r Streatfeild's word for it that 'he professed himself dissatisfied
with it as a whole, and always intended to rewrite, or at any rate, to
revise it.' We could have deduced as much from his refusal to publish
the book. The certainty of commercial failure never deterred Butler from
publication; he was in the happy situation of being able to publish at
his own expense a book of whose merit he was himself satisfied. His only
reason for abandoning _The Way of all Flesh_ was his own dissatisfaction
with it. His instruction that it should be published in its present form
after his death proves nothing against his own estimate. Butler knew, at
least as well as we, that the good things in his book were legion. He
did not wish the world or his own reputation to lose the benefit of
them.

But there are differences between a novel which contains innumerable
good things and a great novel. The most important is that a great novel
does not contain innumerable good things. You may not pick out the
plums, because the pudding falls to pieces if you do. In _The Way of all
Flesh_, however, a _compere_ is always present whose business it is to
say good things. His perpetual flow of asides is pleasant because the
asides are piquant and, in their way, to the point. Butler's mind, being
a good mind, had a predilection for the object, and his detestation of
the rotunder platitudes of a Greek chorus, if nothing else, had taught
him that a corner-man should have something to say on the subject in
hand. His arguments are designed to assist his narrative; moreover, they
are sympathetic to the modern mind. An enlightened hedonism is about all
that is left to us, and Butler's hatred of humbug is, though a little
more placid, like our own. We share his ethical likes and dislikes. As
an audience we are ready to laugh at his asides, and, on the first night
at least, to laugh at them even when they interrupt the play.

But our liking for the theses cannot alter the fact that _The Way of all
Flesh_ is a _roman a theses_. Not that there is anything wrong with the
_roman a theses_, if the theses emerge from the narrative without its
having to be obviously doctored. Nor does it matter very much that a
_compere_ should be present all the while, provided that he does not
take upon himself to replace the demonstration the narrative must
afford, by arguments outside it. But what happens in _The Way of all
Flesh_? We may leave aside the minor thesis of heredity,