by a verb or preposition, and a word that cannot. I have therefore, in the first instance, introduced the learner to a considerable number of the so-called pronouns, under the head of adjectives. These latter I have distributed into the three classes of Qualitative, Quantitative, and Demonstrative Adjectives. It is very perplexing to a beginner to have his notions of an adjective derived from the Qualitative class exclusively, and then to be left to deal with the rest as he best can. Indeed, many writers of grammars have perplexed themselves as much as their pupils, and have put such words as all, many, &c., and even the numerals, into the class of pronouns. It appears to me a most unfortunate misuse of terms, when, instead of keeping to the simple and exhaustive classification of nouns and adjectives, the latter are called nouns adjective. The Latin grammars offend most pertinaciously in this respect. The grammatical affinities of words are greatly obscured by this error. An adjective is not a name. Moreover, it will be seen from the classification of notions and their verbal representatives, which is given in the course of the present work, that the adjective and the verb are more closely related to each other, than the adjective and the noun, since they both express

attributive notions.

The scheme of tenses which I have adopted agrees in its main features with the classification of all the best modern grammarians. It is simpler, more exact, and in every way better than such awkward, ambiguous, and unmeaning terms as pluperfect, prior perfect, progressive forms first future, second future, with which most English grammars abound.

The adverb is a part of speech which has suffered much ill usage at the hands of grammarians. Its domain has been very improperly restricted, and many words which are genuine adverbs in their relation to verbs, adjectives, and other adverbs, have been set down as mere conjunctions. In the classification which I have adopted, I have merely endeavoured to apply carefully the acknowledged truth, that a word which indicates any of the conditions of time, place, manner, degree, cause, or circumstance under which an attributive notion is connected with an object of thought, is an adverb. Some will perhaps demur at first to the truth of the statement that such words as than, as, therefore, &c. are adverbs. Before they finally reject it, however, they should examine and compare what is said in §§ 260, 264, 266, 267, 292,

408, with the examples of the analysis of compound and elliptical sentences. It is important to observe that in continuous speech thoughts may be connected with each other by the simple sequence, or juxtaposition of sentences, without the existence of any formal bond of connection. In this way demonstrative words of various kinds may refer the mind back to something previously mentioned, although there is no structural connection between the sentence in which they occur and the preceding sentence. The relative pronoun is rightly called a connective word, but the pronoun he carries the mind back to some antecedent name, quite as much as the relative does. Yet no grammarian would class he amongst the connective words. Who is a connective word not through its relative force, but through the structural connection which it establishes between two clauses. Through want of attention to this distinction many merely demonstrative adverbs have been set down by grammarians as conjunctions. Becker offends as much as any in this respect. Further remarks on this point will be found in §§ 408, &c. of the present work.

In treating of Conjunctions I have adopted the classification indicated by Becker, rejecting many of the details, which, for reasons indicated above, appeared to me to mar the whole scheme. In a note on § 286, enough has been said to justify the disuse of the stupid old names, copulative conjunctions and disjunctive conjunctions, the former of which involves an unmeaning tautology, while the latter is simply self-contradictory. The division into co-ordinative and subordinative conjunctions has at least the advantage of being based upon a well-established classification of compound sentences, of exhibiting structural distinctions which the old-fashioned division obliterates, and of presenting the only distinction which really has a grammatical import. It is one of the merits of the new Public School Latin Primer, that it adopts this simplified division of conjunctions. Let us hope that that venerable old impostor, the Disjunctive Conjunction, will soon be extant only in a fossil state. If its modern counterfeit, the Adversative Conjunction, shares the same fate, I shall be well satisfied.

The syntactical portion of the present work derives many of its leading features from the principles developed by Becker in his German Grammar. The publication of that work may well be regarded as an epoch in the history of grammatical science. Its

leading doctrines are incontrovertibly sound and philosophical, though the same unqualified praise is by no means to be bestowed on the details of their development. The latter abound in capricious distinctions and arbitrary generalizations.

In this edition I have introduced a classification of words based upon that of Becker, but with some important alterations. His treatment of Relational words appears to me to import into the subject considerations utterly foreign to grammar, and to make a number of very questionable metaphysical distinctions override the most obvious grammatical affinities. The whole question cannot be discussed here, but the more I examine the matter, the more decisively I reject a classification which throws together auxiliary verbs, articles, pronouns, numerals, prepositions, conjunctions, and adverbs, and treats them as relational words, denoting partly the relation of some notion to the speaker, and partly the relation of one notion to another, while verbs, substantives and adjectives, are (rightly enough) set down as words that express some notion. It is an utter mistake in grammar to make the collateral signification of a word override its grammatical functions. The declension and syntactical structure of bonus 'good,' and meus 'my,' show that the latter belongs (grammatically) to the same class of words as the former. To make the non-grammatical consideration that meus involves in its meaning a reference to the speaker, which bonus does not, the ground for assigning the former the class of notional words, and the latter to the class of relational words, is subversive of all sound principles of classification. It seems obvious enough too, that the relation to the speaker, which is involved in meus, is part of the notion expressed by the word.

