EVOLUTION OF FEELING AND EMOTION 293 
generally presuppose the existence of certain specific ten- 
dencies. ‘The anger produced in a dog by taking away its 
bone presupposes the specific appetite for food. The anger 
produced in it by interfering with its young presupposes the 
specific tendency to guard and tend its offspring. So the 
presence of a rival who interferes with its wooing causes 
anger because of the pre-existence of the sexual impulse.” In 
general, we may say that emotional states are, under natural 
conditions, closely associated with behaviour of biological 
value—with tendencies which are beneficial in self-preserva- 
tion or race-preservation—with actions that promote survival, 
and especially with the behaviour which clusters round the 
pairing and parental instincts. The value of the emotions in 
animals is that they are an indirect means of furthering sur- 
vival. But how has the close association between emotional 
condition and the biological end it furthers been established? 
Again, we must say that under natural conditions it is not the 
sort of thing which could be acquired. And again we must 
urge that natural selection through survival is, apart from 
some theory of pre-established harmony, the only hypothesis 
in the field on which the close association can be explained. 
There is one more point to which attention may be drawn. 
If there be one thing, and there certainly are not many, on 
which all writers on the emotions are agreed, it is as to their 
vagueness. They do not readily submit to definition, and 
cannot be described in a sentence. ‘This is not due to any 
indefiniteness of biological end, nor to much indefiniteness in 
the mode of “expression ;” it is due, rather, to an inherent 
dimness and haziness of psychological outline. We seem 
unable to focus them and get a clear-cut result. This is, no 
doubt, in part due to the complexity of emotional states. But, 
may it not be largely due to the fact that there is no necessity 
for definiteness? They fulfil their purpose just as well if they 
are vague. It is quite necessary for the dog to have a clear- 
cut impression of his antagonist ; and, on the cognitive side 
of consciousness, meaning must be in some degree definite 
to be of real value. But, so long as the emotion raises the 
