raise an artist's temperature above boiling-point. A painter may go into
the woods, get his thrill, go home and fetch his panel-box, and proceed
to set down in cold blood what he finds before him. No good can come of
it, as the gloomy walls of any official exhibition will show. Realistic
novels fail for the same reason: with all their gifts, neither Zola, nor
Edmond de Goncourt, nor Mr. Arnold Bennett ever produced a work of art.
Also, a thorough anarchist will never be an artist, though many artists
have believed that they were thorough anarchists. One man cannot pour an
æsthetic experience straight into another, leaving out the problem. He
cannot exude form: he must set himself to create a particular form.
Automatic writing will never be poetry, nor automatic scrabbling design.
The artist must submit his creative impulse to the conditions of a
problem. Often great artists set their own problems; always they are
bound by them. That would be a shallow critic who supposed that Mallarmé
wrote down what words he chose in what order he pleased, unbound by any
sense of a definite form to be created and a most definite conception to
be realized. Mallarmé was as severely bound by his problem as was
Racine by his. It was as definite--for all that it was unformulated--as
absolute, and as necessary. The same may be said of Picasso in his most
abstract works: but not of all his followers, nor of all Mallarmé's
either.

Was he really a great painter? A new generation is beginning to ask the
question that we answered, once and for all as we thought, ten years
ago. Yes, of course, the _douanier_ was--a remarkable painter. The man
who influenced Derain, and to some extent Picasso, is not likely to have
been less. But a great painter? For the present, at any rate, let us
avoid great words.

In 1903, when first I lived in Paris, Rousseau appeared to be very much