unseen empty house which I see
in my mind’s eye. It materially eases the [48] situation with regard
to scientific theories which deal not merely with what has
happened and is happening but also with what would have
happened, might be happening, will be happening and would be
happening if various other situations which might have existed or
may be existing or might still come into being, did so. But it has
formidable difficulties of its own:
(1) The argument of Professor Stout and Mr Hardie.
(2) The argument of Professor Moore.
46
(3) The argument of Mr Joseph.
1. Professor Stout urges three arguments against the view that
propositions about material objects can be translated into
propositions about sensation and introspection data, without
residue.
a. That we mean by physical object something solid and that by
solidity we do not mean tactile or any other data. Matter is not the
permanent possibility of sensations so much as the permanent
impossibility of sensation. Every physical object has an inside as
much as an outside. Cutting it open only reveals more surfaces or
outsides; what is inside is ex hypothesi what is not given to sensation
and never can be. Nevertheless we have a kind of a priori concept
of what it is to be like that. What he is in effect saying is that what
we mean by an onion is not an infinite collection of infinitely thin
skins which can be peeled off in an infinite series of operations. To
say that the onion is simply a globular-looking, hardish entity
which causes one to weep when smelt and so forth, is merely to
describe the symptoms of the onion or its appearances. A thing
could have all these properties and yet not be a ‘real onion.’ We
mean [49] some thing when we say that a thing is thick and solid,
and not what we mean by saying more self-consciously ‘I feel a
pressure on my finger different from that which I feel when I
press upon the surface of a bowl of water’. A spherical object,
according to Mill or Berkeley, is according to this view but a shell
of an object, a kind of a ‘balloon-like entity’ with the superficial
characteristics of a sphere, but ‘hollow within’, and in no
fundamental way different from a system or succession of after
images or hallucinatory images which are real enough in the sense
that they were not invented by us, and so can be accurately
described as an item in experience, but not a physical object
precisely because they are mere images – a hollow shell. Now this
view deserves more sympathetic consideration that is normally
given it, because it does really say something which Locke and
common sense have tried to say, but in the face of the onslaught
of the clever philosophers found themselves unable either to
defend or to abandon.
Anti-Stout.
i. Are all real objects so very hard and solid? What about rivers,
mists, sprays, clouds, and rainbows. Is sight different from the
other senses for this purpose? If not, is it not difference of
‘symptoms’? And is difference of ‘object’ and ‘image’ or sense
47
datum, not merely a matter of ‘feel’ – real shape, but not different
in principle. – It is [ ] and it is [ ].
ii. If critic[ism?] persists, does he mean what Berkeley means by
‘nothing’? Dialogues, page L 137: ‘If you find people saying that
there is a country where we pass scatheless through fire, they turn
out to mean water by “fire”. Attributeless indescribable entities
perhaps equal nothing’. Hume said ‘Berkeley’s arguments admit of
no answer and produce no conviction’. Explain the invisible
elephant fallacy.
[a.] Analogy from other non-translateables: Say dreams and
reality: As how high above my dream mountain is the real ceiling?
Or how my dream horns [compare] to a real horn? Is the sound in
my mind louder than one in reality? So is how my S.I. = [ ] ‘object’
b. More formidable argument: that hypothetical entities cannot
cause actual effects. Hardie says that if I have a magnet in my
pocket but I cannot see it and look at the compass needle [50] in
the palm of my hand and see that it turns towards my pocket, and
then say that ‘obviously the magnet is attracting this [ ] wherever, if
I had looked I should have seen an iron-coloured elongated
datum, and if I had touched I should have felt something cold and
hard etc., and the compass needle appears to point in a certain
direction’. I describe this as a magnet deflect[ing] a needle. But I
did not look and did not touch and nothing is known about the
apodosis of the conditional clause. Indeed, I imply strongly that
when the protasis is false, the conditional sentence is in mid air,
description of no actual entity. How then can its referent be a
cause? Take the proposition, ‘If Hannibal had marched on Rome
he would have taken it’. Now that does not describe anything that
happened. Would it not be eccentric to say that the cause of
something, say Rome’s fear of Hannibal, was that if Hannibal had
marched he would have taken it[?] The causes of actual events
must surely be themselves actual. An event cannot be caused by
something which would have been the case if something else were
the case which perhaps wasn’t. That is the paradox. This is an
important objection and needs scrupulous attention.
It arises from the confusion of propositions about physical
