“A good Receipt to purge”: The Medical Recipes in London, Wellcome Library, MS 8086

Recipes, which have a primary instructional focus, are frequently encountered in scientific texts in the English vernacular written in all periods of the language. Research on recipes in English historical texts has mostly accounted for the linguistic and the structural patterns that they show, often intending to characterising not only recipes themselves, but also the type of text that includes them and/or the potential audience of the text. In line with this research, this article focuses on the recipes in London, Wellcome Library, MS 8086, which holds an Early Modern English recipe collection. The collection is analysed with particular attention to medical recipes, and its features are explored to put it into the broader perspective of recipes and recipe collections in the Early Modern English period. These findings are in turn used to examine the potential readership and character of the text.


Introduction
Recipes have long formed part of the English vernacular writing tradition.Notwithstanding this, they have not changed much across time, as Görlach remarked in his pioneering contribution about the cookery recipe (1992: 756).A common trait to all recipes (regardless of their type: medical, culinary, magical, etc.) is their instructional purpose (Taavitsainen 2001: 86;Carroll 2004: 187;Quintana-Toledo 2009: 24), insofar as guidelines are offered to prepare medicine, a meal or some other utility.Recipes have been usually analysed from a genre and/or text-type point of view, a perspective that is adopted for the scope of this article, since attention is paid to the linguistic traits (i.e.text-type features) as connected to the structural patterns and stages (i.e.genre peculiarities) displayed. 3In doing so, we follow previous research like that carried out by Alonso-Almeida (2013) and de la Cruz Cabanillas (2016,2017) on Early Modern English (EModE) recipes, and also Carroll (1999) and Taavitsainen (2001) for the Middle English (ME) period.
The focus of this article lies on medical recipes, which are attested since the Old English (OE) period (Carroll 2004: 175) when they typically appeared in remedy books.Later on, in ME, some distinctive features were found between recipes in learned and surgical treatises on the one hand, and in remedy books on the other (Taavitsainen 2001: 106-107;Carroll 2004).In EModE, some differences are also perceived according to the potential audience of the text, i.e. whether they are intended for lay or learned audiences (Mäkinen 2011;Marttila 2011).
It must be added that whereas medical recipes in ME have received considerable attention (Jones 1998;Carroll 1999 and2004;Alonso-Almeida 1999-2000;Taavitsainen 2001;Quintana-Toledo 2009;Marqués-Aguado 2014), research on EModE medical recipes still has much to offer, since "[r]elatively little attention has been paid to recipes from the […] Early Modern English period, especially to the analysis of their structure" (Bator and Sylwanowicz 2018: 103).For instance, some corpus-based studies have been conducted using the Early Modern English Medical Texts corpus (hereafter, EMEMT) that have addressed efficacy phrases (Mäkinen 2011), interpersonal strategies (Marttila 2011) and the structure of recipe headings (Bator and Sylwanowicz 2018).Alonso-Almeida has carried out another corpus-based study to determine the structure of Modern English recipes (including medical, culinary and magical ones) (2013: 69), and de la Cruz Cabanillas (2016,2017) has surveyed the structure of EModE recipes in various recipe books written by women.
A final remark should be added concerning the format of the EModE recipe materials used in such studies, since some draw on printed material and others on man-3 Taavitsainen has again recently accounted for the concepts of genre, text-type and register and applied them specifically to recipes.Accordingly, she remarks that, while the abbreviation for recipe or the opening verb form take are "enough to trigger expectations of a text belonging to a genre whose function is to instruct in preparing something", "text type features include imperative forms of verbs, measurements, and an optional efficacy part that may be realized in various ways" (2016: 275).See also the descriptions in de la Cruz Cabanillas (2016: 88) and Bator and Sylwanowicz (2017: 21-22).uscript evidence.There are apparent differences between both formats since at the time printing fostered a standardisation and fixation that handwriting did not facilitate that much.The EMEMT (Taavitsainen, Murray Jones, Pahta, Hiltunen, Marttila, Ratia, Suhr and Tyrkkö 2011), for instance, is compiled for the most part from printed rather than manuscript material.In turn, Alonso-Almeida's study is based on evidence extracted from the Corpus of Early English Recipes (CoER), which draws from both manuscript and printed sources (2013: 69), while the analyses by de la Cruz Cabanillas are described as being entirely based on manuscripts from various sources (2016: 80; 2017: 13).
In the light of this, in this article the medical recipes contained in London, Wellcome Library, MS 8086 (hereafter, W8086) are analysed regarding their structural and linguistic features.The text held in this manuscript forms part of The Málaga Corpus of Early Modern English Scientific Prose project,4 which is compiled from transcriptions of EModE texts.Being a manuscript-based study, our findings will contribute to expanding the information available on recipes in this format.To fulfil our objective, the text, which is anonymous, is first described regarding the sources it shows and its possible authorship (section 2), as this might be relevant to understand certain features in the configuration of recipes.Then, the structure of the recipe collection is discussed, and the types of recipes included are presented (section 3).Next, section 4 examines the recipe elements (or stages) alongside the linguistic patterns linked to them.Finally, section 5 presents the conclusions.

