Roger Stritmatter

In a Winter 2022 Shakespeare Oxford Fellowship Newsletter article, “Who Wrote George Peele’s “Only Extant Letter,” Robert Prechter conducts an analysis claiming to establish that a 1595 letter sent to William Cecil, describing a literary work written by Peele and signed by him, was ghostwritten — for somewhat obscure reasons — by the Seventeenth Earl of Oxford.

Having studied forensic handwriting analysis for many years, and both blogged on or written journal articles about the topic, I went about too see if I could replicate Prechter’s methods in order to verify his conclusions.

In short, I could not.

For larger context, Prechter’s claim for Oxford’s ghostwriting Peele’s letter is part and parcel with his extravagant “discovery,” comprehensively documented in his multi-volume Oxford’s Voices, that Oxford wrote works wrongly attributed to many other Elizabethan writers, among them Thomas Nashe, John Lyly, Robert Greene, and Peele.

It should be acknowledged that the earl of Oxford may very well be responsible for literary works surviving under names other than Shakespeare. In fact, having focused my own study on Oxford’s lyric and narrative poems, I’m pretty sure he is. For example. during the 1580s and 1590s (mostly) he wrote poems under the two allonyms “Shepherd Tony” and “Ignoto.” I also agree with Elizabeth Appleton that Oxford wrote pamphlets under the pseudonyms “Pasquill Cavaliero of England” and “Cutbert Curry-Knave.” And there are probably others, as well. . .

The essential question Prechter ignores is this: granting that there are or may be such works, by what means do we determine who wrote which of them, especially when there is significant competing evidence? In other words, what counts as evidence? What’s the other side of the argument? How rigorous is the basis of our own claims?

In part the answer depends on what we mean by “he wrote them.” In literary studies, proofs are linguistic, and when we have “enough” of them (however we define that) we have identified an author, but not necessarily the writer of the document, who could be an amanuensis or copyist.

In attributing the writership of a document according to the practices of the forensic document examiner, the proofs consist of the shapes and orientations of characters and lines. If we can show that they are “the same,” then we may have identified the writer.

This is the first of a two part reply analyzing the methods of forensic handwriting analysis used by Mr. Prechter in his SOF Newsletter article. For the most part, the wording here reprints my critique as published in the SOF Newsletter (Vol. 58, No. 4: Fall 2022). In a subsequent post I will consider Prechter’s reply to this critique and further document the irresponsible character of his forensic methods.

In this article, in pursuit of a more nuanced and reliable interpretation of the evidence, we will review some principles and best practices for forensic handwriting analysis – that branch of handwriting analysis that is explicitly concerned with answering the question of whether any two samples of handwriting are by the same or different writers. [1]

As a preliminary, it should be noted that, after extensive inquiry and debate, forensic handwriting analysis is recognized as a science by American courts, in part because studies show that trained analysts make only 1/6 as many errors as amateurs without training (Srihari 2002). Here, accordingly are five useful principles that will aid even the untrained analyst in evaluating Prechter’s claims to arrive at a more reliable conclusion.

Principle #1: Fit the Quality of the illustrations used in the argument to the seriousness of the inquiry.

The first problem of Prechter’s treatment is that the quality of his reproductions is often insufficient to support his sweeping generalizations. Accurate forensic analysis requires access to original documents if possible, and if not possible, documents reproduced with sufficient resolution and clarity to draw conclusions based on observable analysis.  A forensic argument unsupported by visible evidence is usually an argument from authority, either explicitly or covertly so.

In keeping with this principle, here are two reasonably high-resolution images of the samples of Peele and Oxford reproduced in Greg’s English Literary Autographs (1925). The reader may already observe several generic discrepancies between Oxford’s hand and Peele’s.

Figure 1: Samples of Oxford and Peele handwriting from Greg‘s (1925) English Literary Autographs.

Oxford: Highly regular, constrained, conservative hand with broad spacing and minimal ornamentation or flourish.

