Third. In the history of Indo-European flection we do wrong if we assume the separate existence of a period of construction and one of decay.
CHAPTER XX.
THE DIVISION OF THE PARTS OF SPEECH.
The division commonly adopted of the parts of speech in the Indo-European language is convenient as a classification; but it must be borne in mind that it is not logically accurate, nor is it exhaustive. It is indeed impossible to divide words into sharply defined categories, seeing that, however we may divide them, we shall find it difficult to exclude some from each category which may fairly claim to be registered under some other category or categories, basing their claim upon at least certain uses.
The accepted grammatical categories have had their form determined mainly by the consideration of three points: (1) by the meaning of each word taken by itself; (2) by its function in the sentence; (3) by its capacity for inflection, and the part it plays in word-formation.
As regards the meaning of the word, we may notice that the grammatical categories of substantive, adjective, and verb correspond to the logical categories of substance, quality, and activity, or, more properly, occurrence. But here, at the outset, we find that the substantive is not confined to the denotation of substance, as there are also substantives denoting quality and occurrence as, ‘brightness,’ a ‘rise.’ There are also verbs which denote continuous states and qualities; as, ‘to remain,’ or the Latin ‘cande’ = ‘to be white.’ Pronouns and numerals again have a right on the score of meaning to be separated as classes from substantives and adjectives: but these, again, must be separated from each other in their substantival as against adjectival use (e.g. each as against each man; Six went and six stayed as against Six men, etc.; this and that as against this book and that one), which forbids us to simply co-ordinate the classes: substantive, adjective, pronoun, numerals. And, on the other hand, it must follow that, if pronouns and numerals are to be regarded as distinct species of the noun class, the same separation must be extended to the adverb class: since badly, there, twice, are related to each other just as bad, this, two.
To come to the connecting words. The lines that define the class of the conjunctions are quite arbitrary; where, for instance, is called an adverb even in passages like this:[201] “Where, in former times, the only remedy for misgovernment real or supposed was a change of dynasty, the evil is now corrected at no greater cost than a ministerial crisis.” As and while, again, are called conjunctions. In the simple sentence, the test usually applied to distinguish prepositions from conjunctions is case-government. But it certainly is entirely illogical to call words like before, since, after prepositions when they occur in simple sentences, and to call them conjunctions when they connect sentences;—for this function is in both cases exactly the same; cf. before my interview with you, and before I saw you.
If we wished to classify words according to their function in the sentence, it might seem obvious to divide words (1) into those which can of themselves form a sentence, (2) into those which can serve as members of a sentence, and (3) those which can only serve to connect such members.
In the first division we might, then, place the interjections, which, when isolated, are really imperfect sentences. But these also occur as members of a sentence, sometimes with and sometimes without a preposition; as, Woe to the land! Out on thee! Oh my!
The finite verb in its original use better fulfils the idea of a perfect sentence. But in its present use it appears—if we except the imperative—as a mere predicate attached to a subject separately denoted. And the so-called auxiliaries are mainly used as mere connecting words.