Ass‐troll‐ogical Nashe: Revisiting Two Dangerous Comets and A Wonderful Prognostication

This article revisits the authorship of and relationship between three mock‐prognostications published pseudonymously in 1591, drawing on contextual, bibliographical, and stylistical analysis to attribute Two Dangerous Comets and A Wonderful Prognostication to Thomas Nashe. The article also considers the significance of these findings for studies of Nashe, satire, and Elizabethan print culture.

a group' of pseudonymous mock-prognostications 'issued in apparent rivalry' that year. 4The first of these, a pamphlet by 'Francis Fairweather', was recorded in the Stationers' Register on 25 February 1591 under the publisher William Wright but is now lost; 5 the second, A Wonderfull, strange and miraculous Astrologicall Prognostication by 'Adam Foulweather', like Two Dangerous Comets, was not entered into the Stationers' Register but copies of both texts are extant. 6Given the connections between these titles, any persuasive arguments about their putative authorship 'must take account of the whole group'. 7Thus, drawing on extensive contextual research as well as new findings from the computational analysis, we revisit the authorship of-and relationship between-these pamphlets, concluding that Two Dangerous Comets and A Wonderful Prognostication are both by Thomas Nashe.While A Wonderful Prognostication has long been associated with Nashe, the attribution of Two Dangerous Comets affords new perspectives onto his experimentation across genres during the early part of his literary career, the development of his satire and his engagement with Elizabethan print culture.

NASHE AND MOCK-PROGNOSTICATIONS
As a genre, the mock-prognostication was relatively new: earlier in the sixteenth century, François Rabelais authored five mock-almanacs which 'in many ways set the stage for later English versions of the trope of playing with astrological discourse and belief'. 8Though these were not translated into English until the seventeenth century, Nashe's penchant for Rabelaisian style had been noticed by his contemporaries as Gabriel Harvey describes a young Nashe haunting 'Aretino and Rabelay the two monstrous wittes of their languages'. 9Huntingdon Brown notes that the Wonderfull Prognostication resembles Rabelais' Pantegrueline Prognostication (1532) 4 R. B. McKerrow (ed.),The Works of Thomas Nashe, 5 vols (London: A.H. Bullen, 1904-5), V: 139. 5 Edward Arber (ed.),A Transcript of the Registers of the Company of Stationers of London; 1554-1640 A.D., 5 vols (London: privately printed, 1875), II: 576 (Register B, fol.271v). 6 'strikingly throughout' and is the 'unacknowledged source' of Foulweather.10Anne Lake Prescott notes that John Wolfe entered 'Gargantua his Prophesie' into the Stationers' Register in April 1592, which may be a translation of the Pantagrueline Prognostication meaning that Nashe may have been familiar with it in translation. 11Indeed, both Foulweather and Smellknave mix the obvious with plausible-sounding astrological knowledge in the manner of Rabelais: for example, Foulweather's claim that 'olde women that can liue no longer shall dye for age' strongly echoes Rabelais' suggestion that 'old Age shall this year be incurable by reason of the years past'. 12Katherine Walker suggests that Foulweather's A Wonderful Prognostication is 'the first notable English version' of the mock-almanac genre, 13 a genre which continued in popularity into the seventeenth century.An early English example is A merie and p [leasant] prognostica [tion] Deuised after the finest fashion, 14 printed in 1577 and advertised as the work of 'fower wittie doctors'-Spendall, Whoball, Doctor Deusace and Will Sommers-though much of the material is derived from an earlier mock-prognostication published 33 years prior, A mery p[ro]nosticacion, which reinforces its comedic credentials on the title page by incorporating a woodcut of a fool, complete with ass's ears and tail. 15The alliterative 'pleasant prognostication' may have inspired the title of Fairweather's pamphlet, which is listed as 'Francis Fairweather's pleasant prognostication, &c. in 4° 1591' in Francis Daniel Pastorius' commonplace book, known as the 'Bee-Hive' manuscript, begun in 1696. 16To our knowledge, this is the only record of a title for Fairweather's pamphlet.It is likely that the writer or writers of the three mockprognostications published in 1591 were familiar with both the new genre and its genuine counterpart-indeed, the partial title of Fairweather's pamphlet suggests a conscious engagement with other mock-prognostications.
