PhilosophyThemes and ThinkersJW Phelan

To my parents

C A M B R I D G E U N I V E R S I T Y P R E S S

Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo

Cambridge Universi ty PressThe Edinburgh Building, Cambridge CB2 2RU, UK

www.cambridge.orgInformation on this title: www.cambridge.org/9780521537421

© Cambridge University Press 2005

This book is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements,no reproduction of any part may take place without the written permission of Cambridge University Press.

First published 2005

Printed in the United Kingdom at the University Press, Cambridge

A catalogue record for this publication is available from the British Library

ISBN-13 978-0-521-53742-1 paperbackISBN-10 0-521-53742-8 paperback

AC K N OW L E D G E M E N T S

Bridgeman Art Library/Getty Images, p. 329t; Bridgeman Art Library © Munch Museum/Munch-Ellingsen Group, BONO, Oslo, DACS, London 2005, p. 318; Fotomas Index UK, pp. 5, 19, 24, 29, 56,64, 78, 109, 117, 150, 168, 201, 212, 252, 265; Hulton Archive/Getty Images, pp. 68, 93, 129, 224, 294,317; Time & Life Pictures/Getty Images, pp. 48, 108, 133, 287; Tracey Emin © Tracey Emin, courtesy Jay Jopling/White Cube (London), photo Stephen White, p. 329b

Cover design by Tim Elcock, cover image The Philosopher’s Trouble by Giorgio de Chirico © Alinari Archives/ CORBIS

Illustrations on pages 8, 84, 105 and 126 by Gerry Ball

ContentsAuthor’s acknowledgements vi

Introduction 1

Part 1 Themes1 Theory of knowledge 4

Knowledge 4Justification 9Scepticism 14Perception 17

2 Moral philosophy 23Normative ethics 23Practical ethics 32Cognitive meta-ethics 38Non-cognitive meta-ethics 43

3 Philosophy of religion 51The meaning and justification of religious concepts 51Arguments for the existence of God 55Faith and reason 63Implications of God’s existence 66

4 Philosophy of mind 72Dualism 72Physicalism 77Behaviourism and functionalism 83Problems in the philosophy of mind 90

5 Political philosophy 99Political ideologies 99Freedom 104Law 108Authority 113The state 116

6 Philosophy of science 120The scientific method 120Scientific development 126The aims of science 133The objectivity of science 139Natural and social science 143

Contents iii

Part 2 Thinkers

7 A commentary on Plato’s Republic 148The preliminary arguments 149The nature of an ideal society (Books III–IV) 150Platonic epistemology (Books V–VII) 152Platonic education (Book VIII) 162Platonic psychology (Book IX) 162Platonic aesthetics (Books X and XI) 164

8 A commentary on Aristotle’s Nicomachean Ethics 167Eudaimonia: the object of life (Book I) 167The mean and moral goodness (Book II) 173Action and akrasia (Book III) 176The importance of prudence (Book VI) 180Pleasure, happiness and contemplation (Book X) 183

9 A commentary on Descartes’ Meditations 188The First Meditation 188The Second Meditation 192The Third Meditation 195The Fourth Meditation 198The Fifth Meditation 199The Sixth Meditation 202

10 A commentary on Hume’s Enquiry Concerning HumanUnderstanding 206Of the origin of ideas (Section II) 207Of the association of ideas (Section III) 211Sceptical doubts concerning the operations of the understanding (Section IV) 213Sceptical solution of these doubts (Section V) 215Of probability (Section VI) 216Of the idea of necessary connection (Section VII) 217Liberty and necessity (Section VIII) 220

11 A commentary on Nietzsche’s Beyond Good and Evil 223On the prejudices of philosophers (Part I) 223The free spirit (Part II) 227The religious nature (Part III) 229On the natural history of morals (Part V) 232We scholars (Part VI) 237What is noble? (Part IX) 238

12 A commentary on Marx and Engels’ The German Ideology 242Preface 242Feuerbach: opposition of the materialist and idealist outlook (Part I) 243Selections from Parts II and III 256

iv Contents

13 A commentary on Mill’s On Liberty 261Introductory (Chapter 1) 261Of the liberty of thought and discussion (Chapter 2) 264Of individuality, as one of the elements of well-being (Chapter 3) 269Of the limits to the authority of society over the individual (Chapter 4) 272Applications (Chapter 5) 274

14 A commentary on Russell’s The Problems of Philosophy 277Appearance and reality (Chapter 1) 278The existence of matter (Chapter 2) 279The nature of matter (Chapter 3) 281Idealism (Chapter 4) 282Knowledge by acquaintance and knowledge by description (Chapter 5) 283On induction (Chapter 6) 285On our knowledge of general principles (Chapter 7) 286How a priori knowledge is possible (Chapter 8) 287The world of universals (Chapter 9) 288On our knowledge of universals (Chapter 10) 289The limitations of philosophical knowledge (Chapter 14) 290

15 A commentary on Ayer’s Language, Truth and Logic 293The elimination of metaphysics (Chapter 1) 293The function of philosophy (Chapter 2) 298The nature of philosophical analysis (Chapter 3) 301The a priori (Chapter 4) 304Critique of ethics and theology (Chapter 6) 305Ayer’s hallucination 307

16 A commentary on Sartre’s Existentialism and Humanism 310Nausea (1938) 310Existentialism and Humanism: an introduction 311The paper knife argument 312The human condition 313A brief history of existentialism 314Anguish 317Abandonment 318Despair 319Intersubjectivity 320Bad faith 322Two heroines 323Final thoughts 323

Part 3 Appendices

Appendix 1: introduction to aesthetics 328

Appendix 2: critical thinking 332

Appendix 3: an alphabet of advice on study skills for philosophy 335

Index 343

Contents v

1Only a shallow man doesn’t judge by appearances. (Oscar Wilde)

AimsOn completion of this chapter you should be able to:

■ evaluate rationalism

■ evaluate empiricism

■ distinguish between a priori and a posteriori knowledge and account for the tensionsbetween these categories

■ evaluate the foundationalist, coherentist and reliabilist theories of justification

■ outline and evaluate the tripartite definition of knowledge

■ explain the difference between ordinary and philosophical doubt

■ evaluate traditional and modern sceptical arguments

■ evaluate the naive realist theory of perception

■ evaluate the representative realist theory of perception

■ evaluate the idealist theory of perception

■ evaluate the phenomenalist theory of perception.

