University of Virginia Library

III
A Directional Study of the Variants

To resolve these ambiguities, we must use other kinds of evidence. The dates on the manuscripts put all of them before 1646, and S is the only one with a definite date, September 1637, one month after the poem was written. But even there Cranfield's note indicates that manuscripts were circulating "about London" in September of that year. Furthermore S itself has the signs of a corrupt version; thus an early date means little. As is commonly known in manuscript study, a late manuscript can easily have been copied from an authoritative early one, whereas an early manuscript can be very corrupt, with many intermediaries between it and the archetype. A similar difficulty comes with study of the accidentals of the witnesses. We can seldom be sure that a scribe did or did not use a holograph manuscript as his copy, for scribes stamp a manuscript with their own habits of spelling and punctuation even more distinctly than compositors of printed books. And again, we can not be sure that another manuscript did not intervene between the scribal copy and the archetype. For simplicity we assume that one has not. Although Suckling's habits of spelling are known from his autograph letters, distinctive authorial spellings do not bleed through the work of these scribes. In the case of the printed edition of 1646, a kind of negative proof has been made, above, to show that authorial papers were not used as copy for the poem.

One other kind of evidence remains: the inherent nature of the variants themselves, the directional variants. We are looking for signs of originality and derivation so as to decide which witnesses stand closer to the archetype than the others. If original readings occur in more than one manuscript, they must descend independently from the archetype, but if original readings appear in only one witness, the descent of the manuscripts is radically altered (as in diagrams D, E, and F). The signs of an original reading or directional change, listed by Dearing, vary considerably in their reliability. In the discussion above, S has been shown to be more sophisticated, simplified, and smoothed out than the others: this is one of the best kinds of proof of a derivative text. A misreading that can go only one way is another strong proof. Also the reading that makes sense rather than nonsense is likely to be more original, but this kind is often reversible, for a scribe may make sense out of a corruption in his archetype. On the other hand, a more difficult reading is not necessarily the more original; nor is the "better" reading or more poetic reading necessarily more original.

A number of directional variants establish M and Ha as most probably closer to the archetype than any of the other witnesses. The strongest case can be made for the format of the stanzas, in conjunction with significant omissions that would not have arisen unless the archetype were arranged in the pattern of the Malone and the Harvard manuscripts. These manuscripts separate the four-line stanzas by spaces, and they link pairs of stanzas by a system of monosyllabic words suspended between them. S and C group

the stanzas in eight lines, C indenting the last four lines of each octave, while neither uses the suspended words between. At the other extreme, E uses four-line stanzas without suspended words, and does not pair any of the stanzas. In placing the monosyllables, Hn and 46 seven times agree with MHa and the other eight times with E; but 46 and Hn do not always suspend the monosyllable between the same pairs of stanzas. In other words, 46 suspends in the middle of octaves 1, 2, 3, 4, 5, 6, and 13, while Ha suspends in octaves 2, 3, 4, 5, 6, 7, and 8. The sense of the poem clearly falls into eight-line units, so that either SC or MHa is likely to have been in the author's manuscript. If we assume that Suckling was careless with such details and wrote his original as we see E, it does not seem probable that a scribe would invent such an unusual scheme as in MHa, and we cannot easily explain the regular presence of the linking words at the beginning of the fifth line of each octave.

Next assume that Suckling started to write the verses with the suspended words but he dropped the habit about halfway through where Hn or 46 leave off. It is possible that a scribe who liked consistency would have finished out the scheme, to produce the format of M and Ha. This possibility is demonstrated by the mistaken addition of a transitional word between lines 56 and 57 in Ha. However the facts contradict such an hypothesis, for at line 53 and 107 all the texts except 46 retain the transitional word, in either the suspended position or as the first word of a line. Why should such omissions occur in 46 at these points, unless an ancestor of 46 had the words in their suspended positions and they were overlooked by the copyist? A similar omission arose independently in 46, M, and Hn at line 93, where only two lines of a stanza survive. C and E read Then as the first word of the line, S reads Next, and Ha suspends Then between the stanzas. The distributional study showed that M is not closely related to 46Hn, nor are C and E and Ha closely related. Therefore the archetype must have read Then and the word must have been suspended between the stanzas, or else it could not have been so easily overlooked by M, Hn, and 46.

