According to David Hume, whose hypothesis is a kind of supplement to the former two, that which "when the sorrow is not softened by fiction, raises a pleasure from the bosom of uneasiness, a pleasure which still retains all the features and outward symptoms of distress and sorrow, is that very eloquence with which the melancholy scene is represented." In reply, Dr. Campbell has shown that the aggravating of all the circumstances of misery in the representation, cannot make it be contemplated with pleasure, but must be the most effectual method for making it give greater pain; that the detection of the speaker's talents and address, which Hume's hypothesis implies, is in direct opposition to the fundamental maxim, that "it is essential to the art to conceal the art;" and that the supposition that there are two distinct effects produced by the eloquence on the hearers, one the sentiment of beauty, or of the harmony of oratorical numbers, the other the passion which the speaker purposes to raise in their minds, and that when the first predominates, the mixture of the two effects becomes exceedingly pleasant, and the reverse when the second is superior, is altogether imaginary. According to Hawkesworth,[21] the compassion in question may be "resolved into that power of imagination, by which we apply the misfortunes of others to ourselves;" and we are said "to pity no longer than we fancy ourselves to suffer, and to be pleased only by reflecting that our sufferings are not real; thus indulging a dream of distress, from which we can awake whenever we please, to exult in our security, and enjoy the comparison of the fiction with the truth." This hypothesis is evidently too gross to need reply. Dr. Campbell has answered the preceding hypotheses at great length, and quite satisfactorily. I regret to say that his own is as worthless, as well as remarkably confused and unintelligible. To Burke, who wrote at a later period, it falls to my lot to reply at greater length.