It is characteristic of philosophy to step backfrom familiar facts and figures and to talk inmore general terms. Plato (c. 427–347 BC),

one of the first philosophers, describes thisapproach in The Republic as ‘the study ofessential natures’.1 The theory of knowledge liesat the very heart of the philosophical enterprise,which is in essence a quest for meaning. Theessential questions surrounding knowledgeconcern its nature and whether we can categorisedifferent types of knowledge, as well as thepossibility of there being limits to what we canknow. Epistemology, as the study of knowledgeis called, inevitably connects with thephilosophies of mind, science and language. Letus begin by examining the notion of knowledgeitself.

KnowledgeThe history of western epistemology (a word thatderives from the Greek word episteme, whichmeans ‘knowledge’) is the history of two schoolsof thought that seek to define knowledge in

different ways. Rationalism sees knowledge asthe product of reasoned reflection capable ofoperating independently of experience.Empiricism, on the other hand, views knowledgeas the product of sensory experience. Rationalistsbelieve that we arrive at knowledge by thinkingabout relations between concepts. We know, forexample, that there are 60 seconds in a minute bylearning the meanings of the terms ‘second’ and‘minute’. In the same way, we know that there are60 minutes in an hour, from which we candeduce that there are 3,600 seconds in an hour.This type of knowledge is called a prioriknowledge. One does not even have to leave thecomfort of one’s armchair to discover it, as theoperation involves thought alone. In contrast, weacquire empirical or a posteriori knowledge fromexamination and observation. Thus we knowfrom sensory experience that barium burns withan apple-green flame: all we need to do isperform the experiment and note the results.

The phrase a priori derives from the Latinmeaning ‘from what comes before’ and the phrasea posteriori means ‘from what comes after’.

4 Part 1 Themes

Theory of knowledge

Theory of knowledge 5

Rationalists, like Plato, believe that human beingsare born with a body of knowledge already storedin the mind that enables us to recognise truthswhen we unearth them. Plato tells the story ofMeno’s slave to illustrate the point that each of ushas an innate capacity to reason.2 The narrator,Socrates, draws a number of geometric figures inthe sand in front of Meno’s slave who, whenasked a series of questions, recognises the rightanswers despite never having done anymathematics before. Empiricists, on the otherhand, believe that all knowledge needs to belearnt and that there is no such thing as innateknowledge. The most famous empiricist, JohnLocke, described the human mind as a tabularasa, a blank slate upon which facts acquiredthrough experience are written.

Attempts have been made to bring togetherrationalism and empiricism. The most famouswas made by the German philosopher ImmanuelKant (1724–1804), who saw the role of reason asone of organising phenomenal experience – thatis, experience gained from the senses. The post-

Kantian rationalist tradition has refined itsstandpoint from the days of Plato. It accepts thatwe acquire genuine knowledge from experience,but holds that we are born with the innatecapacity to reason, which means that we knowwhat to do with the information received fromexperience. A rationalist would agree with anempiricist that we cannot know that there are3,600 seconds in an hour without having beentold that that there are 60 seconds in a minuteand 60 minutes in an hour. But a rationalistwould argue that in order to have reached thecorrect conclusion we require certain innatecerebral capacities such as an understanding oflanguage and an ability to work out therelationship between the concepts that weacquire.

A priori knowledge is that body of informationwe know to be true from thinking about theconcepts involved: from a logical examination wecan conclude that there is something permanent,universal and necessary about this knowledge.Contrariwise, a posteriori knowledge is that body

Born in the Somerset village of Wrington in the summer of 1632,John Locke became one of the most prominent English empiricists ofhis age, as well as writing influential pieces on political philosophy.Locke’s contribution to empiricism consists in his idea that all ourknowledge is accrued through the senses from birth. He explainssensory perception in terms of ‘primary and secondary qualities’ ofobjects. Primary qualities include size, shape and movement, andbelong to objects irrespective of whether or not they are beingperceived, while secondary qualities are imposed on objects by theobserver and include colour, smell and taste.

Locke’s father wanted his son to study for the religious ministry, butthe gaunt, scholarly student pursued various courses in medicine andphilosophy in his intermittent time at Oxford, which lasted somethirty years. Locke was interested in politics, and during the English

civil war supported the parliamentary party, who sought to reduce the power of the monarch andmake him rule with parliament. Locke, as a supporter of the parliamentarians, fled to Holland butreturned to England after William and Mary were brought to the throne during the GloriousRevolution. Locke’s famous political work, the Two Treatises of Civil Government (1690), waspublished after this turbulent time, although recent scholarship suggests an earlier authorship.Having returned to England, Locke suffered bad health and died in 1704.

Significant worksEssay Concerning Human Understanding (1690)Two Treatises of Civil Government (1690)

John Locke(1632–1704)

6 Part 1 Themes

of information acquired through sensoryexperience: from observation we can accuratelyreport something about the world we live in. Atthe heart of the rationalist–empiricist debate isthe question of whether knowledge is innate orgained from experience. But why do we need todecide between these two categories? Manyphilosophers of knowledge reject such a starkchoice.

Hume’s forkThe Scottish philosopher David Hume(1711–1776) was the first thinker to propose anexhaustive account of knowledge that embracedboth the a priori and a posteriori categories. Heoutlined his thoughts in his Enquiry ConcerningHuman Understanding (Section IV, Part 1) and hismodel is often referred to as Hume’s fork. This hastwo prongs that are designed to spear either oneof two types of truth claim – termed ‘relations ofideas’ (a priori) and ‘matters of fact’ (a posteriori).Hume’s outline of knowledge consists in a clear-cut division between the self-evidently true(tautological) and the empirically verifiable(proved through experiment). Hume’s analysismay be represented as a diagram. Hume is acommitted empiricist, yet admits to a category ofknowledge gained from reason alone. This facthelps us to refine our definitions of rationalism

and empiricism. A rationalist is any philosopherwho believes that knowledge gained throughreason is more immune from sceptical attack andtherefore more fully ‘knowledge’, whereas anempiricist is any philosopher who argues thatknowledge gained through the senses isinformative and therefore more worthy of the title‘knowledge’ than that gained from reason alone.

Humes’ fork divides the categories ofstatements with its two prongs, so that anystatement must fall into all three categories onthe same side: it must be either a priori, analyticand necessary, or it must be a posteriori,synthetic and contingent. The semanticcategories ‘analytic’ and ‘synthetic’ refer to thekinds of statement. Hume holds that claims toknowledge are either true by definition (analytic)or can be tested by an experiment (synthetic). Anexample of an analytic statement is ‘All bachelorsare unmarried’ whereas an example of asynthetic statement is ‘Potassium burns with alilac flame’.