Authorship and Sources
W8086 is catalogued in the online Wellcome Library Catalogue as an early to mid-17 th -century recipe book that contains mostly medical but also veterinary and household recipes (see section 3 below).5Such mixture of recipes has been reported to be a typical trait of remedy books or materia medica (e.g.Mäkinen 2011: 160).
The manuscript, which is in overall good condition, contains three front flyleaves, 247 pages with recipes, a 16-page partial index and a couple of blank leaves at the end.The fact that it is classified in the Library Catalogue as an anonymous MS is a common situation encountered with many recipe books and collections, as also referred by Herbert (2014: 105).Several arguments may be used to suggest that a woman could have been the author of the volume.First and foremost, there is a wealth of recipes that tackle women's complaints (as indicated in the online Wellcome Library Catalogue) or that account for other women-related issues, like boiling ladies' cornets (p.242), plus others that place children as the patients to whom the remedies are addressed.We cannot forget that in the EModE period "the most important locus for medical care was the home, and women were the principal providers of remedies" (Taavitsainen, Murray Jones, Pahta, Hiltunen, Marttila, Ratia, Suhr and See http://archives.wellcomelibrary.org/DServe/dserve.exe?dsqIni=dserve.ini&dsqApp=Archive&dsqDb=Cat-alog&dsqCmd=show.tcl&dsqSearch=(RefNo==%27MS8086%27).Tyrkkö 2011: 20). 6Other arguments supporting female authorship are, first, that the recipes in W8086 are not surgical, but can rather be easily circumscribed to the household sphere (due to the ingredients or the instruments that are used); and, second, that various women (such as a Lady Rayes, a Lady Shrewsbury or a Mrs Barrett, to mention but a few) are also cited as sources for recipes to prepare waters, oils and other remedies.
The online Wellcome Library Catalogue further indicates that the author was "possibly of the aristocracy", and in connection with this Herbert suggests the name of a Lady Baesh as the possible compiler of W8086 (2014: 109). 7There are no marks of ownership in the codex stating her authorship (or any other, for that matter), but this possibility should not be underscored, as her name appears among the authorities cited in the compilation of W8086.In fact, two recipes on pp.172 and 173 read "by myself M. B.", initials that could correspond to Mrs M. Baesh.It should not be forgotten, though, that first person references could have been borrowed along with the recipe from the source from which it was copied -the use of "I" is also attested, for instance, on p. 99 with "J Thomas Potter". 8  This is tightly connected with the use of sources and authorities -quoting a variety of sources and medical authorities was used as a strategy to reinforce the reliability of the remedies put forward and to convince readers (Marttila 2011: 148-149).In W8086 a wide range of authorities are referred to not only in the titles of the recipes (as accounted for in section 4.1.2)but also as marginal notes and within the body of recipes (see also section 3).Indeed, certain blocks of recipes are labelled in the manuscript as taken from various sources, including Mr Isham's book (recipes on pp.105-157) or Lord Connaway's book (recipes on pp.[185][186][187][188][189][190][191][192][193][194][195][196][197][198][199][200].This complies with Marttila's statement that "[m]any of the recipe collections are not original works but rather derivative collections of remedies acquired by the 'author' from various sources" (2010: 102).

The Recipe Collection
In W8086 there is a total of 775 titles, or headings, which would mean an equal number of recipes in the collection.Yet, titles alone cannot determine the number of recipes due to various reasons.First, nine recipes lack a title, although other cues may be used to argue that a heading is missing (e.g.there is a space wider than usual 6 References to women and medicine in this period are manifold.According to Leong, women were viewed "as medical caregivers in many early modern households" (2014: 559).Likewise, Wear highlights the importance of "remedies in the female culture of medicine" (2000: 47) and Slack states that remedy collections were used especially by women (1979women ( : 260, in Marttila 2011: 137): 137).Nonetheless, Hunter indicates that these recipe collections were "frequently shared by many members of a household" (2009: 20).

