In his Wissenschaftslehre, Bolzano spoke of “truths in themselves” and, more generally, of “statements in themselves”, in contradistinction to those (subjective) thought processes by which a man may think, or grasp truths; or, more generally, grasp statements, either true or false.
Bolzano’s distinction between statements in themselves and subjective thought processes has always seemed to me of the greatest importance. Statements in themselves can stand in logical relations to each other: one statement can follow from another, and statements can be logically compatible or incompatible. Subjective thought processes, on the other hand, can only stand in psychological relations. They can disquieten us or comfort us, can remind us of some experiences or suggest to us certain expectations; they can induce us to take some action, or to leave some planned action undone.
The two kinds of relations are utterly different. One man’s thought processes can neither contradict those of another man, nor his own thought processes at some other time; but the contents of his thoughts—that is, the statements in themselves—can of course contradict the contents of another man’s thoughts. On the other hand, contents, or statements in themselves, cannot stand in psychological relations: thoughts in the sense of contents or statements in themselves and thoughts in the sense of thought processes belong to two entirely different “worlds”.
If we call the world of “things”—of physical objects—the first world, and the world of subjective experiences (such as thought processes) the second world, we may call the world of statements in themselves the third world. (I now293 prefer to call these three worlds “world I”, “world 2”, and “world 3”; Frege sometimes called the latter the “third realm”.)
Whatever one may think about the status of these three worlds—I have in mind such questions as whether they “really exist” or not, and whether world 3 may be in some sense “reduced” to world 2, and perhaps world 2 to world 1—it seems of the utmost importance first of all to distinguish them as sharply and clearly as possible. (If our distinctions are too sharp, this may be brought out by subsequent criticism.)
At the moment it is the distinction between worlds 2 and 3 which has to be made clear; and in this connection, we will come up against, and must face, arguments like the following.
When I think of a picture I know well, there may be a certain effort needed to recall it and “put it before my mind’s eye”. I can distinguish between (a) the real picture, (b) the process of imagining, which involves an effort, and (c) the more or less successful result, that is, the imagined picture. Clearly, the imagined picture (c) belongs exactly like (b) to world 2 rather than to world 3. Yet I may say things about it which are quite analogous to the logical relations between statements. For example, I may say that my image of the picture at time t1 is incompatible with my image at time t2 and even perhaps with a statement such as: “In the picture only the head and shoulders of the painted man are visible.” Moreover, the imagined picture may be said to be the content of the process of imagining. All this is analogous to the thought content and the process of thinking. But who would deny that the imagined image belongs to world 2; that it is mental, and indeed part of the process of imagining?
This argument seems to me valid and quite important: I agree that within the thinking process some parts may be distinguished that may perhaps be called its content (or the thought, or the world 3 object) as it has been grasped. But it is precisely for this reason that I find it important to distinguish between the mental process and the thought content (as Frege called it) in its logical or world 3 sense.
I personally have only vague visual imaginings; it is usually only with difficulty that I can recall a clear, detailed, and vivid picture before my mind. (It is different with music.) Rather, I think in terms of schemata, of dispositions to follow up a certain “line” of thought, and very often in terms of words, especially when I am about to write down some ideas. And I often find myself mistaken in the belief that I “have got it”, that I have grasped a thought clearly: when trying to write it down I may find that I have not got it yet. This “it”, this something which I may not have got, which I cannot be quite certain that I have got before I have written it down, or at any rate formulated it in language so clearly that I can look at it critically from various sides, this “it” is the thought in the objective sense, the world 3 object which I am trying to grasp.
The decisive thing seems to me that we can put objective thoughts—that is, theories—before us in such a way that we can criticize them and argue about them. To do so, we must formulate them in some more or less permanent (especially linguistic) form. A written form will be preferable to a spoken form, and printing may be better still. And it is significant that we can distinguish between the criticism of a mere formulation of a thought—a thought can be formulated rather well, or not so well—and the logical aspects of the thought in itself; its truth; or its truthlikeness in comparison with some of its competitors; or its compatibility with certain other theories.