The category ‘modal’ describes logicalpossibilities and is described in terms of possibleworlds. A possible world is the same kind ofthing as the world we are experiencing at themoment but with room for conjecture andimaginary differences. I might imagine a possibleworld, for instance, in which I am made of

Field

Epistemological

Semantic

Modal

Category

a priori a posteriori

analytic synthetic

necessary contingent

Figure 1.1 Hume’s fork

Theory of knowledge 7

potassium and burn with a lilac flame! Possibleworlds entertain the most weird and wonderfuloptions. Consider the possible world of cartoonswhere it may be possible for a character’s head toexpand and reduce in a second or for someoneto run as fast as a speeding train. Yet even in thismost creative of fantasy worlds, the laws of logiccan never be compromised. Even the mostinventive animator cannot draw a square circleor represent a married bachelor. Anything that isthe case in this world but doesn’t have to be truein all worlds is called ‘contingent’ whereasanything that has to be the case in all possibleworlds is termed ‘necessary’.

Such a statement as ‘All sisters are female’ isnecessary, analytic and a priori. Indeed Humeassumes that all necessary, analytic truths are apriori. Similarly a statement such as ‘Sodiumburns with a yellow flame’ is contingent,synthetic and a posteriori as we have gained thisknowledge through observation and canentertain the idea of sodium burning with adifferent coloured flame. Other thinkers,however, have attempted to criticise thiscategorisation as being overly simplistic. Let usexamine two such cases.

Bending Hume’s forkThe most famous attempt to bend Hume’s fork isto be found in Kant’s Critique of Pure Reason(first edition, 1781). Kant draws a similardistinction between analytic and syntheticjudgements to Hume’s. He draws our attentionto the fact that basic sentences are made up ofsubjects (what the statement is about) andpredicates (words used to describe the subject).Thus the sentence ‘Some roses are red’ features‘Some roses’ as its subject and ‘are red’ as itspredicate. Kant held that analytic statements arethose in which the subject contains the predicateand which do not supply any additionalinformation; an example would be the sentence‘Red roses are red’. Synthetic statements do nothave predicates which are contained in thesubject and as a consequence give us newinformation; an example would be ‘This rose isred’ (it would still be a rose if it were anothercolour).

Kant, however, suggested that there are certainsentences that can be established as true by theoperation of thought alone (a priori) and inwhich the predicate is not contained in the subject(synthetic). He gives the following examples:

7 + 5 = 12

Every event has a cause.

There is nothing inherently present in theconcept of ‘7’ that makes it evident that when itis added to ‘5’ the result will be ‘12’. Likewise,there is nothing present in the concept of ‘anevent’ that makes us immediately aware that ithas been caused. Kant called such sentencessynthetic, a priori statements, and thiscategorisation set the tone for a ferocioussubsequent debate. Much of the argumentcentres on the clarity of the distinctions drawnand the issue of whether the above categories ofa priori and a posteriori are mutually exclusive.Kant’s notion of the synthetic a priori issignificant as it challenges the belief that thereare two mutually exclusive accounts ofknowledge termed ‘rationalist’ and ‘empiricist’that can never be reconciled. Kant argues thatsense-data are important building blocks ofknowledge but points out that reason has anequally important role in organising this data. Asimilar position was developed by W. V. O.Quine, who is discussed later in the chapter.

A more recent protagonist who has steppedinto the ring to challenge Hume’s fork is theAmerican philosopher Saul Kripke (born 1940).He tries to prove the existence of a necessary, aposteriori category. Many of Kripke’s examplesfocus on the nature of proper names. Kripkeholds that names behave in a very different wayto straightforward descriptions such as ‘red’ or‘female’.

● Proper names do not appear in dictionaries.

● Proper names possess only one bearer(signified by the intention of the speaker).

● Proper names are ‘rigid designators’ as distinctfrom general descriptions.

● Proper names once learnt have a necessary, apriori standing as one knows what is referredto without investigation.

8 Part 1 Themes

Despite it only being a contingent fact that myparents baptised me by a certain name (theymight have chosen another one), I cannevertheless travel through any number ofpossible worlds and know that the inhabitants ofeach possible world are referring to me whenthey address me by means of my name. Myproper name is a rigid designator and when usedin context can only point to me. My propername provides an example of a necessary, aposteriori truth. It is necessary because I retainmy identity in all possible worlds, but it is aposteriori because you have to meet me first inorder to find out who I am. Kripke suppliessome scientific examples to act as exceptions toHume’s fork. His favourites include:

Water is H2O.

Gold has atomic number 79.

Kripke argues that we need to perform anexperiment in order to discover suchinformation (so the truth is a posteriori), but

water is the name of the substance H2O, and sothe truth that water is H2O is necessary; similarlygold is the name of the substance with theatomic number 79. Like Kant’s argument, Kripkeattempts to show that Hume’s position is notwatertight.

Review questions1 Briefly explain the difference between the

rationalist and empiricist accounts of knowledge.

2 Briefly explain the difference between a prioriand a posteriori knowledge.

The twentieth-century logician and philosopherWillard Van Orman Quine was one thinker whoquestioned the profitability of a war betweenrationalism and empiricism. Indeed, thephilosopher’s lexicon offers the verb ‘to quine’ asa denial of any absolute distinction. Quinesought a coherent philosophy of knowledge thatincorporated logical truths and scientificdiscoveries. He attacked the false dichotomybetween the analytic and the synthetic on the

Quine has emerged as one of the most significant intellects of thelater twentieth century. He was born in Akron, Ohio, and majored inmathematics in Harvard where he gained a Ph.D. in only two years.Renowned as a logician and mathematical philosopher, he developedmore original ideas in the years after his influential paper ‘TwoDogmas of Empiricism’ (1951) in which he challenged thesynthetic–analytic distinction. Quine held that to suppose there is aboundary to be drawn between the synthetic and analytic is to adoptan unwarranted ‘article of faith’. His philosophy, which is firmlyrooted in the Anglo-Saxon analytic school, was greatly influenced byBertrand Russell. He sought to unify the scientific and philosophicalenterprises, and believed that epistemology, rather than being someelevated discipline, was firmly rooted in mathematics and science.Quine served as a code-breaker for the navy in the Second World

War before returning to a fellowship at Harvard. He wrote in a crisp, elegant style and lectured in afast, machine-gun-like monotone. He enjoyed Dixieland jazz and travelling; indeed his interest inthe latter is well attested in his autobiography The Time of My Life with details of routes andtimetables. Quine’s other hobbies were stamp-collecting, map-reading and tracing the origins ofwords (especially words beginning with Q).