7
Assuming an aristocratic background might derive from the fact that at this time "[t]here is a clear shift of interest from more elite professional writing to the literature that reached the heterogeneous 'ordinary' readerships, new women audiences and 'the poor'" (Pahta and Taavitsainen 2011: 5).In other words, learned publics increased, including women with "increasing literacy" (Taavitsainen and Pahta 2011: xviii).

8
The examples offered are taken from The Málaga Corpus of Early Modern English Scientific Prose website (http://modernmss.uma.es).They preserve the original manuscript punctuation and line division, spelling (including the positionally-conditioned alternation of <u> and <v>, for instance) and capitalisation, while italics stand for expanded abbreviations and / \ enclose insertions.The only exception in this article to the reproduction of original line division is the set of short quotations from the text (unnumbered examples), where MS line division is indicated using the vertical bar |.
with regards to the previous recipe, recipe elements start anew, marginal annotations indicate a new recipe, etc.).Second, not all titles signal the beginning of a new recipe.It is occasionally the case that the ingredients, the preparation, the applications or the virtues are presented in separate paragraphs, each of them carrying a title of its own.This happens with 26, typically long, recipes, like "The Manner how to Compose á Choyce| Artificiall Balsome./" and "The manner to Compose the| said Balsome./" (p.203) and "The Vertues of the said Artificiall Balsome| and the manner of Application thereof./"(p.204).This means that a total of 29 titles need to be detracted from the initial figure.Finally, in 4 cases one single title subsumes various remedies that cater for the ailment under discussion, which adds other eight recipes.Therefore, although there are 775 titles, the total number of recipes is 763.
As stated above, the collection in W8086 gathers mostly, but not exclusively, medical recipes, since household and veterinary recipes are also found.According to Mäkinen, medical recipes "need to fulfil three criteria: they must have 1. a passage stating the recipe's medical purpose; 2. a passage specifying the ingredients; 3. an instance of a VP in the imperative denoting take (or a word to that effect) either in Latin or in English" (2006: 23;cited in Mäkinen 2011: 160).This set of criteria allows him to distinguish recipes proper from "recipe-like passages […] or other renderings of recipe information, which have been called recipe paraphrases elsewhere" (Mäkinen 2011: 160).Using his criteria, up to 97 recipes (out of the total of 763 in the collection) would not count as medical (thus leaving 666 medical recipes, and 687 medical recipe headings), either because they lack reference to the purpose they serve, or because there is no passage including verbs in the imperative.Some of these recipes actually address the preparation of a water or oil that may ultimately be used in a medical remedy, but which is not formulated as a remedy.Others are related to household elements (preparing a bath, for instance, as on p. 219) or to veterinary issues (as with the recipe for lame sheep).On the other hand, some recipes only include verbs in the passive, as the recipe to wash a tetter, which reads "Prym leaves.Allom.boyld together in Running water./"(p.128).
The medical recipes in W8086 do not follow any particular order, either by the condition addressed, the body part in need of healing, the type of patient, etc. (see Marttila 2010: 103); not even alphabetically.Thus, the classical and widely used capite ad pedem order is not observed. 9It is true that successive recipes tackling the same problem are commonly found, such as those with headings like "another for the same", which appears, for instance, on pp. 1 and 2 with four consecutive recipes against ague.However, recipes addressing the same condition may also be found interspersed throughout the codex and without cross-references.Such is the case of those that specifically address stone according to their titles, which are found on pp. 4 (two recipes), 11, 50 (two), 54 (three), 56, 57 (two), 60 (three), 104, 106, 107 (five), 108, 109 (two), 123 (two), 124 (three), 138, 145, 158 (three), 159 (three) and 177.Some recipes are even repeated, as with that for the "falling sickness", which appears on pp.135 and 136 with very few differences regarding both content and linguistic form.This is all suggestive of a continued habit of noting down recipes, with no pre-established order or arrangement.As Alonso-Almeida states, "[t]he recipe book is conceived as an artefact likely to be updated to meet the needs of their users" 9 See, for instance, Carroll (2004: 184) for an analysis of this ordering in ME texts.Notice however that, according to Taavitsainen, this arrangement "applies to a limited extent only" (2001: 106).
(2013: 69).In quite similar terms, de la Cruz Cabanillas describes the recipe collection as an "active, dynamic compilation" (2016: 82).This would, in turn, point at W8086 being a lay, rather than a professional text, an option that ties nicely with the use of common household ingredients and the absence of surgical procedures.More arguments along this line are provided below.

