Page:Folk-lore - A Quarterly Review. Volume 3, 1892.djvu/56

48 can be proved. Hence the value of the instance I have just examined.

Another lesson that may be learnt from this instance is the invalidity of an argument dear to many students of history, the argument ex silentio. I believe that even in historical investigation proper a most unwarranted use is often made of this argument; in folk-lore research it should never be used save with the utmost caution. Could we apply a universal phonograph to the entirety of living oral tradition we should even then be far from justified in dogmatising about what may or may not have taken place formerly. But, as every folk-lorist well knows, it is but fragments of tradition that have been recorded and published. Every now and then a fresh fragment comes to light, and, like the Gaelic märchen of Gold-tree, opens up new lines of investigation, and compels us to seek in new directions for the solution of our problems.

.