¹ Percy Simpson perhaps best represents those who believe the punctuation of the Folio rhythmic or elocutionary: see his Shakespearean Punctuation (Oxford, 1911). C. C. Fries's article, Univ. of Mich. Studies in Shakespeare, Milton, and Donne, Vol. i (New York 1925), well represents the other view.

² The Manuscript of Milton's Paradise Lost, Book I, edited by Helen Darbishire (Oxford, 1931), Introduction, pp. xxxvi-xlii.

³ Simpson, p. 16: “One result of a rhythmical as contrasted with a logical system of punctuation is the use of fewer stops … A natural result is the wider employment of the comma.” There are no fewer than 158 medial punctuation marks in the edition of 1645 not in the manuscript—150 commas, 3 semicolons, 4 periods, and 1 question mark. Of the 289 medial punctuation marks in the manuscript, 240 are commas. Only 30 are periods.

⁴ Thus in line 2, the where following the verb invites the reader to go on without pausing, expecting an adverbial clause. But the clause is adjectival, and there is a strong caesura to be indicated. This the comma does. Similarly in line 6, the running reader may well neglect the pause which the verse and the syntax both require before & in this construction, especially since the preceding line gives us “above the smoake & stirre of this dim spot,” where we must not pause. Milton makes sure by using the comma in one instance and not in the other. The comma in line 15 is similarly explained. The period in line 18 after the short sentence: “but to my taske.”, like the commas we have noted, is demanded both by the verse and by the sense. It is, in the first place, a sentence (in everything except the explicit possession of a subject and predicate, which absence makes it the more necessary that it be indicated as such). And certainly it is the heaviest medial stop in the verse so far, the strongest cæsura, and the most complete break in the thought. The period so indicates.

⁵ In the whole poem only 24 punctuation marks present in the manuscript are omitted in the text of 1645: two periods (492, 538), the second of which is present in 1637; 22 commas (46, 50, 58, 118, 184 (two), 219, 312, 313, 418, 474, 485, 489, 511, 514, 542, 586, 616, 620, 708, 710, 890, 934). Of these, 4 (219, 620, 708, 934) are present in 1637. Three others are reinstated in 1645 after omission in 1637 (42, 661, 704). One exclamation point (510) omitted in 1637 reappears in 1645.

Twenty-four punctuation marks in the manuscript are replaced in the text of 1645 by lighter stops: 11 periods become commas (36, 485 (two), 512, 517, 561, 583, 600, 605, 652, 661); 4 periods become semicolons (118, 719, 746, 818); 4 periods become colons (148, 325, 462, 667); 1 question mark becomes a comma (491); 3 semicolons become commas (184, 225, 503); and one colon becomes a semicolon (1010). In all but two of these instances the editions of 1637 and 1645 are in agreement: in line 462 the period which becomes a colon in 1645 appears as a semicolon in 1637, and in line 661 the period which becomes a comma in 1645 remained a period in 1637. In line 164 a period which became a semicolon in 1637 is restored in 1645 and in line 374 a manuscript colon becomes a comma in 1637 but is replaced by a period in 1645.

Twenty-four stops in the manuscript are replaced by stronger stops in 1645. Four commas become colons (82, 251, 301, 646, of which that in 301 appears as a semicolon in 1637 and that in 646 as a comma). Eight commas become semicolons (130, 481, 585, 643, 658, 660, 701, 748, of which those in lines 481 and 643 remained commas in 1637). A comma in line 841 appears as a semi-colon in 1637 but is restored in 1645. Four commas become question marks (51, 496, 665, 680, of which two—496, 665—remained commas in 1637). Four commas become periods (176, 279, 300, 329, that in 300 remaining a comma in 1637). One comma becomes an exclamation point (582), one semicolon (150) a period, in both editions. One colon (374) appears as a comma in 1637 but becomes a period in 1645.

In the manuscript itself there are some 289 medial punctuation marks, to which, with the omissions noted, the text of 1645 adds 158. It also adds to the 30 terminal punctuation marks in the manuscript (of which 21 are periods) 545 marks, of which 88 are periods, 23 question marks, 10 colons, 21 semicolons, 4 exclamation points, and 399 commas. One hundred and thirty of the commas and 2 of the periods do not appear in the 1637 edition, in which there are about a dozen terminal punctuation marks which are not reproduced in 1645.

A study of these figures will show that there is a fairly consistent progression from the manuscript to the edition of 1645 toward fuller and heavier stopping. While there is a good deal of addition, there is little omission. Where stops occur in both editions, but are not alike, the heavier stop is usually in that of 1645.

My references are to the line numbers of the Columbia Milton.

⁶ Shakespearean Punctuation, p. 18.

⁷ In the manuscript 485 reads: “heav'n keepe my sister. agen, agen & neere.” Verity prints, “Heaven keep my sister! Again, again, and near!”

⁸ The manuscript omits all terminal punctuation in this passage, and the following marks besides: 420 after that, 427—no punctuation in the manuscript, but there is a variant reading here, the manuscript reading “yea even where …,” 430 after pride. The manuscript also has a comma after which, 418, omitted in 1645. Verity punctuates the passage as follows: see Comus, edited by A. W. Verity (Cambridge, 1927):

Eld. Bro. I mean that too, but yet a hidden strength,

Which, if Heaven gave it, may be termed her own.