objects with propositions about our sensations. Though both
Stout and Hardie can be answered, they can be answered only on
certain unproven assumptions which we must now proceed to
make. (That it is unproven must be shown later.) The assumption
is that whatever we are to say about causal propositions, existential
48
or predicative propositions about physical objects can be analysed
into propositions about certain sorts of experiences – such as
someone might have whatever the world consists of – involving
only spatiotemporal and directly sensed and remembered relations
like similarity, contiguity, regular recurrence – what Price calls
‘family relation [ ]’. If we grant this, then we say a [51] physical
object is a collection or set of actual and hypothetical experiences.
If I am allowed to say ‘This table exists’ can be analysed into ‘If an
observer were to look in a certain direction he should see such and
such colours etc.’ ‘This is a log’ ‘When I did look I did see such
and such. And if I had applied a match I would have seen such
and such smoke data’. If I can say this, I can say that ‘The log
exists’ means that there is a situation such that when it exists these
true hypothetical propositions describe it. If it doesn’t exist, then
some or all of them are false. Now when I say ‘The log is such and
such actual hypothetical data’ I mean no more than that categorical
propositions about the log can be translated without residue into
hypothetical propositions of this sort. It is a statement about the
equivalence of two ways of using words, and not a statement about
the physical or chemical constitution of the log. The proposition
‘The log is really a collection of visual etc. experiences’ is quite
different in kind from the statement ‘the table is four legs and a
top’. I can take away the legs, but I cannot literally take away the
sense-data. If the table is burnt to ashes, that will entail that its legs
are burnt to ashes, but it would be absurd to say ‘My sensible
experiences were burnt to ashes yesterday’. The table does not
consist of sensible experiences, but propositions about the table
do, if you like, consist of propositions about my or someone’s
experiences. Now if I say ‘The invisible magnet caused the
deflection of the visible compass-needle’ I mean by saying ‘There
is an invisible magnet’ that there is a difference between [52] a
situation in which if I look in my pocket I will feel a magnet and
one in which if I feel in my pocket I will not feel a magnet. Now I
say ‘When the first situation is present and this is not hypothetical
but actual, then the compass needle datum points to where my
pocket appears to be. If the compass needle doesn’t incline in that
direction, then I think that if I were to feel in my pocket I should
not in all probability experience the sensation of feeling a magnet.’
Now whether the proposition: ‘There is a magnet in my pocket’ is
exhaustively analysed by saying that if I look there then I shall find
it, and this is the proper description of an actual situation, or
49
whether something more, something to describe our sense of
‘thinghood’ [is] needed to complete it, is a much acuter and more
fundamental question. But if we allow ourselves to assume that a
phenomenalist analysis of ‘This is a material object’ is [ ] for causal
propositions correct, then the causal proposition to which Stout
and Hardie take exception will not cause special difficulties. They
object that the existence of actual sense-data is said to be caused
by hypothetical data. But this is not so. The existence of actual
physical objects is caused by actual physical objects, and since
physical object statements may be translated into partially
hypothetical sense-datum statements, that is the proper language
into which to translate them. You must not cross the languages.
‘The magnet defects the needle’ – physical object statement. If,
every time a needle looks deflected we institute groping-in pocket
sensations, and these are succeeded by magnet-touching
sensations, then the phenomena are called causally interconnected
sense-datum statements. There is no difficulty here, but the real
difficulty is in whether we can analyse physical object statements in
this way at all.
[10Stout’s view of causality is different from that of most people,
and curiously like Berkeley’s. But he attributes to physical objects
precisely what Berkeley denies them, namely causal efficacy (here
develop push-pull view of causality and parallel case Locke’s and
commonsense boggling about reducing physical objects – so solid
– to clouds, i.e. difference between say, a table and say, rain or
wind.)]
Some difficulties about single Phenomenalism.
1. Statements about the past.
2. General propositions.
3. Logical relation between physical object proposition and
sense-datum proposition and difficulty (a) of enumerating and
specifying the latter. (b) giving name to relationship. Is it
entailment? Or [ ]? Or ‘absurd to say what [ ] absurd to present?’
4. Antedeluvian creatures (p. 44 of manuscript).
5. Moore’s objections to Mill.
(a) Material thing in the apodosis.
(b) Material thing in the protasis i.e. observers.
(c) Does family relation exist?
10 This paragaph is marked ‘Omit’.

