This paper aims at briefly recalling the presuppositions the European and Indian grammatical approaches present regarding word-classification. From its very beginning up to the most contemporary trends in linguistics, the European approach knew essentially only one classification of words: the canonical list of eight parts of speech which comes from Dionysius Thrax and which was developed, restructured, sub-divided, etc., throughout the course of history, according to the aims of particular scholars or schools. Within Vyâkarana, native Sanskrit grammar, the situation is not so different: the Sanskrit grammarians generally alternated between only two classifications (the four-fold Yâskan division of parts of speech and the two-fold division proposed by Pânini); some of them like Nâgesa having even tried to reconcile the two.