are now perhaps for the first time offered with a critical discussion
of their historical sources; lastly, because the old legends seem to
show how the fancy of periods less critical than ours dealt with such
facts as are now reported in a dull undramatic manner. Thus (1) the
Icelandic ghost stories have peculiar literary merit as simple
dramatic narratives. (2) Every one has heard of the Wesley ghost, Sir
George Villiers's spectre, Lord Lyttelton's ghost, the Beresford
ghost, Mr. Williams's dream of Mr. Perceval's murder, and so forth.
But the original sources have not, as a rule, been examined in the
ordinary spirit of calm historical criticism, by aid of a comparison
of the earliest versions in print or manuscript. (3) Even ghost
stories, as a rule, have some basis of fact, whether fact of
hallucination, or illusion, or imposture. They are, at lowest, "human
documents". Now, granting such facts (of imposture, hallucination, or
what you will), as our dull, modern narratives contain, we can regard
these facts, or things like these, as the nuclei which our less
critical ancestors elaborated into their extraordinary romances. In
this way the belief in demoniacal possession (distinguished, as such,
from madness and epilepsy) has its nucleus, some contend, in the
phenomena of alternating personalities in certain patients. Their
characters, ideas, habits, and even voices change, and the most
obvious solution of the problem, in the past, was to suppose that a
new alien personality--a "devil"--had entered into the sufferer.

Again, the phenomena occurring in "haunted houses" (whether caused, or
not, by imposture or hallucination, or both) were easily magnified
into such legends as that of Grettir and Glam, and into the
monstrosities of the witch trials. Once more the simple hallucination
of a dead person's appearance in his house demanded an explanation.
This was easily given by evolving a legend that he was a spirit,