 Peele: Large Flourishes often extending far below the baseline (e.g. k, g) or in reverse (counterclockwise) arcades (e.g. d), and some letters (e.g. s) showing strong cursive elements.

Even at this level of comparison, Peele’s hand appears to be more expansive and flourished and Oxford’s more conservative. However, these are very general terms of comparison and are not sufficiently particular in their descriptions to furnish anything like a proof.

Principle #2: Employ multiple exemplars of letters for comparison.

Samples from all writers exhibit to a greater or lesser degree the feature known as natural variation – a term of art in forensic study that describes how particular letters and combinations of letters (exemplars) deviate from their statistical norms within the known population of that writer’s production. Showing multiple exemplars is necessary to draw reliable conclusions about range of variation and determine if this variation matches or diverges in any two samples used for comparison.

A systematic approach that balances the search for common patterns with close observation of potentially relevant differences is absolutely necessary when conducting such analysis. At least half a dozen letters or combinations of letters should be extracted for comparison, with at least three exemplars of each. Above all, a positive identification cannot be made on the basis of a few characteristics that are alleged to be identical when other forms in the sample that are discrepant have been ignored.

Instead of a systematic and self-critical approach, Prechter cherry-picks a handful of letters – e.g., capital E – which he believes are sufficient to demonstrate his point that the Earl of Oxford is responsible for the letter signed “George Peele” and accepted by Greg and other authorities as by Peele, not Oxford. More dispassionately analyzed, however, the evidence copiously contradicts this conclusion.

Let’s start with the letter S, always a useful letter on account of its various types and large range of variation, both natural and systematic. Figure 2 illustrates the three main forms this letter takes in the Earl of Oxford’s surviving handwriting samples.

Taking these forms one at a time, let’s start with comparing Oxford’s medial long-s to the medial long-s forms in Peele (Figure 3).

Oxford almost always uses a classic long-s in the medial position. His use of the short-s is confined to the terminal position, where he almost always uses it. By contrast Peele consistently prefers a cursive long-s or a small-s in a medial position.

Bigrams – combinations of any pair of adjacent letters – reveal characteristic forms of connection between letters that are also subject to various individualized patterns (Figure 4).

The forms seen in this bigram illustrate dramatic differences in construction and emphases. Not merely the form of S, but the form of the connector and the styling and emphasis of the t exemplars are quite distinct and are highly unlikely to be by the same writer. Oxford’s bigram involves two complete and disconnected strokes starting with a long-s followed by an un-ligatured t. Peele begins this combination with a small-s and then joins it with a pronounced ligature to the t. It is not that Oxford never uses an st bigram with a ligature, but that he does so only rarely, while Peele does so as a matter of course.

Like the s and the form of connection, Peele’s t is also wildly different in multiple ways from Oxford’s: The letter includes in every instance a broad foot at the base, an equally broad crossbar, and a loop at the top where the ligature crosses over itself before descending to form the descender of the t.

Double-s can be even equally revealing. In Tudor Italic handwriting, this is a complex and highly variable construction that occurs in at least five major sub-types identified by handwriting analysts such as Fairbanks and Dickens (Figure 5).

Oxford consistently uses Types 3 or 4, a long & short combination, with or without a ligature.  By the 1590s, these forms were becoming archaic and are often superseded by the looped or cursive form joined to a short-s as seen in Type 5 and below (Figure 6) in two exemplars of Peele’s hand.

As in his single medial s, Peele consistently favors a looping or cursive initial s when forming double-s. This double letter is formed with only one continuous stroke, i.e., using an entirely different pattern of hand motions from Oxford’s, which are composed most commonly of three strokes – one for each letter and a third for the ligature, as indicated by the thickening of the lines where two strokes overlap (Figure 7):

Principle #3: Test your Assumptions and Prejudices

Under conditions of constraint (writing that is very small, written in the margins of books, or written under the influence of great emotion)[2] Oxford sometimes employs the reverse-loop d. Ignoring such evidence may result in confirmation bias and premature erroneous conclusions. In this case, including exemplars of these forms (Figure 8) strengthens the argument distinguishing the two samples by showing that even when Oxford does employ these forms, they are, typically, easily distinguishable from those used by Peele:

