CHAP. V. SECT. IX.

Inference does not regard concepts but objects.

there is an à priori principle latent in the mind which serves for the major premise of the syllogism. Against this theory we have already adduced objections.* And the overthrow of this theory renders it impossible to put into syllogistic form the inferences referred to. There are, then, particular kinds of inference which do not proceed from general judgments, but from particular objects of knowledge.

§ 194. With reference to the second point, it is maintained by some that inference has to do primarily with concepts and the relation of concepts. But, as individual objects undoubtedly form subjects of reasoning, concepts are taken to be not merely the representation of the qualities of a class, but also of an individual. To this doctrine, that reasoning refers exclusively, or even principally, to mental representations, we must take exception. Every mental representation which we make, either of an individual object or of a class, has significance to us only as related to the real objects of which it is a concept. When we make use of the term man, we mean, not our mental representation of human attributes, but the real external objects possessing human qualities which we denote by that name. And when we say man is mortal, we mean that the real beings whom we call men will all die. Thus, if we are to depend upon what consciousness reveals to us, inference has reference primarily, not to concepts, but to the real objects or classes of which the * § 184.

concepts are a more or less complete representation. And the predication of the inference asserts or denies a certain relation to exist between these objects.

CHAP. V.

SECT. IX.

in in

§ 195. What is this relation? Is it only that of Relations the container and the contained, the mutual inclusion predicated or exclusion, in whole or in part, of two objects of ference. thought? A reference to particular cases of inference will soon decide the question. From a comparison of certain angles and triangles we infer

"The angles at the base of an isosceles triangle are equal to one another." This is certainly not a relation of inclusion, either total or partial. From a variety of considerations, we infer that the rapid cooling of the atmosphere is the cause of the deposition of dew. It would be difficult to put this inference into the stereotyped form. From many experiences we infer, when we see smoke, that combustion is going on somewhere near at hand. The relation expressed in this inference is certainly not one of inclusion. In short, since inferences consist of predications, and since predications assert any of the relations which may be perceived in intuition as existing between two objects of thought, it is manifest that reasoning must occupy itself with all these relations. This the syllogism, in its present form, does not do; and hence, in its present form, we maintain that it is the type of only an accidental, although important, variety of the reasoning process. § 196. Having decided against the claims of the syllogism to be the universal form of reasoning, we

as to form of reasoning post

Question

poned.

CHAP. V.
SECT. X.

might inquire if there is any other form which can legitimately make similar claims. But we shall be better able to consider this question after we have studied more fully different reasoning processes with which we are familiar. We have already divided the sphere of inference into the subordinate sphere concerned about (1) space and time, with their various divisions and combinations, (2) the nonphenomenal causes of things sensible, and (3) the existence and relations of phenomena. But in these different spheres the grounds upon which the inferences rest may vary greatly in their nature and extent, and the inferences themselves may have every variety of probability attached to them. Upon the principle of advancing from the more concrete and particular to the more abstract and general varieties of inference, the following appears to be a suitable division: (1) Inferences regarding the occurrence of particular events, depending upon evidence or testimony; (2) Inferences founded upon analogy and induction, either particular or expressed in a less or more certain general law; (3) Deductive inferences from general principles.

Inference depending upon evidence.

SECTION X.

EVIDENCE; INDUCTION; DEDUCTION.

§ 197. It is frequently necessary, in the common affairs of life, to form opinions regarding the occurrence or non-occurrence of events, or the nature of

alleged facts which have not come under our own observation. These opinions are of the nature of inference depending upon what is called Evidence. We may divide the facts or events regarding which evidence is taken into two general classes: (1) contemporary facts or events, alleged to exist or occur in our own time, within the memory of persons now living; and (2) historical facts or events alleged to exist or occur in some past generation. The evidence on which inferences regarding the first of these classes of facts are based is most frequently human testimony, although, in the absence of that, or when it is not thought completely trustworthy, a great variety of circumstantial evidence may be resorted to. Inferences regarding historical facts or events depend upon tradition, historical records, the existence of commemorative monuments and other things. We shall refer to certain principles by which we are guided in drawing inferences in each of these cases.

CHAP. V.

SECT. X.

contempo

rary facts."

§ 198. The most direct evidence of contemporary Regarding occurrences which we have not ourselves observed is the testimony of others. But in receiving the testimony of others there are various things which must be considered before placing implicit reliance upon it. We have found by experience that some persons Testimony. are not truthful in relating what they have seen or heard; that some persons are incapable of accurate observation, mixing up their own fancies or prejudices with the facts which they are observing; that some persons are very liable to be governed by

CHAP. V.

SECT. X.

Circumstantial evidence.

Value of evidence.

interested motives so as deliberately to misrepresent in their own favour the facts which they relate. These and other considerations have led to the establishment of various legal customs connected with the taking of evidence, such as the imposition of an oath, cross-questioning, and others. For the purpose of correcting human testimony it is found, also, that a variety of circumstantial evidence is of very great importance. The majority of the alleged occurrences regarding which inferences are to be drawn leave some permanent marks or traces behind them which may be afterwards observed; and these, in the hands of a skilful interpreter, frequently form the basis of important and true conclusions. There is, however, great difference in the skill of different persons in discovering the significance of permanent marks or signs. The barbarous tribes of an uncultivated country usually show remarkable acuteness in reading the meaning of the traces of past events; while the natural or acquired skill of the professional detective is often equally wonderful.*

§ 199. In determining the value of evidence, there are, generally speaking, two things to be considered; the nature of the evidence itself, and especially the character of the persons by whom it is given, and the nature of the fact testified to. The mind frequently refuses assent to alleged facts in consequence of the improbability of these facts occurring. When

* For a specimen of this acuteness and skill the reader is referred to Edgar A. Poe's remarkable description 'The Murder in the Rue Morgue.'