Language and the Pursuit of Truth
First published 1956
This book attempts something new: to introduce the general public to semantics, the study of linguistic communication. This study goes beyond grammar, literature, style and so on, to the study of the meanings and the logical classification of words and statements, and the means of verification of statements.
All the most important problems, the problems of religion, morals, politics and sociology, can only be solved via the use of words.
The discovery of truth and the attainment of knowledge necessarily depend on a good understanding of language: in particular, of the notions of meaning and verification.
The subject requires only the patience to master a particular technique.
Those unacquainted with modern philosophical literature will find books in the first group make easier reading than those in the second.
STUART CHASE, The Tyranny of Words.
GEORGE ORWELL, 1984.
JOHN WILSON, Thinking with Concepts.
L. S. STEBBING, Thinking to some Purpose.
R. H. THOULESS, Straight and Crooked Thinking.
- - -
ALASDAIR MACINTYRE, A Short History of Ethics.
JOHN WILSON, Reason and Morals.
A. J. AYER, Language, Truth and Logic.
P. H. NOWELL-SMITH, Ethics.
G. RYLE, The Concept of Mind.
C. K. OGDEN and I. A. RICHARDS, The Meaning of Meaning.
J. L. AUSTIN, How to Do Things with Words.
- 1 Chapter I. WORDS
- 2 Chapter II. STATEMENTS
- 3 Chapter III. TRUTH
Chapter I. WORDS
A. THE FUNCTION OF WORDS
• Words are tools.
• A word acts as a sign - something which conveys meaning and can be interpreted.
• It is misleading to talk of words standing for things or 'having meanings': they have uses.
B. TYPES OF WORDS
• Proper nouns name, or 'stand for', things; but class-nouns do not stand for anything.
• Adjectives, verbs and adverbs do not stand for things, but their shared use depends on regularities in experience.
• We are not tempted to think that prepositions, pronouns, interjections, conjunctions and articles refer directly to experience.
• The following are basic uses of words.
1. Descriptive words
• These give information about the world.
• They vary in concreteness: compare 'The cat is on the mat' and 'The Norman way of life began to establish itself in England in 1066'.
• Some describe tendencies or relationships, or the properties of other words.
• Words expressing scientific concepts are descriptive in a peculiar way. Their meanings are determined partly by experience, partly by the role they play in theory; therefore gravity and electrons are not things in the same sense that billiard-balls and falling bodies are things.
2. Evaluative words
• Evaluative words praise or blame, commend or criticize
• 'Good' and 'bad' are almost purely evaluative, but according to the context in which they're used, they will imply that certain criteria are satisfied: good coal burns well, a good horse runs well (in a conversation between jockeys) or pulls well (in a conversation between farmers).
• Other words such as 'murder' and 'theft' are mixed descriptive and evaluative.
• 'Merciful' and 'just' are both mixed; both commend, but they describe contrary types of action.
• 'Courageous' and 'rash' could describe the same action, but have contrary evaluations.
3. 'Pointer' words
• These are vitally important in sentences, but do not describe or evaluate; they include conjunctions, articles, pronouns.
• These are expressions of feeling. They may also be descriptive, but don't have to be.
We must not judge utterances by the outward form of what is said but by the use to which it is put.
C. MISTAKES ABOUT WORDS
Most of our failures in communication caused by the abuse of single words (rather than of statements) result from inability to distinguish the proper uses of language. We must continually ask "What am I trying to do with these words? For what purpose am I using them? And is any of them out of place?"
• We often seem to treat words as if they had magical power: blasphemy can be illegal; obscene words are banned by law and convention.
• Words are used ritualistically: 'Communist', 'Fascist', 'bourgeois', etc. are used as ritual denunciations or praise-words, variously in various parts of the world.
• Perhaps the most important mistake is to take abstract nouns as naming things. Nouns used as slogans, such as 'Communism' and 'democracy', are misleading; but the worst offenders are terms such as 'will, 'conscience' and 'soul', whose meanings are very unclear.
• These words are dangerous not because they are abstract but because they have magical force.
• They have a use of some kind; but not to 'stand for' something.
• Defining your terms is not enough: a slide in meaning can still occur. For example, someone might argue:
Most people commit adultery, so it's natural.
What's natural is right.
Therefore adultery is right.
The arguer might define 'natural' as 'normal' and still maintain the argument. But both 'natural' and 'normal' have both evaluative and descriptive meaning. There is a slide in the argument from the descriptive to the evaluative use.
• We can only avoid such fallacies by detecting mixtures of descriptive and evaluative meaning.