Almanacs were extremely popular and 'were bought, read and used by men and women across the social spectrum'.17Cheap and annually produced, once their original purpose had been served almanacs would be used 'for lining pie dishes, for lighting tobacco, as toilet paper' and so relatively few survive. 18erhaps this ephemerality was attractive to writers of mock-prognostications, that even the genuine article was not expected to last afforded satirists some measure of protection.Writers of mock-prognostications tend to use pseudonyms that not only signal the genre, such as Fairweather or Will Sommer, but the practice is also in line with genuine almanacs and prognostications, such as the prognostications of Erra Pater (also known as the Prognostications of Esdras).Little is known about Erra Pater but his anonymity provides the prognostications, which were reprinted at least 12 times between 1536 and 1639, 19 with authority as he is described on the title page as 'a Jewe borne in Jewery, a Doctour in Astronomye, and Physycke'. 20In A Wonderful Prognostication, Foulweather adopts a similar method of establishing authority by describing himself as 'student in astronomy' or, in the unique British Library copy (STC 11209), the more jovial 'student in Asse-tronomy'.In this, Foulweather follows other genuine almanac writers such as Gabriel Frend, who describes himself as 'student in Astronomie' in his almanac and prognostication of 1592. 21Nashe has a particular issue with Frend because he believed him to be 'no Frend, but my constant approued mortall enemie Gabriell Haruey'. 22he form itself offers opportunities for creativity and to target specific groups.Both Foulweather and Smellknave take issue with various professions, such as tailors who 'shall steale nothing but what is brought unto them' 23 and who 'shall haue more conscience, for where they were wont to steale but one quarter of a cloak, they shall haue due Commission to nick their customers in the Lace, and take more then enough for the newe fashion sake, beside theyr old fees' and, in taking aim at tailors, they also criticise frivolous fashion and those who pursue it. 24Nashe reproduces the form of a mock prognostication so perfectly that, at times, it would be difficult to tell the difference between his text and a real prognostication.For example, Frend lists so many potential diseases as a product of winter, including 'Rewmes, Catarres, Coughes […] joynt Aches, goutes and such lyke', 25 that it is almost a guarantee that the reader should be afflicted with one of them.In derision of the form, Smellknave couches similarly obvious predictions in profound terms: 'it shall be wonderfull to behold, (through this sinister influence) howe men that are deafe, shall heare no more than those that are dead: and such as are without teeth, shal chewe as little as babes newe borne', 26 whereas Foulweather claims that 'manye shall goe soberer into Tauernes then they shall come out'. 27isually, mock-prognostications employ many of the aesthetics of genuine almanacs: they were the same size, sold cheaply and some employed other generic signals on the title page such as horoscopes.Adam Smyth notes that 'mock almanacs are very close to the originals: there is a curious alignment of parody and original; a sense, even, of satire lagging behind its object', 28 and so unsuspecting readers could, and evidently did, mistake mockprognostications for the genuine article.Thus, Pastorius might be forgiven for listing Fairweather's pamphlet along with other titles of 'Books treating of Magical Arts better to be burnt than sold'. 29t is easy to see why the mock-prognostication genre would have appealed to Nashe.It gave him licence to continue developing his satire under a pseudonym, something he was used to as one of the writers employed by the archbishops to write the anti-Martinist pamphlets.Indeed, Simon Smellknave recalls his previous alter-ego, 'Cutbert Curry-knave', who appeared the year before in An Almond for a Parrot. 30In 1589, under the pseudonym 'Pasquill of England', Nashe promised Martin Marprelate-and his readers-an almanac: Pasquill hath vndertaken to write a very famous worke, Entituled, THE OWLES ALMANACKE: wherein the night labours and byrth of your Religion is sette downe: the ascent and descent of the Starres that fauour it, is truelie calculated: the aspects of the Planets raigning ouer it, are expressed, with a iollie coniecture drawne from the iudgement of the Theame, what end your Religion is like to haue. 31ough The Owl's Almanac did not appear as part of the anti-Martinist programme, Nashe clearly signals a future intention to engage with the genre at this early stage.A publication of the same title appeared in 1618 (anonymously, once attributed to Thomas Dekker but now thought to be by Thomas Middleton) and the title's reference to the promised publication of 30 years previous is clear. 32Although ephemeral, mock-prognostications are far from inconsequential pieces of literature: they are self-referential, which helps to ensure their longevity and prolong discussion about the very things they satirise long after they have lined pie dishes alongside their genuine counterparts.As a young satirist emerging from the Marprelate controversy, with a literary nemesis to target, and with the genre favoured by Rabelais and Aretino, how could the anonymity, ephemerality and self-reflexivity of mockprognostications not have appealed to Nashe?PREVIOUS ATTRIBUTIONS, 1778-2023

R. B. McKerrow included A Wonderful Prognostication in
