The following paper was read by DARREN FREEBURY-JONES at ‘Authorship and Attribution in Early Modern Drama: John Marston and Others’, January 23rd 2016 at Birkbeck, University of London.
© Darren Freebury-Jones 2016
FOR ENQUIRIES PLEASE EMAIL: Darren_f.email@example.com
As many of you, no doubt, will not know who I am, I shall introduce myself and my specific research interests. My name is Darren Freebury-Jones and I conducted my PhD research at Cardiff University. My thesis investigated the arguments and counterarguments for and against an expanded Thomas Kyd canon, and also examined Shakespeare’s early plays, as well as the works of Thomas Nashe. I am currently working with fellow attribution scholars Marcus Dahl and Lene Petersen on a monograph-style book, which, rather like this workshop, explores different methods for identifying early modern dramatists’ hands in extant texts, and I will also be serving as editor and authorship consultant for a new edition of Kyd’s plays. I must say that I consider it a real privilege to be here, and I am incredibly grateful to Martin Butler and Matthew Steggle for inviting me. I was very excited when I saw the list of contributors, and I have long admired the pioneering work conducted in the field of authorship studies by the speakers here.
I must concede that I do not profess to be a Marston expert. Put simply, before I was invited to this event, I had never read any of his plays. I am pleased to say that I have thoroughly enjoyed all the works of his I have been able to lay hold of, and I cannot agree with Ben Jonson’s evaluative statement that Marston’s plays were ‘barren trash’. I am certain that this academic collection will be truly special.
I was originally going to give a paper on The Family of Love, before I was requested to turn my attentions towards perhaps the most complicated authorial case discussed today, Lust’s Dominion. Incidentally, I ran a couple of tests on The Family of Love, which yielded some thought-provoking results, and I would happily send my data to anyone who is interested, after this workshop. I found overwhelming verbal evidence for Lording Barry’s hand, but a number of matches with The Dutch Courtesan also.
I shall present my preliminary evidence regards the possibility of Marston’s hand in Lust’s Dominion, in the hope that I can be of some use. I’ve made some interesting discoveries over the last month or so that, I believe, with a little more time, could provide solid grounds for an attribution, or, indeed, deattribution. First, I should like to give a very brief survey of previous ascriptions of this text to Marston.
Lust’s Dominion is a revenge tragedy, printed in 1657. Gustav Cross made a case for Marston’s having been involved in the authorship of the play in 1958. He argued that the play exhibited Marston’s idiosyncratic vocabulary and could be identified with The Spanish Moor’s Tragedy, for which Philip Henslowe paid Thomas Dekker, John Day, and William Haughton in 1600, and that it was also the unnamed tragedy for which Henslowe paid Marston in September 1599. In 1980, Cyrus Hoy agreed that Lust’s Dominion and The Spanish Moor’s Tragedy were the same play, but that Marston began a revision of an older play for Henslowe in 1599, which was continued by Dekker, Haughton, and Day the following year. In 2001, Charles Cathcart argued that the play originated with Marston, was revised by Dekker, Haughton, and Day, and perhaps went through a subsequent limited revision, most likely in 1606. I refer audience members to Cathcart’s work on the play for a considerably more detailed survey of the external evidence for Marston’s hand.
Before conducting any tests on Lust’s Dominion, I contacted Martin Wiggins, for he is very much an expert in canon and chronology. This is what he had to say:
The people who were paid for The Spanish Moor’s Tragedy (i.e. Lust’s Dominion) were Dekker, Haughton, and Day; nobody has ever attempted to differentiate their shares. The payment, made on 13 February 1600, was a part-fee of £3, paid on completion of the work but with more to come (from some Admiral’s Men pot of money not controlled by Henslowe). The hypothesis that Marston had anything at all to do with it depends solely on the fact that Henslowe paid him an initiating ‘in earnest’ £2 on 28 September. This was (a) nearly six months earlier, and (b) wouldn’t leave much more if anything to be paid on top of the 13 February remittance as the standard total fee paid for most plays was between £5 and £6. For a £2 in earnest you would expect Marston to be quite well advanced in the play, so there would be abundant internal evidence of his hand in the finished play, even if he then left the project to be finished off by the other three. I doubt you’re going to find it, and I interpret the 28 September payment as a fee for a play of which no other trace survives.