Analysis of the Recipes
To carry out the analysis of the recipes in W8086, the prototypical structure of recipes during the EModE period has to be first reviewed.As Alonso-Almeida claims, "the majority of stages follow the medieval fashion both in form and content, although the recipes have undergone some formal update to include new ways of thinking and new aspects concerning cultural and social behaviour" (2013: 72).Medieval recipes normally followed Stannard's pattern of so-called recipe constituents (1982: 60-65), which includes first necessary constituents (purpose, ingredients, procedure and equipment, application and administration) and then optional ones (rationale and incidental data).Some modifications to this configuration need to be introduced to respond to the changes suggested in EModE, as shown in Figure 1: Differences with regards to Stannard's classification are apparent since only the ingredients are now obligatory -that is the reason why this element is not enclosed in parentheses.Additionally, the last two components are formally added here; according to Alonso-Almeida, "[t]hese stages are not new to the Modern English period, but they have received little scholarly attention" (2013: 72).Moreover, as the asterisk indicates, the order of all stages is variable, with the obvious exception of the title.Hence, our analysis follows Alonso-Almeida's proposal outlined above, and the analysis of its linguistic features accompanies the description of each stage. 10Since the recent literature has focused more specifically on the linguistic traits and/or formulae of two of these stages (titles and evaluation/efficacy), special attention is drawn to such features by adding complementary quantitative evidence from the text.Comments are also added as to features that deviate from expected and reported patterns (e.g. the vagueness or imprecision attested in some stages).

Title (heading)
The title, or heading, conveys typically the content and purpose of the recipe found beneath, and in doing so it helps to visually separate recipes, which, as mentioned 10 Contrariwise, Marqués-Aguado (2014) and de la Cruz Cabanillas (2016) analyse recipe elements and linguistic traits separately.
above, are usually presented as one-paragraph blocks (i.e. each paragraph contains all the stages).By the same token, from the reader's point of view, the heading potentially facilitates the search for the intended recipe.As discussed in section 3, headings are almost systematically used in W8086, with only nine exceptions, where this absence is not always compensated by means of cues such as a dividing line, as Alonso-Almeida occasionally found in his study (2013: 72). 11Attention is paid in this section not only to the linguistic patterns used in headings but also to the means used to strengthen the validity of the remedies they head, which include evaluative adjectives and references to authorities.

Linguistic patterns
As with titles in the ME period (Taavitsainen 2001: 99;Carroll 2004: 181), recurring linguistic patterns in the EModE have been categorised.The headings of all the recipes in W8086 have been first classified according to Alonso-Almeida's taxonomy (2013: 73), which identifies up to 8 different patterns.These are listed in Table 1 along with the frequencies of each in W8086 and a couple of examples: In their study of medical recipes, Bator and Sylwanowicz found that only around 5% of the recipes analysed lacked a heading (2018: 109).In W8086, this represents less than 2% of the total of recipes.
The figures in Table 1 confirm a clear tendency towards the use of patterns including for and a NP (fourth and sixth in Table 1), while others are mostly neglected, especially for to + infinitive + NP, for followed by a verb in the -ing form, and how to + infinitive + NP (which amount to 12× in all). 12Some of these results agree with those obtained by de la Cruz Cabanillas (2017: 16).The general tendency shown here is then a clear preference for nominal rather than infinitive structures.
Notwithstanding, the linguistic patterns in Table 1 do not cover all the structures in W8086: the frequencies add up to only 606 (total of headings: 775).Many of the missing patterns in Table 1 tend to be longer (and, thus, more informative), as Alonso-Almeida also suggests regarding 17 th -and 18 th -century recipes (2013: 73).
Attention is now focused on the headings of medical recipes only (amounting to 687 in W8086, as explained above).For the purposes of our analysis, these are categorised according to the patterns used by Bator and Sylwanowicz (2018: 109-116), who have recently analysed the structure of recipe headings (both medical and culinary) in EModE recipe texts taken from EMEMT.Their typology of possible structures is more varied and specific, and is, therefore, more likely to adapt to the reality of W8086.
According to their classification, headings can be divided into titles (i.e."the name of the medicament to be prepared or the name of the ailment to be cured") and statements of purpose (SP), which "specif[y] the health problem for which a prescribed medicament is to be used" (2018: 110), although a third set, labelled 'other', groups whatever patterns are not amenable to this classification, including the relatively frequent heading "Another".Whereas titles are only realised as noun phrases (NP), a variety of patterns can be found for SP: prepositional phrases (PP) (labelled here SP1), noun phrases (SP2), infinitives (inf.)(SP3) and clauses (SP4).
Our results for the three types of headings correlate nicely with theirs, as shown in Table 2. Table 3, in turn, shows the same agreement when exploring the internal structure of titles and SP, since there are only minor divergences in the slightly higher percentages of NPs and 'other', and the lower rate for infinitives, a tendency that goes in tune with the results in Table 1: By the time this recipe collection was written, the for to infinitive was already obsolete (Rissanen 1999: 288)."Doctor Turners powder for the Rume| to sprinckle on the mould of the head."(SP2.ivplus inf.)

