as the exclusion from his sentences of irrelevant or overburdening material is concerned.

A glance at the table will show that the greatest "actual sentence-length," 36.7 words, is recorded against The Rambler. This simply means that it took Johnson 36.7 words during the Rambler period to express one complete thought. The shortest, 18.6, as one would imagine, belongs to the expository sections of the Lives of the Poets. Between the two styles there is all the difference between leisureliness and incisiveness. It is interesting to note, too, that between the time of The Rambler and that of The Idler his "actual sentence-length" decreased 33 per cent; and between that of the Idler and that of the Lives of the Poets diminished further by about 25 per cent.

The second part of this table is based directly on this same matter of "actual sentence-length," or number of main clauses, but from a new point of view. It involves a double study: A-the numerical occurrence of main clauses that are logically interlocked with the preceding main clauses and the number that are not. And for this purpose, I wanted to ascertain whether, as his style developed, associated clauses, those closely related, increased relatively in proportion to the number illogically joined. The presumption was that they would increase, since looseness of clause relationship accompanies lack of emphasis, and is the mark of a less sophisticated style. Such main clauses as were closely articulated I brought under Type I; such as were not, under Type II. The other side of the study, B, had to do with the relative proportions of Type I to Type II. This was important for purposes of chronological comparison. The presumption here was that the early style, with its greater sentence-length and its tendency to overcrowd the periods, would produce more disassociated main clauses,-clauses of Type II, that is, than the later style. This proved to be borne out by the facts, as the table will show.

Another point. The table concerns itself with two styles of writing, narrative and exposition. Johnson's narrative style was always looser and more disjointed than his ex

pository. And this fact can be clearly seen in every entry save that for the first ten Rambler papers, the most ponderous of all his well-known work.

And finally the table is further divided into Major and Minor sections. The latter division is in the nature of an appendix. It may prove interesting to note his advance in sentence swiftness in the same field over twenty years of writing. Taxation No Tyranny is one of his strongest, most trenchant pieces of writing.

Table VI, besides borrowing from Table V in the matter of "indicated" and "actual" sentence-length, undertakes an analysis of Looseness and Periodicity, of Predications, and

of the percentage of Simple Sentences. When I say looseness and periodicity I have in mind only exaggerated occurrences. Looseness here is used in an adversely critical sense and refers only to such sentences as trail on and on until just this side of infinity. Very few such will have less than fifty words or fewer than four or five trailing elements imposed on the period after the completion of the thought. I use the term "periodicity" with its usual signification but record as examples only those sentences, rather elaborate in their architecture, that release the thesis of the period only after fairly lengthy subordinate elements have been advanced. No sentences of under thirty words have been included for the simple reason that such would not be ambitiously periodic. It is axiomatic in connection with Looseness and Periodicity that the excessive presence of the former will be indicative of weakness or at least carelessness, and that the presence of the latter will indicate strength. The following passage from the Life of Savage will furnish instances of both types.

That this act of generosity may receive its due praise, and that the good actions of Mrs. Oldfield may not be sullied by her general character, it is proper to mention what Mr. Savage often declared, in the strongest terms, that he never saw her alone, or in any other place than behind the scenes.

At her death he endeavoured to show his gratitude in the most decent manner, by wearing mourning, as for a mother; but did not celebrate her in elegies, because he knew that too great profusion of praise would only have revived those faults which his natural equity did not allow him to think less, because they were committed by one who favoured him; but of which, though his virtue would not endeavour to palliate them, his gratitude would not suffer him to prolong the memory, or diffuse the censure. In his Wanderer, he has, indeed, taken an opportunity of mentioning her; but celebrates her not for her virtue, but her beauty, an excellence which none ever denied her: this is the only encomium with which he has rewarded her liberality; and, perhaps, he has, even in this, been too lavish of his praise.

Life of Savage.

I am using the term "predications" in its traditional sense of trailing participial and gerundial phrases, and clauses, but with this restriction: although I have counted all clauses that

follow completed sense, I have recorded participial and gerundial phrases only when they were long enough to impress a reader as really loose elements. And I have recorded, as well, long, dangling double predicates. For instance, in the above quoted passage from Savage there are three sentences. In the first I find no predications, as may be expected where there is strong periodicity. In the second sentence there are seven predications, in the third, five. The record of a writer's predications will, of course, be a registry of his looseness. In all the matters just discussed Johnson varied chronologically both in narrative and exposition.

I define "simple sentence" rather strictly as a group of words expressing complete thought through the medium of one clause. This will exclude even subject and object clauses.

1 Note the coincidence of the expository section from the Lives of the Poets having but half the number of predications recorded against The Rambler.

Perhaps the most persistent stylistic device employed by Johnson was his conscious rhetorical use of minor and major balanced elements. By minor elements I mean triads, or duplicated couplings or phrases, with or without homologous modifiers for instance, "his lilies and his roses, his satyrs and his dryads"; "to amusements of greater pleasure, or to studies of better prospect." By major elements I have in

mind balanced clauses. This device was ever present in Johnson's prose as it has been present in the work of most who employ artifice to gain effect. Lyly and Swinburne notably employ it, as the appended sentences will show.

This I gather from your answer, that beauty may have fair leaves, and foul fruit, that all that are aimable are not honest, that love proceedeth from the woman's perfection and the man's follies, that the trial looked for is to perform whatsoever they promise, that in mind he be virtuous, in body comely, such a husband in my opinion is to be wished for but not looked for.

John Lyly.

The points of contact and sides of likeness

between

William Blake and Walt Whitman

are

so many and so grave

as to afford some grounds of reason to those who preach the transition of souls or transfusion of spirits.

(and)

The pinchbeck structure of Hibernian fiction and
the pasteboard outworks of Hibernian faction.

Both from Swinburne.

Johnson is not so artificial as either of these writers, for, although he employs parasonic parallelism, he very infrequently has recourse to transverse alliteration. Indeed he is never strongly alliterative.

This trick of Johnson's was of great importance in determining the quality or type of his rhythm. I should say that it was the most important single factor.

These instances of balance vary both in length and in artifice. I was myself interested in trying to ascertain whether his tendency to employ them increased or diminished, and whether-no matter how the case held in regard to frequency -the balanced elements, by lengthening, grew more artificial. It was, in a way, a labor of supererogation. I did learn, though, what I had only suspected, that both numerically and artificially he was more given to the device in 1780 than in 1750. Table VII will show to what extent.