Page:Mind (New Series) Volume 12.djvu/225

 HEDONISM AMONG IDEALISTS. 211 very rough feelings of magnitude into something more like estimating the degree in which, say, a number of architects' designs meet the requirements in view of which they have been framed. " The degree," it may be replied ; " then your comparison is quantitative after all." This example I think extremely significant. Suppose there is a competition of designs, and you give marks for the degrees in which re- quirements are fulfilled ; or, indeed, we may take the case of any examination in which marks are given. This is a rough way of symbolising the relation of performances to requirements ; but it it is not the result of a calculation, or true handling of quantities, except in so far as requirements are subdivided, separate marks assigned for conformity to each, and subsequently added together. But we know that the more this is done, the less reliable the result becomes ; and a highly skilled assessor or examiner, if compelled to use marks instead of reporting in detail, is inclined, I suspect, to make sure of his totals first, and subdivide them afterwards, i.e., to " cook " his marks for details. And the reason is that in each case you are translating the fulfilment of concrete conditions into the bare form of quantity, and the more the arithmetical element enters in the more is the bareness of this form perceptible. If I prefer this design very greatly to that, I may simply give the one 200 marks and the other 100 ; but it would have made no serious difference if I had said, instead, 180 and 100 respectively. I convey, roughly, the fact that I think the one a good deal better than the other. But if I take 200 as full marks and try to divide the require- ments to be satisfied into ten heads with twenty marks each, and assign marks on this hypothesis, and sum them into totals, I shall probably find my total fail to express, even roughly, my true preferences, unless I have as above sug- gested adjusted the subtotals to the total required. And the reason is that the process is not a result based throughout on the handling of quantities. The relation of each character in the design to a requirement, and of each requirement to the whole, is concrete and individual, and needs to be repre- sented in the intelligent language of a detailed report ; these relations are not quantities ; and the reduction or rough translation of the mere fact of preference into quantity, as a memoria technica for comparison ad hoc, has an accidental ele- ment. In a single preference this matters little, because re-translation is easy ; but in the arithmetical handling of a number of preferences it tends to monstrous errors. Or a simpler case may put the point clearly enough. Let the question be which of two pocket knives, or guns, or micro-