The Works of Thomas Nashe, which remains the standard edition, 'as it seems to be commonly regarded as a genuine work of Nashe', though he was doubtful of the attribution: 'I have been unable to discover any reason whatever for so considering it, nor have I been able to learn by whom or on what grounds it was first attributed to him'. 33Printing the pamphlet among Nashe's 'Doubtful Works', the earliest association McKerrow could locate comes from the 1807 auction catalogue of Isaac Reed's library, in which A Wonderful Prognostication is listed alongside other putative Nashe titles. 34The catalogue evidence, however, cannot be taken at face value, given that the same also ascribes several works penned by Gabriel Harvey and others to Nashe. 350 Grosart situated the pamphlet among the 'Harvey-Greene Tractates' as an example of 'Nashe [coming] to the rescue of the dishonoured memory of his deceased friend'. 41To round out the century, George Saintsbury anthologised A Wonderful Prognostication in his Elizabethan & Jacobean Pamphlets of 1892, ascribing it to Nashe as 'composed in direct imitation of Rabelais' and only 'indirectly an attack on the Harveys'. 42hough the precise origins of Nashe's association with A Wonderful Prognostication remain a mystery, the validity of the attribution has not gone uncontested.In his 1907 Leipzig University doctoral dissertation on Nashe's polemics, for example, Arno Piehler doubted the attribution for several reasons, including authorial style: The diction is not Nashe's in his other prose writings; everything characteristic of his language is missing.The personal, which is always in the foreground with him, recedes here completely.After such a severe challenge as Nashe had received from the Harvey brothers, he would certainly have strayed from his subject several times and made insults to his offenders.The fine irony without any personal barbs completely contradicts his biting mockery, which, mixed with strong insults, never leaves anyone in doubt who he is hitting.with a writer with so distinctive a style, not to be convinced that it is his is almost to be certain that it is not'. 44By contrast, more recent commentators have deemed the attribution plausible.For Willem Schrickx, not only is 'the very vocabulary' of A Wonderful Prognostication 'entirely consistent with that of Nashe's genuine writings', but its 'complaints of pennilessness' and topical references are tantamount to Nashe having 'written his own signature'. 45Charles Nicholl suggests A Wonderful Prognostication could have been written by Nashe 'to tilt at Richard Harvey and his much-ridiculed Astrologicall Discourse', and, with its references to the Marprelate controversy and intimate knowledge of the city, 'might perhaps anticipate the sharply realized London ambiance of Pierce Pennilesse'. 46Anne Lake Prescott also tentatively ascribes authorship to Nashe, 47 though other critics such as Swapan Chakravorty put the attribution in the past by saying it was 'once attributed to Nashe'. 48Twenty-firstcentury scholarship continues to ascribe authorship to Nashe, though with some uncertainty.For Matthew Steggle, A Wonderful Prognostication 'lurks at the edge of the Nashe canon', reflecting 'how close the conceit of a mockastronomer is to Nashe's general satirical repertory'. 49Walker also treats the text as Nashe's by virtue of its relationship to the Marprelate controversy. 50he attribution of Two Dangerous Comets shares a similar historical trajectory with A Wonderful Prognostication, though the text itself has attracted considerably less scholarly attention.The same early attribution of A Wonderful Prognostication that escaped McKerrow's attention also associates Two Dangerous Comets with Nashe: as reported by Edmond Malone, George Steevens identified Two Dangerous Comets as 'by Nashe, in ridicule of Gabriel Harvey' in his commentary on Shakespeare's Henry IV, Part Two. 51Had he known this, McKerrow may not have proposed Anthony Munday as Two Dangerous Comets' likely author on the scantest evidence: 'joking on the days of the week, in which Monday is said to be the best of all'. 52Consequently, Two Dangerous Comets was excluded from the edition and mention of it limited to establishing a chronology-since Smellknave opens his 'Epistle to the Reader' by describing 'Adam Fowle-wether' as his 'familiar friende' and crediting his 'learned Astonomicall discourse' as the impetus for his own offering, 53 it must have appeared after A Wonderful Prognostication-and glossing an allusion to a ballad in Nashe's Have with You to Saffron Walden. 54he text has never appeared in a modern edition.Ironically, McKerrow's attribution of Two Dangerous Comets to Munday seems also to have gone largely unnoticed, 55 with what little discussion the pamphlet has garnered content to treat it as an anonymous work.Instead, critical conversation has focused almost exclusively on Thomas Middleton's revision of Two Dangerous Comets into The Penniless Parliament of Threadbare Poets, 56 'changes and additions' which, although 'quantitatively marginal', nonetheless 'converted the relatively obscure text of Smellknave into a work which had at least six editions between 1601 and 1649'. 57hile stylistic similarities between the 1591 mock-prognostications have been observed, 58 only Donna N. Murphy has gone so far as to make a positive attribution based on phrasal matches in Two Dangerous Comets and Nashe's other acknowledged works. 59In short, consensus about Nashe's authorship of A Wonderful Prognostication has yet to be reached or, in the case of Two Dangerous Comets, seriously considered.