Becker distinguishes three relations in which words stand to each other:-1. The Predicative; 2. The Attributive; 3. The Objective. About the first two of these there is no difficulty. In place of the third I have introduced two separate relations, the Objective Relation and the Adverbial Relation. Practically this makes but little difference, for Becker subdivides Objects into Objects that complete the predicate (to which the term object is commonly applied in grammar), and objects that determine or individualize the general meaning of the verb or adjective, without completing it, with reference to which he uses the term adverbial relation. My objection to Becker's classification is

directed mainly to his endeavour to bring both these under the one head of Object. That term has such a very definite and intelligible sense in grammar, that it appears to me a most arbitrary and unnatural use of it, to say that the place, the manner, nay even the cause of an action, is an object of the action. I also object to the distinction that he draws by his use of the term completing relation. In such sentences as "He strikes the ball," "He runs across the meadow," the verb strikes expresses the action referred to at least as completely as the verb runs; and the phrase across the meadow completes the notion in the latter case, quite as much as ball does in the former. This consideration will become still more obvious when we consider that the original force of the accusative case was to denote motion to an object.

The mode in which I have treated the terms predicate and copula (§ 347) agrees with that of Dr. Kennedy in his Latin Grammar (§ 101), though it was not borrowed from him, as, at the time when my grammar first appeared, I knew Dr. Kennedy's Latin Grammar only by name, and was quite unaware of the mode in which he treats the subject. The omission of the verb sum as a copula has also the weighty sanction of Madvig (Lat. Gr., § 209). It is also rejected by Mr. Roby (Lat. Gr., § 143) and the authors of the Public School Latin Primer. The obstinate vitality of grammatical errors is something wonderful. Grammars of repute (both English and Latin) will be found in which the learner is taught that the predicate of a sentence may be a verb, an adjective, or another substantive, as though an adjective could by any possibility be the equivalent of a verb. One advantage that will follow from the adoption of the view here taken will be that we shall get rid of a difficulty, which, if not quietly ignored (as is often the case in systems of grammatical analysis), is sure to lead to an anomaly. If, in the sentence He is rich, rich is the predicate, and is the copula, why, in the sentence He becomes rich, should we not call becomes the copula? The notion of becoming has quite as good a right to be considered copulative as the notion of being. The difficulty is removed, and the anomaly obviated, when we regard neither be nor become as a copula, but treat them both as verbs of incomplete predication (see § 392). And now ensues another advantage from discarding Becker's use of the term completion of the predicate, as applied to the object of a transitive verb. We can apply it, or some equiva

lent term, in the case of verbs which really do express only an incomplete notion. To avoid confusion, I have adopted the term complement instead of completion.

In the present edition the alphabetical system of English has been discussed at greater length. The sections in which the Pronouns are treated of are arranged in a way which will render the subject easier to the learner; and some important additions have been made in this portion of the work. In the conjugation of the verb an alteration has been made in the tenses of the Subjunctive Mood. Forms which before had been classed as future tenses are treated as secondary forms of the past tenses of the subjunctive. This makes them the equivalents in name, as well as in force, of the tenses which to a considerable extent they have supplanted. On comparing them with the tenses still classed as futures, an important difference in the forms for the second and third persons will be observed. The view taken of gerunds and participles (§§ 197-202, 212) is, in the main, that of all the best authorities. Some modifications in the sections on Analysis are explained and justified in the notes. As regards the mode in which subordinate clauses are bracketed and denoted, although it is by no means essential to the analysis of a sentence, I can affirm from experience that it is attended with great advantage, and that pupils of ordinary intelligence master it without difficulty.

For some improvements introduced into this edition I am indebted to suggestions derived from Mr. Abbott's excellent Shakspearian Grammar;" from Mr. Morris's " Specimens of Early English;" and from the scholarly work on the English Language by Dr. Ernest Adams, a book which should be in the hands of all students who desire to investigate the sources, structure, and affinities of our language more deeply than is attempted in the present work.

I have to thank some friends for criticisms which have enabled me to correct one or two oversights. I also take this opportunity of expressing my sense (by no means a grateful one) of the liberal manner in which subsequent writers and compilers (to whose attention I would commend the etymology of the term applied to them) have availed themselves, sometimes with and sometimes without acknowledgment, of the contents of the present work.

DENMARK HILL, July, 1870.

C. P. MASON.