The truth is, the line of separation is altogether superfluous, unless it indicates a real distinction in terms which might otherwise be confounded with each other. Now, between verbal and non-verbal terms employed in predication, a real distinction appears in analysis and construction; but none whatever between verbs and verbs, or between verbs and participles.

PROOF OF THE PRINCIPLE AS TO THE LINE OF SEPARATION.

The truth of this, and the justness of the general criticism may be readily seen by instituting in tabular form, an actual analysis of the two species of predicates, as follows:

TABULAR ANALYSIS OF EXAMPLES.

* This tabular analysis is drawn out in detail, partly for the purpose of exhibiting the tabular method of verbal construction.

Here it is at once seen, that so soon as, in the analysis, we reach the distinction as to parts of speech, there is a marked difference in the construction of the two predicates. Is reading continues a unit throughout. There is no separation whatever in construction. Hence, predicates, of however many terms composed, if purely verbal, neither require nor admit the line of distinction in the diagram. But predicates composed of verbal and nonverbal elements, in their word construction, necessarily diverge, the one part from the other, each pursuing its own line of analysis, and, in the tabular statement of that analysis, requiring to be separated from the other, by the vertical line seen in the example. Of this line, the line of distinction or separation in the diagram is the proper symbol. Hence, only in such predicates should that line* be inserted, as has been affirmed in rule seventh.

A peculiar interest attaches to this demonstration, inasmuch as it beautifully illustrates the fact before urged; namely, that the diagram system is neither fanciful nor arbitrary, but is everywhere permeated by a consistent philosophy.

2. ATTACHMENT OF ADJUNCT FIGURES AS TO ORDER.

Secondly. In some instances, adjunct figures are so attached in the diagrams, as not to represent the true relations of the elements enclosed. Take, for example, the adverb How in the following diagram :

* In most of these figures the line of separation is too indistinct. It should be longer and should be more boldly drawn.

mar, page 57, is truth.

See Gram

pre

Here, how is attached to the verbal section of the dicate figure, as if it related to are, and the sense were how are. The true sense is, however, how dear. It depends, then, wholly upon dear. Its figure should, therefore, (as well as that of the phrase following) have been attached farther to the right, so as to bring it under the adjective section of the predicate. The line of separation is, of course, required between are and dear.

Again, take the adverb not, and the phrase to tell what, in the following:

Here, both adjuncts are unmistakably made to modify the verbal auxiliary is. Their relation is, however, solely that of modifiers of the predicate adjective able. They should, then, like the adjuncts in the former example, have been attached to the predicate, and on the right of the line of separation.

The true mode is fairly shown in the predicate in the following:

FAULTY ATTACHMENT OF A GENERAL ADJUNCT.

The last diagram but one is identical with the one often applied to the sentence,

"I speak not to disprove what Brutus spoke."

Insert the sentence in the diagram, and not, while construed in pleasant accordance with the profound philosophy of the old-style "parsing," is nevertheless squarely at variance with a just analysis. It is logically false. It not only misrepresents the speaker's meaning, but makes him assert a manifest untruth. His statement is not, I speak not; for he is, on the contrary, actually speaking. His real declaration is, I do speak, but not to disprove what Brutus spoke. Not, is, therefore, not a predicative adjunct, but a general phrase adjunct modifying to disprove, and it should have been attached, accordingly, beneath the relative and the subsequent, after the manner of almost in the following example.*

This case is especially important as representative of a large class of adjuncts, with reference to which teachers are often in doubt, and, in good part, because, under the old system, they were, in general, so falsely construed.

3. ATTACHMENT AS TO GRAMMATICAL SUBORDINATION.

Thirdly. A still graver error occurs in the treatment of infinitive phrase adjuncts, as partially exemplified in to be wondrous wise, in the following:

* The example is referred to, purely for the diagram, not because this construction of almost is correct. For the proper construction, see the first diagram, page 97.

Accepting this phrase as a mere adjunct, which will hereafter be shown to be not the truest classification, it is found treated as a modifier of the predicate. A closer analysis, however, shows that if it be an adjunct, it is strictly objective and dependent on him.

PROOF OF THIS OBJECTIVE RELATION OF INFINITIVES.

This will be seen from the following demonstration. The logical reference of every verbal term is, either more or less directly, to some substantive term employed as representative of the agent. Hence, verbs in the sentential modes, with one exception, relate to their subjects. Verbal terms in the two phrasal modes, however, cannot have subjects. But they may have in grammatical connection, terms logically representative of the agent. To these logical agents, they as truly relate, as their corresponding sentence forms relate to their proper subjects. Take, for example, this parallelism,

I wish that he would go,

I wish him to go.

Now the relation of would go is no more distinctly to he, than is the relation of to go to him. He and him, equally though differently, represent who is to perform the required act; and would go and to go, equally though differently, represent a similar act to be performed by the same agent. Hence, the proper form can only be the following: