great weight, but it also insists that they shall not be garbled. "In the proof of confessions, as in the case of admissions in civil cases, the whole of what the person said on the subject at the time of making the confession should be taken together. It is not reasonable to assume that the entire proposition, with all its limitations, was contained in one sentence. . . . Unless the whole is received and considered. the true meaning and import of the part which is good evidence against him cannot be ascertained."*

On the same principle it is a rule of evidence that if a witness tells about a part of any conversation the lawyer who cross-examines him has a right to ask about any other part of the same conversation.

The difference between the fallacy of Accent and the fallacy of Accident in the broader sense of each is this: the former misinterprets a writer by confusing incidental statements with essential; the latter confuses aspects of things or situations (or statements about such aspects) with the things or situations (or statements about them) as a whole. †

* Greenleaf, op. cit., Sec. 218.

Whether we should call the over-emphasizing of some one aspect of the moral code accent or accident would thus depend upon whether we regarded the law as a revelation each part of which should be interpreted with reference to the whole, or as an analysis of conduct into various good and bad aspects, several of which may be combined in the complex whole. An aspect of a law taken for the whole law is accent; an aspect of an act taken for the whole act is accident.

CHAPTER IV.

DIVISION AND CLASSIFICATION.

To avoid confusion in the use of names we must define them; but all definition of names involves a classification of objects. If the words 'animal', 'red', 'verte- Relation to

brate

have any definite meaning at all there definition. must be some things to which they can be properly applied and some things to which they cannot, and the things to which any one of them can be applied must all have the qualities or relations which the name implies, and therefore resemble each other in this respect, while the things to which it cannot be properly applied must all resemble each other in not having these qualities or relations. Hence every

time we use a name we imply the existence of two classes of things: those that have the quality or relation which the name implies and those that have not. To define a name is to distinguish between these two classes, and the more clearly we understand this difference between the things the more clearly we can define the word. Hence we shall stop speaking about words for a little and speak about the principles of Division or Classification. At the end of the chapter we shall return to the discussion of words and their interpretation.

Principles of division

When we have made two classes to one or other of which every object in the world can be assigned according as it has or has not some given quality or relation, each of these classes can be subdivided subdivision. with reference to some other quality or relation; and this

and

process can be carried on indefinitely, as in the following

This division and subdivision of everything into precisely two classes according to the presence or absence of some given mark (technically called Division by Dichotomy) is often useful, and in the example just given seems very appropriate, but there are cases in which it is cumbersome and rather absurd, e.g.:

In this case it would be better for almost every practical purpose to divide directly into Red, Orange, Yellow, Green, and the other colors. Hence instead of dividing according

to the presence or absence of some given quality (e.g., redness), we may divide according to the determination in some given respect (e.g., color), making as many co-ordinate classes as there are different determinations in that respect; and then we can proceed to subdivision, with reference to other respects, if such subdivision is necessary, immediately and without so much confusion.

Whether each class should or should not be subdivided in the same respect as every other (e.g., each color into light and dark), depends altogether upon the purpose of the division. Often it is impossible. Creatures with a nervous system, for example, could be divided according to its arrangement or development. Those without one could

not.

The one great rule for division is that each of the objects divided shall have one place in the system, and only one. Thus, if we should divide all human beings into Americans, Europeans, and uncivilized peoples, we should commit a double blunder, for some peoples, such as the Japanese, would not fall under any head, and some of the American Indians would fall under two.

When any object falls into each of two co-ordinate classes— as in this example-there is said to be a Cross-division. This is always likely to occur when we classify objects in more than one respect at a time. If we had first Crossdivided peoples according to their geographical division. distribution and afterwards subdivided each group according to their civilization, or vice versa, this could not have occurred, e.g.,

To be sure, the uncivilized Indians still fall into two classes, Americans and uncivilized Americans, but this is now perfectly proper, because the classes are no longer co-ordinate; the latter is subordinate to the former.

Cross-division is a particularly bad logical blunder, because of the mental confusion which is shown by the indiscriminate use of several principles of classification (technically called Fundamenta Divisionis) at once.

"Overlapping, however, may be unavoidable in practice, owing to the nature of the objects. There may be objects in which the dividing characters are not distinctly marked, objects that possess the differentiæ of more than one group in a greater or less degree. Things are not always marked off from one another by hard and fast lines. They shade into each other by imperceptible gradations. A clear separation of them may be impossible. In that case you may allow a certain indeterminate margin between your classes, and sometimes it may be necessary to put an object into more than one class." *

* Minto's "Logic," p. 95.