• The use of such words is not to clarify meaning but to pronounce judgement.
• We need to keep description separate from evaluation.
In the 1930s some American students challenged the Nazis to distinguish 'Aryan' from 'non-Aryan' blood in a number of samples that they provided. The Nazis were unable to do so, because there is in fact no special quality in the blood of Aryans which distinguishes it from the blood of other races.
Many people, perhaps including the American students, took the Nazis' assertions quite literally. But intelligent supporters of Hitler would probably have said that talking about the superiority of Aryan blood here was metaphorical, a way of talking about the superiority of Aryans. If this had been recognized, they would have had to face the question of what the actual use of 'blood' was, and to answer this before making any assertions about it. It was only by preserving the ambiguity of the word that they were able to make and believe in these curious assertions.
Animal intelligence and pain
Arguments about animal intelligence revolve primarily around what it is that descriptive words are supposed to describe: A thinks that "intelligence" describes certain types of actions, like barking to go out for a walk, bringing home the newspaper, etc.; B thinks that intelligence must describe the ability to talk and to understand.
Similarly, A says "Animals certainly feel pain - they squeal and writhe when they're hurt"; B answers that they can't feel, because they're not conscious at all.
- A and B are arguing about the proper use of words: for they may be in perfect agreement about the observed facts. They evidently differ about what descriptive word to use. Until they realize the purely verbal nature of their argument, it is likely to be inconclusive.
- But people think that there are 'things' labelled 'justice', 'intelligence' and so on, and that all they have to do is to get other people to see the thing and read the labels.
- Once we see that it does not matter what words we use to describe what, provided that we agree about the uses, these arguments seem rather foolish.
- But all fields of discussion are full of vague descriptive words:
- instinct, heredity, discipline, sanity, complex
- state, law, rights, freedom
- will, soul, conscience, sin, grace, God, hell, heaven
- universe, life, time, space
- Definition of a word does not necessarily overcome magic and ambiguity: the words used in the definition may be as vague as the words defined. But people usually resent having their views subject to analysis.
Chapter II. STATEMENTS
A. THE FUNCTION OF STATEMENTS
1. Poetic communication
- This is communication where the literal and prose sense of the words has either secondary importance or no importance at all. They are expressive.
- Communications are often mixed. Religious statements are often literal statements of faith, while also being poetic.
2. Prose communication
- This is communication which we are supposed to understand with our reason, not appreciate with our feelings. It is the type we ought to use in arguing, discussing, solving problems and discovering truth.
The best method of distinguishing one type of statement from another, and throwing light on each type, is the important verification principle.
The 'method of verification' is simply the way in which you find out whether something is true.
- Verification is a guide to meaning and a guide to truth.
- To find out whether a certain statement is true, we should have to go through the following process:
- (i) Discover the meaning of the statement - i.e. its use and what sort of thing it is intended to communicate.
- (ii) Agree about how to discover whether it is true.
- (iii) Consider the evidence and decide.
Most arguments are fallacious or inconclusive because (i) and (ii) are overlooked.
- (i) Verification is a guide to meaning: the disagreement about animals' intelligence described above arose because the disputants were not agreed about the proper way to verify the statement 'animals are intelligent'.
Sometimes we can use verification to observe that a statement has no meaning at all.
- (ii) Verification is a guide to truth, because we must know how to verify something to set about answering the question of its truth.
Nearly all important arguments turn on the method of verification of the statements involved.
We can now investigate particular types of statements, paying particular attention to their method of verification.
C. TYPES OF STATEMENTS
1. Imperatives and attitude-statements
- Used to give commands or to express the speaker's wishes, hopes, desires and so on. They are of no possible value in argument or discussion.
- What appear to be other kinds of statements may really be imperatives or attitude-statements in disguise - for example, "All men are equal".
- Most statements about liberty, equality and fraternity are disguised attitude-statements.
- We can always discover how a statement is to be treated by discovering how it is to be verified.
2. Empirical statements
- An empirical statement is one which gives information about the world, based on our experience of it. We verify it by tests conducted in terms of our experience: and ultimately, in terms of our sense-experience.
- Empirical statements are neither true nor false on logical grounds alone.
- We can classify empirical statements according to their probability, as judged from the evidence for them, which we cannot do for other kinds of statements.
3. Analytic statements
- Analytic statements do not give information about the world; their verification is not to be found in experience but in seeing whether or not they obey certain logical rules. Their use is to show how we have agreed to relate the meanings of verbal signs to one another.
