§6. We could imagine that the language of §2 was the whole language of A and B; even the whole language of a tribe. The children are brought up to perform these actions, to use these words as they do so, and to react in this way to the words of others.

An important part of the training will consist in the teacher's pointing to the objects, directing the child's attention to them, and at the same time uttering a word; for instance, the word "slab" as he points to that shape. (I do not want to call this "ostensive definition", because the child cannot as yet ask what the name is. I will call it "ostensive teaching of words".—I say that it will form an important part of the training, because it is so with human beings; not because it could not be imagined otherwise.) This ostensive teaching of words can be said to establish an association between the word and the thing. But what does this mean? Well, it can mean various things; but one very likely thinks first of all that a picture of the object comes before the child's mind when it hears the word. But now, if this does happen—is it the purpose of the word?—Yes, it can be the purpose.—I can imagine such a use of words (of series of sounds). (Uttering a word is like striking a note on the keyboard of the imagination.) But in the language of §2 it is not the purpose of the words to evoke images. (It may, of course, be discovered that that helps to attain the actual purpose.)

But if the ostensive teaching has this effect,—am I to say that it effects an understanding of the word? Don't you understand the call "Slab!" if you act upon it in such-and-such a way?—Doubtless the ostensive teaching helped to bring this about; but only together with a particular training. With different training the same ostensive teaching of these words would have effected a quite different understanding.

"I set the brake up by connecting up rod and lever."—Yes, given the (5) whole of the rest of the mechanism. Only in conjunction with that is it a brake-lever, and separated from its support it is not even a lever; it may be anything, or nothing.


  1. Why does W. start to discuss A teaching and training B (Master and Novice) - to which he often returns? Why should how language is acquired be something that the philosopher should care about?  It doesn't seem to be of much intrinsic philosophical interest. He hopes that it will illuminate how language actually works - in effect, how it must work, or at least how it is plausible that it works, given that this is the way that it is acquired. For example, perhaps, if language works the way that he thought it did in his Tractatus, how could it possibly be learned? How could teaching a child a meaning be possible, if it is process which is apparently public, but is supposed to establish an association between a word and an entity which is essentially private? And if the Tractatus approach is right, wouldn't teaching be, from the beginning, a rather simple process of pointing and naming - of ostensive definition? Yet, he says, this is not what happens. Learning is not atomic, like this, but always involves context.

  2. For example, the very first teaching of an association between the word 'Rover' and an object (the family dog) is not ostensive definition of a name. The father points at the dog and says "Rover". The dog comes into the room, and the child says "Rover!". The father sticks a yellow Post-It note on the poor dog, which has 'Rover' written on it. At this early stage, the word 'name' has not appeared in the child's language; the child cannot ask "Daddy, what is the name of that?" (pointing at the cat).

  3. When the father says "Rover", perhaps without the dog present, a private "image" of the dog may appear in the child's mind. The appearing of this image may be a purpose of language, of the use of the word 'Rover', but it is not the only purpose; indeed, a complete model language could work with such images completely absent. A says "Slab!"; B goes over and fetches the appropriate object. Unconscious machines with voice synthesisers and microphones could operate like this. A conscious person who found that 'slab' also, as it happens, evokes an image might be able to operate in the form of life better - but this is merely a bonus. Yes, and beyond that, such a person might find it much easier to draw an absent slab, which then... But this is not directly relevant. W.'s target is his Tractatus view that language works with such images essentially. He doesn't need to assert that the images are totally unimportant for a FoL, just that they are unnecessary to the working of language. He is going to argue that the images, the feelings, the atmosphere, that may, or may not, be called up when A says "Slab!", or "Now he is reading", or "Now I understand how to go on", or "Now I know what 'love' means!", are amusing, where they exist; they may be intense; they may be the most important things in our lives; but they are irrelevant to the operation of our language.

  4. It is nice, with W., to imagine a language whose purpose is to evoke images and sensations entirely, and nothing else. A says "Sad", and B, hearing it, recalls that particular sensation; that is all. B says "Mouse", and A imagines a little furry, scurrying, creature; again, that is all. When people actually are sad, they don't use the word; when people actually see the little creatures, they don't use the word. This is a good example of W.'s use of devised simple model languages to illuminate aspects of our multi-faceted actual one.

  5. W. use of 'whole' is significant, since he is inclined to move from the atomic simplicity of the Tractatus to a description which is more holistic. 'Red', for example, cannot possess a meaning by simply being linked to <. It is linked with the 'colour' of <. Therefore the word 'red' is inextricably tied up with he word 'colour'. 'Colour' here is like the support, that is needed before the piece of metal ('red') becomes a brake-lever, being linked to < (the brake). Without the support of 'colour', red can be doing anything, or nothing.