conditions (physique, temperament, state of health), the educational influences, the social conditions to which he was subject. The history of literature has accustomed us to researches of this kind.
Among the habits of a man it is necessary to determine his fundamental conceptions relating to the class of facts in which his influence was felt, his conception of life, his knowledge, his predominating tastes, his habitual occupations, his principles of conduct. From these details, in which there is infinite variety, an impression is formed of the man's "character," and the collection of these characteristic features constitutes his "portrait," or, to use a favourite phrase of the day, his "psychology." This exercise, which is still held in great esteem, dates from the time when history was still a branch of literature; it is doubtful whether it can ever become a scientific process. There is perhaps no sure method of summing up the character of a man, even in his lifetime, still less when we can only know him indirectly through the medium of documents. The controversies relative to the interpretation of the conduct of Alexander are a good example of this uncertainty.
If, however, we take the risk of seeking a formula to describe a character, there are two natural temptations against which we must guard: (I) We must not construct the formula out of the person's assertions in regard to himself. (2) The study of imaginary personages (dramas and novels) has accustomed us to seek a logical connection between the various sentiments and the various acts of a man; a character, in literature, is constructed logically. This search for coherency must not be transferred to the study of real men. We are less likely to do