_The constantification of conditions_ now consists in standardizing the present instance, thus: Whenever similar circumstances are given, i.e.: the same instruments of evidence are present, the evidence of guilt is established. Now the accompanying changes with regard to the effect, i.e.: proof of guilt through evidence, have to be tested--therefore the individual conditions--i.e.: the individual sources of evidence have to be established and their values to be determined and _varied_. Finally, the accompanying change in effect (conviction by evidence) is to be tested. The last procedure requires discussion; the rest is self evident. In our business isolation is comparatively easy, inasmuch as any individual statement, any visual impression, any effect, etc., may be abstracted without difficulty. Much harder is the determination of its value. If, however, we clearly recognize that it is necessary to express the exact value of each particular source of evidence, and that the task is only to determine comparative valuation, the possibility of such a thing, in at least a sufficiently close degree of certainty, must be granted. The valuation must be made in respect of two things--(1) its _reliability_ (subjective and relative); (2) its _significance_ (objective and absolute). On the one hand, the value of the evidence itself must be tested according to the appraisement of the person who presents it and of the conditions under which he is important; on the other, what influence evidence accepted as reliable can exercise upon the _effect_, considered in and for itself. So then, when a testimony is being considered, it must first be determined whether the witness was able and willing to speak the truth, and further, what the importance of the testimony may be in terms of the changes it may cause in the _organization_ of the case.
Of greatest importance and most difficult is the variation of conditions and the establishment of the changes thereby generated, with regard to the _effect_,--i.e.: the critical interpretation of the material in hand. Applied to a case, the problem presents itself in this wise: I consider each detail of evidence by itself and cleared of all others, and I vary it as often as it is objectively possible to do so. Thus I suppose that each statement of the witness might be a lie, entirely or in part; it might be incorrect observation, false inference, etc.--and then I ask myself: Does the evidence of guilt, the establishment of an especial trial, now remain just? If not, is it just under other and related possible circumstances? Am I in possession of these circumstances? If now the degree of apparent truth is so far tested that these variations may enter and the accusation still remain just, the defendant is convicted: but only under these circumstances.