I have tried in this book to distinguish clearly two kinds of question, which moral philosophers have always professed to answer, but which, as I have tried to shew, they have almost always confused both with one another and with other questions. These two questions may be expressed, the first in the form: What kind of things ought to exist for their own sakes? the second in the form: What kind of actions ought we to perform? I have tried to shew exactly what it is that we ask about a thing, when we ask whether it ought to exist for its own sake, is good in itself or has intrinsic value; and exactly what it is that we ask about an action, when we ask whether we ought to do it, whether it is a right action or a duty. But from a clear insight into the nature of these two questions, there appears to me to follow a second most important result: namely, what is the nature of the evidence, by which alone any ethical proposition can be proved or disproved, confirmed or rendered doubtful. Once we recognise the exact meaning of the two questions, I think it also becomes plain exactly what kind of reasons are relevant as arguments for or against any particular answer to them. It becomes plain that, for answers to the _first_ question, no relevant evidence whatever can be adduced: from no other truth, except themselves alone, can it be inferred that they are either true or false. We can guard against error only by taking care, that, when we try to answer a question of this kind, we have before our minds that question only, and not some other or others; but that there is great danger of such errors of confusion I have tried to shew, and also what are the chief precautions by the use of which we may guard against them. As for the _second_ question, it becomes equally plain, that any answer to it _is_ capable of proof or disproof--that, indeed, so many different considerations are relevant to its truth or falsehood, as to make the attainment of probability very difficult, and the attainment of certainty impossible. Nevertheless the _kind_ of evidence, which is both necessary and alone relevant to such proof and disproof, is capable of exact definition. Such evidence must contain propositions of two kinds and of two kinds only: it must consist, in the first place, of truths with regard to the results of the action in question--of _causal_ truths--but it must _also_ contain ethical truths of our first or self-evident class. Many truths of both kinds are necessary to the proof that any action ought to be done; and any other kind of evidence is wholly irrelevant. It follows that, if any ethical philosopher offers for propositions of the first kind any evidence whatever, or if, for propositions of the second kind, he either fails to adduce both causal and ethical truths, or adduces truths that are neither, his reasoning has not the least tendency to establish his conclusions. But not only are his conclusions