|
Article Excerpt
I
A LOVER'S COMPLAINT was first published in 1609 at the end of Thomas Thorpe's famous quarto of Shakespeare's Sonnets. Until the early 1960s this narrative poem of 329 iambic pentameter lines had been neglected by Shakespeareans, who tended to reject it as spurious or disdain it as an unsatisfactory product of Shakespeare's youth. Then Kenneth Muir and MacD. P. Jackson independently argued in favor of the poem's authenticity and a seventeenth-century composition date. (1) Most editors have accepted their case, which has been supplemented by other scholars. (2)
Muir and Jackson overlooked one signficant point. Samuel Daniel's sonnet sequence Delia (1592), an obvious influence on Shakespeare, had closed with The Complaint of Rosamund, and before Shakespeare's Sonnets appeared a convention of completing a book of sonnets with a long poem had been firmly established. Scholars such as John Kerrigan and Katherine Duncan-Jones have regarded A Lover's Complaint as an integral part of Thorpe's volume. (3) Recent commentators have sought to show not only that the quarto preserves Shakespeare's own arrangement of his sonnets, but that he intended A Lover's Complaint to be the third movement in a sonata-like structure preceded by the sections devoted to Fair Friend and Dark Lady, and helping to resolve their contradictions, or at least to put the experience embodied in them into a new perspective. (4) The status of A Lover's Complaint--whether it is Shakespearean or non-Shakespearean--is thus directly related to the question of the authority of the 1609 quarto's text and the order in which sonnets are presented and numbered. A spurious A Lover's Complaint would undermine trust in Thorpe's volume; a Shakespearean A Lover's Complaint tends to authenticate it.
However, even as criticism has been discovering a rationale for the authorial inclusion of A Lover's Complaint within Shakespeare's Sonnets, research by the Claremont-McKenna Shakespeare Clinic, run by Ward E. Y. Elliott and Robert J. Valenza, has been casting fresh doubts on Shakespeare's responsibility for the poem. (5) Elliott and Valenza evolved a variety of tests for Shakespearean authorship, starting with his undisputed plays and establishing, for each of the linguistic phenomena counted, a range within which rates of occurrence for any authentic play should fall. Works generally accepted as wholly Shakespeare's exhibited Shakespearean rates on all but a very few tests, whereas suspect, collaborative, and apocryphal works failed large numbers of tests. Elliott and Valenza found that fourteen of the tests thus validated on plays were also usable on poems, and that when the poems (Venus and Adonis, The Rape of Lucrece, and the Sonnets) were broken into approximately 3000-word blocks, none of the fourteen undisputed Shakespearean blocks failed more than two tests and most failed none, but that A Lover's Complaint failed six. The same Claremont-McKenna methodology that emphatically, and rightly, rejected Donald Foster's ascription of A Funeral Elegy to Shakespeare also questions the authenticity of A Lover's Complaint. (6)
Elliott and Valenza are sensibly cautious about their findings, conceding that the new methods they and other computer users have adopted are so experimental that "it is foolish to expect any of them to be the last word on the subject at this stage." (7) It is a pity that none of the six "blocks" of sonnets tested comprises the last twenty-five or so of those to the Fair Friend: from the number of iambic pentameter lines given for each block, it appears that Block 4 consists of Sonnets 85-112 and Block 5 of Sonnets 113-140, so that the sonnets that can most plausibly be assigned to the seventeenth century and were thus most likely to have been roughly contemporary with A Lover's Complaint (approximately numbers 100-126) are mixed with much earlier ones. (8) However, even Shakespeare's last-written group of sonnets might display a rather different profile from a 329-line complaint written in rhyme royal. A major difficulty is that, whereas the profile for Shakespeare as dramatist rests on a wide range of plays, that for Shakespeare as poet rests merely on two early narrative poems and the Sonnets: these may not provide firm enough grounds for inferring Shakespeare's likely practices in a complaint narrative composed around 1604-6, even though, in setting Shakespearean parameters, Elliott and Valenza also take usages in the plays into account.
One test that A Lover's Complaint fails concerns the ratio of "no" to "no" plus "not," which is lower than in any undoubted Shakespeare work. Yet it is also lower than in any dramatic work by anybody else listed; and of eighty-one blocks, of approximately three thousand words each, from non-Shakespearean poems only the Funeral Elegy and one block from Heywood's Oenone (1605) are as low or lower. (9) The figure of 120 for A Lover's Complaint is thus anomalous for any author, and within-author variation may evidently be considerable: the six Heywood samples range from 111 to 667. Content would seem to be the most significant influence in such ratios. Moreover, there is a certain arbitrariness about the piecemeal creation of tests of this kind. The ratio of "no" to "no" plus "not" is, of course, a function of the tallies for each individual word, and two more instances of "no" would have lifted the ratio for A Lover's Complaint into the "Shakespearean" range. A safer way of dealing with such function words may be to investigate all those that occur in Shakespeare's works above a certain level of frequency and compare samples by principal component analysis.