50
(d) Queer sense of ostensibly simple terms like ‘round’ and
‘red’. (This
objection very trivial).
Moore says in ‘Contemporary British Philosophers’, Volume 2,
‘There is no reason for thinking propositions about [54] things
entail or are causally connected with propositions about mental
data’, e.g. “Dinosaurs existed a million years ago” does not entail,
and is not causally connected with, anything about mental data. It
does not entail that anyone has ever had a thought in the world. In
other words, ‘Dinosaurs existed’ – ‘If you had looked you would
have seen dinosaurish data’, but if you had not looked, then
nothing follows, and yet it is the apodosis of the conditional in
whose analysis we are interested. It is like saying ‘X is pink’ entails
‘If you say “X is pink” you’ll be saying what is true, and if you are
silent X is still pink, whatever you may think or not think’. There is
certainly something in this objection, since what it shows is that
the simple translation of categorical into hypothetical statements
somehow fails to work: if I say that X is Y, even though it be a
physical object I cannot exhaust the meaning of what I say by
saying that if certain conditions which may or may not obtain
certain consequences which otherwise might not follow will
follow. Yet this is what Berkeley’s analysis reduces us to. This is a
difficulty not for Berkeley alone, but for all phenomenalists. When
I see a table, I do not indeed see its back, and what I am told is the
proposition that it has a back is simply the proposition that if I
were to look from another place I should see what I now call back
as its front. Yet the form of expression seems to make the
existence of the back of the table dependent on or identical with
my hypothetical data. But I wish to be able to say that [55] the
back of the table exists and exists now when ex hypothesi I cannot
be observing it since I am looking at the front and that this does
not mean though it may entail that if I were looking from some
other position I should have been seeing it. Very likely I should,
but even if I did not this would not, as we normally use words,
mean that the back does not exist. To see this consider the following
argument: supposing I say ‘It is very cold outside this building’.
This is interpreted to mean: ‘If I or any observer were outside, they
would be colder than they are now when they are inside’. Now it
seems clear that I cannot be both outside and inside the building at
the same time. Therefore I can experience only one set of
experiences at a time. I can not know what it would be like to have
51
been outside at the moment when in fact I am inside. If there were
only one observer in the world therefore (and don’t let us assume
that there are more at this stage) he could never assert ‘I know that
it is cold outside and warm inside’ because the experience of
verifying each of these experiences are not both possible.
Although he is [ ] a giant, something will remain unknown –
whatever is alternative to whatever he is doing. But if you can
never know, then what you could never experience could never be.
‘To be’ is to be experienced, or to be capable of being experienced.
Robinson Crusoe could not simultaneously experience the warmth
of his hut and the cold outside it; therefore when he was in the hut
the temperature outside it, whatever that phrase might mean, could
not be said to be cold in the sense in which the hut was warm.
Alternatively, if you say that ‘it is cold outside’, = ‘if Robinson had
stooped outside he would have shivered’, then the actual
experience which he was having, namely experiencing the indoor
warmth, would [56] not have been occurring. Therefore certain
perfectly significant statements such as ‘It is cold outside now
although I am warm’ turn out to be meaningless. This means there
is something wrong with the analysis. This is sense of Moore jibe
about train wheels.
(b) Moore’s other objection, that in analysing propositions
about physical objects as hypothetical propositions about
sensations I keep re-introducing my own body, is not so
formidable. His statement of it goes as follows: ‘Though in general
when I know such a fact as “This is a hand” I certainly do know
some hypothetical facts of the form: “If these conditions had been
fulfilled I should have been possessing a sense-datum of this kind,
which would have been a sense-datum of the same surface of
which this is a sense datum”. It seems doubtful whether any
conditions with regard to which I know this, are not themselves
conditions of the form “If this and that thing had to be in those
positions and conditions”.’ This objection seems to me invalid. If
some method of translating propositions about material objects
into those of sense-data can on other grounds be found,
propositions about my body need not offer special difficulties
(Develop this).
(c) Moore’s third objection is, in language which I do not intend
to reproduce, the doubt whether there is any such relation between
sensations which enables me to say that they belong to me and the
same physical object. Moore casts a doubt on this, but does not
52
develop it. But there is no ground for believing this doubt to be
serious. When faced with the practical [57] question whether what
we observe is one or two trees in the quad, we do in fact know
then what kinds of series of tests we need to go through in order
to become reasonably certain that there is only one tree, or that
there are two as the case may be. The tests may be refined to any
given point, but they remain the same in character as the crude
tests which we normally employ.
(d) Finally, Moore objects that the sense in which a material
object is ‘round or square’ would necessarily be utterly different
from that in which we refer to what we sense is round and square,
and this seems to him untrue. But he is mistaken. This sense is
different. A penny is called by us ‘round’ only if [we] think it is true
to say that it often looks oval. If the sensations of it are always
round, we should not call the penny round, we should deny the
penny to be round, but would suppose ourselves to be having
hallucinations connected with aberrations of perspective. And we
call Earth ovoid or egg shaped though [ ] looks egg shaped at all,
but on the contrary, flat, hilly etc.: it means that something would
look egg shaped – the appearance of the earth if we
circumnavigated it from suitable distance and with telescope etc.
And we have reason for thinking that for the proposition ‘Earth is
ovoid’ – (looking ovoid from spaceship being only one of the
entailed propositions of ‘X is ovoid’) more true and interesting
propositions follow than from proposition ‘Earth is flat’: but that
is all: all that’s meant by call[ing] this hypothesis true: Earth looks
ovoid certainly false: and if ovoid = looks ovoid, Earth is ovoid is
false: and in poetry and novels and history is so.[58]
Moore’s Refutation of Idealism.