- They include mathematical statements: arithmetic and algebra are simply manipulations of symbols which we ourselves have contrived. Geometry is even more obviously not dependent on what we experience: we never meet with perfect triangles, circles and squares.
- If an analytic statement is not true it is self-contradictory.
- Although the truth of analytic statements does not depend on experience, they enable us to transform one piece of empirical knowledge into another.
- Sometimes it is not clear whether a statement is analytic or not: and then we have to discover how the speaker verifies it.
- Analytic statements often appear in disguise: for example,
- Everyone's motives for acting are always selfish.
- If someone says, "What about martyrs who die for their cause, or people who sacrifice their own happiness and desires for the good of others?", the speaker replies "They're doing what they desire to do, and that's selfish." The speaker is using "selfish" in a curious way: in a way from which it logically follows that his original statement was true. He will not accept any evidence against motives being selfish, because to him motives are selfish not in point of fact but because of the meanings of the words. Hence his statement is analytic.
- Some apparent statements of value are analytic statements in disguise. "You ought to do your duty" may be analytic, if "duty" means merely "what one ought to do". But you may mean by "duty" a number of specific things, such as fighting for king and country or obeying your parents. In that case the statement may be non-analytic and helpful. It depends on how you would verify what counts as "duty": if you would do this in the same way as you would verify what counts as "what you ought to do", then the meanings of the two terms are the same and the statement is analytic.
4. Value statements
- A value statement includes evaluative words; it commends something or somebody.
- This commendation is more than just expressing an attitude: "Joe is a good man" says more than "Good old Joe!"; it has criteria, means of verification. In making a value statement it is not only or even our main object to express an attitude: we are also assigning value on the basis of criteria.
- The most important thing with value statements is to agree on the proper method of verification. Value judgements about men, actions, motives and societies are essentially the same as value judgements about knives.
- We all have the same criteria of what makes a good knife, but we do not always agree about the criteria for what counts as a good man. What counts as good behaviour is different in different times and places. If I argued with a caveman, it would become clear that we had different criteria for good behaviour.
- We do not have to give up there: it is possible to discuss the criteria themselves – for example, to discuss with the caveman whether stealing is compatible with general happiness. But this will not work if we do not share the higher-level criteria by which we judge this (if, for example, he does not value the general happiness).
- "Poppies are red" is an empirical statement; we are agreed about what counts as red, and if somebody calls poppies green he is not just expressing a difference of opinion, but misusing words. But if one person calls Abraham Lincoln good while another calls him bad, while they agree on all the details of Lincoln's life, neither of them is misusing words: evaluative words are not tailored to fit any one system of verification. People agree about the reasons for describing things, but have different reasons for valuing things.
- If two people hold the same set of criteria, the argument turns simply on whether the subject of the value statement does or does not satisfy the criteria.
5. Metaphysical statements
- Statements about whose verification we aren't agreed, or which even seem to lack any method of verification, we can call metaphysical.
- We can't dismiss them as meaningless: we may come to agree about their method of verification, or even give them one.
- Cf. "There are men on Mars" and "God will save the righteous". The latter we can class as metaphysical because there is complete disagreement about how to verify it; even the individual terms 'God', 'save' and 'righteous' are vague and their use is not established.
- Most metaphysical propositions are religious, but not all: statements about 'ideals' such as justice, beauty, truth and so on, or about natural rights, the self, the will and so on, are so obscure as to count as metaphysical.
- Often these statements can be classified as attitude statements or evaluative statements.
Only those statements that can be reclassified as empirical are informative – in the sense that they give us important information that had escaped our notice.
- Most metaphysical statements must retain that status for the present. Even as simple-seeming a statement as "God exists" cannot be answered, although many fallacious arguments in favour of it have been offered. It's not nonsense but we have been unable to give it an agreed sense, or to agree a method of verifying its truth.
We have now covered all types of statement. Important problems remain. There are no grave problems with empirical, analytic and attitude statements; but there are problems about value statements – for example, how we come to have our value criteria, and whether we can change them - and for metaphysical statements - whether there is any special method for finding meaning and verification for them. It is a good thing to acknowledge the obscurities in these statements rather than to be blind to them.
Chapter III. TRUTH
A. THE CONDITIONS OF TRUTH
We speak of the importance of 'truth', 'certainty' and 'knowledge'. Let us begin with 'truth'. What does it mean to speak of something's being 'true'?
- Truth has to do with statements. If I am to say correctly that a statement is true, I must be able to do three things:
- i) Know what the statement means
- ii) Know the right way to verify it
- iii) Have good evidence for believing it
- The same analysis applies to "I am certain that ..." and "I know that ..." (except perhaps that someone might be certain of something without satisfying iii) - i.e., not having good evidence; but this is strange).