Mechanisms to strengthen the validity of the recipe
As Alonso-Almeida has remarked, headings may also contain evaluative elements, which include adjectives like excellent or good (preferably the former in the EModE recipes he assesses) and references to sources, although he further adds that "[t]here is not a fixed rule concerning the frequency of all these evaluative elements in our corpus" (2013: 74).
13 Structures such as this or "To make One sleepe."(p.113), conform to the most common pattern for causative make followed by the object and an infinitive, which is the bare infinitive (Rissanen 1999: 287).The use of the bare infinitive, according to Rissanen, became more restricted after the 16 th century (1999: 286) .
In the case of W8086, 123 out of the total of 775 headings (115 out of 687 medical ones) contain evaluative adjectives, which is a substantial amount (approximately a sixth of totals): good and best (36×), along with very good (12×); excellent (27×) and very excellent (1×); approved (13×); special (6×); easy (4×); sovereign (4×); comfortable (3×); precious (3×); rare (2×); gentle (1×); and very true (1×). 14These figures indicate that the frequencies of use of good and excellent rank the highest, although the former is more usual than the latter, against Alonso-Almeida's evidence.This might suggest that these recipes do not completely fit mainstream EModE patterns, by showing either more innovative options or more traditional, ME patterns.
Additionally, two of these adjectives appear in the superlative, thus intensifying the quality conveyed: most approved and most excellent (1× each).The same effect is achieved when using two evaluative adjectives together: excellent and good co-occur once, and excellent and approved twice.There are even more copious or elaborate formulations, like "Mister Bonds potion for the shapenesse of| Vrine.and often making water with great| difficulty.by which he found wonderfull| ease."(p.81), "The flower of Oyntments./"(p.155) or "A medicine which never fayles" (p.243).Some occurrences of these evaluative adjectives form part of structures with multiple adjectival premodifiers.According to Tyrkkö, such multi-adjectival sequences in EModE could be used "as a means of giving weight to emphatic and affective arguments", and that non-professional writers were more prone to using such sequences compared to professional ones with the intention of "evoking strong responses in readers" (2014: 158-159).In W8086, such structures (of the type showing two adjectives and a noun), only amount to 6× in recipe headings, as with "excellent good water" (p.175), "Comfortable purging Broath" (p.175) or "Excellent Burnt wyne" (p.207).Three of the six instances feature the adjective excellent.There are three additional cases in which one of the adjectives precedes the head noun, and the other follows (the so-called 'ambilaterial placement' that was still possible in EModE [Rissanen 1999: 208]).
In turn, names of sources occasionally feature in recipe titles, which are typically intended to add prestige to the remedies presented, as mentioned in section 3. 34 different people are mentioned only in headings (amounting to 46 citations), plus many others in the course of the recipes and even in marginal annotations.These authorities range from vague nouns like "an Egiptian" (p.46) to common names such as "Adrian Gilbert" (p.199), to doctors like Dr Butler (p. 198) or Dr Tuner (p.21), and to people from the high ranks of society such as "Sir Walter Raleigh" (p.201), to mention but a few.
The use of such authorities can be viewed from the standpoint of the interpersonal strategies and their relation to the potential audience of the text.Such strategies address "the interaction between readers and writers, concentrating on how the writ-14 It is worth noting that the adjective good appears both pre-and postmodifying the head noun, as in "A Medicine good for the Rume.for disgestion| for the spleene.the mother.and against ye| Consumption […]" (p.13), "A Salve made by an Egiptian good| for any wound./"(p.46) or "Mistress Mulsho her Receipt of Stomaticall| pills very good for an ill Stomack./"(p.125).Adjective postmodification was no longer usual in EModE (Rissanen 1999: 208-209) and was normally reserved for special contexts (Fischer and van der Wurff 2006: 123-124), none of which seems to apply here.These examples might rather be understood in the context of the long complementation this adjective may take.
ers more or less overtly assume and assign roles for themselves and their readers and attempt to influence the readers' behaviour" (Marttila 2011: 135).Following Marttila, different categories of authorities may be employed, i.e. classical medical, modern medical, the writer him-/herself, the reader, an 'inclusive we', general public or laypeople, and divine. 15 The authorities mentioned in the headings of our collection belong to only two of the categories in Marttila's study: modern medical authorities and general public or laypeople, with the second more than doubling the first group. 16There are no direct references to the writer (even when this is taken to be one of "the most frequent sources or validators of information" [Marttila 2011: 150]), to the reader or to God (as opposed to the invocation of divine authority in the efficacy section, as shown in section 4.5.1).
As Marttila notes, references to laypeople appear in both professional and non-professional writings, but classical references are only attested in the former (2011: 149-150).As we can see, then, W8086 accommodates to the pattern for lay texts.The group of laypeople includes both men and women (as opposed to the first group) and a wide range of people from different ranks, including royal or aristocratic people along with (supposedly) eminent authorities.In some cases the references take the form of anecdotes or example cases, with modern patients being mentioned (Sylwanowicz 2018: 83), thus strengthening the positive views on the recipe.

