_Swear by your sword--swear on your sword._--Which of these two expressions is right? This depends on what the speaker means. If he mean _make your oath in the full remembrance of the trust you put in your sword, and with the imprecation, therein implied, that it shall fail you, or turn against you if you speak falsely_, the former expression is the right one. But, if he mean swear _with your hand upon your sword_, it is the latter which expresses his meaning. To take a different view of this question, and to write as a rule that {394} _verbs of swearing are followed by the preposition on_ (or _by_) is to mistake the province of the grammar. Grammar tells no one what he should wish to say. It only tells him how what he wishes to say should be said. Much of the criticism on the use of _will_ and _shall_ is faulty in this respect. _Will_ expresses one idea of futurity, _shall_ another. The syntax of the two words is very nearly that of any other two. That one of the words is oftenest used with a first person, and the other with a second, is a fact, as will be seen hereafter, connected with the nature of _things_, not of words. s. 469. The following question now occurs. If the history of forms of speech be one thing, and the history of idioms another; if this question be a part of logic, and that question a part of rhetoric; and if such truly grammatical facts as government and concord are, as matters of common sense, to be left uninvestigated and unexplained, what remains as syntax? This is answered by the following distinction. There are two sorts of syntax; theoretical and practical, scientific and historical, pure and mixed. Of these, the first consists in the analysis and proof of those rules which common practice applies without investigation, and common sense appreciates, in a rough and gross manner, from an appreciation of the results. This is the syntax of government and concord, or of those points which find no place in the present work, for the following reason--_they are either too easy or too hard for it_. If explained scientifically they are matters of close and minute reasoning; if exhibited empirically they are mere rules for the memory. Besides this they are universal facts of languages in general, and not the particular facts of any one language. Like other universal facts they are capable of being expressed symbolically. That the verb (A) agrees with its pronoun (B) is an immutable fact: or, changing the mode of expression, we may say that language can only fulfil its great primary object of intelligibility when A = B. And so on throughout. A formal syntax thus exhibited, and even devised _[`a] priori_, is a philological possibility. And it is also the measure of philological anomalies. {395}