Significant worksFrom a Logical Point of View (1953)Word and Object (1960)Web of Beliefs (1978)Quiddities (1987)

W. V. O. Quine(1908–2000)

Theory of knowledge 9

grounds that there is no such thing as pure apriori knowledge, and believed that everything iscapable of being revised in the light ofexperience. Quine held that if one squeezes ananalytic statement such as ‘All bachelors areunmarried’ hard enough then it turns into a setof synthetic propositions: that some men are notmarried, and that these men are collectivelygiven the English name ‘bachelor’. At somemoment in time someone had to point out anunmarried man and tell us that the English word‘bachelor’ is used as a rigid, syntheticdescription. As he declares in ‘Two Dogmas ofEmpiricism’, ‘the lexicographer is an empiricalscientist’. Synthetic descriptions remain open torevision at some time in the future: another termmight be used or a new discovery made. LikeKant, Quine finds a role for reason in organisingsensory perceptions into what he terms ‘a web ofbeliefs’. Each belief gains its meaning from theother beliefs to which it stands in relation andeach individual thinker spins and repairs theirown web of beliefs in order to catch the truth.Quine’s most famous attack on theanalytic–synthetic distinction appears in ‘TwoDogmas of Empiricism’. The two unfoundedbeliefs or dogmas are: an unwavering convictionin the existence of pure, analytic truths and thebelief that all meaningful statements can bereduced to verifiable sense-data. Both dogmasare mistaken in assuming that one can easilyseparate language and experience.

JustificationKnowledge has traditionally been defined as‘justified true belief ’. This idea is called thetripartite definition of knowledge (see pp. 13–14)and finds its origins in Plato’s Theaetetus. Platoargues that knowledge cannot just consist in truebelief alone and a further ingredient is required,namely justification. He gives the example of ajury who have a hunch that the accused is guiltyand are in fact correct in this assumption butstates that we would not be right in condemningthe accused without some form of evidence.3

Plato declares that the jury’s true belief does notcount as knowledge without proof, warrant orjustification. It is this notion, that of justification,

that concerns us in this section.Knowledge is not based on guesswork or

chance, but on having a good reason for holdinga belief. Thus guessing a stranger’s namecorrectly would not count as knowledge. This iswhat is meant when we say that a belief is‘justified’. There are three principal explanationsas to what constitutes a good reason for holdinga belief: foundationalism, coherentism andreliabilism. Let us examine each in turn.

FoundationalismFoundationalists divide all beliefs into two maingroups:

● basic beliefs – i.e. beliefs that supportthemselves and are, therefore, not inferredfrom other beliefs (non-inferential);

● non-basic beliefs – i.e. beliefs that need to besupported by other beliefs, which are inferredfrom other beliefs (inferential).

Foundationalism is often seen as a superstructureof non-basic beliefs that are built on a firmfoundation of basic beliefs; such an image has ledphilosophers to call this theory of justification‘the pyramid model’. Foundationalism isfavoured by empiricists, who are intent onbuilding a structure of scientific knowledge. Anempiricist model that uses scientific facts verifiedby sense-data as the candidate for basic beliefs

Thebelief

about how my radio works

Types of wavelength

Electromagnetism

A belief in molecules

A belief in the existence of atoms

may run as follows:Each level of beliefs builds on the previousstratum; one might say that each belief isinferred from what has already been established.In this example, the pinnacle belief acts as a testof the basic beliefs. Thus, if one is engaged in the

10 Part 1 Themes

activity of manufacturing some device, the mainendorsement for the validity of your beliefs is ifthe end product works.

Foundationalism is not an exclusivelyempiricist model. Descartes (1596–1650) is onerationalist who constructs a pyramid from what

before it is built on. One might argue, however,that no basic belief is truly simple. So, forexample, something as seemingly fundamental asa white sense-datum nevertheless implies aspectrum, an object to hold the property ofwhiteness or an area in which to be white. Basicbeliefs are subject to two problems: either thebasic belief is so basic that nothing of interestcan be inferred from it or it yields too manypossible but indeterminate beliefs.

Criticisms of foundationalismThere are two further strands of criticism ofthe foundationalist theory. The first we shall call the regress argument and it may be summarisedas follows:

1 In the foundationalist model, each beliefrelies on another belief for its justification.

2 Basic beliefs cannot be treated as anexception and must in turn rely on a furtherset of beliefs for their justification.

3 There is no end to this process of justification,which must continue ad infinitum.

Philosophers term such a never-ending sequencean ‘infinite regress’, and consider it to beextremely problematic in any explanation.

The two pyramid models cited above createinfinite regresses: the Cartesian model begswarrant for a belief in God; the empiricist modelassumes a correspondence between sense-dataand reality. Foundationalists respond either byaccepting that a regress occurs ad infinitum inany case of justification or by denying this fact.The second line of defence (denying infiniteregress) sees hard-line foundationalists arguingthat basic beliefs, by definition, do not need anyfurther justification. Soft foundationalists viewbasic beliefs as lying beyond reasonable doubtand argue that accurate observation is the bestwe can hope for as the foundation to ourknowledge.

The second major criticism offoundationalism rests on the fallibility of thesense-data (i.e. the possibility that sense-data arewrong). The English philosopher BertrandRussell (1872–1970) argues in the followingexample that a more sophisticated way ofdetermining truth and falsity is needed asopposed to an uncritical reliance on sense-data:

Facts

Theorems

External world

Clear and distinct ideas

God / cogito ergo sum

he holds to be a priori beliefs:Descartes uses as his foundations the premises‘God exists’ and ‘I think therefore I am’ (cogitoergo sum). He holds that such clear and distinctideas are self-evidently true and from this firmbasis we can demonstrate the existence of anexternal world of objects and even go as far as to

TaskInvent your own pyramid of basic and non-basicbeliefs to illustrate the foundationalist theory ofjustification. Draw on any subject area, e.g.economics, history or the sciences, and set it out as above.

discover laws governing how they work.For non-basic beliefs to be justified they need torest on basic beliefs. In many cases, these basicbeliefs appeal to present sensory experiences. Forinstance, an empiricist might use as a basic beliefan observation such as litmus paper changingcolour or the atomic structure of a substance asseen through a microscope. Underlying this viewis the conviction that sense-data are reliable, tothe point of being privileged as basic beliefs. Inturn, non-basic beliefs are justified on a scale ofprobability in relation to the concrete, infalliblefoundations of empirical data. For this reason,foundationalism has been termed an error-avoidance theory: foundationalists seek to avoidfalse beliefs and may make use of scepticism (seepp. 14–17) to rid the superstructure of anydubious beliefs that may lead to its collapse.Every level of the pyramid needs to be secured

Theory of knowledge 11

Domestic animals expect food when they see theperson who usually feeds them. We know that allthese rather crude expectations of uniformity areliable to be misleading. The man who has fed thechicken every day throughout its life at last wringsits neck instead, showing that more refined views asto the uniformity of nature would have been usefulto the chicken.4

Russell’s sentiment, inherited from Humeanscepticism, is that one cannot rely on anynecessary connection between sense-data. This isa particular problem for empiricism. UnderlyingRussell’s concerns is the fact that just becausesomething happens in a certain way does notmean that it always will (see section on theproblem of induction, pp. 123–24). No belief canrealistically be termed ‘infallible’, and as aconsequence one cannot use basic beliefs as anindubitable foundation as even they might turnout to be unjustified. In response to this, somefoundationalists argue more humbly that basicbeliefs can at least provide us with a usefulworking model.

