re comparative method linguistics karl teeter mistake cannot classify language basis phonological correspondence lexical item treat english french relate vium norman borrowing idea method novel classify lamnguage relate write comparative grammar etc many language family is basis classification available lack kind morphological complexity belove indo europeanist algonkianist those ist happen enjoy blessing language b danger confuse borrowing cognate alway real easy lexical connaection english french transparent phonologically less central semantically those dutch hence recent hence borrowing c nothing novel since method example edward sapir establish uto aztecan language really family rather three family d cannot possibly realistically expect normal spend write comparative grammar language previously show relate classification must first hand perfectly easy write crazy comparative grammar random group language e g french english perhap most important point reconstruction morphological system badly comparison lexical item both yield correct result both produce garbage p s lexical item deliberately seem language family where morphology compare most case howveer bind morpheme one compare those sapir case uto aztecan far cry reconstruct kind intricate morphological pattern older indo european language example absent many language family 