Wittgenstein is often referred to as a very difficult philosopher, but I think it is more accurate to say that he is simple and yet not obvious. Once you have grasped his overall approach, you would not refer to him as difficult in whole, but you may still find some of his specific examples and analogies to be difficult.
Part of the problem in Wittgenstein studies is that he never published any of his best work formally, and was adverse to creating narrative order from his aphorisms. There are, however, works such as his lectures which in places provide useful introductions to Wittgenstein’s work using a soft and comfortable narrative.
The excerpts that I have transcribed here are obscurely placed in one of the middle of his lecture series, and yet I think that they serve a really excellent basis to anyone who is working through the beginnings of understanding Wittgenstein to cement the most important points together regarding his technique. The text below is an amalgamation of lectures ten and eleven of Michaelmas Term, 1934. Comments [interpolated] are by me.
The question has been raised how far my method is the same as what is called description of meaning by exemplification. That sounds as if I had invented a method, a means of giving a meaning which is just as good as definition. The point of examining the way a word is used is not at all to provide another method of giving its meaning. When we ask on what occasions people use a word, what they say about it, what they are right to substitute for it, andin reply try to describe its use, we do so only insofar as it seems helpful in getting rid of certain philosophical troubles. We seem to be asking about the natural history of human beings. Yet you know that in some obvious sense we are not interested in natural history. Nevertheless, when I say that a word is as a matter of fact defined in such-and-such a way, or ask whether people might accept a certain definition, I seem to be talking about natural history. But it is not natural history to invent languages of our own, as I have done, and lay down rules for such languages as, for example, the chemists of the 19th century did with the language of chemistry. We are interested in language only insofar as it gives us trouble. I only describe the actual use of a word if this is necessary to remove some trouble we want to get rid of. Sometimes I describe its use if you have forgotten it. Sometimes I have to lay down new rules because new rules are less liable to produce confusion or because we have perhaps not thought of looking at the language we have in this light. Thus we may make use of the facts of natural history and describe the actual use of a word; or I may make up a new game for the word which departs from its actual use, in order to remind you of its use in our own language. The whole point is that I cannot tell you anything about the natural history of language, nor would it make any difference if I could. On all questions we discuss I have no opinion; and if I had, and it disagreed with one of your opinions, I would at once give it up for the sake of argument because it would be of no importance for our discussion. We constantly move in a realm where we all have the same opinions. All I can give you is a method; I cannot teach you any new truths. It is the essence of philosophy not to depend on experience, and this is what is meant by saying that philosophy is a priori.
One could teach philosophy solely by asking questions.
When we give a description of the use of a word we do so only so far as it seems helpful in removing certain troubles. For example, people are troubled by the assumption that there must be something common to all uses of the word “good”. They say, “There is one word; therefore there must be one thing common to all its uses”. Every philosophical problem typically contains one particular word or its equivalent, the word “must” or “cannot”. The word “must” in the present case means that one is misled into supposing that because there is one word there must be one thing in common. One can be obsessed by a certain language form. One can think for years about a certain problem and make no progress because one never thinks of making up a new language. A philosophical trouble is an obsession, which once removed it seems impossible that it should ever have had power over us. It seems trivial.
The obsessions of philosophers vary in different ages because terminologies vary. When a terminology goes some worries may pass, only to arise again in a similar terminology. Sometimes a scientific language produces an obsession and a new language rids us of it. When dynamics first flourished it gave rise to certain obsessions which now seem obsolete. Something may play a predominant role in our language and be suddenly removed by science, e.g., the word “earth” lost its importance in the new Copernican notation. Where the old notation had given the earth a unique position, the new notation but lots of planets on the same level. (It might be said that Copernicus discovered certain facts about the planets, and that it was the discovery of these facts which removed the obsession about the earth and not the change from Ptolemy’s notation. But the new facts might still have been expressed, in a complicated way, in Ptolemy’s notation and the obsession not removed. On the other hand, the obsession might have been removed had Copernicus made up a notation with the sun as centre, even though it had no application. Of course Copernicus did not think about notation but about planets.) Any obsession arising from the unique position of something in our language ceases as soon as another language appears which puts that thing on a level with other things. When there was only one dynamics philosophers asked how they could reduce everything to one mechanism, and became obsessed. With the discovery of several other dynamics, the obsession disappeared.
In the case of the confusion between “is” and “equals”, philosophers noted the use of “is” in “2 + 2 is (equals) 4” and “The rose is red”, and went on to ask whether the rose equals, or is identical with, red, and so on. When logicians like Frege and Russell introduced the symbol “epsilon”, a difference in the uses of “is” was brought out in a way which was not brought out between “is” and “equals” in our ordinary language. Our ordinary language is in tremendous flux, so that it is difficult to make distinctions in it. Their notation removed the temptation to treat different things as identical. I invented a notation to get rid of the identity sign as used in “A = A”, because nobody ever says a chair is a chair; and the difficulties connected with this use vanish.