Famous doctrine responsible for British realism resting on the bald
assertion that we detect a distinction between the object and the
consciousness of it, that the relation of the object to the
consciousness is different from that which blue has to glass in the
case of blue beads, or blue has to hair in the case of blue beards.
Moore goes so far as to say that a sensation of blue is a knowing that
there is something blue: that the element of consciousness can be
isolated from the element blue although it is very diaphanous, and
that it is difficult but not impossible to discover it in introspection.
It is really what the particle ‘of’ stands for. A sensation of blue is
different from a blue sensation, whatever that might be. If I sense
53
something blue, that is quite different from saying that I sense
‘bluely’. In some cases no doubt the object can not be
distinguished from the so-called act, e.g. ‘I heave a sigh.’ ‘I dreamt
a dream.’ ‘I dance a waltz.’ These are said to be quite different
from: ‘I touched the cup’; ‘I saw the moon’, since it is clear that if I
had not heaved there could logically have been no sigh, and if I
had not danced there could have been no waltz, and if I had not
jumped there would have been no jump. But although I did not
touch there might still have been a cup, and even if I had not
looked there would still be a moon. This is valid on the
assumption that hypothetical propositions are excluded. Although
this argument does throw useful light on the different senses in
which such propositions as ‘I felt a pain’ and ‘I felt the door’ are
used, [59] it rests on a false psychological premise about the
diaphanous medium.
Leads to view that ‘knowing’ is undistorting: but ‘knowing
distorts’ is meaningless: ergo ‘knowing does not distort’ is
meaningless. [60]
Matter as cause
Berkeley’s examination of objectivity.
Berkeley correctly sees that the notion of material object entails:
1. That it is in causal relations with other objects.
2. That it is in some sense objective rather than subjective.
3. That there is some sense in which when we speak of objects
we use symbols with ‘outer-reference’ rather than ‘inner reference’.
He then argues (paragraphs 25, 27, 28) that matter cannot be a
cause because it is inactive. This serves to emphasise his view of
causes as active pushings and pullings (contrast volitional view of
causality with regularity view). Berkeley’s view that matter cannot
cause is then analytically deduced from his distinction between
active and passive.
Objective versus subjective means what I can control as
opposed to what is forced upon me. I cannot order sense ideas like
those of imagination. They are strong, lively, have order,
coherence, steadiness, and follow the rules which can be acquired
by experience. But if rules, then will is at work, i.e. God. This is
half illuminating, half obscure. The difference between sensations
which I cannot myself regulate and laws which I do not invent but
discover and those which I can conjure up and for which I can
invent rules, like chess, is absolutely radical and is the basic
54
concept whereby we distinguish subjective and objective, what we
can help and what we cannot. On the other hand, the deduction of
a will at work from the existence of laws implies a teleological view
of laws deriving from earlier metaphysics and which Spinoza and
Hume successfully destroyed. Berkeley declares that the laws of
Nature show forth no observable necessary connection,
anticipating Hume, hence a will at work. This amounts to ‘all
events must have causes, all causes are active, only minds are
active, therefore all events caused by minds’. But all events are not
caused by my mind, therefore by God’s mind. This could lead to
Pantheism, or Theism equally.
Perception = the action of mind and not matter. (See Dialogues
p. 406-407 and Manuscript P. 21). Another argument used by
Berkeley to prove that matter cannot cause ideas is the proposition
produced in the Second Dialogue that only Entity which has ideas
can give them to us. Matter is passive and cannot think, i.e. have
ideas. Hence only a spiritual entity can communicate ideas to us.
This rests on Cartesian and Scholastic nonsense about eminent
causes. Causality is giving something to something. I cannot give
what I have not got, and so the spark must contain the explosion.