CoherentismThe coherentist theory of justification stressesthe consistency, connectedness and cohesion ofjustified beliefs and is sometimes referred to asholism for this reason. A belief coheres with ageneral pattern of beliefs if:

● it is based on adequate evidence;

● it is not disproved by one’s current pattern of beliefs.

Thus I believe that an acid turns litmus paperred as I have observed the reaction take place andit fits in with my other convictions aboutchemistry. Coherentism may be compared to araft on which each proposition is a length ofdriftwood bound together in order to float on asea of uncertainty. Foundationalists, in contrast,set out to construct an oil-refinery-likesuperstructure. The coherentist aim is to sail on aself-contained body of knowledge in which:

● each proposition is to be judged on its ownmerits;

● each proposition is connected to at least oneother proposition;

● no proposition can stand separately from the rest.

Beliefs in a foundationalist model have anasymmetrical structure, i.e. belief A is justifiedbecause it is based on basic belief B.Coherentism, contrariwise, operates asymmetrical relation between beliefs, i.e. belief Ajustifies belief B as much as belief B justifiesbelief A. For example, Eddington used planetarymotion to prove Einstein’s predictions, yetEinsteinian physics can equally be employed toexplain the position of Uranus. The two beliefsare mutually explanatory. Coherentists arereluctant to change the mutually reliantstructure, as its main aim is to stay afloat. Forthis reason, the model has been called anignorance-avoidance theory. It seeks to offer anexplanation, accepting the likelihood of somefalsehoods in order to obtain beliefs about theworld around us and avoid drowning. As the setof justified beliefs expands, each belief in turn isbetter explained by the additions. An expansionof the set of justified beliefs must result in animproved, more comprehensive explanation thanwas previously offered. Cosmology is oneexample, where in ancient times poetic creationstories were told in order to fill an unsettlingvoid. In due course, a more refined ‘big bangtheory’ has replaced mythology as it dovetailswith other scientific theories. It, nevertheless,remains a conjecture.

Three strands of coherentism may beidentified: positive, negative and mixed. Positivecoherence involves beliefs being justified if thereis evidence for their truth. This approach isfavoured in hard sciences such as physics andchemistry. Negative coherence involves beliefsbeing accepted until one unearths evidenceagainst them. One might think of anarchaeologist devising a likely explanation for agiven feature that may be revised in the face ofcontrary evidence that comes to light. Inpractice, most theories are mixed.

Criticisms of coherentismLet us examine two criticisms of coherentism.The first argument runs as follows:

12 Part 1 Themes

1 Coherentists seek to offer a complete andcomprehensive explanation made up ofjustified beliefs.

2 There might be more than one complete andcomprehensive explanation made up ofthese justified beliefs, given that each beliefhas an equal and opposite possible belief.(For example, for my belief that Nelson’sColumn is in London, there is an equal andopposite belief that Nelson’s Column is notin London.)

3 Coherentism does not provide adequatecriteria with which to judge rivalexplanations.

4 Therefore, coherentism is an inadequatetheory of justification.

This accusation is a powerful one: if every beliefhas an equal and opposite belief floating on thesea of uncertainty, how does one choose betweenthem? One would have to argue in defence thatthe truth is somehow unique and that a truthfulanswer is the only answer where each beliefcoheres properly.

The argument from plurality may be refinedin a second criticism. According to coherentism,any system of beliefs works if, and only if, itcoheres. However, the entire system could bewrong or misinterpreted. Surely the justificationof a belief needs to be linked to the truth of thatbelief. Yet no criteria are supplied by coherentiststhat enable us to test for truth. How do we judgerival explanations for the origins of the universeif each appears to cohere with what we alreadyknow about physics?

Review questions1 Identify and explain two differences between

foundationalism and coherentism.

2 Briefly explain the view that knowledge isjustified true belief.

ReliabilismThe third theory of justification examined in thischapter is termed reliabilism. This account holdsthat a belief is justified if and only if it is theproduct of a reliable method. Reliabilists are notas much concerned with the ins and outs of howa method comes to be reliable as with the

realisation that it is. Consider the followingexamples of reliable methods:

● Testimony: historians are able to work withprimary sources to generate valuableinformation (such as ‘The battle of Hastingswas fought in 1066’). They work with first-hand testimonies to elucidate facts and figuresabout past events.

● Memory: doctors use memory to rememberthe symptoms of various diseases in order todiagnose accurately.

● Perception: scientists use sensory perception intheir day-to-day experiments. We learn thatcopper burns with a blue-green flame throughobservation.

● Inference: logicians use inference as a fail-safemethod for deducing a valid conclusion. Thismay take the basic form, ‘Either p or q: not p:therefore q’ (see appendix 2 on critical thinking).

In practice, many reliable methods may be usedtogether. Consider the example of a person whodiscovers a potentially valuable painting in herattic. She takes it along to an expert, who tellsher that it is indeed a priceless work. Ondiscovering another painting in this style she iseven more delighted. This person relies on theart expert’s testimony. He is able to rememberdetails and recognise them in works of art. Inturn, given that one work is valuable and asecond bears sufficient resemblance to the first,one may infer that the second piece is alsovaluable. One might express the conviction thatthe second work of art is of value in terms ofprobability. Indeed reliabilism is linked toprobability:

1 It is reasonable, all things considered, tobelieve p (where p is some proposition).

2 p makes q (another proposition) highlyprobable.

3 So, all things considered, it is reasonable tobelieve q.

Reliabilism differs from foundationalism andcoherentism in that it is concerned with theprocesses involved in generating justificationrather than the relationship between beliefs. Forthis reason, it is termed an externalist theory of

Theory of knowledge 13

justification. Reliability does not depend onhuman beings but on an external process that isnoted by them. Foundationalism andcoherentism are termed internalist theories ofjustification. To illustrate the difference, considerthe scenario where an internalist and anexternalist enjoy a day at the races. Theinternalist spends all day considering the form,the going, which trainers and jockeys are in luck,the age, weight and breeding of the runners. Theexternalist, however, spends an hour in the bartalking to his contacts, who he knows will givehim a winner. The reliabilist or externalist doesnot understand how his contacts come up withso many successful tips at each meeting, but iseternally grateful that they do. Note that bothpunters might bring home equal winnings.

