of philologists. The "personal equation," as the astronomers call it, the allowance for difference of temperament, endowment, and skill, has to be applied, certainly not less rigorously, in estimating the observations and deductions of linguistic scholars than those of the labourers in other sciences. There is, on the one hand, the class of facile and anticipative investigators, whose minds are most impressed by apparent resemblances; who delight in construction, in establishing connections, in grouping together extensive classes, in forming grand and striking hypotheses; who are never willing to say "I do not know:" and, on the other hand, there is the class of less ardent and more phlegmatic students, who look beneath superficial resemblances to profounder differences; who call always for more proof; who are ever ready to confess ignorance, and to hold their judgment in suspense; who refuse their assent to engaging theories, allowing it to be wrung from them only by cogent and convincing evidence. Each class has its advantages: the one furnishes the better explorers, the other the sounder critics; the one is the more numerous and the more popular, the other is the safer and the more strictly scientific.
A notable exemplification of this temperamental difference of authorities is furnished us in connection with one of the families of which we have already treated. We saw reason, in the last lecture, to regard with some doubt the genetic relationship claimed to exist between the five great branches of the Scythian family, as being founded too little on actual correspondence of linguistic materials demonstrably derived from a common source, and too much on mere analogies of linguistic structure—analogies, too, which were able to consist with such important differences as separate the jejune dialect of the Manchus from the rich and almost inflective languages of the Finns and Hungarians. We could not pronounce it certain that the family will be able to maintain its integrity in the light of a more thorough and comprehensive investigation. But, on the other hand, we were unable to deny that it may succeed in doing so; and farther, it is altogether possible that recognizable evidences of ultimate