The lost play Wiggins mentions could be The King of Scots, which E. K. Chambers proposed was a collaboration between Marston, Dekker, Chettle, and Jonson, although it is worth noting that, just as in Lust’s Dominion, Henslowe did not actually link Marston’s name with the authorship of that play. Wiggins nevertheless doubts that Marston had anything to do with Lust’s Dominion but, notably, he suggests that if it is the same play that Marston was paid an initiating fee for, there would be considerable internal evidence for his hand. I’m inclined to agree, and so I conducted entirely objective tests on this text with the purpose of either eliminating or legitimising Lust’s Dominion.
During my Kyd researches, I analysed verse style, prosody, intensifiers, colloquialisms, prefixes, suffixes, function words, rhyme forms, linguistic idiosyncrasies, compound forms, plot, characterization, and overall dramaturgy. Of course, being new to Marston, and he new to me, and given the limited timescale I had to examine this play, I was unable to immerse myself in his dramaturgy and combine all of these approaches for Lust’s Dominion. The method I have been asked to talk about today is collocation matching. This methodology has been utilised, in conjunction with other authorship tests, for a number of play editions. Brian Vickers and Marcus Dahl have used plagiarism software during their work on the John Ford corpus, with highly reliable results; Richard Proudfoot for Edward III and Fair Em, which he has been editing; and I understand Martin Wiggins has recently combined collocation evidence with analysis of function words to ascribe a play to Dekker.
Ian Lancashire notes that collocations ‘are the linguistic units we work with most: they fit into working memory and resemble what we store associatively’. We all have certain phrases we repeat, chunks of words that recur in our everyday conversations, such as, to give a common example, the interrogative ‘How are you’, or, to give a rare example here in the UK: ‘Nice weather we’re having’. In our writings, these combinations often take the form of contiguous word sequences, called N-grams. Two contiguous words are known as bigrams; three consecutive words are known as trigrams, or triples; four, known as tetragrams etc. Longer word strings, such as tetragrams and pentagrams (five-word sequences), are statistically rarer in early modern drama and can be a good indicator of either common authorship or plagiarism.
There can be little doubt that dramatists such as Marston, Dekker, Day, and Haughton, also consciously and unconsciously repeated combinations of words and ideas. In fact, the constraints of writing in ten-syllable pentameter lines for plays, with character dialogue written in blank verse and prose, and the haste with which dramatists had to provide playing companies with material, means that these writers were particularly prone to self-repetition.
We can therefore get an idea of whether Marston had a hand in Lust’s Dominion by identifying, as Lancashire puts it, ‘idiolect markers’, which are ‘combinatorial, embedded in an author’s long-term memory, and repeated. We recognise them by unconscious pattern matching similar to what enables us to quickly make out a face in a crowd’. Vickers elaborates that:
Where earlier linguistic theories held that users of natural language selected single words to be placed within a syntactical and semantic structure, it now became clear that we also use groups of words, partly as a labour-saving device, partly as a function of memory. Such verbal economy is particularly prevalent in the drama written for the public theatres, where constraints of time demand speedy composition, characters fall into a set of roles with attendant speech patterns, and the verse line easily admits ready-made phrases. It is hardly surprising that many dramatists frequently repeat themselves.
In order to highlight consecutive clusters of words shared between electronic documents, I employed anti-plagiarism software called WCopyfind, which was developed by Lou Bloomfield, Professor of Physics at the University of Virginia, in order to identify students who were guilty of borrowing feathers. I am not alone in using WCopyfind, which has been my favourite software for the last four years spent examining Kyd’s plays. In fact, I am in very good company: Vickers uses it frequently, and MacDonald P. Jackson has used it in order to search for three and four-word units shared between Double Falsehood and its authorship candidates. I have always found that a minimum of three words is the optimum amount for identifying authors’ hands. No doubt a search that admitted bigrams would reveal a substantial number of matches, whereas a search limited to large phrasal structures might filter out statistically significant findings. I thus set the software to highlight every consecutive three-word phrase shared between old spelling versions of plays drawn from ProQuest, and then checked the matches, which were highlighted within a fraction of a second, using the database Literature OnLine. I double-checked results using the database Early English Books Online for variant spellings. The plagiarism software automatically highlights all verbal repetitions, whether commonplace in drama or distinctive. I followed MacDonald P. Jackson in that my criteria for establishing whether a matching utterance was rare or not was that it occurred not more than five times in plays performed in London, during the period that the four primary candidates for authorship were writing contemporaneously for the theatres, namely, in this instance, 1590 to 1610. Marston, of course, had ceased writing plays and was ordained deacon by the end of September 1609. This is what Jackson would refer to as ‘an appropriate timeframe’. What I like about Jackson’s criteria is that it allows one to compare matching phrases that, say, appear in both Marston and Dekker, but one can ascertain which writer gives the more approximate example, through an examination of prosodic characteristics, grammatical usage, or whether the phrase also provides a thought-parallel. Even if a word sequence is not unique to a writer, it can still demonstrate an author’s individuality when analysed closely. I thus adhere to David Lake’s categorisations of ‘grammatical or semantic’ patterns, and, as Muriel St Clare Byrne put it in 1932, ‘parallelism of thought coupled with some verbal parallelism’.