Over time the analyst will begin to observe that handwriting is composed less of discrete letters than of combinations of types of strokes that tend to replicate themselves across various letter forms. Oxford’s letter k is distinguished by a very modest termination which often meets or runs into the next letter, especially when it is an e. By contrast (Figure 9), the termination on Peel’s k forms a sweeping flourish that extends out under the successive letter(s), even in the ke combination:

This expansive motion in the formation of Peele’s k replicates the dynamic motions seen in the  sweeping backward loop of Peele’s d or the cursive form of his long-S. Such individual elements when repeated in various letters comprise the particulars that allow the generalization above in Figure 1 contrasting Peele’s more expansive hand to Oxford’s more conservative forms.

Finally, let us conclude with the letter M (Figure 10):

Capitals are often better for comparison than lower case, either because they are larger, often involve more strokes, and have a wider range of features that can exhibit structural patterns of consistency (indicating shared writership) or variation (indicating more than one writer).

Principle #4: Examine as many letters as needed to form a definitive conclusion.

If natural variation is variety in the way letters are formed within a body of writing by a single writer, systematic variation is the type of difference that indicates two samples are by two different writers. In every case the above examples show variation of this second type. Were the evidence more ambiguous, further comparisons might be warranted or even required. As it is, examining these eight letters and bigrams allows a definitive conclusion that W.W. Greg was not mistaken when he included documents written by both Peele and Oxford without raising any question about the possibility that the two men were somehow the same or that one was responsible for writing the letters of the other.

Principle #5: Begin with a Literature Search and Study the Prior Scholarship and Analysis.

An appendix to my 2002 University of Massachusetts dissertation, prepared with the assistance of Mark K. Anderson, includes a line-up of available letters from the annotations of the de Vere Bible to the handwriting of Oxford in his letters, comparing both to exemplars from two other contemporary poet-playwrights, John Lyly (Greg XVII) and George Peele. Anderson and I chose for these writers for comparison precisely because their samples correspond more closely to Oxford’s hand than any others in Greg’s book, which remains the most authoritative and comprehensive study of early modern English handwriting by literary writers.

Resemblance is not identity, and the close comparisons of the tables of the dissertation appendix  (pp. 517-537) already make it obvious for many reasons that Oxford’s hand is not at all the same as Peele’s. Mr. Prechter is unfortunately so far behind  in his literature survey that he failed to consult critical resources, widely available on the internet and in print, that could have saved him the trouble of making such a spurious argument.

*************************

In a reply to my critique, published in the same issue of the Newsletter, Prechter accused me of “cherry picking” the examples used to present my analysis. In my next blog entry on this topic, I will reproduce Prechter’s argument and evidence for this claim and show that his charge is a baseless projection.

Bibliography

Davis, Tom. (2007). “The Practice of Handwriting Identification,” The Library: Transactions of the Bibliographical Society 8.3, 251-276.

Greg, W.W. (1932). English Literary Autographs 1550-1650. Oxford University Press.

Huber, Roy A. and A.M. Headrick. (2018). Handwriting identification: Facts and Fundamentals, 2nd edition, edited by Heidi Harralson and Larry S. Miller.

Osborne, Albert S. ( 1929) Questioned Documents. 2nd edition. Nelson Hall: Chicago.

Srihari, Sargur N. et al. (2002) “Individuality of Handwriting,” Journal of Forensic Science 47:4, 1-17.

[1] For a classic introduction to the study, please see Huber and Headrick (2018 2nd ed). More generally, although technologically obsolete,  Osborn (1929) remains the most important study of general forensic method. Srihari et al. (date) was a landmark in establishing the scientific validity of forensic analysis, and Tom Davis’s “The Practice of Handwriting Identification” (2007) is the most sophisticated consideration of handwriting analysis from the perspective of the humanities scholar.

[2] e.g., Oxford’s 1583 “I am that I am” letter to Burghley, from which these exemplars are drawn.