Criticisms of reliabilismOne frustration that is encountered in pursuingthe reliabilist theory of justification is that nomethod of justification is infallible (i.e. reliable)in every instance. Your watch may have alwaysprovided reliable information about the time inthe past but might stop tomorrow! In addition,one might imagine a historian being misledwhen looking at the Bayeux tapestry depiction ofthe battle of Hastings, thinking that Harold wasshot in the eye with an arrow when, in fact, thecharacter with an arrow in his eye might nothave been Harold. Similarly, we might beforgetful or be impeded so we are unable toperceive accurately. Lastly, inference may supplya valid method, though not necessarily trueconclusions. For example, one might argueflawlessly as follows without realising theornithological error in the first sentence (not allswans are white!):

1 All swans are white.

2 Frank is a swan.

3 Therefore, Frank is white.

A second criticism of reliabilism stems from thecircular nature of the reliabilist reasoning. Theargument runs:

1 Reliable methods generate knowledge.

2 Knowledge is what is generated by reliablemethods.

When laid out like this it becomes evident thatthe theory does not present clear criteria againstwhich knowledge or reliability can be judged.Taken together these criticisms from error andcircularity prove difficult for reliabilists to argueagainst. Yet, as we have seen, criticism can bewaged against each of the three theories ofjustification: foundationalism, coherentism andreliabilism. Is it then ever possible to achievecertainty? One group of philosophers whobelieve that every form of justification andknowledge can be challenged are called ‘sceptics’.We shall meet some of their arguments later inthis chapter.

The tripartite definition of knowledgeSo far we have surveyed theories of knowledgeand justification. We shall now look at therelationship between these concepts and explainthe tripartite definition of knowledge. The term‘tripartite’ means divided into three parts andrefers to the traditional conception of knowledgeas justified true belief. The definition may beexpressed more formally in the following way:

For person X to know that p:

1 p must be true.

2 Person X must believe that p is true.

3 Person X must have some reason forbelieving (evidence or assurance) that p istrue.

I can be said to know that Mercury is the nearestplanet to the Sun because it is, I accept that it is,and have assurance every time I scan the skywith my telescope.

The most famous objection to the tripartitedefinition of knowledge comes from E. L.Gettier’s short paper ‘Is Justified True BeliefKnowledge?’ (1963). Gettier (born 1927)maintains that in certain circumstances all threeelements are present, yet the resulting intellectualstate cannot be counted as a genuine instance ofknowledge. Such scenarios have been called‘Gettier counter-examples’. Imagine yourself at aparty; you gaze across the dance floor and noticethat your friend Ben happens to be dancing. Infact, it is Ben’s identical twin brother Bill, butcoincidentally Ben is also dancing close by:

14 Part 1 Themes

1 It is true that Ben is dancing.

2 You believe Ben is dancing.

3 You have a good reason to believe that Ben isdancing.

Yet the above instance, according to Gettier, doesnot warrant the description of knowledge.5

Gettier’s examples dwell on the fact that thejustification does not relate to the truth of thesituation. Is there a clause that we can insert torescue the tripartite definition from the scrap-heap of invalid arguments? Perhaps one mighttry the following to counter the effect ofcoincidence:

The justification for p must correspond to the truth of p.

Thus, one might add to the previous examplethat you are justified in believing that Ben is onthe dance floor if, and only if, the person youhave observed is indeed Ben. In practice, it willbe extremely difficult to prove this link for everysupposition one makes.

such as the mathematical truth 2 + 2 = 4,appears to be beyond the need for justification.

ScepticismDoubtWhen we say a person is sceptical, we usuallymean that that person is not easily convincedand doubts any claim that is not well founded.Scepticism is also the term used for aphilosophical school of thought that has itsorigins in ancient Greece and which has hadmany famous advocates over the centuries.Sceptics believe that if they apply a systematicmethod of doubt to what is taken for grantedsome indisputable truths will be established.Philosophical doubt is used to discover theindubitable, i.e. that which cannot be doubted.Philosophical doubt differs from ordinary doubtin a number of ways:

● Philosophical doubt is concerned with the bigquestions in life.

● Philosophical doubt examines the justificationof truths.

● Philosophical doubt does not need to have anypractical benefits.

● Philosophical doubt challenges our idea ofwhat counts as knowledge.

One of the earliest sceptics was the ancient Greekphilosopher Pyrrho of Elis. He concluded thatwe could not know anything for certain and thatthis yielded a calm detachment from life knownas ataraxia. Another ancient Greek philosopher,and follower of the Pyrrhonian school, was thestoic Agrippa. He challenged people, in whatbecame known as ‘Agrippa’s trilemma’, to find aclaim to knowledge that did not involvedogmatism, regress or circularity:

● Dogmatism is the entrenched commitment to abelief even in the face of contrary evidence.For example, members of some ancient tribesretain their beliefs about the nature of theworld even when faced with contrary evidence.

● Regress is an endless sequence of justificationwhere successive beliefs justify the next adinfinitum (we have already encountered thisproblem when looking at foundationalism).

TaskInvent your own Gettier counter-example. Explainhow each of the three elements of truth, beliefand justification are present and why your beliefcannot, according to Gettier, be classed asknowledge.

Other wounds have been inflicted on thetripartite definition of knowledge. As we havepreviously seen, there is little agreement as towhat constitutes justification. The tripartitedefinition does not specify which account ofjustification (foundationalism, coherentism orreliabilism) is to be accepted, nor does it specifywhat counts as truth.

There are rival epistemological theories as tothe definition of truth (e.g. coherence,correspondence, performance and pragmatism).The tripartite definition of truth makes noattempt to resolve this wrangling.

Finally, one may accuse exponents of thetripartite definition of knowledge of overcookingthe definition. In many cases we take knowledgeto be self-evident and do not need to carry outlaborious examinations. Self-evident knowledge,

Theory of knowledge 15

● Circularity occurs when a set of beliefs justifyeach other. One might imagine a religiousperson arguing that they believe in Godbecause God’s existence is documented inscripture, which should be believed because itis the word of God.