Treating language as we have here brings with it the puzzle: What becomes of the rigidity of logic? We have the impression that logic is not a thing within our control. There seems to be a way of explaining why it is not by thinking that logicians make up an ideal language to which our normal language only approximate. I once said that logic describes the use of language in a vacuum. Games or languages which we make up with stated rules one might call ideal languages, but this is a bad description since they are not ideal in the sense of being “better”. They serve one purpose, to make comparisons. They can be put beside actual languages so as to enable us to see certain features in them and by this means to get rid of certain difficulties. Suppose I make up a language in which “is” has two meanings. Is it better? Not from the practical point of view. No ordinary person mixes up the meaning of “is” in “The rose is red” and “2 + 2 = 4”. It is ideal in the sense of having simply statable rules. Its only point is to get rid of certain obsessions—it does not do more. One might of course suggest a notation or language which would be better for some practical purpose, but this would be accidental. It is not part of our design.
It is characteristic of obsessions that they are not recognised and at certain stages are not even recognisable. These are attacked as scientific problems are, and are treated perfectly hopelessly, as if we had to find out something new. The problems do not appear to concern questions about language but rather question of fact of which we do not yet know enough. It is for this reason that you are constantly tempted to think I am giving you some information, and that you expect from me a theory. In using the words “I think so-and-so” it looks as if I were discussing the problems of a science called metaphysics.
I shall now go on to discuss propositions. All sorts of definitions of “proposition” have been given. But when I am asked about propositions I explain by examples. The examples are usually sentences in some known language, and that produces the strong feeling of there being something common to all such sentences. Some people have said that a substantive and verb are what is common. Hardly anyone will give as an example of a proposition a command such as “Take this chair to Mr. Smith” or “Walk in this direction”, uttered in conjunction with a gesture. (Is the gesture combined with the command part of the command?) The proposition is usually considered to consist of words. But where there are no words, say in a line drawn like this, __|```|__|```|__, a person might understand that he is to walk in a certain direction. Is this a sentence? If it is, then we would not say that what is common to sentences is a substantive and a verb.
Logicians have had the obsession that the life of a proposition is the copula “is”. But they know as well as we do that not all English sentences have a subject, a copula, and an adjective. They have said that every sentence can be “reduced” to such a sentence. The fact that it can be reduced is analogous to the following: that every closed curve is said to be a circle. To the objection that a certain curve is not a circle, the reply is that it could be projected into a circle. To twist “Every closed curve can be projected into a circle” into “Every closed curve is a circle” is exactly paralleled by “Every proposition can be transformed into one of the subject-predicate form”. But in neither case have I said anything about the method of projection, and until I do, I have said nothing. I could have said that I was going to use a symbolism in which every close curve would be represented by a circle. That is a rule: to replace “closed curve” by “circle”. Similarly, I could say that I would transform every propositional form into a subject-predicate sentence. The statement that every proposition has a subject and predicate could assert a fact of natural history; there could be a language composed entirely of such sentences. On the other hand, in uttering that statement I might be laying down a rule, a rule about the jingle of my sentences. But the rule does not enlighten us unless I say how I propose to transform all propositions into this form, i.e., what is to be done with other words than substantives, copula, and adjectives.
Difficulty is created by the fact that we have invented an enormously complicated language for our use and we are all grown-ups. The philosophy of a child would be quite different from ours, but in a different sense than the physics of a child would be. The physics of a child would be different because it does not know various physical facts, but its philosophy would be different because its language is simpler. It will be very valuable to study more primitive examples of language, what I call “language games” (synonymous with “primitive languages” for the most part). These will bear the same relation to our language as primitive arithmetic bears to our arithmetic. It is a fallacy to suppose these languages are incomplete. Primitive arithmetic is not incomplete, even one in which there are only the first five numerals; and our arithmetic is not more complete. Would chess be incomplete if we knew another game which somehow incorporated chess? It would be merely a different game. To think otherise is to confuse mathematics with a natural science. If mathematics were a science of numbers as pomology is a science of apples, then a mathematics which did not include irrational numbers or numbers after 5 would be counted incomplete, just as would a treatise on pomology which left out reference to one sort of apples. And the latter would be incorrect if it invented kinds of apples which did not exist. But mathematics is not natural science.
* * *
Surrounding the blurred whole of ordinary language there are the special languages, e.g., the languages of chemistry and meteorology. I shall consider a language as such a conglomerate. In our language we find a mixture of descriptions, hypotheses, questions, orders, etc., but any list we made of these would be entirely inadequate. Let us compare it with a simple language in a tribe in which only orders are given. We, who talk about the tribe having this language, call these “orders” because the rule these words play in the life of the tribe is that of orders. The word “order” is not in their language, nor is there any such thing as conversation. The whole object is communication between a builder and his workman. The builder orders “Brick!”, for example, upon which the workman brings him a brick.
