dramatist should aim at _being_ logical without _seeming_ so.[9]

It is sometimes said that a playwright ought to construct his play
backwards, and even to write his last act first.[10] This doctrine
belongs to the period of the well-made play, when climax was regarded as
the one thing needful in dramatic art, and anticlimax as the
unforgivable sin. Nowadays, we do not insist that every play should end
with a tableau, or with an emphatic _mot de la fin_. We are more willing
to accept a quiet, even an indecisive, ending.[11] Nevertheless it is
and must ever be true that, at a very early period in the scheming of
his play, the playwright ought to assure himself that his theme is
capable of a satisfactory ending. Of course this phrase does not imply a
"happy ending," but one which satisfies the author as being artistic,
effective, inevitable (in the case of a serious play), or, in one word,
"right." An obviously makeshift ending can never be desirable, either
from the ideal or from the practical point of view. Many excellent plays
have been wrecked on this rock. The very frequent complaint that "the
last act is weak" is not always or necessarily a just reproach; but it
is so when the author has clearly been at a loss for an ending, and has
simply huddled his play up in a conventional and perfunctory fashion. It
may even be said that some apparently promising themes are deceptive in
their promise, since they are inherently incapable of a satisfactory
ending. The playwright should by all means make sure that he has not run
up against one of these blind-alley themes.[12] He should, at an early
point, see clearly the end for which he is making, and be sure that it
is an end which he actively desires, not merely one which satisfies
convention, or which "will have to do."

Some dramatists, when a play is provisionally mapped out, do not attempt
to begin at the beginning and write it as a coherent whole, but make a