Agrippa reached the conclusion that we cannever assert a genuine claim to knowledge. Manycenturies later, however, the French philosopherRené Descartes believed that, through usingsceptical arguments as a tool, he could engineer awhole epistemic structure founded on certainty.Unlike the previous Pyrrhonian scepticism,which has as its end uncertainty, Cartesianscepticism aims for truths that cannot bedoubted. Note, however, that both Pyrrho andDescartes are global (radical) sceptics in thatthey cast doubt over all claims to knowledgerather than over claims in a localised area, suchas morality, religion or science.

Cartesian scepticismDescartes used a system of methodological doubtfor a positive end in order to establish what isindubitable and to extinguish any trace ofuncertainty. His conclusions sound surprising tothe modern ear, which is used to hearing wordssuch as ‘scientific’ and ‘empirical’ as bywords forassurance and authoritativeness. Instead,Descartes held that we cannot be certain ofscientific truths and can only be sure about theexistence of God, the existence of our souls andself-evident truths such as those found inmathematics. He used three main scepticalarguments to reach this conclusion. Thesearguments can be phrased in the form ofquestions:

● How do we know that our senses are deliveringreliable information to our brains about whatreality is like? (the argument from illusion)

● How do we know that what we experience isnot a dream? (the argument from dreaming)

● How do we know that all that we believe is notdeliberately created in our minds by an evildeceiver? (the argument from deception)

Descartes concluded that even if we were beingdeceived we could still be sure of our own

existence, as it is impossible for someone to bedeceived and not exist. He went on from thisbasic conclusion to derive certainty about theexistence of God and the truths of mathematics.

Modern ‘brain in a vat’ arguments are directdescendants of Cartesian doubt, and take theform of asking the disquieting question ‘How doyou know that you are not a brain in a vathaving your experiences controlled by somesuper-psychologist?’ Hilary Putnam (born 1926)describes the scene thus:

The person’s brain (your brain) has been removedfrom the body and placed in a vat of nutrientswhich keeps the brain alive. The nerve endings havebeen connected to a super-scientific computerwhich causes the person whose brain it is to havethe illusion that everything is perfectly normal.6

Every brain wave that we experience, whether itis a memory, personal experience or feeling, isthe result of controlled electronic impulses and itis possible that every human being is fed theseimpulses artificially, as in the film The Matrix.Putnam operates in the same way as Descartes inusing scepticism for an anti-sceptical purpose.Like Descartes, Putnam hopes to establish apositive conclusion. His brain in a vat argumentdiffers from the Cartesian account of an evildeceiver in some notable ways. Putnam dispenseswith a malign genius whose intent is to feed falsebeliefs into people’s brains. Instead Putnam talksabout all sentient beings suffering from acollective hallucination that is controlled by aself-refuting system. That we are all brains in avat of nutrients would be admittedly a strangequirk of nature, but not inconceivable. Havingsketched out the scene, Putnam goes on to arguethat the expression ‘I am a brain in a vat’ is infact a self-refuting proposition – that is, onewhich if true proves to be false. Putnam belongsto the analytic school of philosophy that definesmeaningful statements as those which are eitherobviously true or provable by experiment. ForPutnam, the sentence ‘I am a brain in a vat’ isfactually meaningless as it suggests somethingthat goes beyond the evidence for or against. Theidea of being a brain in a vat has no externalreference point that we can investigate and assuch goes beyond belief and disbelief. If we are to

16 Part 1 Themes

refer to some object or property in the physicalworld with any degree of meaning then we needto establish a causal link between our descriptionand the thing itself. It is as if our descriptionexists as a balloon that needs to be tied tophysical evidence. Such a declaration as ‘We existas brains in a vat’ is like an untied balloon orfree-floating belief that exists beyond our reachand beyond the realm of reasonable doubt.Expressed more formally Putnam’s anti-scepticalarguments runs:

1 I know that I am here talking to you.

2 Sitting here talking to you entails that youare not a brain in a vat.

3 Therefore, I know that you are not a brain ina vat.

Review questionIdentify two ways in which a sceptic can doubt thatwhat we see is what there is.

Humean scepticismDescartes was one of several philosophers tosuccessfully use sceptical methods to establish acertainty (that thinking entails existence).Scepticism has also been used in the empiricisttradition to clarify the nature of our knowledge.Two powerful arguments put forward by Humeillustrate the limits of the scientific enterprise.Hume states that just because we have observedthe same effect follow an event on a myriad ofprevious occasions it does not mean that there isa necessary relationship between the cause andthe effect. A sceptic is justified, therefore, inquestioning such seemingly indubitable notionsas that the sun will rise tomorrow by assertingthat there is no logical justification for supposingthat the sun will rise rather than that thedarkness will remain.

The philosopher’s second argument warns ofthe dangers of classifying general laws asindisputable, owing to the fact that we can nevercomplete an infinite number of observations. Wemight embark on a number of experiments totest the hypothesis ‘All frogs can swim’. Even if allthe frogs in our experiment took to the waterwith consummate ease, it is still a leap of faith tomove from the premise ‘All observed frogs can

swim’ to the conclusion ‘All frogs can swim’. If wecan theoretically conceive of an amphibian stillwearing frog-sized armbands that hasn’t learnt toswim yet, then we cannot say that we havediscovered a certainty about all frogs.

Russellian scepticismBertrand Russell was another empiricistphilosopher who added to the mêlée of scepticalchallenges. He focused on the distinctionbetween appearance and reality (see chapter 14)and his argument can be laid out as follows:

1 Appearances differ from reality.

2 We are only acquainted with theappearances of things.

3 Therefore, we can never pronounce withcertainty on the reality of things.

Russell’s argument is classed as sceptical as itsows the thought in our minds that reality mightbe different from the world as depicted by oursense-data. He concludes that we can only talkmeaningfully about our sense-data and notabout a separate reality, but he remains open tothe possibility that there is such a separate reality.

In summation: the various strands of scepticismmentioned above all focus on the inadequacy ofsense-data. Cartesian scepticism doubts thatknowledge gained from sensory evidence couldbe as assured as knowledge gained from reason.Humean scepticism acknowledges informationgained through the senses but doubts that thisinforms us of necessary truths. Russellianscepticism doubts that sense-data give us a truepicture of what lies behind the world of everydayexperience.

SolipsismDescartes, Hume and Russell attempted toachieve a positive result from their scepticalchallenges. For Descartes, it was the assurance ofhis own existence through the activity ofthinking; for Hume, it was contingent truthsgained from observation of the world around us;and for Russell, it was the fact that we can onlymeaningfully describe sense-data. The mostextreme form of doubt is called solipsism, andyields a different conclusion. Solipsism is the

Theory of knowledge 17

view that you alone are the only object of realknowledge and source of existence – that is,other objects only exist when and because youperceive them. The term is coined from twoLatin words: solus meaning ‘alone’ and ipsemeaning ‘self ’.