Fallacy view to identification of causing partly with giving partly
with ground and consequence relation when ground contains
consequence. (Explain this a little).
Argument proceeds: Ideas cannot be my mind alone (why not:-
because unalterable by me). This brings us back to notion that
ideas are consciously willed, if not by me by somebody. Nothing
happens de facto: everything springs from a motive. Confusion
‘how’ and ‘why’. Hylas wants to know if God suffers pains and
blue patches which He gives to us. The answer is: No. Pain is part
of [62] corporeal nature: God is not corporeal. Still, either God is
so different from us that He ‘perceives nothing from sense as we
do’: in which case what happens to the unseen tree? And if He
does perceive as we do, how exactly does this help us? My looking
at yonder tree does not seem to make a difference to your seeing
it, why should God’s?
Having proved that matter is not needed to act for our
perception of physical objects, as he thinks, which now turn out to
be sets of sensations injected into us by the orderly will of God
Berkeley runs through all the other things which matter might
possibly be thought to mean.
55
(a) Some people desperately think it to be mere existence. Bare
existence. Para 17. This can be proved by demonstrating that
existence is not a predicate. A thing could not merely exist and
have no attributes.
(b) It cannot be necessarily connected with sensations and have
no other attributes.
(c) Since it is unknown it cannot explain uniformity (Para. 20.
Read aloud). If it exists unknowable. But this will not do if he
thinks it is meaningless. If it is meaningless it cannot be
condemned as useless. Just like a thief arguing that the bank note
he is accused to having stolen is:
i. Non existent = a hallucination of the owner. Never was there.
ii. Is really forged, useless, cannot buy anything.
If the notion of matter is unintelligible it cannot also be [63]
dismissed as useless.
(d) But if all matter is really ideas, can we be said to eat ideas
and clothe us in ideas[?] May we no longer say ‘Fire burns’ only ‘a
mind causes an idea fire to be followed by the idea ashes’[?] Or
may we not say ‘the sun sets’ but ‘the idea of’, etc.? Here Berkeley
gives a very early and very original analysis of logical constructions,
saying that it may sound queer to speak of eating ideas or of the
sun setting inside our heads, but that if adequately translated this is
no longer eccentric. He repeats that ideas are not thoughts, but
actual data. The idea of fire is not hot if by the idea is meant
thought.
(e) How can we say that things 10 miles off are in our minds?
He argues that distance at any rate [is] not visual but tactile.
(f) If distance is an idea, mind is extended, which is absurd. But
this is the fat oxen argument.
(g) What are we to make of Newton’s Physics?
Answer: Newton does not mean matter, only the regular
motion. Newtonian physics could be deduced from dream-data.
‘All we need is uniformity in the production of natural effects’. I
am not for changing things into ideas but rather for changing ideas
into things’. (p. 462 of Dialogues). (h) What about the general
consensus of opinion in favour of matter[?]
i. Majorities can be wrong.
ii. Question ‘If not I who is responsible for my data’? is to him a
perfectly correct question. Matter is a meaningless answer.
(i) Why should God need all this intermediate mechanism? Why
not [64] direct hypnotism by God?
56
Answer: Because ideas are Divine language. Curious view of
natural symbolism. Where the symbols refer to themselves or each
other. Arguments for design; the beauty of the world etc. – If God
exists and speaks this would be his language: but does not prove this.
Ideas are Creator’s language which in miracles He occasionally
uses exotically or paradoxically. Fire tells us we will be burnt but
does not burn us. One event is a dependable symptom for or
testimony for another. But this confuses symptoms and causes. A
cause is sometimes identical with a symptom where we cannot
alter things, e.g. Astronomy. But where we can, it is very different.
The needle of the thermostat tells me how hot it is, but I cannot
alter the heat by tampering with the needle. Bell-ringing is not the
cause of men’s leaving work, nor the day the cause of the night,
