§2. But in thus stating the problem we are assuming that the latter term of this comparison can be satisfactorily defined and sufficiently developed; that we can frame with adequate precision a system of rules, constituting the true moral code for human beings as deduced from Utilitarian principles. And this seems to have been commonly assumed by the school whose method we are now examining. But when we set ourselves in earnest to the construction of such a system, we find it beset with serious difficulties. For, passing over the uncertainties involved in hedonistic comparison generally, let us suppose that the quantum of happiness that will result from the establishment of any plan of behaviour among human beings can be ascertained with sufficient exactness for practical purposes even when the plan is as yet constructed in imagination alone. It still has to be asked, What is the nature of the human being for whom we are to construct this hypothetical scheme of conduct? For humanity is not something that exhibits the same properties always and everywhere: whether we consider the intellect of man or his feelings, or his physical condition and circumstances, we find them so different in different ages and countries, that it seems prima facie absurd to lay down a set of ideal Utilitarian rules for mankind generally. It may be said that these differences after all relate chiefly to details; and that there is in any case sufficient uniformity in the nature and circumstances of human life always and everywhere to render possible an outline scheme of ideal behaviour for mankind at large, But it must be answered, that it is with details that we are now principally concerned; for the previous discussion has sufficiently shown that the conduct approved by Common Sense has a general resemblance to that which Utilitarianism would prescribe; but we wish to ascertain more exactly how far the resemblance extends, and with what delicacy and precision the current moral rules are adapted to the actual needs and conditions of human life.
Suppose, then, that we contract the scope of investigation, and only endeavour to ascertain the rules appropriate to men as we know them, in our own age and country. We are immediately met with a dilemma: the men whom we know are beings who accept more or less definitely a certain moral code if we take them as they are in this respect, we can hardly at the same time conceive them as beings for whom a code is yet to be constructed de novo: if, on the other hand, we take an actual man---let us say, an average Englishman---and abstract his morality, what remains is an entity so purely hypothetical, that it is not clear what practical purpose can be served by constructing a system of moral rules for the community of such beings. Could we indeed assume that the scientific deduction of such a system would ensure its general acceptance; could we reasonably expect to convert all mankind at once to Utilitarian principles, or even all educated and reflective mankind,. so that all preachers and teachers should take universal happiness as the goal of their efforts as unquestioningly as physicians take the health of the individual body; and could we be sure that men's moral habits and sentiments would adjust themselves at once and without any waste of force to these changed rules, then perhaps in framing the Utilitarian code we might fairly leave existing morality out of account. But I cannot think that we are warranted in making these suppositions; I think we have to take the moral habits, impulses, and tastes of men as a material given us to work upon no less than the rest of their nature, and as something which, as it only partly results from reasoning in the past, so can only be partially modified by any reasoning which we can now apply to it. It seems therefore clear that the solution of the hypothetical Utilitarian problem of constructing an ideal morality for men conceived to be in other respects as experience shows them to be, but with their actual morality abstracted, will not give us the result which we practically require.
It will perhaps be said, ``No doubt such an ideal Utilitarian morality can only be gradually, and perhaps after all imperfectly, introduced; but still it will be useful to work it out as a pattern to which we may approximate.'' But, in the first place, it may not be really possible to approximate to it: since any particular existing moral rule, though not the ideally best even for such beings as existing men under the existing circumstances, may yet be the best that they can be got to obey: so that it would be futile to propose any other, or even harmful, as it might tend to impair old moral habits without effectively replacing them by new ones. And secondly, the endeavour gradually to approximate to a morality constructed on the supposition that the non-moral part of existing human nature remains unchanged, may lead us wrong: because the state of men's knowledge and intellectual faculties, and the range of their sympathies, and the direction and strength of their prevailing impulses, and their relations to the external world and to each other, are continually being altered, and such alteration is to some extent under our control and may be felicific in a high degree: and any material modifications in important elements and conditions of human life may require corresponding changes in established moral rules and sentiments, in order that the greatest possible happiness may be attained by the human being whose life is thus modified. In short, the construction of a Utilitarian code, regarded as an ideal towards which we are to progress, is met by a second dilemma:---The nature of man and the conditions of his life cannot usefully be assumed to be constant, unless we are confining our attention to the present or proximate future; while again, if we are considering them in the present or proximate future, we must take into account men's actual moral habits and sentiments, as a part of their nature not materially more modifiable than the rest.
Nor, again, can I agree with Mr. Spencer in thinking that it is possible to solve the problems of practical ethics by constructing the final perfect form of society, towards which the process of human history is tending; and determining the rules of mutual behaviour which ought to be, and will be, observed by the members of this perfect society. For, firstly, granting that we can conceive as possible a human community which is from a utilitarian point of view perfect; and granting also Mr. Spencer's definition of this perfection---viz. that the voluntary actions of all the members cause ``pleasure unalloyed by pain anywhere'' to all who are affected by them---;  it still seems to me quite impossible to forecast the natures and relations of the persons composing such a community, with sufficient clearness and certainty to enable us to define even in outline their moral code. And secondly, even if it were otherwise, even if we could construct scientifically Mr. Spencer's ideal morality, I do not think such a construction would be of much avail in solving the practical problems of actual humanity. For a society in which---to take one point only---there is no such thing as punishment, is necessarily a society with its essential structure so unlike our own, that it would be idle to attempt any close imitation of its rules of behaviour. It might possibly be best for us to conform approximately to some of these rules; but this we could only know by examining each particular rule in detail; we could have no general grounds for concluding that it would be best for us to conform to them as far as possible. For even supposing that this ideal society is ultimately to be realised, it must at any rate be separated from us by a considerable interval of evolution; hence it is not unlikely that the best way of progressing towards it will be some other than the apparently directest way, and that we shall reach it more easily if we begin by moving away from it. Whether this is so or not, and to what extent, can only be known by carefully examining the effects of conduct on actual human beings, and inferring its probable effects on the human beings whom we may expect to exist in the proximate future.