Most authors, however, who have any real gift for character-creation probably fall more or less under this illusion, though they are sane enough and modest enough to realize that an illusion it is.[4] A character will every now and then seem to take the bit between his teeth and say and do things for which his creator feels himself hardly responsible.  The playwright’s scheme should not, then, until the latest possible moment, become so hard and fast as to allow his characters no elbow room for such manifestations of spontaneity.  And this is only one of several forms of afterthought which may arise as the play develops.  The playwright may all of a sudden see that a certain character is superfluous, or that a new character is needed, or that a new relationship between two characters would simplify matters, or that a scene that he has placed in the first act ought to be in the second, or that he can dispense with it altogether, or that it reveals too much to the audience and must be wholly recast.[5]

These are only a few of the re-adjustments which have constantly to be made if a play is shaping itself by a process of vital growth; and that is why the playwright may be advised to keep his material fluid as long as he can.  Ibsen had written large portions of the play now known to us as Rosmersholm before he decided that Rebecca should not be married to Rosmer.  He also, at a comparatively late stage, did away with two daughters whom he had at first given to Rosmer, and decided to make her childlessness the main cause of Beata’s tragedy.

Perhaps I insist too strongly on the advisability of treating a dramatic theme as clay to be modelled and remodelled, rather than as wood or marble to be carved unalterably and once for all.  If so, it is because of a personal reminiscence.  In my early youth, I had, like everybody else, ambitions in the direction of play-writing; and it was my inability to keep a theme plastic that convinced me of my lack of talent.  It pleased me greatly to draw out a detailed scenario, working up duly to a situation at the end of each act; and, once made, that scenario was like a cast-iron mould into which the dialogue had simply to be poured.  The result was that the play had all the merits of a logical, well-ordered essay.  My situations worked out like the Q.E.D.’s of Euclid.  My characters obstinately refused to come to life, or to take the bit between their teeth.  They were simply cog-wheels in a pre-arranged mechanism.  In one respect, my two or three plays were models—­in respect of brevity and conciseness.  I was never troubled by the necessity of cutting down—­so cruel a necessity to many playwrights.[6] My difficulty was rather to find enough for my characters to say—­for they never wanted to say anything that was not strictly germane to the plot.  It was this that made me despair of play-writing, and realize that my mission was to teach other people how to write plays.  And, similarly, the aspirant who finds that his people