The position divides reality into twoconceptual frameworks or matrices. The firstmatrix consists in a familiarity with one’s ownthoughts, experiences and emotions. The secondcomprises everything else, including otherminds, the external world, time and space. Whilewe can have certain knowledge of the firstmatrix, references to the second cannot be heldas claims to knowledge.

Solipsism is in turn subject to the followingcriticisms:

● Language is too complicated to be invented byany one person and similarly could not beauthentically maintained by only one user.

● Solipsism seems to offer a poor account of ourexistence in that even if it were true it does notsupply an account of where our thoughts,experiences and emotions come from.

● Solipsism poses the wrong sort of questions.While in the past philosophers investigatedthings like consciousness and existence, todaysuch topics belong in the remit of scientificinvestigation.

No mainstream philosophical movement hasever adopted a seriously solipsistic standpoint,perhaps due to the perceived lack of an audience,but it is nevertheless an interesting notion toexplore! Solipsism raises uncertainties about thetrue nature of our perceptions. How do we knowthat what we see is what there is? The last sectionof this chapter supplies an outline of the fourmain accounts of perception.

PerceptionWe gain knowledge of the external world fromsensory data: touch, sight, smell, hearing andtaste. For this reason philosophers are interestedin the mechanisms of sensory perception andpose the question, ‘What is going on whensomething is perceived?’ Four possible accountsare explored below: naive realism, representativerealism, idealism and phenomenalism.

Naive realismAny theory of perception that is described as‘realist’ holds that we perceive the world as itreally is and that there is no difference betweenappearance and reality. Naive realists believe thatwe have direct access to a naked, unmediated andunadorned reality. It is the commonsensical viewthat what we see is what there is; thus, observinga bowl of ripening fruit involves nothing morethan opening my eyes and seeing what is reallythere. In addition, the term ‘naive’ has a technicalmeaning in the philosophy of perception, namelythat objects contain the same properties whenthey are not being perceived as when they arebeing perceived. So, for instance, if I were toleave the kitchen for a moment or two, theoranges in the fruit bowl would continue beingorange and the bananas would still be banana-shaped. In summary, the two principles of naiverealism are as follows:

● the principle of direct realism, i.e. the beliefthat the relevant physical object and yourimmediate perception of it areindistinguishable in every way;

● the principle of innate properties, i.e. the beliefthat properties intrinsically belong to objectsand that objects hold their properties evenwhen they are not being perceived – this is todeny any distinction between primary andsecondary qualities raised by representativerealists (see below).

Naive realism runs into problems as there isclearly a complex process involved in perceptionthat consists in an image of an object beingpresented to the mind’s eye. Science explains thisin terms of light rays hitting the retina andproducing brain waves. To argue that we haveunmediated access to the external world is toignore the role of the sense organs as instrumentsof perception. Secondly, naive realism seems tohold that whatever is immediately perceived mustimmediately exist. This is not true in cases of timelag, as when I see a star that ceased to exist yearsago. Thirdly, arguments from illusion successfullycounter the theory by showing how mistaken onecan be when perceiving the world. The mosteffective arguments of this type involve cases ofhallucination, where there is a radical discrepancy

18 Part 1 Themes

between our perceptions and relevant physicalobjects – there are no relevant physical objectspresent when a hallucination takes place.

Naive realism may be summarised in thefollowing diagram:

Perceiver → World

Representative realismRepresentative or indirect realism seeks toincorporate the scientific analysis of perceptionby talking about two worlds: the world aroundus and the world of images or representationsthat are presented to me. To use the example ofsight: light reflects from objects in the externalworld (perhaps a bunch of grapes from ourearlier example), my eyes are stimulated and abrain wave is produced. As a result I enjoy avisual experience of the grapes draped over thefruit bowl. Representative realists talk in terms ofthe actual grapes as they exist in the real worldand the visual experience, representation orperception I have of the fruit.

The English empiricist John Locke advances afurther notion concerning how we perceive thefeatures of objects in the real world. Locke arguesthat some features (termed primary qualities)belong to the objects themselves and includeshape, size and movement while other features(termed secondary qualities) are added by oursensory systems and include colour, texture andsound. Returning to our fruit example, a pearpossesses the primary quality of being pear-shaped while also possessing the power to inducethe secondary quality of a sweet taste asexperienced by the perceiver. This secondaryquality of sweetness is perceived by the personwho eats the fruit but it is not said to belong tothe pear itself.

Representative realists are left with animportant question as to how the mental imagecorresponds to what exists in the real world.Some philosophers, called idealists, have evendoubted that the real world exists, arguing thatall that exists are our minds and theirperceptions. For the moment let us summariserepresentative realism in the following diagram:

Perceiver → World I → World II (in our mind) (in reality)

IdealismIdealism (and phenomenalism) may be describedas anti-realist theories of perception in that theyhold that there is no independent reality ‘outthere’. The most prominent idealist was theAnglo-Irish bishop George Berkeley(1685–1753), who posed the question ‘How dowe know that a material world exists?’ Berkeley’sconclusion was that there was no good reason forbelieving in the existence of material objects andthat the notion of matter as an independentlyexisting substance was groundless andunintelligible. For Berkeley, we can only be sureof three things: the existence of our own mind,the existence of ideas and the existence of God. Itis God who keeps ideas in existence when theyare not entertained in our minds. The idealistphilosophy of existence and perception consistsentirely of non-physical entities. Thus you maybelieve that the book you are currently readingexists in solid material form, but it is in fact aseries of mental images perceived by your mind’seye and stored in the library of God’simagination when not ‘in use’, i.e. beingperceived. Berkeley summed up this notion in hisfamous epigram esse est percipi – to be is to beperceived.

The idealist theory of perception has mademany enemies down the years since its firstformulation, but one must admire Berkeley’soriginality. He is not questioning the existence ofmatter in order to be irritating but raises thequestion, ‘What good reason do we have forbelieving in a material realm?’ One of Berkeley’scontemporaries Samuel Johnson famously kickeda stone outside Berkeley’s church and declared, ‘Irefute it thus’. His point was that it is obvious thatthe material world exists, yet his refutation is norefutation at all, as the stone, the act of kickingand Johnson’s exclamation would, according toidealism, all belong to the immaterial world ofideas. A more successful criticism may be madethrough countering Berkeley’s question ‘How dowe know that matter exists?’ with a furtherquestion, ‘Is there a good reason for supposingthat matter does not exist?’ This return firesuggests that idealism itself lacks justification andis, as a consequence, unlikely.