Ingredients
This section shows the substances (i.e."herbs, animal and fish products" [Alonso-Almeida 2013: 75]) to be used in the preparation of a remedy.This stage has been said to frequently appear together with the preparation (Alonso-Almeida 2013: 75), and this is also the case of W8086, especially when the preparation implies several steps and each of them requires different ingredients.An example is shown in (1), where a broth is first prepared with certain ingredients, and then it is subsequently mixed with other products: (1) Take the bone end of a neck of mutton or veale.cleane washed.sett it on the fire to boyle in three pints offaire water.and when it is cleane skummed.thenput into it of the Rootes of fennell and parsely cleane wash't scraped and brused. of either three Rootes, of Camomile and Mallowes of each an handfull.Lett all these boyle together till halfe be consumed, then strayne it.take .3.quarters of á pinte of this broath of browne suger Candy . 2 .ounces . of oyle of flexe= seed .2. ounces, mingle all these and take it for á Glister bloud warme.(p.28) In fact, only rarely do the ingredients stand alone under one single title, as in (2): 15 See also Sylwanowicz (2018: 81-83) for study a with a slightly modified (i.e.reduced) typology.
(2) A purging drinke for the winter.Take of sene .7 .ounces.sarsaparilla .5 .ounces, Hermadactillis .4 .ounces.Pollypodie . 2 .ounces, Epithimye .3 .ounces, Red Rubia tinctora . 1 .ounce.Bayberries halfe an ounce, Cinamon too dramms licorish . 1 .ounce, nuttmeggs . 1 .ounce, fennellseeds halfe an ounce, mace . 2 .dramms.Camapitis halfe an ounce./(p.168) The previous examples also show that ingredients (or groups of them) are commonly followed by the amount or quantity required.However, the measures are not always precise, and the same vagueness sometimes extends to the very need of using certain ingredients, as shown in ( 3) and ( 4) below (see also de la Cruz Cabanillas 2016: 86-87 for a similar case).Though not infrequent in many recipe collections, this vagueness might indicate a lack of thorough medical treatment and procedure, thus pointing to a not very professional (or learned) text.Sometimes, the addition of certain ingredients (like sugar) or the variable quantities required seem to be conditioned by good taste rather than by medical efficacy, as in (3): (3) and put to them fine sugerbeaten and searced.as much as you shall think to be enough to please your taste.(p. 5) (4) please) of Ambergreece: you may if you please put alsoe some yellow or white Amber with the Rest of the things aboue said; (p. 196) Linguistically speaking, the imperative verb form take normally introduces ingredients, although other, more specialised verbs are sometimes used (e.g.choose), along with general verbs like put.Regarding the terms referring to ingredients, common substances are found next to other modern ingredients, like sarsaparilla, giacum or chinaroot, which were not imported into England until the second half of the 16 th century (Wear 2000: 71; see also de la Cruz Cabanillas 2017: 17).These lexical choices thus reflect the modern character (and dating) of the text.