NASHE, THE HARVEYS AND ASTROLOGY
One of McKerrow's concerns regarding the authorship of Comets is that it was intended as a parody of Richard Harvey's Astrological Discourse, 60 arguing that 'a parody of an ephemeral production which had appeared eight years earlier would be but a pointless jest'. 61Indeed, parodying a text that was 8 years old does seem small-minded and pointless.But this is Nashe we are talking about, from whose pen flowed 'mortall Aconite' to his enemies, 62 and to whom the lapse of 8 years would be nothing were his satire and ridicule still to be recognised.Richard Harvey's Astrological Discourse is a prediction of the effects of the Great Conjunction of 28th April 1583, a conjunction of Saturn and Jupiter which only takes place approximately once every 20 years and so of some astrological significance.This particular Great Conjunction was 'of a very rare and significant kind' because these two planets, the highest planets, were to be conjoined not only in a new sign but also in a new triplicity or trigon as they moved from the watery to the fiery trigon, marking the end of an 800-year astrological cycle and the birth of a new one. 63Richard's predictions, then, were less likely to be forgotten than for a more pedestrian astrological event.Prophecies about the end of the world were common and people's fear was all too real.Richard Harvey 'went far beyond the ordinary discretion practised by any sensible astrologer', 64 however, and the wording of his prediction seemed to suggest that strong winds would begin to blow at exactly 12 noon on 28th April: The 28. of April being Sunday, about high noone, there shal happen a very greate and notable Conjunction of the two superiour and weightie Planets Saturne and Jupiter, which Conjunction shall be manifested to the ignorant sort, by many fierce & boysterous windes then sodenly breaking out & continuing certain daies before and certaine dayes after the same Conjunction. 65ven the heightened anxiety surrounding the Great Conjunction of 1583, Richard's confident assertion quickly became ludicrous when nothing happened.Not even a breeze, auspicious or otherwise.7 Thus, an astrological prediction of 8 years previous was still worth satirising and was very much alive within cultural memory, including erroneous predictions which, previously feared, became the source of amusement and ridicule.Richard Harvey was clearly still an easy target when Nashe penned his mock-prognostications.
The link between the Harveys and prognostication in the public mind was no doubt refreshed when Richard's brother, John Harvey, published an almanac in 1589. 68John had previously entered the fray to clarify Richard's prognostication from anticipated misconception, producing an Astrological Addition prior to the Great Conjunction 'to adde vnto [Richard's] Astrological Disco [u]rse, what I (vpo[n] some co[n]ference) thought, might reasonablie be demaunded, as therein requisite'. 69Margaret Aston suggests that 'it was clear that a main motive for the Addition was the defense of Richard from his critics, to stop the mouths of his envious and carping enemies'. 70With the aid of horoscopes, notably missing from Richard's pamphlet, John's Addition provides further support for the conjunction 'of the two most waightye Planets, Saturne and Iupiter, in the ende of the Trigonisme, namely in the third and last face of Pisces, and the 21.de[…] the same signe.Anno 1583. the 28.day of Aprill, a little before high […]'. 71Though the text is obscured (see Fig. 1), John is referring to high noon or midday, which corroborates his brother's prediction.