although they are symptoms. Anyhow, if sensible ideas are
language, what does it say? Why should there not be an infinity of
languages, the object of each of which is a symbol of the one
below.
(j) Is matter the score of God’s music? A pretty but silly idea.
God the musician, the ideas are the occasion.
(k) Maybe matter exists, but we are all blind to it as the blind
are blind to colours. But if so, it is either unknowable, in which
case we mean nothing by it, or else it is given to a sense which we
do not possess and does not need a substratum any more than the
ordinary senses do.
(1) What about miracles? Christ and the wine in Cana? Did [65]
Christ merely make people think they drank wine? No, says
Berkeley, if public testimony agrees the liquid looks, smells like
wine, and has the effect of wine, then by God it is wine. Hence the
miracle. Interesting point about public testimony as involved in
objectivity.
(m) God may not reliably keep uniformity going. How could we
be sure, etc. Why should matter be more reliable? But he allows
that the language of the Maker may vary. ‘A man may be well read
in the language of nature without understanding the grammar of
it’. If this is literal, then there literally are sermons in stones, books
in the running brooks, etc.
(n) The reality of unperceived Time and Space. Assertion of
absolute time and empty space is meaningless. Space is relation
between bodies in terms of each other, or of forces. This is
consistent with phenomenalism. If objects may be analysed
sensations, space will relate sensations.
57
(o) What about mathematical entities? – [ ] Answer: The study
not of reality but of symbols. Conventional arrangements and not
objective relations. Extraordinary view of Geometry: ‘There is no
such thing as one ten-thousandth part of an inch’; but because an
inch represents a line of anything length, you can use it to divide it
into any number you like. But if so, the symbol has the property of
what it symbolises, and is infinitely divisible after all. Mathematics
either about the use of symbols or empirical.
(p) Finally we come to selves or spirits. Cannot be ideas [66]
since they think, will, perceive that we know them because we
know ourselves, in some sense different from that in which we
know passive ideas. We understand ideas in the minds of others by
analogy with our own. Introspection is a form of sensation which
does not yield to self. Hume right. But the letter ‘I’ does not stand
for an idea. Self-awareness is not ordinary cognition. Something in
this:
i. We obviously cannot eliminate selves from even Phenomenal-
ist analysis.
ii. I am not like a table: hence the real table which does not
behave like a table is ultimately silly; whereas behaviourist analysis
of selves no less silly in the contrary direction.
iii. Something meant by privacy of data.
iv. Bishop Butler wrong in saying ‘my body as external to “me”
as any other external thing’, since if so, do I cease to exist in
dreamless sleep? Is a madman literally a different person from his
previous self? Supposing I wake up and look like a cockroach, am
I still I? Or am I not I as a cockroach, but still I as a madman? We
know others by analogy (Paragraph 145) and infer the wills of
others from the behaviour of the data. We do not see other men
or God, but infer both (148).
‘Spiritual substance,’ says Philonous in the Third Dialogue, is
meaningful since I know what I do when I think, will, etc. Selves
define in terms of activity. Interesting connection with later French
Ideological school and ‘Volo ergo sum’.
[67]
Notions as opposed to idea of self.
For distinction of individual selves see Manuscript, p. 41, on
Identity. Thus we get a general picture of identical selves, each
with its own discontinuous sensations, flickerings of colours, odd
prickly and soft sensations, succession of ticklings, itches, odours,
fluty sounds, rainbows and iridescences, each induced by the
58
archetypes inspected by God. Tactual data are inferred from visual,
and vice versa, although he implies that the tactile are more
primary than the visual. Visual presented, tactual inferred.
Perception emerges as inference.
Notions are of spirits, mental operations and relations.
Are relations imposed? Like ‘all’, ‘if’, and ‘any’, in acc. to what
principle?
© The Isaiah Berlin Literary Trust 2005
Posted 24 November 2005
59