Once I had arrived at this stage I found that I had to people my world 3 with inmates other than statements; and I brought in, in addition to statements or theories, also problems, and arguments, especially critical arguments. For theories should be discussed always with an eye to the problems which they might solve.
Books and journals can be regarded as typical world 3 objects, especially if they develop and discuss a theory. Of course the physical shape of the book is insignificant, and even physical nonexistence does not detract from world 3 existence; think of all the “lost” books, their influence, and the search for them. And frequently even the formulation of an argument does not matter greatly. What do matter are contents, in the logical sense or world 3 sense.
It is clear that everybody interested in science must be interested in world 3 objects. A physical scientist, to start with, may be interested mainly in world 1 objects—say, crystals and X-rays. But very soon he must realize how much depends on our interpretation of the facts, that is, on our theories, and so on world 3 objects. Similarly a historian of science, or a philosopher interested in science, must be largely a student of world 3 objects. Admittedly, he may also be interested in the relation between world 3 theories and world 2 thought processes; but the latter will interest him mainly in their relation to theories, that is, to objects belonging to world 3.
What is the ontological status of these world 3 objects? Or, to use less high-sounding language, are problems, theories, and arguments “real”, like tables and chairs? When some forty-four years ago Heinrich Gomperz warned me that I was, potentially, not only a realist in the sense of believing in the reality of tables and chairs but also in the sense of Plato, who believed in the reality of Forms or Ideas—of concepts, and their meanings or essences—I did not like the suggestion, and I still do not include the left-hand side of the table of ideas (see section 7 above) among the denizens of my world 3. But I have become a realist with respect to the world 3 of problems, theories, and critical arguments.
Bolzano was, I think, doubtful about the ontological status of his statements in themselves, and Frege, it seems, was an idealist, or very nearly so. I too was, like Bolzano, doubtful for a long time, and I did not publish anything about world 3 until I arrived at the conclusion that its inmates were real; indeed, more or less as real as physical tables and chairs.
Nobody will doubt this as far as books are concerned, and other written matter. They are, like tables and chairs, made by us, though not in order to be sat upon, but in order to be read.
This seems easy enough; but what about the theories in themselves? I agree that they are not quite as “real” as tables and chairs. I am prepared to accept something like a materialist starting point according to which, in the first place, only physical things like tables and chairs, stones and oranges, are to be called “real”. But this is only a starting point: in the second place we are almost bound to extend the range of the term radically: gases and electric currents may kill us: should we not call them real? The field of a magnet may be made visible by iron filings. And who can doubt, with television such a familiar phenomenon, that some sort of reality has to be attributed to Hertz’s (or Maxwell’s) waves?
Should we call the pictures we see on television “real”? I think we should, for we can take photographs of them with the help of various cameras and they will agree, like independent witnesses.294 But television pictures are the result of a process by which the set decodes highly complicated and “abstract” messages transmitted with the help of waves; and so we should, I think, call these “abstract” coded messages “real”. They can be decoded, and the result of the decoding is “real”.
We are now perhaps no longer quite so very far removed from the theory in itself—the abstract message coded in a book, say, and decoded by ourselves when we read the book. However, a more general argument may be needed.
All the examples given have one thing in common. We seem to be ready to call real anything which can act upon physical things such as tables and chairs (and photographic film, we may add), and which can be acted upon by physical things.295 But our world of physical things has been greatly changed by the content of theories, like those of Maxwell and Hertz; that is by world 3 objects. Thus these objects should be called “real”.
Two objections should be made. (1) Our physical world has been changed not by the theories in themselves but, rather, by their physical incorporation in books, and elsewhere; and books belong to world 1. (2) It has been changed not by the theories in themselves, but by our understanding of them, our grasp of them; that is, by mental states, by world 2 objects.
I admit both objections, but I reply to (1) that the change was brought about not by the physical aspects of the books but solely by the fact that they somehow “carried” a message, an informative content, a theory in itself. In response to (2), which I regard as a far more important objection, I admit even that it is solely through world 2 as an intermediary between world 1 and world 3 that world 1 and world 3 can interact.