- When we're clear about i) and ii), then finding evidence is largely a matter of patience and hard work. Our scientific progress has outstripped our progress in morals, politics and religion because we have long been agreed about the meaning and verification of scientific statements. By contrast, few people can even distinguish value statements and metaphysical statements among the statements of morals, politics and religion, let alone get clear about them and make progress on agreeing about them.
- So the question of truth and falsity does not arise for imperatives and attitude statements, and the method of determining truth is straightforward for analytic statements and empirical statements.
- The reason why such departments of knowledge as history, chemistry, psychology and mathematics flourish and produce useful results is partly that they use empirical and analytic statements. The job of psychology, in particular, is to produce laws which explain matters of fact. If it cannot yet flourish as it ought, it is probably because psychologists have not yet collected enough evidence.
- The 'branches of knowledge' in which there is a disturbing lack of progress are ethics and metaphysics. However, we should concern ourselves not with these large and mysterious fields but with ethical and metaphysical statements.
- These are the very ideas about which we feel most strongly and ought to want to learn the most. Our difficulties with them arise because they do not satisfy the first two conditions of truth – their meaning and method of verification are not clear.
B. THE BASIS OF MEANING AND VERIFICATION
- Our shared agreements about the meaning and use of words were mostly not reached deliberately and consciously, but grew up. Sometimes however, as in the sciences, agreement has been reached consciously, and the clarity that results is partly the cause of the progress in these fields.
- If we are to reach agreement about the meaning and use of value and metaphysical statements, we shall have to reach it consciously, for we still aren't achieving it after thousands of years by trial and error.
- The verification of "The Earth is round" changed after Galileo [sic]. The phrase "unconscious desires" can be given a definite meaning following Freud.
- We reach such agreements because i) we have similar experiences and ii) we find such agreements useful.Do we have common experiences in regard to value and metaphysical statements?
C. VALUE STATEMENTS
- Where we do not share criteria of value it is because we do not yet have sufficient experience - as with modern artists, for example - or we are prejudiced.
- For example, criteria of good sexual behaviour vary widely from society to society and from age to age. This is because we have little experience of the good and bad effects of different kinds of sexual behaviour and because most people react strongly to behaviour that does not accord with their own criteria.
- Again, the value of one sort of society versus another is not agreed because few people have the experience of both, and few are without prejudice.
- We disagree about the criteria for human values because we do not know much about human beings, and because we do not admit our ignorance.
- The sciences that are supposed to help are psychology, sociology, anthropology and history. They are not yet sufficiently developed to provide the information we need. Even so, they have already done much to bring agreement on values closer.
- As for prejudice: most people feel hostile towards those sciences; and many people who feel themselves to be 'moral' feel strong disapproval towards those who hold other criteria.
- Only a scientific approach can help to solve problems of value. A 'scientific approach' means the attitudes of open-mindedness and an absence of prejudice, and the methods of observation, experiment and hypothesis.
- For example, until recently the rightness of capital punishment was almost universally accepted. Therefore no-one took the trouble to find out if capital punishment is an effective deterrent to murder. Today many people doubt the wisdom of capital punishment and realize the importance of sociological research to arriving at a considered and unprejudiced opinion.
- Scientific research into such questions as "What sort of treatment best reforms criminals, and what sort most effectively deters evildoers?" or "What effect does prolonged physical punishment have on the minds of children?" is a threat to strongly held prejudices and so abhorrent to most people.
- We have every hope that criteria for all values will gradually come to be uniformly accepted throughout mankind.
D. METAPHYSICAL STATEMENTS
Many metaphysical statements are, as we have seen, statements of attitude, so the question of truth does not arise; or statements of value, so can be dealt with as in the preceding section. What we are interested in now is the type of metaphysical statement that might be treated as conveying special information of its own – a type analogous to empirical statements.
Just as our sense-impressions allow us to build a framework of verification for empirical statements, and our common experiences and desires enable us to agree about the verification of value statements, so it may be possible to discover or notice experiences of a special kind which we might use as a basis for the verification of metaphysical statements.
There seems to be no doubt that these special experiences do exist. We have different sorts of experience when:
• we use our senses
• we consider whether we want to commend an action
• we consider whether a work of art is good
• we appreciate the nature of logical and mathematical rules
- and for each of these we have a different framework of meaning and verification. So it is logically possible that special 'metaphysical' experiences exist. And the experiences of the religious when they worship, of mystics when they contemplate, and of ordinary people when they feel 'in tune' with the world seem to be instances of a type of experience that exists in its own right and cannot be assimilated to any other type.