The assumption that SC have the original format is destroyed by the same kind of evidence, for both of them omit And at line 21, and C omits But at line 29. By elimination, the only hypothesis which fits the evidence is that M and Ha have the original format. Therefore, we must conclude that the repeated copying of the poem caused the gradual omission or rearrangement of the suspended words, a theory which is nicely confirmed by the way the compositor of Merry Drollery, 1661, dropped out or realigned suspensions which he found in his copy of 1648, making the format of the stanzas like that in E.

The title contains another directional variant, for "The Witts" is probably the original title of the poem that was sung before the King in New Forest, in August 1637. George Garrard, in a letter to Thomas Wentworth,

called it "a Ballad made of the Wits,"[8] and Lionel Cranfield wrote on the back of his copy (S), "Of som Poetts / Of som Wittes." Only 46 has "A Sessions of the Poets"; ESC have no title; and M, Ha, and Hn have "The Witts." The word sessions probably crept into 46's title from the first line, "A Sessions was held the other day," although it does seem peculiar that 46, along with S, reads Session in the first line and Sessions in the title.

Other directional changes depend upon the plain sense of some readings over others. Thus 46 appears to be corrupt in

instead of the majority reading Hn, E, and 46 are probably corrupt at line 53, because "Sullenly" is a mood appropriate to the disappointed poet, and "Suddenly" was probably introduced because the scribe saw that Will Barkeley had smiled a few lines before. C changed Sullenly to Silently, possibly for the same reason.

A peculiar substantive change at line 88 may have occurred because of a misreading.

M follows neither, with the word silene. Silent may be the original reading because it involves more complicated grammar; but did M mean to put a t instead of an e or did his copy read silence and he left out the c? It cannot be determined either way. The probabilities favor silent as the harder reading.

A similar puzzle is at line 29.

Reading (a) is most likely original because M and Hn have been shown to be independent witnesses (in the distributional study above), and it seems unlikely that two scribes would have added the same word soe at the same point, but it is possible for Ha46E to have independently dropped the word in order to improve the meter. And (c) is derivative because it appears to be another attempt to make iambic meter of the basically anapestic rhythm.[10]

A few errors can be explained as transpositions or slips of the eye over similar words in nearby text.

The second of the readings at line 25 may be original because the word there can be seen in exactly the same place in the line beginning the following stanza. Here But has been retained in the eye of the compositor of 46, substituted for and in the second line, as it reads in Σ. At line 42, the scribe of Ha wrote "in travelling France" rather than "travelling in France" (Σ). Similar errors are: In the last example, (a) is probably derivative because one and attracted the other. It is probable that (c) is more original than (b) because of its occurrence in independent witnesses MHaSC. Thus (b) is a transposition. However, this argument hangs on the assumption that MHaSC do not derive from the same faulty ancestor. Considering this variant in isolation. we could just as reasonably conclude that the ancestor of HnE46 read as in (b) and had the original reading. But in the light of all the previous evidence, (c) is more probably original.

In the examples of directional variation, M and Ha most frequently have original readings, and neither SC alone nor 46HnE alone have original readings; only when they agree with M or Ha do they contain such variants. An editor puts special weight on the variants of format of the poem, because there the argument for original readings is irreversible. Therefore M and Ha stand closer to the archetype, and we can resolve some of the ambiguities of the distributional formula. Trees (D), (E), and (F) are out of the question, for no manuscript has a significant number of original readings among its type 1 variants. Trees (B) and (C) are not likely to be right because M and Ha do not have a prominent enough position in them, and CS, which have no original readings alone, are given undue weight. Therefore tree (A) is the best.