To begin with, I ran the plagiarism software on undoubted Marston plays to get an idea of his habits of self-repetition within that ‘largely self-contained phenomenon’ that is early modern drama. There are a number of highly individual word combinations shared between Antonio’s Revenge and The Dutch Courtesan, which I’ve replicated in the handout. The first rare match
Best art Presents, not what it can, but should (Dutch)
Tis praise to doe, not what we can, but should (Antonio’s)
shows us that Marston, like his contemporaries, often repeated phrases in the same position in the verse line, such as the formulaic line-ending ‘can but should’. Moreover, the phrase ‘not what’ in this discontinuous six-word sequence serves the same semantic and syntactical purpose. Matches of this kind are thus useful indicators of common authorship. The match marked number 6,
I must be enforced to forget all men (Dutch)
I must be forced to conclude the tyring man (Antonio’s)
is also useful as an authorship marker, with the pentagram ‘I must be forced to’ utilised by Marston in both passages as a formulaic line-opening, while ‘men’ and ‘man’ co-occur at the end of the respective verse lines. The slight variations in this match, such as ‘forced’ and ‘enforced’, ‘men’ and ‘man’, indicate that these lines stem from a single author’s mental repertoire of verbal combinations. Number 5 suggests Marston’s complex networks of association, as he links the three-word unit ‘and no question’ with the notion that men, or husbands, perfect women, or their wives.
If we compare Antonio’s Revenge to a play closer in genre, The Malcontent, we find that many of the matches are fairly unobtrusive, such as example 1, the line-opening ‘O heauen O’, or example 3, which provides strong evidence for a single author’s verse cadences. Less noticeable phrases are not as likely to be copied, imitated, or parodied by another author, and therefore provide strong evidence for common authorship. In my experience, same author play pairs share a substantial number of unobtrusive matches, along with longer word strings, such as examples 4 and 5. These texts by Marston, as we can see, share some striking word sequences, but also a number of seemingly trivial phrases picked up by the plagiarism software. A reader would be unlikely to notice the latter, and I dare say Marston was not conscious that many of these phrases constituted his idiosyncratic lexicon.
Now, let us turn to verbal matches between undoubted Marston texts and Lust’s Dominion. If we examine parallels with Dutch Courtesan, we find that they are hardly striking, influential phrases likely to have been appropriated or parodied. As David Lake noted in his examination of verbal parallels differentiating the hands of Dekker and Thomas Middleton, ‘the collocations themselves are so unremarkable that imitation is very unlikely’. Interestingly, the match marked number 1 shares similar prosodic characteristics, with the triple ‘eldest child of’ serving the exact same purpose in the verse line, preceded by the initial iambic foot. Example 2, the four-word unit ‘I’ll play the devil’, also suggests a single author’s pattern matching, for this sequence is followed by a conjunction in both instances. The most interesting match between these plays, in my opinion, is number 4, with its slight variation in syntax, indicative of a single author’s word associations. But it is the verbal affinities with Antonio and Mellida and Antonio’s Revenge that provide truly compelling evidence for Marston’s hand.