'Tis chastity, my brother, chastity:

She that has that is clad in complete steel,

And, like a quivered nymph with arrows keen,

May trace huge forests, and unharboured heaths,

Infamous hills, and sandy perilous wilds;

Where, through the sacred rays of chastity,

No savage fierce, bandite, or mountaineer,

Will dare to soil her virgin purity:

Yea, there where very desolation dwells,

By grots and caverns shagged with horrid shades,

She may pass on with unblenched majesty,

Be it not done in pride, or in presumption.

Some say no evil thing that walks by night,

In fog or fire, by lake or moorish fen,

Blue meagre hag, or stubborn unlaid ghost,

That breaks his magic chains at curfew time,

No goblin, or swart faery of the mine,

Hath hurtful power o'er true virginity.

⁹ 6, 15, 42, 55, 58, 60, 109, 118, 124, 143, 171.

¹⁰ See also lines 60, 219, 261, 474, 620, 673, 708, 710, 934, 974.

¹¹ I have examined 60 adjective clauses which occur in the first 500 lines of Comus. Thirty-three of them (5–6, 9, 12, 13–14, 21–22, 30, 54, 84–85, 93, 111, 130, 196–197, 207, 229, 232, 236–237, 270–271, 293–294, 298–299, 330–331, 367, 380, 382, 404, 414–415, 418–419, 420, 431, 433–434, 447, 450, 457, 472) are restrictive modifiers, in my judgment, and the other 27 non-restrictive (2, 23–25, 37–39, 46–47, 50–52, 56–58, 59, 65–66, 84–86, 112–113, 126–127, 129–130, 158–159, 209–210, 216–217, 252–255, 264–265, 319–320, 321–322, 324, 333–334, 375–377, 378–379, 441–442, 447–448, 478–479, 493–495).

¹² Of the 27 instances cited, thirteen of the clauses begin in the middle of a line. Of these, 10 (2, 46, 51, 66, 129, 210, 319, 324, 333, 493) are preceded by commas in the manuscript. The 1645 text supplies commas for two of them, but omits the comma after Bacchus (46) which we have already discussed. For example, line 216

²¹⁶ . that he the supreme good to' whome all things ill

²¹⁷ . are but the slavish officers of vengeance

is entirely unpunctuated in the manuscript, but the editions of 1637 and 1645 add commas after he and good, setting off the appositive. We should not be able to connect this with the adjective clause if the 1645 edition did not also add a comma at the end of 217, where the clause ends. The other clauses, which begin with the beginnings of the lines in which they occur, are preceded by no punctuation in the manuscript. Three of them, we have seen, remain unpunctuated throughout. Four more (254, 375, 441, 478) are without the preceding commas in 1637 but not in 1645.

¹³ Of 33 restrictive clauses, none are preceded by commas in the manuscript, 3 in 1637 and 1645 (5, 84, 207) and 2 (196, 433) only in 1645. Four of these 5 are terminal commas, which would not normally be in the manuscript. The fifth is medial.

²⁰⁴ . … a thousand fantasies

²⁰⁵ . begin to throng into my memorie

²⁰⁶ . of calling shaps, & beckning shadows dire

²⁰⁷ . and ayrie toungs that syllable mens nams

²⁰⁸ . on sands, & shoars, & desert wildernesses.

Here the punctuation is the same in all three versions except for commas added in the two editions after dire (206) and after toungs (207), before the restrictive clause that syllable mens nams. The comma after dire is clearly necessary in the series, and is terminal. That after toungs marks a strong cæsura, which, ignored, would change the character of the line and make it fit much less well in the context of verses much broken by sense pauses. (Lines 431–435 contain a restrictive clause similarly punctuated for similar reasons.) The commas at the ends of lines 5, 84, and 196 precede clauses technically restrictive but which at the same time have the force of the half-parenthetic non-restrictive clause. All the rest of the restrictive adjective clauses in the first 500 lines of the poem remain unpunctuated.

¹⁴ Line 573 illustrates one of the most interesting of Milton's practices with regard to elocutionary punctuation—the punctuation of appositives. More of them are set off in the printed text than in the manuscript, where Milton uses commas only if there is a strong possibility of confusion otherwise, but even in print he was apparently governed only by the need for pauses, not by the need to indicate the grammatical structure. In the manuscript, only 3 (of 25)—212, 240, 521—are set off by commas. The 1645 edition sets off 14 (68–69, 212, 213, 216, 229, 240, 261–262, 459–460, 520–521, 563, 693, 825, 829, 1009–1010) and leaves 11 unpunctuated (50–51, 128–130, 349, 376, 440–441, 447, 573, 737, 821–822, 827, 1003).

¹⁵ The 1637 edition omits the commas at the ends of 78 and 80. The manuscript omits all terminal punctuation and has a comma instead of a colon after do in 82.

¹⁶ J. W. Mackail, in The British Academy Warton Lecture on English Poetry, xv, Bentley's Milton, reprinted for the British Academy by Humphrey Milford, Oxford, on page 19 lists several errors of interpretation into which Bentley was led by the bad punctuation of the text with which he worked.