The conversion of low frequencies into rates can make the differences between samples seem larger than in fact they are. A second test failed by A Lover's Complaint is for rates of "with" as the penultimate word of a sentence, where a sentence is defined as a sequence of words ending in a full stop, question mark, exclamation mark, colon, or semicolon. Although Elliott and Valenza relied on edited texts, we may grant that there is likely to be some consistency in the punctuation of a single modern Shakespeare edition, and that this punctuation will reflect authorial sentence-structure. The rates for "with" as the penultimate word of a sentence are presented as ranging from 6 to 34 in fourteen approximately 3000-word blocks from Venus and Adonis, The Rape of Lucrece, and the Sonnets, and as being for A Lover's Complaint. But the figure 6 in each of the first two Venus and Adonis blocks and the second Lucrece block represents only a single instance, as does the figure 7 in the third, fourth, and fifth blocks from the Sonnets. The raw figures for the fourteen blocks work out at 1, 1, 5, 2, 1, 3, 3, 3, 2, 3, 1, 1, 1, 2. This yields an average of just over 2, with an estimated standard deviation for the whole population of 1.2. There is a better than one in twelve chance that a 3000-word sample from a genuine Shakespearean poem will have no examples of "with" at the end of a sentence--and A Lover's Complaint happens to consist of only 2579 words.
A Lover's Complaint also fails two "Thisted-Efron" tests. These, as Elliott and Valenza remark, are "novel, fancy, high-tech tests" devised by statisticians Ronald Thisted and Bradley Efron, and difficult to explain in "nontechnical terms." (10) But in essence they concern the rates of use of words unique or rare in the core Shakespeare canon. A Lover's Complaint has "too many." Here I think we have a case where mechanical counting is less informative than hands-on analysis of the poem's vocabulary, with an eye to the kinds of rare words it contains and to Shakespeare's habits of word formation. (11) A Lover's Complaint reads like an experiment in the adaptation of Shakespeare's most condensed mature style to nondramatic poetry, and, if my dating is correct, it was composed at about the same time as the most lexically rich and inventive of his plays. His early narrative poems use "once-used words" at much higher rates than his plays of the same date. (12) We might expect A Lover's Complaint, packed with images and associations, to be the boldest of all Shakespeare's works in its contributions to his vocabulary.
The other two tests failed by A Lover's Complaint are metrical. It has too few "leaning microphrases" of both the "proclitic" and "enclitic" types, as these are defined by Marina Tarlinskaja, who has recently studied the prosody of A Lover's Complaint in detail and found it to be at variance with Shakespeare's at any phase of his career. (13) The issues raised are too complex to be treated here. Tarlinskaja is an expert analyst of meter, and her previous work has cast light on problems of authorship and dating. Yet the technical anomalies identified in the verse of A Lover's Complaint do not prevent it from sounding Shakespearean when read aloud--at least to my ear. Tarlinskaja's data for plays do not distinguish between blank verse and rhymed lines: the 285 rhymed lines in Ali's Well That Ends Well might conceivably yield figures closer to those for A Lover's Complaint.
Elliott and Valenza distinguish between "green-light" testing, in which the investigator finds "quirks" that link a disputed work with the Shakespeare canon and then seeks to establish the genuine rarity of these quirks within non-Shakespearean controls, and their own "red-light" testing, which begins by setting Shakespearean boundaries outside which, for any one test, rates for at least 10 percent of non-Shakespearean controls fall, and goes on to demonstrate that the non-Shakespearean works flunk significantly more tests than any of the canonical ones. And they judge "red-light" testing to be superior: "with less-than-perfect identifiers, strong exclusionary evidence normally trumps strong inclusionary evidence." (14) Elliott and Valenza are right in claiming that their "red-light" procedures more effectively safeguard against "false positives"--against the identification as Shakespeare's of material that is not in fact his. But those procedures may be more apt to suffer from the obverse flaw of sometimes registering "false negatives."
Ideally, from a statistical point of view, Elliott and Valenza would have divided the noncontentious Shakespearean...
|