In authorship studies, it is a basic principle that, for most reliable results, a practitioner should compare like to like. Indeed, the matches between Lust’s Dominion and Marston’s tragedy Antonio’s Revenge reveal no discernible differences, in terms of quality, with many of the matches I have discovered between Antonio’s Revenge and the undoubted Marston texts. The first example, the recurring tetragram ‘Why do you frown’, varies slightly in terms of pronoun choice, but serves the same function at the beginning of the verse line. The second example, the match
‘Tis done, and who gainsaies it is a Traitor (Lust’s)
Tis done, and now my sowle shal sleep in rest (Antonio’s)
also suggests Marston’s verse habits. On the surface, the repetition of the triple ‘Tis done and’ seems trivial, but examining the phrase in context gives us a possible insight into Marston’s associative memory. This phrase occurs during the King’s speech in Lust’s Dominion. The King is responding to Philip’s interrogative ‘Dare sons presume to break their fathers will?’. If we analyse the matching sequence in Antonio’s Revenge, we discover a further verbal link in Andrugio’s subsequent line ‘Sons that reuenge their fathers blood, are blest’. The association of this unremarkable three-word unit with the topic of sons at the behest of their fathers would seem to derive from a single author’s verbal memory. There are also indications of Marston’s recurring thought patterns in example 5. In Lust’s Dominion, Maria speaks of a ‘powerful drug’, which is instilled ‘through all’ the King’s ‘spirits’. Similarly, Antonio speaks of ‘hell-straid iuyce’ which is ‘powred’ through ‘vaines’. Antonio’s moralising speech also shares the tetragram ‘More than a devil’ with Lust’s Dominion, as we can see in example 11. Looking at all of the verbal data I have collected, the exchanges between Philip, the Cardinal, and the Queen Mother near the conclusion of Lust’s Dominion have a particularly high concentration of Marstonian phrases, while match number 6,
And since I liv’d for her, for her I’l die (Lust’s)
And since I cannot liue with him, I dye (Antonio’s)
is also noteworthy. The triple ‘And since I’ embraces ‘live’ and ‘I die’, placed in almost identical positions in the pentameter line, which once again suggests a single author’s habits of incorporating idiosyncratic phrases into his verse structure. We might note that both lines contain pronouns and give us an unmistakeable thought parallel. In Lust’s Dominion, the King delivers this line following the news that Maria is dead, while Mellida’s line is in response to the fallacious report that Antonio has drowned. Additionally, I have discovered another thought-parallel, coupled with verbal parallelism, between these lines and Antonio’s Revenge, through traditional reading-based methods (it was not picked up by the plagiarism software). At the conclusion of the tragedy, the revenger vows to remain faithful to his dead lover: ‘She lives in me, with her my love is dead’.
Number 8 is perhaps significant in that the line
One stratagem that in despite of fate (Lust’s)
gives us a double match with both Antonio and Mellida’s
Weele sing yet, faith, euen despite of fate (Mellida)
and Antonio’s Revenge:
Let’s all turne sighers. Come, despight of fate. (Antonio’s)
Such recurring phrases could provide evidence for Marston’s verbal habits. We might also note that, in example 10, the triple ‘before thou dost’ serves a similar contextual purpose. Isabel speaks of killing herself, while Balurdo also associates this phrase, which is hardly reliant on that context, with death.
I propose that there are two possibilities as to why Lust’s Dominion shares so many distinct clusters of words and ideas with Antonio’s Revenge. Either Marston was drawing from his mental repertoire, to serve the same generic purpose, or the dramatists responsible for Lust’s Dominion were deeply familiar with the verbal texture of that play, and were either consciously or unconsciously replicating the phraseology of Marston’s text. Wiggins dates the two Antonio plays 1599, and it is conceivable that Marston was working on them around the same time as he was working on the text we now know as Lust’s Dominion. As Charles Crawford put it in 1906, ‘works of the same date by the same writer invariably repeat each other more often than others that are separated by longer intervals of time’. If either Dekker, Day, or Haughton were recycling verbal details from Marston’s play in 1600, it seems they must have had access to a manuscript, for many of these N-grams reproduce verse structure. They do not strike me as the products of aural memory, due to the dramatists having seen these plays during performance. Given that my research on Lust’s Dominion is somewhat in its salad days, I shall resist asserting, apodictically, whether these N-grams are the result of reading knowledge or common authorship.
Nonetheless, I should also like to consider the matches between Lust’s Dominion and The Malcontent. We find that example 4 is indicative of Marston’s specific idiolect, for the trigram co-occurs with another Marston text. Conversely, the match marked number 8 is, I think, closer to Dekker in terms of thought, while the last example, the triple ‘In true contrition’ is highly distinctive, and is utilised as a formulaic line-opening in both examples.