Unfortunately for John, this diagram was replicated on the title page of Smellknave's pamphlet (Figs. 2 and 3), which shows that Nashe was directly engaging with the Harveys' contributions to the astrological predictions for the Great Conjunction of 1583.
The use of this image cannot have been an accident; it serves to create a direct visual link between the two texts while also allowing the 67 Robert Greene, A quip for an upstart courtier (London, 1592; STC 12300), sigs.E3v-E4r.Greene had this passage removed during the printing process, possibly due to the death of John Harvey in 1592.On the textual history of the Quip, see Edwin Haviland Miller, 'Deletions in Robert Greene's A Quip for an Upstart Courtier (1592)', Huntington Library Quarterly, 15.3 (1952) mock-prognostication to imitate the genuine article at first glance.There is also a slight variation between the two extant copies of Comets: whereas the centre of the horoscope in the Bodleian copy is blank, the words 'Twelve a clocke at midnight' are printed in the British Library copy (Fig. 3).This addition to the British Library copy points more specifically to the Harveys' astrological pamphlets and recalls Richard's ill-judged precision in predicting high winds to occur at 'high noone'.Even if the Harveys were not the sole intended target of the mock-prognostications, Nashe's reuse of this horoscope, visually and textually, makes the link clear.Furthermore, John's almanac of 1589 has a horoscope on its title page, an unusual feature for almanacs which were distinctive in their use of black and red type.While it is different to the horoscope in the Addition, by making the unusual visual choice to include a horoscope on the title page Smellknave is also purposefully imitating John's 1589 almanac.That the writer of the mockprognostications knew of John's publication is further confirmed by Foulweather's claim in the opening letter that 'Astrologie is not so certaine, but it may fayle', 72 which seems to be contradicting John's assertation that 'the knowledge of many thynges by Astronomie, doth much good to many'. 7372 Foulweather, Wonderfull Prognostication, sig.A2r. 73Harvey, An Almanacke, sig.B2r.  it is certainly possible that Scarlet or Charlewood could also have printed the lost Fairweather pamphlet and that Wright may have been the unnamed publisher of A Wonderful Prognostication.
John Busby has gained a reputation as a rogue printer among scholars, but Gerald D. Johnson notes that he was 'an exceptionally law abiding member' of the Stationers' Company from his admittance in 1585 until he turned his attention to drama in 1599. 75Johnson describes Busby as a 'successful procurer of authorized manuscripts' during the 1590s-for instance, he acquired Nashe's Pierce Penniless in 1592. 76His publishing of Two Dangerous Comets thus marks an early stage of his relationship with Nashe.In addition to printing for both Wright and Busby, Charlewood also had connections to Nashe, having printed The Anatomy of Absurdity, Pierce Penniless and several of the anti-Martinist texts attributed to Nashe. 77According to MacDonald P. Jackson, Charlewood also printed the unauthorized first quarto edition of Sir Philip Sidney's Astrophil and Stella, for which Nashe supplied a preface. 78Jackson identified Charlewood on the basis of his compositorial habits, including the use of a 'highly distinctive head-piece', a 'decorative six-point star and three dots within parentheses' and the setting of fleurons. 79The same 'highly distinctive head-piece' is present in Two Dangerous Comets, 80 which suggests that Charlewood may have participated in-or at least overseen-the printing of the pamphlet.
If Wright published A Wonderful Prognostication, why did Nashe seemingly engage the services of another publisher for Two Dangerous Comets?Writers in this period often worked with printers and publishers with whom they had existing relationships; that the mock-prognostications may have different publishers might, therefore, suggest they were written by different authors.Without Fairweather's pamphlet or a suggestion as to their identity, it is impossible to know whether they, too, were working with publishers and printers with whom they already had established relationships.One possible explanation is that Nashe enlisted the services of different publishers for the same reason he adopted different pseudonyms in the mock-prognostications: what appears to be 'apparent rivalry', 81 in other words, may be entirely manufactured.As Don Cameron Allen remarks, Smellknave's 'debts' to Foulweather 'are never verbal steals': whenever Smellknave 'follows the tune' of Foulweather, 'he has always the good taste to alter the lyric'. 82Where Allen sees an author 'mindful of the intellectual property of others', we might instead see an artificially constructed tête-à-tête, a playful but no less strategic composition with two characters, but one mind at work.

