there are at least four different ways of writing a book on dreams. there is, for instance, the literary method. in this way one goes to books or to the memories of other people for one's material, and so collects a great number of more or less wonderful stories. i have rejected this method, for it is entirely untrustworthy. dreams are elusive at the best; only a very careful observer can set down a dream faithfully, even directly after it has occurred, and no one can safely entrust a dream to memory.

there is, again, what i may call the clinical method of studying dreams by the personal observation and collection of facts, with summation and analysis of the results. on a large scale, with the aid of the questionnaire, this method has been especially carried on in the united states, notably at clark university under the inspiration of dr. stanley hall. a strict and scientific adherence to the clinical method of studying dreams has resulted in professor sante de sanctis's book i sogni (first edition 1899), which is, on the whole, the best book on dreams published in recent years.

then there is the experimental method, which, not content with mere objective study of the phenomena, endeavours to interfere with them and to find out the[vi] results of interference. this method may be combined with other methods of studying dreams. in its pure form it has in recent years been especially practised by the late mourly vold. its results are not without interest, but they do not cover a large part of the field, and they are not altogether reliable. dreaming activity is so fluid and suggestible—and this is notably so when experimenter and subject are the same person—that interference with the phenomena deforms them, and we cannot be sure that by experiment we have really learned much about the life of dreams.

there is, finally, the introspective method. this may be said to be the earliest of the more scientific methods of studying dreams. maine de biran was here a pioneer, and maury, in his famous book, le sommeil et les rêves (1861), which inaugurated the modern study of dreams, adopted a mainly introspective method, though he was not always quite successful in avoiding the fallacies of that method. it is in france that this method has been most frequently and most successfully cultivated.

professor sigmund freud's die traumdeutung (first edition, 1900), may be said to belong to the introspective class, though to a special division which freud himself terms psycho-analytic. this is undoubtedly the most original, the most daring, the most challenging of recent books on dreams, and is now the text-book of a whole school of investigators. it is not a book to be neglected, for it is written by one of the profoundest of living investigators into the obscure depths of the human soul.[vii] even if one rejects freud's methods as unsatisfactory and his facts as unproved, the work of one so bold and so sincere cannot fail to be helpful and stimulating in the highest degree. if it is not the truth it will at least help us to reach the truth.

the little book now presented to the reader belongs mainly to the introspective group of dream studies, though not to the psycho-analytic variety. it is based on data which have accumulated beneath my hands during more than twenty years, and some of the ideas developed in it were put forward in a paper 'on dreaming of the dead,' psychological review, sept. 1895; in 'a note on hypnagogic paramnesia,' mind, no. 22, 1896; and in 'the stuff that dreams are made of,' popular science monthly, april 1899. the book is not the outcome of experiment or of any deliberate concentration of thought on dreaming. i have simply noted down dream experiences,—most often in myself, less often in immediate friends,—directly they have occurred, usually on awakening in the morning. the few unimportant exceptions to the rule are duly noted. by maintaining this rule i have been able to satisfy myself that everything i have set down is reasonably accurate. such a method certainly tends towards the exclusion of peculiar and exceptional dreams. this i do not greatly regret. i am chiefly interested in the problems of normal dreaming; they are sufficiently puzzling and mysterious and they properly present themselves for explanation first. i do not wish it to be understood that i question the[viii] existence of telepathic and other abnormal dream experiences. that is not the case. but it so happens that under the conditions i have laid down i have not met with any dreams that clearly and decisively belong to this abnormal class. such few possible examples as have come under my immediate observation (in no case as personal experiences) are slight, and, moreover, sometimes of too intimate a character for full exposition.

thus my contribution to the psychology of dreaming is simple and unpretentious; it deals only with the fundamental elements of the subject. i do not make this statement entirely in a spirit of humility. it seems to me that in the past the literature of dreaming has often been overweighted by bad observation and reckless theory. by learning to observe and to understand the ordinary nightly experience of dream life we shall best be laying the foundation of future superstructures. for, rightly understood, dreams may furnish us with clues to the whole of life.

havelock ellis.