 | |  |
II
A Distributional Study of the Variants
Texts N and 61 can be eliminated from the analysis immediately as derivative versions which reproduce readings that are peculiar to their extant ancestors, Hn and 48 respectively. In the same fashion we can ignore all the other descendants of 46 and 48, as mere reprints, each of which successively introduces more corruption. And 48 itself is a derivative of 46 even though 48 contains certain new readings which cannot be the work of a mere compositor. The corrections and alterations in 48 have come from some other manuscript traditions, separate from the MS. behind 46. Consequently 48 must be treated as a contaminated or mixed text, part derivative and part substantive. Since it is not strictly a collateral text, descending from an ancestor common to the other witnesses, we must for the moment set it aside.
This leaves M, Ha, S, C, Hn, E, and 46 for distributional study. Below is a table of substantive variants. The terms type 1 and type 2 have come from Greg: type 1 is the occurrence of a reading in one text against all the others; type 2 is the agreement of at least two texts against at least two of the others. Among the type 2 variants I have not tabulated elisions, minor omissions, expansions, punctuation, or spellings of proper names, because they could easily have arisen independently.
| Text | type 1 | type 2 with one |
type 2 with two |
shared with one MS in minority |
no. of times in majority |
|||||||
| M | 17 | Ha | 1 | HnC | 1 | Ha | 2 | 19 | ||||
| HaS | 1 | S | 2 | |||||||||
| SC | 1 | C | 2 | |||||||||
| Hn | 1 | |||||||||||
| Ha | 17 | M | 1 | S | 1 | 46E | 1 | M | 1 | E | 2 | 17 |
| Hn | 1 | CHn | 1 | C | 2 | Hn | 2 | |||||
| E | 1 | SC | 1 | S | 2 | 46 | 1 | |||||
| Hn | 20 | E | 4 | E46 | 5 | E | 9 | M | 1 | 10 | ||
| Ha | 1 | CHa | 1 | 46 | 6 | S | 1 | |||||
| S | 1 | MC | 1 | Ha | 2 | |||||||
| 46 | 1 | C | 2 | |||||||||
| E | 19 | Hn | 4 | Hn46 | 5 | Hn | 9 | Ha | 2 | 11 | ||
| Ha | 1 | Ha46 | 1 | 46 | 6 | |||||||
| 46 | 30 | S | 1 | HnE | 5 | E | 6 | Ha | 1 | 16 | ||
| Hn | 1 | HaE | 1 | Hn | 6 | S | 1 | |||||
| S | 70 | C | 5 | MHa | 1 | C | 5 | 46 | 1 | 12 | ||
| Hn | 1 | MC | 1 | M | 2 | Ha | 1 | |||||
| 46 | 1 | Hn | 1 | |||||||||
| C | 29 | S | 5 | MS | 1 | S | 6 | Ha | 1 | 13 | ||
| HnM | 1 | Hn | 2 | |||||||||
| HnHa | 1 | M | 2 | |||||||||
| Totals | 181 | Total type | 2 | 23 |
This kind of statistical study, of course, does not offer ironclad proof, but it is helpful as a preliminary analysis, strictly objective, pointing to some probable conclusions. For instance, the first column of the chart shows that all the texts are terminal, as we should expect, since the derivative ones have been eliminated. The Sackville MS. has an unusual number of type 1 variants, so many as to suggest that it is another version (i.e. an early draft or later revision) or a very corrupt transcription. But an inspection of S's type 2 variants immediately cancels out the former hypothesis, because S is not likely to be another version in another manuscript tradition and at the same time to have type 2 variants that distribute along with the other six texts. How could it be closely associated with C, apparently not another version, and still be a revision or early draft, unless we suppose that the author wrote his revisions on an already derivative manuscript and S is a descendant of it? Therefore, I assume that S is collateral or virtually so with the other texts; and if that is true, S is very corrupt. A number of individual readings in S confirm this hypothesis (if I may depart from a rigidly statistical analysis for this point), because they are regularly "improvements" of an especially obvious sort: smoothing out the verse, simplifying the grammar, and trivializing the thought.
That all Men desired to please the queene S (line 70)
And calling him presently out of the thronge S (91)
Was now to be given to him that deserved S (4)
For his things were workes, the others but plaies S (20)
Of errors continued for many an age S (22)
That concluded of merit uppon a successe. Σ
they did much digress
From truth, that judg'd things by the success. S (51-52)
Must carie the Bayes: At which Ben turn'd about S (27)
Modestly hop't that his handsome muse S (42)
Consider'd he was well hee had a Sewers place S (40)
The type 2 variants in the chart imply that certain manuscripts are
probably related more closely than others: SC against the others, EHn
against the others, and 46EHn against the others. Neither M nor Ha is
closely identified with any single manuscript or group of manuscripts, if we
assume that four or more exclusive readings suggest a close relationship.
If this seems like a small number, we must remember that only twenty-three
type 2 variants are suitable for the analysis. The only variants considered
in this class are those that could not have easily arisen independently. Of
course some anomailes exist, readings that are common to SHn, or HaE in
one instance alone. These are to be expected, particularly with poems which
were memorized, set to music, and passed around in London society.
Consequently the lines of descent of a given manuscript may be
considerably more tangled than we suppose; however there simply is not
enough statistical evidence for a hundred line poem such
as this on which to construct a more complicated hypothesis. Therefore a
critic must make whatever distributional hypothesis is warranted by the
majority of the evidence. The formula SC: M: Ha[46(HnE)] best expresses
the groupings of type 2 variants. Diagramatically, the formula looks like
this:
Several of the complex variants confirm this formula. For instance, these patterns occur:
- SC:MHa:46Hn:E (grouping of stanzas and location of transitions)
- SCM:Ha46E:Hn (placing lines 93-94)
- C:SMHa:Hn46E (line 53)
But the formula is ambiguous, for if the texts are truly collateral the
diagrams could be any of the following:
| | |