I chose to annotate poems VII and XII from The Passionate Pilgrim. Neither one has been confidently attributed to an author. I explain to the reader in my annotations that the first poem bears resemblance to Shakespeare’s Venus and Adonis for its 6-line rhyming structure. I also note that the second, according to Shakespearean scholar Hallett Smith, is “The unassigned poem which readers have shown the greatest inclination to claim for Shakespeare…” but he notes that “…there is nothing to support the attribution.” These annotations about the poems allow me to justify the choice of the two poems, since they are not attributed to Shakespeare with any certainty, but could have been written by him.
The rest of my annotations, other than an introductory note on the original attribution of the two poems to “W. Shakespeare”, fall into the following categories: nuance, definition, authorship, meter, rhyme scheme, analysis, language analysis, structure, and correction. Though they often overlap, these are the tags I have used to distinguish between my various types of annotation.
Some are self-explanatory, like definitions, in which I simply provide the meaning of a word that may be difficult to understand either because it is archaic (e.g., lecher, hie) or because it contains multiple meanings, only one of which contributes to a reader’s understanding of the poem (e.g., care, damask). When I could find a word in a dictionary truer to the time period in which The Passionate Pilgrim was published, I did, relying on John Kersey’s 1708 Dictionarium Anglo-Britannicum. Otherwise, I used the Oxford English Dictionary.
I found it important to include annotations on poetic structure, noting meter and rhyme scheme for both poems, as well as miscellaneous structural aspects. Students and scholars of literature are certainly interested in this information, and though less complex analysis of the works do not require deep knowledge of their structure, the basic information on meter and rhyme that I included should be useful to most audiences.
Analysis of various sorts is included in most editions of the poems, including the Riverside Shakespeare that I consulted for this project. I wavered slightly on adding my own analysis, but in the end decided to add a short analysis of each stanza, as well as additional line or language analysis when it seemed necessary. My analysis is aimed at students like myself who have a fair grasp of the poetic structure and language but might need help grasping meaning.
Finally, there were several locations in which the Project Gutenberg file, where I accessed the full text of the poems, differed from the facsimile version of the text that I found on the Internet Archive. That copy was scanned in 2008 from the 1905 publication of a 1599 copy of The Passionate Pilgrim held at that time in the Christie-Miller library in Britwell. The 1599 copy showed clear capitalization that was not included in the Project Gutenberg file. Initially, this did not seem impactful on the text, but in some cases, it seemed to change the meaning by indicating that some words were proper nouns. For example, in line 11 of poem XII, the original copy indicates that the word ‘Shepherd’ should be capitalized, implying that the ‘Shepherd’ is the specific subject of the poem, the young lover, and not part of a simple turn of phrase.
In my annotations, I did not lose sight of the overall goal of my project, which was to provide an interesting perspective on the question of The Passionate Pilgrim authorship. To that end, with my annotations, I provided context on the authorship question for the two poems I chose and noted where in the poems indications of potential authors arose. For example, in poem XII, the ‘sweet Shepherd’ calls to mind the ‘young man’ who was the subject of many of Shakespeare’s later sonnets. I left it up to the reader whether that is coincidence or correlation, but I found it important to mention.
The primary research question behind this project is the following: Who wrote the poems in The Passionate Pilgrim? That has been the subject of serious scholarship throughout the years, ever since the poems were originally published in 1599. On the original title page, the publisher stated that the poems were by “W. Shakespeare”. Indeed, some of the poems were. I addressed the authorship of the poems with established attribution on the “Introduction” page, 5 of them Shakespeare’s, 2 by Richard Barnfield, and 1 each written by Bartholomew Griffin and Christopher Marlowe. That left 11 unclaimed poems for us to assign authorship.
Shakespearean authorship is a fascinating application of computational and mathematical methods in literature. Such methods have been applied to the authorship of the poems “Shall I Die, As This is Endless”, “Elegy”, and “Timon of Athens”, as well as to question or shore up Shakespeare’s authorship of his most famous works. Furthermore, stylometric analysis has been applied to this particular set of poems before. Elliott and Valenza, at the Claremont McKenna colleges, published a study in 1991 in which they applied their new method of “modal analysis” to identify which of the poems in The Passionate Pilgrim were written by Shakespeare. In this study, we had a published benchmark for our study.
As for why we chose The Passionate Pilgrim itself, and not some of the other apocrypha whose membership in the Shakespearean canon is in doubt, ease was the primary factor. The full text of the poems is available on Project Gutenberg, as well as in facsimile form on The Internet Archive, and definite attribution had been done for 9 of the poems in The Riverside Shakespeare. We adapted the metrics we used to compare the authors and works primarily from metrics used in the Federalist Papers authorship attribution assignment provided to us by Professor Michael Skalak, originally based on Michelle Craig’s assignment at the University of Toronto. We were able to simply test those metrics for accuracy and supplement our selections with additional statistics.
Our goal with this project was to put forth our own contribution to the conversation about Shakespearean authorship. What we found after careful testing, many mistakes, and a little bit of luck, were the following results:
Poem XII was written by Bartholomew Griffin Poem XIII was written by Christopher Marlowe Poem XVII was written by Richard Barnfield Poem IX was written by William Shakespeare. Poems X, VII, and XV were written by Marlowe or Barnfield Poems XIV and IV were written by Griffin or Shakespeare Poem VI was written by Shakespeare or Marlowe Poem VII was written by Marlowe or Barnfield
These results may seem odd, but they come from our particular methodology. Two of the systems of comparisons that we tested, least geometric mean of the error and most categories with lowest error, produced 100% accuracy on our tests. Where those two methods agreed, on poems IX, XII, XIII, and XVII, we propose with increased confidence that those results are correct. Where they disagreed, we propose that the true author is likely one of the two conflicting results. For the sake of argument, on the “Stylometric Analysis” page, I have chosen to adhere to the least geometric mean of the error model.
The significance of our results lies in both observations about the results themselves, and how they compare with the previous scholarly standards. The most fascinating result that our models both agreed on is that Shakespeare was predicted to be the most likely author of the full text of The Passionate Pilgrim. On some level this makes sense. Shakespeare is the author with the most confirmed works in the collection (5) and after our analysis, he still holds that title, with 8 attributed poems.
In a broader sense, however, this result means that a reader familiar with Shakespeare’s work would have likely believed Jaggard, the publisher, when he attributed the works to Shakespeare. These were not Shakespeare’s finest works, nor do many of them resemble his more well-known poetry, but in word length, punctuation, and line length, they were close enough to Shakespeare’s poetic canon for our program to assign him as its author.
I mentioned earlier that Elliott and Valenza had created what they called “A Touchstone for the Bard” using modal analysis and calculated that 2 blocks of 4 poems in The Passionate Pilgrim were likely written by Shakespeare. Our model agreed with their classifications on poems IV, VII, and IX, but did not classify VII, X, XII, XIII, or XV as Shakespearean, while they did.
It makes sense that our results were more varied, since Elliott and Valenza did not consider other possible authors, only whether Shakespeare was a likely candidate. Further enlightening is the fact that in 4 out of the 5 cases where we disagreed, our models identified Marlowe as the most likely author. In fact, these examples that Elliott and Valenza had identified as Shakespearean were the only works that our model attributed to Marlowe. Upon further consideration, this difference is logical. We created, but never used, a cosine similarity function to compare works, which produces a % value for how similar the vectors of statistics we used to compare poems were. We calculated the cosine similarity for Shakespeare and Marlowe to be 99.2%, by far the closest authors to each other.