§3. There is, however, one method of exhibiting a priori the absolute duty of Truth, which we must not overlook; as, if it be valid, it would seem that the exceptions and qualifications above mentioned have been only admitted by Common Sense from inadvertence and shallowness of thought.
It is said that if it were once generally understood that lies were justifiable under certain circumstances, it would immediately become quite useless to tell the lies, because no one would believe them; and that the moralist cannot lay down a rule which, if generally accepted, would be suicidal. To this there seem to be three answers. In the first place it is not necessarily an evil that men's confidence in each other's assertions should, under certain peculiar circumstances, be impaired or destroyed: it may even be the very result which we should most desire to produce: e.g. it is obviously a most effective protection for legitimate secrets that it should be universally understood and expected that those who ask questions which they have no right to ask will have lies told them: nor, again, should we be restrained from pronouncing it lawful to meet deceit with deceit, merely by the fear of impairing the security which rogues now derive from the veracity of honest men. No doubt the ultimate result of general unveracity under the circumstances would be a state of things in which such falsehoods would no longer be told: but unless this ultimate result is undesirable, the prospect of it does not constitute a reason why the falsehoods should not be told so long as they are useful. But, secondly, since the beliefs of men in general are not formed purely on rational grounds, experience shows that unveracity may long remain partially effective under circumstances where it is generally understood to be legitimate. We see this in the case of the law-courts. For though jurymen are perfectly aware that it is considered the duty of an advocate to state as plausibly as possible whatever he has been instructed to say on behalf of any criminal lie may defend, still a skilful pleader may often produce an impression that he sincerely believes his client to be innocent: and it remains, a question of casuistry how far this kind of hypocrisy is justifiable. But, finally, it cannot be assumed as certain that it is never right to act upon a maxim of which the universal application would be an undoubted evil. This assumption may seem to be involved in what was previously admitted as an ethical axiom, that what is right for me must be right for `all persons under similar conditions'.[1] But reflection will show that there is a special case within the range of the axiom in which its application is necessarily self-limiting, and excludes the practical universality which the axiom appears to suggest: i.e. where the agent's conditions include (1) the knowledge that his maxim is not universally accepted, and (2) a reasoned conviction that his act will not tend to make it so, to any important extent. For in this case the axiom will practically only mean that it will be right for all persons to do as the agent does, if they are sincerely convinced that the act will not be widely imitated; and this conviction must vanish if it is widely imitated. It can hardly be said that these conditions are impossible: and if they are possible, the axiom that we are discussing can only serve, in its present application, to direct our attention to an important danger of unveracity, which constitutes a strong---but not formally conclusive---utilitarian ground for speaking the truth. {Note}