But this is obviously the case with Mill's theory.He makes way for a good being by an arbitrary division of nature into two sets of forces.He saves the benevolence by limiting the power of the deity;but then the limits are,by his own admission,utterly unknowable.A power,restrained by unknowable bounds,is a power from which nothing can be inferred.Whatever its attributes,we do not know whether they will affect any state of things.The goodness may be indefinitely frustrated.In fact,on Mill's showing,a power omnipotent but not benevolent,or an indefinite multitude of powers of varying attributes,or a good and a bad power eternally struggling,or,in short,any religious doctrine that has ever been held among men,would suit the facts.Mill's 'plurality of causes'might have suggested this difficulty.I see a corpse.The death may have been due to any one of an indefinite number of causes.What right have I to select one?I am in the same position when I regard the whole of nature as what Hume called a 'unique effect.'The four methods of induction become inapplicable,for there are no other universes and I have no compass to steer by in the region of the unverifiable.
What,then,can be the advantage of any belief where conflicting hypotheses must be all equally probable?The question is partly discussed in the second essay upon the utility of Religion.Here Mill takes up the old argument of 'Philip Beauchamp,'the 'only direct discussion'of the point with which he is acquainted,(148)and endeavours to state the case more fairly and in a less hostile spirit.His argument,however,is in general conformity with Bentham and Grote,and is very forcibly put.One point may be noticed.He virtually identifies 'religion'with a belief in 'the supernatural.'(149)He compares the efficacy of such beliefs with the efficacy of education (which,as he characteristically says,is 'almost boundless')(150)and of public opinion,and shows with 'Beauchamp'that when conflict occurs,these influences are stronger than those derived from supernatural sanctions.Now when we believe in a revelation it is intelligible to ask,What is the influence of a creed?It represents a new force influencing men's minds from without.But when the creed is supposed to be generated from antecedent beliefs,the argument must be altered by considering what are the true causes of the belief.How did it come to prevail?An admirer of Comte might have brought out more distinctly the fact that such beliefs mark an essential stage of progress,that what are now sporadic superstitions were once parts of a systematic religion and represented the germs of science.They were approximate hypotheses which had to be remodelled by extricating or dropping the 'supernatural'element.A full recognition of this would diminish the paradoxical appearance of the statement from which he starts,that 'a religion may be morally useful without being intellectually sustainable.'The truth surely is that we cannot separate the two elements of a creed.Doubtless there were no such beings as the Zeus or Apollo of popular belief;but polytheism may still have provided the only form in which certain truths could be presented;and was,as Comte would have said,a stage in the process from fetichi** towards monotheism and positivism.A discussion of the utility of belief in the 'supernatural'without reference to the place of the supernatural in the whole system of belief must be necessarily inadequate.Mill admits this in substance,and argues that the moral truth may survive the superstitions in which it was bound up.(151)He goes on to argue,as Comte had argued,that the instincts which once found their sanction in the supernatural world might find their embodiment in the 'Religion of Humanity.'(152)This he holds to be not only entitled to the name of religion,but to be 'a better religion than any of those ordinarily called by that title.'It is disinterested and does not tend to cramp the intellect or degenerate into a worship of mere power.Mill says emphatically that the Bentham mode of considering religion as a supplement to police by providing 'sanctions'is inadequate;and that religion,like poetry,is valuable as suggesting higher ideals and gratifying the craving for knowledge of corresponding realities.To the selfish,supernatural religion offers heaven;and to the 'tender and grateful'it offers the love of God.He points out that it does not follow that we must 'travel beyond the boundaries of the world we inhabit'in order to obtain such consolation.(153)And the essay concludes by saying that,though the 'supernatural religions'have always the advantage of offering immortality,the value set upon immortality may diminish as life becomes higher and happier and annihilation may seem more desirable.(154)Yet in the middle of this argument we have the defence of Manichaeism as a possible creed,(155)and in the last essay we seem to reach the true account of his leanings to such a belief.
He still,that is,requires a breathing-space for the imagination.'Truth is the province of reason,'but 'in the regulation of the imagination literal truth of facts is not the only thing to be considered.'(156)Reason must keep the fortress,but the 'imagination may safely follow its own end and do its best to make life pleasant and lovely inside the castle.'
