'We owe to the perspicacity and learned labours of various writers, chiefly German, the reforms made in Latin Syntax. I cannot assign to each his due share. The Grotefends, Krüger, Zumpt, O. Schulz, Ramshorn, Kühner, Madvig, Key, have each their special merits. Of these I place Raphael

Kühner in the first rank; and I am much indebted to Grieben's "Lateinische Satzverbindungen." In cur own country the scholastic study of this part of Grammar was usefully promoted by the Exercise-books of T. Kerchever Arnold.

'These reforms brought into the teaching of Latin Syntax, besides the terms already named, a certain number more, perhaps from forty to fifty, including the names given to the several varieties of the Simple and Compound Sentence, with their subdivisions; including also the terms Protasis and Apodosis in sentences which, like the Conditional, take these parts.

§ 9. As regards the new terms which my own improvements have suggested, three alone have frequent and important practical use; the value of which I insist on as very great. These are, (1) Prolative (Infinitive); (2) Copulative Verbs, introduced first in my "Elementary Grammar "; (3) Suboblique (clause or verb), a convenient abridgment of the necessary phrase "Subordinate to Oratio Obliqua."

'Further, it appeared that the doctrine of copulative predication in Grammar required, for its clear statement, the use of some terminology from which the term predicate itself should be excluded; and this was at length found in the term used by Mr. C. P. Mason, (predicative) Complement.

'I say then, generally, that a new term proposed in Grammar is not to be condemned because it is new; but, if at all, for one of three reasons: that it is superfluous; or that it is inadequate; or because a better term is suggested. As respects myself, I repeat that I have not the least disposition to use hard terms; and I say that those which I have introduced are unjustly so described. But I cannot adopt the poor pedantry which refuses to facilitate and abridge discourse by the use of suitable terms; to write, for instance (after due explanation) "Collective Subject" rather than "Nominative Singular of a Substantive which implies a multitude of persons or things": and "Composite Subject" rather than "two or more singular Nominatives agreeing with one plural Verb."

§ 10. "My "Elementary Latin Grammar," first published in 1843, obtained, after twenty years, approval so wide, that its circulation approached 8,000 copies annually: and, during those years, not a single complaint affecting its terminology was heard either from the public press or from the eminent teachers who used it in their schools. Such attacks broke out when it was adopted as the groundwork of a new school grammar; and their justice may be tested both by this fact, and by comparing the imaginary difficulty imputed to a few new terms in the Primer, with the many and great obstacles existing in its chief predecessor, Lilly's Grammar."

§ 11. In the Preface to the Third and Fourth Editions certain portions of Syntax were discussed. Those discussions, being of signal importance to the right appreciation of Latin Compound Construction, will here be repeated generally: but with partial suppression of some topics and enlargement of others.

I. The Doctrine of Predication.

§ 12. This Doctrine is treated (§§ 100-106) in agreement with the principles now received in all Continental Latin Grammars, and in most Grammars of the English language, but with some slight variations in the mode of treatment. Logic and Grammar are akin to one another; but their spheres are different. Logic is the Grammar of reasoning: it develops 'the laws of thought.' Grammar is the Logic of language it displays the rules and idioms of discourse. The Correlation and the Terms Subject-Predicate are necessary to both sciences. But the scope of these terms is not the same in both.

If we take a Simple Sentence, such as 'beneficium male collocatum nocet (noxium est) hominum societati,' we see that the Logical Subject of this proposition is 'beneficium male collocatum,' but the Grammatical Subject of the sentence is 'beneficium,' of which male collocatum' is an adjunct. Again, the Logical Predicate is 'noxium,' the Grammatical Predicate nocet' or 'noxium est,' of which'hominum societati' is an adjunct. Hence appears the propriety and necessity (if confusion is a thing proper and necessary to be avoided) of distinguishing the terms Subject and Predicate in Grammar by the epithet 'Grammatical.' As for the terms Subject

Predicate themselves, they have now so firm a footing in the science of Grammar that they cannot be excluded from it, if their exclusion were desirable. See 'Predicate' in Index I.

The Subject is 'id quod Praedicato subjectum est': the Predicate is id quod de Subjecto praedicatum est.' The combination of the two (as Kühner says: 'Ausführliche Grammatik der Lat. Spr.,' Part iii. § 1) is rightly called the Predicative Relation, because the Predicate (or Verbal notion) is the kernel of speech, to which the Substantival notion stands in subjection, and is therefore called Subject; often indeed expressed by the endings of the Verb (am-o, ama-s, &c.).

When I was preparing my Elementary Latin Grammar' forty years ago, being in some dread of interference with Logic, I took for my type of simple predication, 'homo est mortalis.' But, when the Primer was compiled in 1866, the four Oxford scholars engaged in that work unanimously held that (in Grammar) Subject and Finite Verb are the true norm (homo moritur), and that Incomplete Predication (of the form homo est mortalis) should be taken afterwards as the large exception. This settled the question then, in accordance (as before noticed) with the practice of all continental writers and a verdict thus authoritatively and generally pronounced is surely entitled to acceptance.