Preparation
In this stage, the manner to combine the ingredients is exposed, hence the presence of verbs in the imperative and in the second person present tense, and of adverbs of time (and similar temporal marks) to sequence the steps.Specialised verbs to refer to the procedures are also found (Alonso-Almeida 2013: 76-77).The preparation stage may be given a title of its own, thus being separated from the ingredients and other stages, as in (5): (5) directions for the making thereof./ffirst you must sett your Rosin on the fire.and melt it with á soft fire.and stir it with á stick á little now and then and when it is melted, put your sheepes tallowe into it.then put in your pitch and when they be all melted put in your Labdanum.and then lett it boyle á little together softly.and then take it of the fire.and put in your fran= kensence.then set it on the fire againe vntill it be ready to boyle.Then take it from the fire againe, and put in your Cummin, Cloves, Mace, and saffron altogether, and then set it on the fire á little againe.then take it off.and spread it vpon an vntamed sheepes skynn.thynn: and soe it must be layd to the soales of the feete.and never taken away so long as they will cleave to them./(p.70) Example ( 5) clearly attests to the linguistic traits exposed above: imperative verb forms like take and put prevail over other more specialised verbs (e.g.melt or stir in this passage, but also bathe, stamp or lay at other points of the text), and sequencing is marked by and, then or when.
Some imprecisions and vague language are again occasionally found, as de la Cruz Cabanillas also remarks (2017: 18), as with "á little now" or "á little againe".The phrase "according to Art" (p.221) is another expression that conveys imprecision, or else assumed knowledge on the part of the audience.

Application
Next, the application (i.e.how much of the product should be applied and when/how long this should be done) is provided.This implies the use of body parts and verbs (precise ones) in the imperative, along with "time and metrical units to indicate both dosage and duration of treatment" (Alonso-Almeida 2013: 77), typically in medical recipes.Also, dosage is said to be more specific in EModE than in ME (see also de la Cruz Cabanillas 2017: 18).
Excerpts ( 6) to (9) may serve as examples of how the application is presented in W8086 (see also the last two lines in example ( 5)): (6) When you vse it, putt as much into the Corner of each Eye as á pinns head, and anoynt the Lidds with it, when you goe to bedd; and wash it oft in the morning with fennell water./(p.188) (7) This water is to be snuft into the Head at ye Nose the Head being leaned back; (p.221) (8) Give as much as á Pease at á Tyme./ (p.229) (9) then sweeten it with fine suger ready beaten.and lett the party take it the last thinge shee takes.being an hower after supper, this do three nights together ./(p.58) Despite the reported tendency towards precision at the time, dosage is occasionally vaguely put forward with expressions similar to present-day English "as it pleases you" or "when you think it best", as shown in ( 10) and ( 11): (10) and take thereof halfe á pinte at one time erly in the morning warme, and sleepe after it.ifyou can.and as much .2 .howers before supper.at your pleasure (p.27) (11) and then drinke whatquantity you thinke good at . 4 . of the clock in the afternoone.and when you goe to bedd.and in ye mor= ning when you Rise./ (p.139) It is interesting to note that several recipes highlight different dosage or administration patterns depending on the age or sex of the patient, with a special interest in children, as shown in ( 12) and ( 13).This, as mentioned above, might be used as an argument in favour of the likely female authorship, insofar as women were in charge of their families' well-being: (12) Take .3 .Bees for á man and but one for á Child./(p.114) ( 13) and soe give it at the full of ye moone to Children of .4 .or .5 .yeares of age, and to young Children at ye increase of the Moone, the bignes of á Nutt at á time./(p.220)

Evaluation/Efficacy
For Alonso-Almeida, "efficacy statements have been variously called 'incidental data', 'statements of efficacy ' and 'efficacy phrases'" (2013: 78).Incidental data and the rationale were categorised as two optional pieces of Fachinformation by Stannard (1982) and referred to anecdotes, arguments to reinforce the validity of a remedy, and so on.
Evaluation and efficacy are assessed in W8086 with reference to efficacy phrases (hereafter, EPs) first, and then to other mechanisms that mark the efficacy of a remedy (excluding, of course, the evaluative devices used in titles, which are surveyed in section 4.1.2above).