Berkeley'S External World: The Isaiah Berlin Virtual Library

Uploaded by

Copyright:

Available Formats

Berkeley'S External World: The Isaiah Berlin Virtual Library

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Berkeley'S External World: The Isaiah Berlin Virtual Library

Uploaded by

Copyright:

Available Formats

The Isaiah Berlin Virtual Library

BERKELEY’S EXTERNAL WORLD

This is a lightly edited transcript of a text in Berlin’s papers. No attempt has

The argument from illusion

1 cf. Gulliver and the Houyhnhms.

(a) echo: doesn’t count: if we hear a note and an echo or a note

Physics cannot answer this – nor does it pretend to. Typical

(1) Tale told by an idiot full of sound and fury, signifying

2 Virginia Woolf, ‘Modern Fiction’, in The Common Reader (London, 1925), p.

Locke’s account of substance and physical matter

(4) Newtonian Physics and Atoms

Berkeley rightly saw in the concept of matter one of the main

6 Principles, introduction, paragraph 3.

8 ‘Tritos anthrōpos’ (‘the third man’).

9Transcript corrected up to this point by Henry Hardy, hereafter by James

10 This paragaph is marked ‘Omit’.

Moore’s Refutation of Idealism.

© The Isaiah Berlin Literary Trust 2005

Posted 24 November 2005

You might also like