II. Complement (of Predication).

:

§ 13. This suitable and useful term was first suggested by Mr. C. P. Mason in his English Grammar,' to designate that which completes the sense of a Simple Sentence when the verb is one of incomplete Predication' (called 'Copulative' in this Grammar, p. 351).

In sentences such as 'homo moritur (est mortalis),' we have seen above that the Grammatical Predicate is (not 'mortalis,' but) 'moritur' or 'est mortalis.' Donaldson's expedient, of using the terms 'primary, secondary, tertiary' predicate, I cannot approve. It confounds confusion, invades the domain of Logic gratuitously, and carries into the rules of Grammar the use of a word (predicate), which, however necessary to the preamble of Syntax, as the correlative of Subject, may be replaced afterwards by the term Finite Verb (or Verb of the Sentence) with great advantage. All confusion is happily avoided by the term 'Complement,' which is wide

enough to include every word or phrase capable of completing the construction of a Copulative Verb, whether finite or infinitive. See the Examples on p. 352.

III. Relations in the Simple Sentence.

§14. Mr. Mason, in his 'English Grammar,' following Becker's 'Organism der Sprache,' treats of the Relations of Words in the Simple Sentence. The 'Public School Latin Grammar' does the same. One of our critics regards these Relations as 'spurious children of Logic and Grammar.' But he has failed to interpret the procedure rightly. It is as purely grammatical as any procedure can be, which admits (what no grammarian can now exclude) the correlations Subject-Predicate and Subject-Object.

Two of Mr. Mason's 'Relations,' the Predicative (I.) and the Objective (III.), are the same, in title and extent, as those of this Grammar. His Attributive' Relation contains the Qualitative (II.), but is more extensive: his 'Adverbial' Relation contains the Circumstantive (V.), but is more extensive.

Mr. Mason was dealing with English, a language of rare inflexions, using Prepositions in their stead. I deal with Latin, a largely inflected language. But even in English the Genitive should not be merged in the Attributive Relation, and the Dative Case in the Adverbial (Circumstantive): much less in Latin. For, true as it is that numerous instances of the Genitive are attributive in character, and that many Datives might be replaced by Preposition with case (ie. adverbially); still there remain very many examples of each case which cannot be so represented, and this fact, combined with that signal distinction between forms of construction, which merits distinct treatment in Grammar, leads to the conclusion that the Dative and Genitive Cases ought to rank as separate Relations. The Dative is therefore classed here under the 'Receptive' (IV.), and the Genitive under the 'Proprietive' Relation (VI.).

Relation VII., that of the Prolative Infinitive,' appeared for the first time in the Public School Latin Primer.' It comprises all the instances in which the Infinitive extends (profert) the construction of words capable of being followed in dependence by a Copulative Infinitive with Nominative Com

plement. See § 180. In the 'Elementary Latin Grammar' the Infinitive with some of these Verbs (soleo, possum, &c.) was called Objective; with others (videor, dicor, &c.) Predicative (ie. complemental). But these shifts never satisfied: for if, in 'soleo errare,' the Infinitive is Object of soleo,' it is an unique Object : and if, in 'videor errare,' the Infinitive is predicatively complemental (which in some sense it is), its character as a Complement' is widely distinct from that of an Adjective or Substantive (which qualify the Subject), and from every other instance in p. 352. And how, on the same principle, can we analyse without the most unpleasant confusion such sentences as these?

Marcus putatur velle fieri philosophus. Sapientis est velle fieri doctiorem.

At length a conviction was reached, that this usage of Grammar (common to all Aryan languages at least) deserves separate classification as a specialty of the Infinitive Verb-noun.

:

Madvig's mode of treating this construction is not essentially different in principle. Under one head (§ 180) this Grammar gives what he sets forth in three places (§ 389, § 393, § 400). He treats in one and the same chapter of the Infinitive in Oratio Recta and Oratio Obliqua. Deeming it right and important to keep Simple and Compound Construction apart, we consider in Chapter III. the Infinitive of Oratio Recta, in Chapter IV. that of Oratio Obliqua. But when Madvig speaks of the Infinitive as joined to these (extensible) Verbs 'um den Begriff zu ergänzen und die Handlung zu ergeben' (to complete the idea and supply the action), this is exactly what is meant by the Prolative Relation of the Infinitive and it is very much the same as the use of the Infinitive, in German and English, with those Verbs which some grammarians have very inadequately called 'auxiliary' (ich will, soll, kann, muss, &c. kommen: I will-shall-can-must, &c. come). The construction belongs also to French, a Romance (latinistic) language. For though French inflects (with Latin) I will come, I would come, by 'je viendrai, je viendrais,' it falls in with Latin, German, English, in saying je peux—je veux—je désire-j'ose, &c. venir. It is unquestionably true that after many of these Verbs the Infinitive may be called an Object by anybody who wishes to do so, as in vincere scis, tu sais vaincre,'' cupis abire, tu désires partir,' &c. The use of the