Second, there is a sense in which, despite everything said above, metaphysical theories are subject to the test of experience. That metaphysics aspires to give an account of the world as a whole means that each metaphysician claims that his fundamental insight illuminates every department of life. It may be that there are no neutral facts to which a metaphysician can appeal to show the shortcomings of his opponents; metaphysicians pronounce on what is to count as fact, and this puts them in the happy position of being judges in their own case. It remains true, however, that everyone who engages in that type of philosophy has the formal task of accounting for all the facts that he recognizes, and this is something that can be done more or less well. The value of different metaphysical insights is sometimes shown in the success with which they are applied. Furthermore, it is not quite true that the metaphysician need consult no opinion but his own when it comes to working out his views. What might be called public opinion has a part to play as well, though it has no absolute right to a hearing. A metaphysician who chooses to dismiss areas of experience or ways of thinking that are commonly accepted as being in order does so at his peril; he reduces the initial plausibility of his own theories the oftener he finds himself in this position. He could, of course, be right and common opinion wrong; no genuine metaphysician is put off by the thought of such a conflict. Though he is not put off, however, he has to be wary all the same. He may be able to say what in the end is to count as fact, but if this involves him in dismissing as illusory what instructed opinion generally takes to be real, his triumph may be hollow. Whether he likes it or not, he has to frame a theory that will carry conviction with experts in the different fields concerned, or, if that is going too far, one that will strike them as not wholly implausible. A metaphysician who exercises his veto past that point is simply failing to do his job.
It must be admitted that the tests one can apply to determine the value of a metaphysical theory are at best unsatisfactory. Often one is driven back onto the expedient of asking if the theory is internally self-consistent; a surprisingly large number of philosophical theories are not. To confute a philosopher out of his own mouth is, perhaps, the most effective form of confutation. If this expedient will not apply, however, the questioner is not quite helpless. Whatever the explanation, it is a well-known fact that a philosopher can purchase consistency at the expense of plausibility; he can put forward theories that evade difficulties by simply declaring them nonexistent. In so doing, he turns his back on what instructed opinion generally takes to be fact. His hope is, of course, to persuade others to see the situation as he does, and there is always the possibility that he will succeed. If, however, after a suitable interval he has not, that must surely count against him. It is by this test that one decides, for example, that the metaphysics of Hobbes is not worth prolonged study, despite the enormous ingenuity of its author; there is too much in this system that seems to be sheerly arbitrary. The same comment could be made of certain forms of Idealism, which are so intent on the omnipresence of spirit that they neglect the materiality of the material order. Admittedly, the test is harder to apply when attention is transferred to the major theories in their most persuasive form, because here the question concerns views that have stood the test of time. It is not, however, entirely inapplicable even there. An individual, at least, may feel that this or that view will not do precisely because it achieves comprehensiveness by turning its back on fact; and, though it is unsatisfactory to fall back on personal judgment in this way, there is perhaps no other alternative in this difficult area. Role of personal or social factors
Some writers on the philosophy of philosophy, such as Dilthey, have suggested that the persistence of a plurality of metaphysical systems is to be explained in terms of personal or social factors. Certain kinds of metaphysical outlook appeal to certain types of human being, or gain currency in social circumstances of this kind or that; to understand why they are accepted, recourse must be had to psychology or sociology or both. In the above account, stress has been laid on the historical background against which a number of famous metaphysical theories got their classical formulations; it is idle to deny that each was originally designed to solve a problem deemed to be urgent at the time. Nevertheless, the problem was, of course, an intellectual problem, and the solution offered claimed to be true, not simply comforting. No doubt wishful thinking is as rife in the field of metaphysics as anywhere; it is all too easy here to confuse what men ought to believe with what they want to believe. Philosophies reveal something about their authors and even about their historical age, as works of literature do; they constitute historical evidence as books on mathematics, perhaps, do not. Yet all this can be admitted without agreeing that metaphysics is merely of psychological or historical importance. Science does not cease to be true because it is shown to be useful. Nor is it true that metaphysical theories always in fact give comfort; there are cases in which men find themselves returning over and over again to possibilities that they would very much like to believe were not realized. A philosopher can commit himself to a view of the world that is not at all to his taste, simply because it seems to him on due consideration that this is how things are. That philosophers are godlike beings able to rise entirely above the limitations of their age seems unlikely. It is equally unlikely, however, that their opinions are determined throughout by nonrational factors, and thus that their thinking can lay no claim to truth. Metaphysical arguments Logical character of metaphysical statements
Metaphysical statements fall into two main classes: statements about what exists and prescriptions about how to take or understand what exists. It might seem obvious that the first is the more important; the metaphysician first lays down what he takes to exist, and then tells how to interpret it. This would be correct if metaphysics were a departmental inquiry like, for example, botany; but, of course, it is not. Metaphysicians possess no special resources for the detection of unfamiliar entities, and in consequence the realities they accept must all be argued for. The fundamental items that fill the metaphysical world are one and all theoretical; they are not so much palpable realities as artificial constructs. That being so, there is less of a gulf between the two types of metaphysical pronouncement than might at first appear. It could indeed be argued that the two go closely together to constitute what may be called a metaphysical point of view, a standpoint whose primary purpose is to provide understanding. In a metaphysical context, to say what exists is itself a step on the way to understanding; it is not something that antedates theory, but part of a theory itself.
It may be asked whether metaphysical pronouncements are empirical or a priori and, if the latter, whether they are analytic or synthetic. They are certainly not straightforwardly empirical, for reasons just set out, and cannot be merely analytic (i.e., true in virtue of the definitions of their terms and of the laws of logic) if metaphysics is to retain any significance. The conclusion that they must be synthetic a priori (i.e., such that, unlike analytic propositions, they convey new knowledge and yet claim complete universality and necessity) seems to follow, and it is just what the opponent of metaphysics wants the metaphysician to adopt. Metaphysics, as he sees it, is a wholly unwarranted attempt to say what the world must be like on the strength of pure thinking, an attempt that is doomed to failure from the start. Before this condemnation is accepted, however, the function that the metaphysician assigns to his principles should be considered. When this is done, it becomes plain that the charge that he claims factual knowledge of a nonempirical sort is false; in one way he recognizes exactly the same facts as anyone else. Where he claims superiority is in knowing how to take facts, and the burden of his message consists in the advocacy of principles that, he alleges, will provide overall understanding. One can describe these principles as synthetic a priori if one chooses. It is probably best, however, to avoid this misleading term and simply say that they are thought of by the metaphysician as applying unequivocally to whatever falls within experience. These metaphysical principles are instructive at least in the sense of having alternatives, and they are certainly treated as being necessary. It is not true, however, that they take the form of statements of fact, even highly general statements of fact; nor is their necessity the same as that which characterizes logical truths. The principles are prescriptions rather than statements, and their necessity arises from the role they play in the constitution of experiential knowledge. It is a necessity that is in one way absolute: nothing that can claim to be real can escape their jurisdiction, because they tell how to take whatever occurs. Nevertheless, in another way the necessity of the principles is merely conditional, for other ways of interpreting the same data can be conceived, and it is admitted that there are circumstances, however hard to specify exactly, in which it would have to be agreed that they do not apply. Logical form of metaphysical arguments