ATTRIBUTION TESTING
Stylometric analysis strengthens the case for Nashe's sole authorship of both extant mock-prognostications. 83In this section, we describe the procedures and results of five different tests using two different methods now standard in attribution study and which together yield better cross-validation and, thus, more accurate results.Attribution testing of this kind requires a corpus of machine-readable texts from which to generate authorial profiles for each candidate to compare with A Wonderful Prognostication and Two Dangerous Comets. 84To ensure the accuracy of these profiles, only sole-authored, well-attributed texts are included-as far as possible, works of collaborative or uncertain authorship, dubious attribution or questionable provenance must be avoided.A pertinent example is Greene's Groatsworth of Wit, a pamphlet salaciously advertised as the deathbed repentance of Robert Greene, 'Describing the follie of youth, the falshood of makeshifte flatterers, the miserie of the negligent, and mischiefes of deceiuing courtezans', 85 which bibliographical examination and stylometric study have suggested instead to be the work of Henry Chettle. 86  8 Since there is no suggestion of collaboration, our testing proceeds on the assumption that each pamphlet represents the work of a single author.We repeated the tests described below using 2,000-word nonoverlapping segments (with any smaller remainder excluded) instead of whole texts-that is the standard procedure when trying to identify individual authorial shares in a collaborative text-and the results confirmed our earlier findings. 84 it safer to exclude it from our analysis. 88We likewise exclude Pap with a Hatchet, 89 one of the racier anti-Martinist retorts now generally thought to be by John Lyly, but which was 'first and in early times most persistently attributed' to Nashe. 90ince genre is thought to constrain an author's style, attribution testing also typically compares only like with like-plays with plays, non-dramatic verse with non-dramatic verse and so on.For this reason, we limit our corpus to works of non-dramatic prose. 91iven the context of the mock-prognostications in relation to the Martin Marprelate controversy and the feud with the Harveys, we draw our pool of candidate authors from those professional writers known to have produced anti-Martinist tracts or associated with them: Robert Greene, John Lyly, Anthony Munday and Thomas Nashe. 92Table 1 lists our corpus of 35 representative sole-authored, well-attributed prose works by these authors, as well as A Wonderful Prognostication and Two Dangerous Comets, along with their source texts and dates of publication.
Whole texts are not used to construct authorial profiles; instead, certain variables (or so-called features) are selected for their power to discriminate between authors.The features chosen for several methods we employ are counts of the most frequent words across the corpus, of which function words-that is, words that function primarily to express grammatical relationships among other words in a sentence and which carry little, no or ambiguous lexical content-make up the bulk.Because they are essential to the structuring of sentences, function words are among the most commonly used in a language and have become one of the most popular and best-understood features selected for authorship attribution. 93In addition to being high-frequency and ubiquitous, function words appear to be less constrained by context (such as genre or subject matter) and perhaps even 'lie outside the conscious control of authors', 94 which may mean they better reflect a consistent authorial style than lexical or so-called content words.Personal pronouns are an exception: several studies have shown their use often to be correlated with genre and, in the case of literary works, the gender of characters; 95 for this reason, we exclude personal pronouns from our feature selection.
Base transcriptions of the texts listed in Table 1 were obtained from the Early English Books Online Text Creation Partnership (EEBO-TCP); conwere then expanded, and function words were annotated to distinguish between (and thereby enable distinct counts to be made for) homograph forms, such as the adverb and preposition forms of the word in. 96As appropriate, youle was expanded as an instance of you and one of will verb , ins as an instance of in preposition and one of his, and so on.Potentially non-authorial material such as prefatory matter was excluded on a case-bycase basis. 97Since our interest is in individual words, we are not concerned with punctuation; however, to facilitate more accurate counting of lexical words, we used VARD 2, a software application designed to automatically normalise spelling variants in historical corpora. 98As an example, the opening sentence of Greene's Menaphon-'After that the wrath of mightie 94  Ioue, had wrapt Arcadia with noysome pestilence, in so much that the ayre yeelding preiudiciall sauors, seemd to be peremptory in some fatall resolution'-is annotated and normalised following this procedure as 'After thatconjunction the wrath of mighty Jove, had wrapped Arcadia with noisome pestilence, in preposition so adverbDegree much air yielding prejudicial savours, seemed to infinitive be peremptory in preposition some fatal resolution'.Table 1 also lists the total words for each text counted as tokens, including homograph and normalised forms. 99ith the corpus prepared and annotated as described, we use a software application called Intelligent Archive to count the frequency of the top 250 most frequent words (excluding personal pronouns) across the corpus in tokens. 100Since the texts vary in size, the frequency counts are recorded as proportions of total tokens to enable direct comparison.The result is a large table with 37 rows (one for each text) and 250 columns (one for each of the most frequent words, excluding personal pronouns). 101Using this table of wordfrequency proportions, we can calculate stylistic 'distances' between the mockprognostications and the representative prose texts by Greene, Lyly, Munday and Nashe, with the assumption that the author whose texts are closest by this measure is the least unlikely author.A standard method for this task is Delta, 102 which begins by transforming the word-frequency proportions into z-scores (i.e. by subtracting the mean and then dividing by the standard deviation of the proportions for each word across all the texts), and then calculating absolute differences (i.e.ignoring whether the figures are positive or negative) between the mean z-scores for each author and the corresponding z-score in the text of uncertain authorship.The absolute differences are then combined to produce a composite measure of difference or 'Delta' distance for each author.
