III
One note which we shall find persistently struck and re-struck in Stevenson’s essays is the memory of childhood. From Child’s Play to The Lantern-Bearers we are confronted by a mass of material regarding one childhood, by which is supported a series of generalisations about all children and their early years. So we proceed to youth, to the story of A College Magazine; and so to Ordered South. Then we return again to An Old Scotch Gardener and The Manse, where again that single childhood, so well-stored with memories, provides the picture. Now it is one thing for Stevenson to re-vivify his own childhood, for that is a very legitimate satisfaction which nobody would deny him; but it is another thing for Stevenson, from that single experience and with no other apparent observation or inquiry, to generalise about all children. While he tells us what he did, in what books and adventures and happenings he found his delight, we may read with amusement. When, upon the other hand, he says, “children are thus or thus,” it is open to any candid reader to disagree with Stevenson. Whether it is that he has set the example, or whether it is that he merely exemplifies the practice, I cannot say; but Stevenson is one of those very numerous people who talk wisely and shrewdly about children in the bulk without seeming to know anything about them. These wiseacres alternately under-rate and make too ingenious the intelligence and the calculations of childhood, so that children in their hands seem to become either sentimental barbarians or callous schemers, but are never, in the main, children at all. Stevenson has a few excellent words upon children: he admirably says, “It is the grown people who make the nursery stories; all the children do, is jealously to preserve the text”: but I am sorry to say that, upon the whole, I can find little else that is of value in his general observations.
It is open to anybody to reconstruct a single real childhood from Stevenson’s essays, and no doubt that is a matter of considerable interest, as anything which enables us to understand a man is of value. Curiously enough, however, Stevenson’s essays upon the habits and notions of children seem to suggest a great deal too much thought about play, and too little actual play. They seem to show him, as a little boy, so precocious and lacking in heart, that he is watching himself play rather than playing. It is not the preliminary planning of play that delights children, not the academic invention of games and deceits; it is the immediate and enjoyable act of play. Our author shows us a rather elderly child who, in deceiving himself, has savoured not so much the game as the supreme cleverness of his own self-deception. That, to any person who truly remembers the state of childhood, may be accepted as a perfectly legitimate recollection; and it is so far coherent. That his own habit should be, in these essays, extended to all other children whatsoever—in fact, to “children”—is to make all children delicate little Scots boys, greatly loved, very self-conscious, and, in the long run, rather tiresome, as lonely, delicate little boys incline to become towards the end of the day. Unfortunately the readers of Stevenson’s essays about little boys have mostly been little girls; and they are not themselves children, but grown-up people who are looking back at their own childhood through the falsifying medium of culture and indulgent, dishonest memory. Culture, in dwelling upon interpretations and upon purposes, and in seeing childhood always through the refraction of consequence, destroys interest in play itself; and if play is once called in question it very quickly becomes tedious rigmarole.
Stevenson’s essays must thus be divided into two parts, the first descriptive, the second generalised. The first division, sometimes delightful, is also sometimes sophisticated, and sometimes is exaggerative of the originality of certain examples of play. The second is about as questionable as any writing on children has ever been, because it is based too strictly upon expanded recollections of a single abnormal model. You do not, by such means, obtain good generalisations.