The Malcontent was probably written a couple of years after Lust’s Dominion. The verbal affinities between these texts therefore cannot be explained by the theory that Dekker, Haughton, or Day were appropriating phrases from Marston; this time, we would have to hypothesise that Marston was imitating Lust’s Dominion. As Cathcart puts it: ‘we would have to posit a Lust’s Dominion which was thoroughly and repeatedly influenced by Marston’s verse satires and which Marston himself then proceeded to plunder’.
I have already begun testing Marston’s plays against those of Dekker, Haughton, and Day. It may seem paradoxical, but identifying an author’s patterns of influence can also be useful for either attributing or deattributing a text to an authorship candidate. For example, Kyd often replicated language from John Lyly’s plays, sometimes irrespective of scenic context or genre. That is not to say that Lyly contributed to Kyd’s plays, but analyses of appropriated phrases, and the nature of such borrowings, can yield useful authorship indicators.
When I tested the Marston texts that had dense verbal relations with Lust’s Dominion against Haughton’s Englishmen for my Money, I discovered that the two dramatists merely shared some common phrases within the theatrical vernacular of their time. I therefore doubt that Haughton was borrowing distinct phrases from Marston, or vice-versa, in Lust’s Dominion. John Day’s The Isle of Gulls also shared few rare phrases with the Marston texts, although I did discover a handful of matches with The Malcontent; however, they seldom exhibited similarities in terms of thought processes or verse structure. Perhaps Day had seen The Malcontent during performance (Wiggins dates that play 1603 and Day’s play 1606), but this would not explain the matches between Lust’s Dominion and Marston’s tragedy, given that The Malcontent had yet to be written when Day collaborated with Dekker and Haughton. Dekker, as we can see in the handout, seems to share a number of phrases with Marston but, for the most part, close consideration of context enables us to distinguish between the writers.
When I compared a selection of Dekker texts against Antonio’s Revenge, The Malcontent, and What You Will, I discovered a couple of matches that suggest Marston and Dekker, who were grouped together in Jonson’s The Poetaster, influenced each other. Nonetheless, I have yet to find parallels in Dekker’s plays of the same nature or quantity as those with Marston’s tragedies in Lust’s Dominion. But, if requested by the editors, I would be happy to continue investigating the verbal relations between these playwrights, perhaps complemented by stylistic studies and examinations of characterization and overall dramaturgy, in order to reach a conclusion.
Another play probably written by Marston within a year or so of Lust’s Dominion, but generically different, is the romantic comedy What You Will. We might note that in example 5, the triplet of words ‘must be when’ embraces the pronoun ‘you’, while the fact that the phrase ‘be proud and’, in example 10, is repeated twice in Lust’s Dominion, could help us to zone in on Marston’s possible contributions to the play. I shall speak further on internal repetition shortly. Example 12 is significant, I think, for the tetragram ‘the circle of thy’ is employed as part of a rhyming couplet in both passages, and the context of magic, or ‘damned charms’, is strikingly similar. One cannot yet be certain whether common authorship provides the answer for this parallel, which almost certainly exceeds the bounds of coincidence, but it is fair to say that something is going on here.
There are also some interesting matches between Lust’s Dominion and The Fawn, particularly number 3, which gives us a parallel of both language and thought. The Queen Mother, speaking of her son, says that he ‘woo’s greatnesse, He takes up Spanish hearts on trust, to pay them’ while, in The Fawn, Hercules speaks of spendthrifts who have ‘treasonably purchased Ladies affections, without being of ability to pay them for it with money’. I must stress that these matches are the results of my preliminary searches; a work in progress, if you will. I have thus far only been able to test Lust’s Dominion against a selection of plays by our four authorship candidates, and no doubt one’s hypothesis regards the possible divisions of shares, which do not appear to be neat, and the extent of Marston’s contributions, if indeed he did contribute, could alter once all the data is collected.
