adding ful to it, without concerning ourselves, as to whether the corresponding phrase, "full of so and so," would or would not make good sense. And when we hear a Scotchman say fearfu', carefu', we both understand him without difficulty, and do not think of inquiring whether he also clips the adjective full to fu'.
The word of opposite meaning, fearless, is not less readily recognizable as a compound, and our first impulse is to see in its final element our common word less, to interpret fearless as meaning 'minus fear,' 'deprived of fear,' and so 'exempt from fear.' A little study of the history of such words, however, as it is to be read in other dialects, shows us that this is a mistake, and that our less has nothing whatever to do with the compound. The Anglo-Saxon form of the ending, leas, is palpably the adjective leas, which is the same with our word loose; and fearless is primarily 'loose from fear,' 'free from fear.' The original subordinate member of the compound has here gone completely through the process of conversion into a suffix, being so divorced from the words which are really akin with it that its derivation is greatly obscured, and a false etymology is suggested to the mind which reflects upon it.
Take, again, such words as godly, homely, brotherly, lovely. Here, as in the other cases, each is composed of two parts; but, while we recognize the one as a noun, having an independent existence in the language, we do not even feel tempted to regard the other as anything but an adjective suffix, destitute of separate significance; it appears in our usage only as an appendage to other words, impressing upon them a certain modification of meaning. What, however, is its history? Upon tracing it up into the older form of our speech, the Anglo-Saxon, we find that our modern usage has mutilated it after the same fashion as the Scottish dialect now mutilates the ful of fearful—by dropping off, namely, an original final consonant: its earlier form was lic. The final guttural letter we find preserved even to the present day in the corresponding suffixes of the other Germanic languages, as in the German lich, Swedish lic, Dutch lijk. These facts lead us naturally to the conjecture that the so-