This is an important point, as will be seen when I turn to the body-mind problem. It means that world 1 and world 2 can interact, and also world 2 and world 3; but world 1 and world 3 cannot interact directly, without some mediating interaction exerted by world 2. Thus although only world 2 can act immediately upon world 1, world 3 can act upon world 1 in an indirect way, owing to its influence upon world 2.
In fact, the “incorporation” of a theory in a book—and thus in a physical object—is an example of this. To be read, the book needs the intervention of a human mind, of world 2. But it also needs the theory in itself. For example, I may make a mistake: my mind may fail to grasp the theory correctly. But there is always the theory in itself, and somebody else may grasp it and correct me. It may easily be not a case of a difference of opinion, but a case of a real, unmistakable mistake—a failure to understand the theory in itself. And this may even happen to the originator of the theory. (It has happened more than once, even to Einstein.)296
I have touched here on an aspect which I have described in some of my papers on these and related subjects as the (partial) autonomy of world 3.297
By this I mean that although we may invent a theory, there may be (and in a good theory, there always will be) unintended and unforeseen consequences. For example, men may have invented the natural numbers or, say, the method of proceeding without end in the series of natural numbers. But the existence of prime numbers (and the validity of Euclid’s theorem that there is no greatest prime) is something we discover. It is there, and we cannot change it. It is an unintended and unforeseen consequence of that invention of ours. And it is a necessary consequence: we cannot get around it. Things like prime numbers, or square numbers, and many others, are thus “produced” by world 3 itself, without further help from us. To this extent it may be described as “autonomous”.
Somewhat related to the problem of autonomy but, I think, less important, is the problem of the timelessness of world 3. If an unambiguously formulated statement is true now, then it is true for ever, and always was true: truth is timeless (and so is falsity). Logical relations such as contradictoriness or compatibility are also timeless, and even more obviously so.
It would be easy for this reason to regard the whole of world 3 as timeless, as Plato suggested of his world of Forms or Ideas. We only need to assume that we never invent a theory but always discover it. Thus we would have a timeless world 3, existing before life emerged and after all life will have disappeared, a world of which men discover here or there some little bits.
This is a possible view; but I don’t like it. Not only does it fail to solve the problem of the ontological status of world 3, but it makes this problem insoluble from a rational point of view. For although it allows us to “discover” world 3 objects, it fails to explain whether, in discovering these objects, we interact with them, or whether they only act upon us; and how they can act upon us—especially if we cannot act upon them. It leads, I think, to a Platonic or neo-Platonic intuitionism, and to a host of difficulties. For it is based, I think, upon the misunderstanding that the status of the logical relations between world 3 objects must be shared by these objects.
I propose a different view—one which, I have found, is surprisingly fruitful. I regard world 3 as being essentially the product of the human mind. It is we who create world 3 objects. That these objects have their own inherent or autonomous laws which create unintended and unforeseeable consequences is only an instance (though a very interesting one) of a more general rule, the rule that all our actions have such consequences.
Thus I look at world 3 as a product of human activity, and as one whose repercussions on us are as great as, or greater than, those of our physical environment. There is a kind of feedback in all human activities: in acting we always act, indirectly, upon ourselves also.
More precisely, I regard the world 3 of problems, theories, and critical arguments as one of the results of the evolution of human language, and as acting back on this evolution.
This is perfectly compatible with the timelessness of truth and of logical relations; and it makes the reality of world 3 understandable. It is as real as other human products, as real as a coding system—a language; as real as (or perhaps even more real than) a social institution, such as a university or a police force.
And world 3 has a history. It is the history of our ideas; not only a history of their discovery, but also a history of how we invented them: how we made them, and how they reacted upon us, and how we, in our turn, reacted to them.
This way of looking at world 3 allows us also to bring it within the scope of an evolutionary theory that views man as an animal. There are animal products (such as nests) which we may regard as forerunners of the human world 3.
And ultimately it suggests a generalization in another direction. We may regard the world of problems, theories, and critical arguments as a special case, as a world 3 in the narrow sense, or else as the logical or intellectual province of world 3. World 3 in a more general sense includes all the products of the human mind, such as tools, institutions, and works of art.
I first lectured on this view of world 3 and on its history in 1960 to my seminar at the L.S.E. (First publication: 1968 (s).)