It is worth mentioning, however, that the affinities I have discovered with Haughton, who seems to be, fortunately for our purpose, quite self-repetitious, suggest that he played a dominant role in the verbal fabric of Act Two. The matches with Dekker I have found so far are interesting (particularly those in the play’s opening act, which provides some of the strongest evidence for Dekker’s hand) in that they suggest he attended to another writer’s draft; they are very different to the matches with Marston and Haughton, in terms of distribution. They frequently occur at moments when characters either enter or exit. Dekker’s phraseology is often found in the first or last speeches delivered by a character on stage, and sometimes paves the way for the entrance of another character. I am particularly suspicious of the play’s opening scenes, which feature a number of hypermetric lines that could indicate another dramatist touching the dialogue up in places (although the text as a whole is generally messy, with prose seemingly printed as verse and vice-versa). That said, Dekker was sometimes prone to writing in a loose, broken verse style, with some of his lines extending to twelve and even fourteen syllables. My verbal evidence also suggests that Dekker was responsible for the dialogue between Philip and the soldiers in Act Four Scene Five. Matches between Lust’s Dominion and Dekker’s The Shoemaker’s Holiday and Satiromastix can be found on the last two pages of the handout.
Overall, the matches I have thus far discovered with Haughton, Dekker, and Day (I have refrained from including all of them in the handout, for the sake of the world’s forests) are hardly different in terms of quantity or quality with those I have adduced from Marston. I will have a better understanding of the distribution of parallels once I have tested this play against all of the works of these candidates. I am currently developing a method of checking internal repetitions, which I detail in a forthcoming article on Shakespeare’s early plays. I divide the text into segments and utilise plagiarism software to highlight repeated N-grams. In my experience, authors in both collaborative and sole-authored texts repeat phrasal units at the forefront of their verbal memories, and I have discovered little evidence that dramatists shared extended verbal details in collaborate works, despite similar scenic contexts, or the fact that they were writing for the same characters, in the same settings. This is hardly surprising, given the time constraints and pressures these writers were under to deliver material. They are unlikely to have had the time to scrutinise each other’s portions in an effort to homogenise their shares. Moreover, when I have discovered shared phrases between co-authors, they are often used in very different contexts and indicate separate authorial cognitive processes, as opposed to formations persisting in a single brain. They also often demonstrate different verse habits. Once all the data regards Marston, Dekker, Day, and Haughton is collected, this method could be of use in distinguishing between the hands involved in Lust’s Dominion, although, as I have noted, the divisions of authorship do not appear to be neat.
That N-gram repetition in early modern plays can help to determine authorship has been demonstrated by Martin Mueller, co-author of The Chicago Homer, which allows direct study of the thousands of N-grams repeated in the corpus of early Greek epic (the famous ‘Homeric Formulae’). Mueller has created a database called Shakespeare His Contemporaries, consisting of over 500 early modern plays dated between 1552 and 1662. These plays are tagged so as to identify all verbal repetitions, ranging from two-word sequences to much longer strings of words. What is most refreshing about Mueller’s database is that we are afforded a corpus that has been created to facilitate the study of N-grams. It produces objective, automated results, which create, as Mueller puts it, ‘a framework of expectations’ within which their evidentiary value can be evaluated. This digital corpus, ‘enriched with linguistic metadata, makes it easy to look at microscopic details of verbal texture across millions of words. The corpus becomes a network of intertextual relations in which changes or differences in the distribution of words or repeated phrases let you contextualize particular textual objects by author, topic, style, genre, or period’. Mueller has set about answering the question ‘If you are interested in the intertextual relationship between one play and another, how many shared n-grams do you need to make a plausible case that something is going on?’ He notes that in his corpus, ‘plays by the same author are likely to share more dislegomena. If we look more closely at shared dislegomena’ (that is a sequence of words with a collocation frequency of 2) ‘by same-author play pairs, we discover that on average plays by the same author share five dislegomena, and the median is four. Roughly speaking, plays by the same author are likely to share twice as many dislegomena as plays by different authors’. Mueller also notes that, in his corpus, longer word sequences are statistically rarer than triples, and that ‘it is quite rare for two plays–texts that are typically between 15,000 and 25,000 words long–to share more than one or two’ unique N-grams of four or more words. I have profited much from Shakespeare His Contemporaries, and Mueller was kind enough to share his data on Lust’s Dominion.