It is not strictly necessary that all people should actually have these special experiences – only that they are available, are the same or similar for all or most people [when they have them] and that we want to describe them in the same way.
I personally believe that it will emerge from investigation that there are at least some experiences which we may all have which we should want to describe by statements now classified as metaphysical. For example, if we all had experiences of what I shall loosely call 'love' and 'power' that could not be accounted for by observation of the natural world, we might wish to describe the source of these experiences by the word 'God'. "God exists" would then have a definite meaning and method of verification, just as "That table exists" has. By discovering these similar experiences and agreeing to use certain signs to describe them, we should be able to build up a secure language of 'metaphysics' (though such statements would not be metaphysical in the sense of having doubtful meaning and verification – we should have to find a new name for them).
Prejudice stands in the way of such investigation. Most supporters of religion or any other type of metaphysical believe that they already have the truth – which they logically cannot have, since the three conditions of truth (III A above) are not satisfied.
I do not wish to commit myself on which statements now classed as metaphysical will survive this investigation. But it cannot now be said with certainty of any metaphysical statement that it is true or false. – it cannot even be said that it is meaningful. But there is at least hope for the future.
The purpose of studying language is to hep us to discover truth. We often have to combat rhetoric and emotional appeals as dangerous enemies. I group these together under the title of 'prejudice', and we shall now see that it is perhaps this factor more than any other that stands between us and the truth.
Almost no-one seems concerned with the importance and difficulty of agreement about the meaning and verification of their metaphysical and value statements. And not only do they use social and religious disapproval to impress on adults and children the necessity of believing such statements – they go further. The mediaeval inquisitors tortured and burned heretics; the Nazis persecuted and burned Jews. The alleged justifications for these acts depended on statements that were entirely unverifiable and possibly meaningless. To torture and kill on this basis seems to border on insanity.
A few observations on the causes for the widespread and occasionally fanatical opposition to discovering meaning and verification:
• Many people have a vested interest in preserving their metaphysical statements intact: for example, the government of Russia [formerly] had every reason to insist that its citizens believe the doctrines of Communism, so that they could be governed more easily. Most religions also fear that too much questioning will prove their houses to have been built upon sand.
• The social and moral behaviour of most people is based on religious or other metaphysical premises, and they are afraid to question them lest the behaviour of others (especially their children) should become immoral and antisocial.
• Men's basic desire is for security; their desire for knowledge is more apparent than real. When people argue they do not do so in the desire for knowledge but in order to persuade others and to increase their own sense of security.
• The obscure statements of morals, politics and religion express beliefs on matters close to our hearts. We are liable to react emotionally to anyone who investigates them with the use of pure reason.
Sometimes people try to cover their weakness here by supporting strange theories about 'faith' or 'intuition' , but the conditions of truth listed above cannot be evaded.
Intuition, conscience, or any other supposed authority must be grounded in reason. This does not diminish their importance: we need faith, but we need to have reasons for it. (We have faith in the engine-driver because we have good evidence that he will bring us to our destination safely.)
Certain conditions must be fulfilled if people are to effectively apply the study of language to the dubious statements offered by politicians, teachers, moralists and so on:
1. We need freedom from government control. Government must not compel citizens to support unverifiable statements by its control of the law and education. In the West we are largely free of the more obvious forms of government control, but the authorities do not encourage children and adults to question the beliefs of their society. (Although it may be justifiable under certain circumstances for governments to prevent such questioning.)
2. We need freedom from the control of religious and public convention. We are fortunate in Britain that there is no single body of moral or religious opinion that is overwhelmingly strong. Nevertheless, there are still many people in our society whose opinions are shaped by the emotional shock-tactics of their immediate environment. One can hardly expect them to apply the study of language to beliefs that they have been forced to hold from an early age.
3. We need freedom from the control of our own emotions. The two previous forms of control largely derive from our fear of doubt. We must recognize our own unwillingness to let go our hold on cherished beliefs.
Most of our fears in this respect are empty ones. We can live a good life without assenting to meaningless and unverifiable statements.
Since this book is for ordinary people, we can suggest one remedy which might help us to overcome these barriers: simply, to get as much practice as possible in applying the study of language to our beliefs - in the classroom, when thinking on our own account, when arguing, when reading and listening to the views of others, and on all the occasions when the truth of statements is in question.
This is not a difficult or specialist activity and does not require outstanding intelligence: it requires only patience and the desire to learn. The results come with surprising speed. Let us at least try it, and see what happens.