To test the method, we conduct 'leave one out' cross-validation using Nashe-that is, we conduct a series of tests, treating each Nashe text in turn as if it were of unknown authorship and, using the remaining corpus, calculate Delta distances for each author from that text.Table 2 gives the results of the cross-validation, with the lowest Delta distance in each run 99 Words are often counted as unique forms (types) or as concrete instantiations of those forms (tokens).For example, 'the cat sat on the mat' contains five word types (the, cat, sat, on and mat) and six word tokens (i.e. two instances of the, and one instance of all remaining word types). 100 With these results, we can confidently exclude Lyly and Munday as candidates; however, further testing using a different method and feature selection is necessary to independently confirm the greater likelihood of Nashe's authorship.'It makes sense that writers have preferences for some words, and a tendency to neglect others', and these authorial habits extend beyond the (possibly unconscious) use of function words and other very common words   Following the same process of feature selection and counting as before, we generate proportion counts for the next 250 most frequent words in the mock-prognostications and texts by Greene and Nashe.These new features are not as ubiquitous as the top 250 most frequent words analysed above nor uncommon enough to be exclusive to one author or another and, unlike the top 250 most frequent words, the vast majority are lexical (e.g.words such as blood, counsel, doctor, gold and melancholy).The resulting table, with 30 rows (one for each text) and 250 columns (one for each of these words), 106 is imported into R, a software environment for statistical computing. 107We then use the 'randomForest' package to train 20 forests of 500 decision trees on the texts of Greene and Nashe with which to classify Two Dangerous Comets and A Wonderful Prognostication. 108In the first run, Random Forests trains 500 trees which correctly classify 26 out of the 28 texts of known authorship during the training process, giving an expected error rate of 7.14%, before assigning both Two Dangerous Comets and A Wonderful Prognostication to Nashe.For illustrative purposes, Fig. 4 gives an example of a single decision tree generated during this run: if a text contains the word ere with a proportion equal to or greater than 0.0373 and a proportion of gave greater than or equal to 0.0230 or a proportion of ready greater than or equal to 0.0348, then the tree votes for Nashe as the author, and so on.
Table 4 summarizes the results of the 20 runs, giving the number of variables tried at each 'split' by the 500 decision trees, the expected error rates (as a percentage of misclassification), predicted classifications of the Greene and Nashe texts and the outcome classifications of Two Dangerous Comets and A Wonderful Prognostication.The Random Forests algorithm consistently classifies A Wonderful Prognostication as a Nashe text, while attributing Two Dangerous Comets to Nashe in all but 4 of the 20 runs (i.e.80% of the time).Although none of the training classifications is perfect, the expected error rates are relatively low, with a maximum of three training misclassifications in any given run: the forests frequently misclassify The Anatomy of Absurdity (all runs) and the Preface to Menaphon (all except for runs 3, 11, 14, 17 and 20), and infrequently misclassify The Unfortunate Traveler (runs 6, 7, 13 and 18) and 3 Cony-catching (run 2).Further investigation into these texts may reveal reasons for their repeated misclassification, but such speculation is outside the scope of the present study.That the forests are evidently more likely to misclassify Nashe texts should allow us to place more certainty in those Nashe classifications that remain.Bearing in mind that every run generates 500 decision trees, each using a random and different sample of the data, the relatively high level of consistency and accuracy in these findings independently confirms the greater likelihood that Nashe authored both mock-prognostications.