Mueller sent me a document listing all play pairs involving Lust’s Dominion, ranked in order of plays with the densest verbal relations to the tragedy. I have limited my searches for this paper to the period 1590 to 1610, in order to analyse verbal repetitions in comparison to the four primary authorship candidates. Mueller’s database, however, allows one to examine the tragedy against over 500 plays written over a hundred year period, with 543 pairwise combinations involving Lust’s Dominion. Mueller explained how he ranked these plays. He
assigned each n-gram the initial value of the number of its characters, and then divided that value by the number of plays in which it occurs. So an octagram restricted to two plays has a value of 4, while an octagram that occurs in eight plays has a value of 1. For each play pair combination I added the values for each n-gram and then computed the rate per 10,000 words in the combined word count of the two texts. This is a very primitive procedure, but it works in the sense that the results are certainly not arbitrary. The ranking you get from this expanded list of n-grams isn’t very different from the ranking that you get from using n-grams that only occur in two plays. I’ve become interested in the general behaviour of n-grams. If an n-gram occurs in eight texts, is its distribution random, or is it more likely to cluster in a genre or an author? What do such n-grams tell us about author habits? Quite a bit, I think.
In the list that Mueller sent to me, the top value for Marston is in the 80th percentile, and his other plays are below that. Marston ‘isn’t even playing at the top of the second division’ for play pairs suggesting characteristic patterns of authorial usage. One would therefore be tempted to rule Marston out, but we must remember that Lust’s Dominion is a very complex authorial case, and that nobody is arguing for Marston’s sole authorship. The top play associated with Marston in this list is Histriomastix, followed by The Malcontent, The Insatiate Countess, Antonio’s Revenge, and a further six Marston texts. Marston might not be playing in the first division, but the plays I have just listed all appear higher in the rankings than Haughton’s ‘comedy of tricks’ Englishmen for My Money, and we can be confident that Haughton had a hand in Lust’s Dominion. John Day also figures prominently in this list (five of his plays are included), with the top ranked play being The Isle of Gulls, but this play is still ranked below the four Marston texts I’ve just named. Dekker outranks all of these authors, with twenty of his plays, albeit many collaborative, making the list, and his Patient Grissel playing near the top. It is notable that Haughton is also supposed to have been involved in the composition of this play. Twelve of Dekker’s plays are ranked higher than the top Marston text. We can thus see that Dekker appears to have the densest verbal relations with Lust’s Dominion out of the four authorship candidates, but Marston is in fact ranked higher than Day and Haughton. In short, according to these rankings, Marston is just as likely an authorship candidate as Day and Haughton, who have been linked with the play according to strong external evidence and, as my research suggests, internal evidence.
I am confident that, if I were given the opportunity to expand my research, perhaps combining this method with other methods I have used previously, or in coordination with some of the approaches discussed today, I could provide strong evidence for the divisions of labour in Lust’s Dominion, and ascertain whether the play should be ascribed in part to Marston. Unlike my work on Kyd, I have not yet been able to identify many of Marston’s authorial traits, although, to instance just a couple of examples, I should be happy to analyse the use of Latinate terminations in Lust’s Dominion, in relation to Marston’s ‘affected diction’, as well as such Marstonian habits, identified by G. K. Hunter, as ‘sudden breaks in speech’ and ‘exclamatory noises’. Further investigations of this kind, combined with a completed analysis of N-gram repetitions, may be of use for future inquiries.
Verbal matches with Marston abound throughout the play, although my findings are in agreement with Professor Hoy, who proposed that he ‘seems mainly to have concerned himself with the […] end of the play’. The strongest evidence for Marston’s hand can be found in Act Four Scene Three, the combat between Philip and the Spanish Moor, and Act Five Scene Six, the last scene of the play. These phrases are woven into the verbal fabric of the tragedy and do not seem to indicate later material added by Marston. I suggest that if indeed the play originated with Marston, he was responsible for some fundamental plot points; I can detect few verbal matches in passages that we might consider superfluous, or ‘padding’. Something is certainly going on here; there is plenty of evidence for verbal relations with Marston plays that preceded and followed Lust’s Dominion.
As Crawford put it:
A man’s vocabulary is the surest test by which he can be judged, for no author can jump out of his own language into that of another without betraying himself. His other work will condemn him, and vindicate the wronged party at the same time. It only means the exercise of much patience and minute inquiry to know “which is which.”
Minute inquiry would indeed be the next appropriate step. For now, I hope that my findings will be of some use, for this play and other attributional puzzles.