We shall suppose that a child learns this language by being drilled. He is given, say, ten words, such as “brick”, “column”, “clay”. In the description of this training is understanding left out? You will say the child must understand the words else he cannot be taught to react to orders. I reply, Certainly, if you like, just as a dog can be taught to look after sheep. A calf or cat cannot be taught; I could go through all the motions with these animals and would not get an appropraiate reaction. Training can be described as consisting of two steps (1) the trainer’s doing certain things, (2) the occurence of certain reactions on the part of the subject, with the possibility of improvement. Teaching a language always depends on a training which presupposes that the subject reacts. If the subject does not react in a given case, that is, does not understand, reference to understanding will be not appear in the description of the training. But nothing is omitted from the description by omitting reference to understanding.
Now there is a certain preliminary exercise to obeying the order, namely, learning what to do when an order, e.g., “Brick!”, is given. This is very close to what we should call “giving the thing a name”. The mother puts a brick on a pile and says “brick”, and then the child does the same thing. Notice that “brick”, said in the presence of the child is not properly an ostensive definition, because in this language we have not yet the question, What is this called? It is a process of naming in a different kind of surroundings.
The question might be raised whether the word “brick” has the same meaning in this language as in ours. You might say that the builder means by it what we mean by “Bring me a brick”. But this would be dangerous. Although these expressions play the same role in the two languages, in the primitive language the words “bring me” do not come in. We could imagine that even in English, although we said “Bring me so-and-so” for everything else, instead of “Bring me a brick” we said “Brick”, as in the orders “Charge” and “Fire” in military usage. Then the word “brick” would play a different role from what it plays in the sentence “There is a brick”.
Consider now another language (2) in which an order consists of two words. Besides the words “brick”, “column”, etc. we have a series of letters A–J or a series of ten notes, say the first ten notes of “God Save the King”. These must be learned by heart, whereas words such as “brick” are not. The order now consists of a word and a letter, say “E brick!”. The child must go to the pile of bricks, take up one brick for each letter through E, and bring E number of bricks to the master. The letters of the alphabet are thus seen to be numerals in this language. The tribe has a very primitive arithmetic in which it can count up to 10, but has no addition or multiplication. Note how different are the function of the words of this language (a) that of the letters of the list of ten, which must be learned by heart, and (b) that connected with actions of bringing something which the builder orders. Although the word “E” and “brick”, a spoken or written, are similar, their functions are in no way comparable.
You will notice that in the two language which we have described there is in a particular sense no “understanding”. There is nothing corresponding to asking for the name of a thing or giving it a name. The philosophical question about meaning would not arise for the philosophers of our tribe.
We now introduce another game (3) having question and answer. We might have, say, twenty-five letters or numerals, and the words “brick”, etc., as before. We suppose the helper can count the bricks against his letters, and that for any number beyond twenty-five he says “many”. The role of “many” then is rather like that of a numeral and yet different. The question in our game might always be “How many?” This would be answered by “J”, or by “Y”, or by “many” for a number over twenty-five.
In the game (1) we had in the training something which was somewhat similar to ostensive definition. For the numerals in game (2) we could have a sort of ostensive definition. When shown three bricks the helper would be taught to say “C” or “3”, instead of learning the numbers by heart, and this might be an ostensive definition of “3”. Here we have a different sense of ostensive definition. Three columns would be “3” as well as three bricks.
Another language (4) might introduce the word “there”, which has a different function still from substantives and numerals. An order would be, for example, “J bricks there!” together with a gesture of pointing, which would be followed by the helper’s putting bricks over there. Look now at the use of the word “there”. One might perhaps say it is the name of a place. But to call it a name is to use the word “name” in a very different sense from that of the name “Charing X” [Charing Cross, in London]. “There” has no meaning unless it is accompanied by a gesture. Are expressions such as “Brick!” and “J bricks there!” sentences? As you like. You can draw the distinction wherever you like, but it is not easy to show why it should be drawn at any particular point.
* * *
Do not make the mistake of supposing that I am showing how language is built up or how it has evolved. Sometimes it is easier to imagine these invented languages of a primitive tribe and sometimes as the actual primitive language of a child. A child does actually begin with such a primitive language. Its language training is mostly in the form of such games. A new game introduces a new element into language, for example, negation. It will be noticed that the elements we have already introduced [some of which I’ve omitted] are of great variety. The difficulty of this method of exhibiting language games is that you think it is perfectly trivial. You do not see its importance.
(Ambrose, Alice (ed.) (1979). Wittgenstein’s Lectures, Cambridge, 1932–1935. From the Notes of Alice Ambrose and Margaret Macdonald. ISBN 0-631-10141-1. pp.96–105)
Sean B. Palmer