Such a tree surely relegates the printed text (46) to an inferior position; but that does not solve all the preliminary problems of an editor, because it is still uncertain which manuscript, M or Ha, has the shorter line of transmission. In other words, which manuscript should be the copy-text for a critical edition? This is a difficult choice, for both seem equally distant from the archetype. Greg has shown that the copy-text is most important for the texture of its accidentals, and an editor should choose the one that offers the best chance of bearing some of the details of the author's manuscript. In our case, neither M nor Ha shows any significant signs of authorial copy immediately behind the scribal transcription. The scribe of M is amazingly uniform in details of spelling and punctuation whether he is copying Suckling or Waller or Godolphin's poems, and the text of "The Witts" shows no deviation from his established patterns. Ha is a little more flexible in his practice, but when he does deviate from his habits while transcribing "The Witts," he moves farther than ever from Suckling's spelling. Suckling writes doe, goe, so, no; whereas the scribe of Ha writes doe, goe, soe and noe more frequently in this poem than elsewhere. M seems to have more examples of Suckling's favorite past participate ending -nd, -vd, -rd, -md, -sd, and more examples of parenthesis in exclamations, direct address, and asides; however the scribe does the same thing in the rest of M. Both M and Ha have a small number of type 1 variants and both appear many times among the "shared" readings on the distributional chart. Thus it seems like a toss-up between the two. The only substantial difference between them is that M appears to be more of a professional job, neater, less prone to literal errors, more regularly punctuated, and generally cleaned up; while Ha is more naive, prone to simple omissions and misreadings, messy, and unprofessional. An editor of a critical text would surely want to favor readings from Ha, if for no other reason than the obviousness of its errors. As it has been often observed, all other things being equal, a messy and careless copy is less likely to have deliberate improvements by a scribe than a highly professional transcript. But the final decision rests with the accidentals: which of the two agrees more with Suckling's known habits? Again Ha has an advantage, by a ratio of forty-eight to twenty-one; therefore Ha is the most suitable basis for an edition.

Before the final editing, we must ask about the mixed text of the poem in Fragmenta Aurea 1648. Whence did its twenty-two corrections come? None of the six type 1 variants appear to be original, and some are clearly corrupt; for instance, hyde bound for hard bound (35) and what made for how came (57). Once it agrees with Ha alone: betwixt for 'twixt (72), an insignificant expansion. Once it agrees with HnE alone, cheared for cleared (115); and twice with S, himself scarce for himself hardly (56) and to any for any (118). The other twelve variants correct type 1 readings in 46, including the stanza omitted in 46, and they bring the text into general agreement with the manuscripts. Consequently the corrections may have

come from a comparison of 46 with a manuscript closely related to S, itself possibly conflated with HnE. Another possibility is that 46 was compared against two manuscripts, one similar to S and one to HnE. There is not enough evidence for either case. At any rate, the text of 48 is considerably better than 46, but one of the ironies of textual transmission is that 48 was reprinted only in the surreptitious edition of 1672(?) and in Merry Drollery 1661. Whereas the corruptions of 46 were perpetuated in Suckling's works of 1658, 1694-96, 1709, 1766, 1770, and 1836. Nor were the mistakes rectified by W. C. Hazlitt's editions of 1874 and 1892, nor in A. H. Thompson's "standard" edition of 1910, which whimsically followed 1646 for a while, then 1648, and once picked up the erroneous reading hide bound (35) from 1648. Hazlitt and Thompson did not even notice that 48 supplied the missing stanza. The only modern edition which contains a substantial number of the corrections of 48 is that by R. G. Howarth in Minor Poets of the Seventeenth Century (1931). But since Professor Howarth was unfamiliar with the manuscripts, he could not know the importance of seven variants in 48, nor could he know of the twenty-six corruptions in both 46 and 48.

In the little edition below, I have followed Ha except where a possibly better reading could be found in other witnesses (if I could explain how Ha fell in error). I have silently modernized u, v, i, j, and long s; expanded abbreviations, italicized proper names, and supplied capital letters at the beginning of lines. In the notes I have recorded only the substantive and semi-substantive variants from sources that could descend directly from the archetype, readings in Ha, M, the ancestor of CS, and the ancestor of 46EHn.[11] Clearly derivative readings from witnesses placed lower on the tree are omitted; thus the reader will not find type 1 variants from C, S, 46, 48, Hn, and E, because such a reading was apparently introduced by a copyist. And although I have altered the accidentals at several points, I have left a complete record of these changes for a full-dress critical edition.