CONCLUSION
With the exception of George Steevens, whose remarks have seemingly gone unnoticed, scholars have not seriously entertained the possibility that Foulweather and Smellknave are one and the same person.Rather than a satirical partnership, collaboration, or rivalry using a popular form, we are left with Nashe writing on his own, playing with form and his own authorial identity as well as establishing relationships with different publishers.Without the third pamphlet by Fairweather, it is impossible to know whether he authored a trilogy of mock-prognostications in 1591 under three separate pseudonyms, perhaps in further mockery of the three Harvey brothers.Attributing both A Wonderful Prognostication and Two Dangerous Comets to a single author in Nashe also alters our framework for interpreting them both: rather than seeing Two Dangerous Comets as a conscious response to A Wonderful Prognostication, it shows the same
Furthermore, Greene targeted Richard Harvey in the first edi-of his A Quip for an Upstart Courtier (1592), a year after the mockprognostications under discussion.In Greene's pamphlet, a ropemaker-the occupation of Harvey senior-discusses the careers of his three sons, the second being 'a Physitian or a foole, but indeed a physitian, & had proued a proper man if he had not spoiled himselfe with his Astrological discourse of the terrible coniunction of Saturne and Iupiter'.
These findings have not gone unchallenged; 87 since the true authorship of Greene's Groatsworth of Wit remains unclear, we deem 81 McKerrow (ed.), Works of Nashe, V: 139. 82Don Allen Cameron, The Star-Crossed Renaissance: The Quarrel about Astrology and Its Influence in England (Durham: Duke University Press, 1941), 222.

Table 1 A
Wonderful Prognostication, Two Dangerous Comets and representative wellattributed, sole-authored prose works by Greene, Lyly, Munday and Nashe.
Centre for Literary and Linguistic Computing, University of Newcastle, Australia, Intelligent Archive, .According to Table2, Nashe has the lowest Delta distance in every run-in other words, the method has correctly identified Nashe as the least unlikely author when holding out and treating one of his texts as if it were of unknown authorship.Satisfied that the method is sound, we repeat the procedure to calculate the Delta distances of Two Dangerous Comets, and then A Wonderful Prognostication, from texts by Greene, Lyly, Munday and Nashe.The results of the test suggest that Lyly is the most unlikely author of Two Dangerous Comets and A Wonderful Prognostication, with Delta distances of 350.15 and 374.38, respectively.Munday is the next most unlikely author, with Delta distances of 342.84 and 348.66.Of the remaining candidates, Nashe emerges as the least unlikely author with Delta distances of 288.59 and 301.49, versus Greene's 327.68 and 312.36.To test these results, we conduct a 'leave one out' cross-validation using the whole corpus, excluding each text from the analysis in turn and repeating the procedure.Table3gives the results, in which Nashe scores the lowest Delta distance from Two Dangerous Comets and A Wonderful Prognostication in all 35 iterations. shaded

Table 2
Delta distances between a hold-out Nashe text and four candidate authors as represented in Table1, using the top 250 most frequent words (excluding personal pronouns).

Table 3
Lowest Delta distance in each run shaded.

Table 3 (
104tinued)to include less frequent but 'strategically chosen lexical words'.103Toconsidertheselessfrequent(but no less characteristic) words, we use a machine learning method called Random Forests,104a classification technique for bioinformatics that has since found success in stylometric analysis and authorship attribution testing.105RandomForests begins by constructing binary decision trees (i.e. a series of Yes/No decisions leading to further decisions or a predicted classification), each derived from different-and random-samples of the data.Hundreds of such trees are constructed (hence 'Forests'), with each tree contributing a single 'vote' to the outcome classification by majority.Roughly a third of the data is withheld from this 'training' procedure to test the predictive power of the decision trees and to calculate an expected error rate when classifying new, unseen data.By design, this process mitigates against the problem of 'over-fitting' the classifiers to the training data and the need for cross-validation.

Table 4
Results of 20 Random Forests classifications of Two Dangerous Comets and A Wonderful Prognostication between Greene and Nashe using the second 250 most frequent words (excluding personal pronouns).