their primitive constituent atoms. Here we have at once
a tacit assumption which governs his method. Philosophy, speaking
roughly, is by some people supposed to start from truths, and thus to
be in some way an evolution of logic. According to Mill it must start
from facts, and therefore from something not given by logic. To state
clearly, indeed, the relation between truth and fact may suggest very
intricate problems. Mill, at any rate, must find a basis in fact, and
for him the ultimate facts must be feelings. The reality at least of a
feeling is undeniable. The _Penser c'est sentir_, or the doctrine that
all 'ideas' are transformed sensations is his starting-point. The word
'feeling,' according to him, includes every 'phenomenon of the mind.'
'Think,' he says elsewhere,[512] does not include all our experience,
but 'there is nothing to which we could not extend the term "I feel."'
He proceeds to infer that our experience is either a knowledge of the
feelings separately, or 'a knowledge of the order in which they follow
each other; and this is all.' We may add that the knowledge is the
feeling. Reid, Kant, and the Germans have indeed tried to show that
there are feelings not derived from the sensations, but this, as
Hartley and Condillac have shown, is a mistake. This is his first
principle in a nutshell, and must give a clue to the various
applications.

The next step is familiar. Hume had distinguished impressions and
ideas. 'Ideas' are copies of previous 'impressions.' It is for
psychology to say what are the laws by which they are related to their
originals. The ultimate origin cannot be explained by psychology
alone. Impressions are caused by the outward world acting in some way
upon the mind; and the psychologist can only classify the various
modes in which they present themselves. Mill therefore begins by the
usual account of the five senses, through which comes all knowledge of
the external world. He adds to Reid's list muscular sensations, and
those derived from the internal organs, to which last Cabanis in
particular had called attention. So far he is following the steps of
his predecessors. He is, he says, simply asserting an 'indisputable'
fact.[513] We have sensations and we have ideas, which are 'copies of
sensations.' We may then consider how far these facts will enable us
to explain the whole series of mental phenomena. 'Ideation,' which he
suggests as a new word--the process by which a continuous serie