Efficacy phrases
EPs have been studied in more detail regarding ME medical texts, especially by Jones, who first described them as tags or phrases that "attest to the value of a recipe, and which are found in this final closing position" (1998: 201).As Mäkinen (2011: 169) and de la Cruz Cabanillas (2017: 18) indicate, the EPs frequently repeat the purpose of the recipe.Further research touching on EPs has broadened the scope and considered additional pieces of information as likely EP materials, including case reports or references to sources and authorities (Hunt 1990: 22-24;Taavitsainen 2001: 104-106) or initial EP material (Alonso and Cabrera-Abreu 2002) (all cited in Mäkinen 2011: 161).
For the EModE period, we count on the recent study by Mäkinen, who, following Jones, classifies EPs into stock, specific and general.Stock EPs are phrases that are "highly formulaic and that indicate […] the potency of the medicine without disclosing the type or the name of the malady" (2011: 161).In turn, specific EPs are those that "will mention the name of the malady, or the vocabulary used will indicate its nature" (2011: 162).Finally, general EPs "are akin to stock phrases in the sense that they do not refer to the malady in question; however, nor are they formulaic" (2011: 162).Mäkinen's study only comprises printed materials that are included in EMEMT, thus leaving aside manuscript evidence.He follows Jones' definition and taxonomy but with some modifications, insofar as he takes into consideration not only recipe-final EPs but also recipe-initial and recipe-medial ones.His study shows that lay texts were more likely than learned texts to include EPs (see also Sylwanowicz 2018: 80), although the former experienced a sharp decrease after 1650.Hence, Mäkinen concludes that through the EModE period EPs are generally in decline, although noticeable differences are found not only between lay and learned materials (in the former, EPs gradually "become rarer, more formulaic", while in the latter the number of EPs increases, with specific ones outnumbering formulaic ones [Mäkinen 2011: 177-178]), but also within the EModE period, which he divides into 50-year subperiods.
In order to draw comparisons to Mäkinen's results, his procedure is reproduced here.Determining the types and frequencies of EPs in W8086 will allow us to relate our findings to the kind of text it represents.The results obtained from W8086 show the following distribution of EPs into types (Table 6): Table 6.Types of EPs in W8086.

Type of EP Frequency
Stock 55× Specific 81× General 10× TOTAL 146× As Table 6 shows, 146 of the 666 medical recipes in W8086 contain EPs (almost 22% of the total of recipes), so they cannot be considered an extremely frequent device.From this table we can also gather that specific EPs (exemplified in ( 14) and ( 15)) outnumber stock EPs (illustrated in ( 16) and ( 17)) and especially general EPs (see example ( 18)).De la Cruz Cabanillas also finds that free formulas are more common than set phrases (2017: 18): ( The results obtained by Mäkinen indicate that in the period 1601-1651, when W8086 can be dated to, around 80% of the recipes in the lay texts of his corpus include an EP, as opposed to 20% of those in learned texts; yet, for the period 1652-1700 the figure for lay texts falls to around 30% (2011: 172-173).Therefore, the figures for W8086 place it closer to learned rather than to lay texts, but the distribution into types (as shown in Table 6) does conform to Mäkinen's distribution of EPs in lay texts (with specific EPs ranking first, followed by stock and finally general EPs) (2011: 173-174).
In addition to these EPs, we find nine instances of stacks of EPs, which are made up of two EPs.According to Mäkinen, they "are not uncommon, and the effect achieved or attempted, as with, for example, multiple negation, is one of emphasizing the efficacy statement" (2011: 170) -a similar effect to that observed regarding superlatives and stacks of evaluative adjectives in titles (see 4. As shown in ( 19) and (20) above, Latin is still used in certain EPs in the period, especially in so-called 'proof'-phrases; i.e. "translations or derivations of the Latin probatum est" (Mäkinen 2011: 168).Latin is more frequent in lay than in learned texts, too, although in this period there is a clear decay in the use of this language (Alonso-Almeida 2013: 80).In W8086, only 14 stock phrases (whether isolated or in stacks) are rendered in Latin, thus showing that it is not the preferred option for stock EPs either.
A last note should be added regarding EPs: the invocation of divine authority (i.e.God).Following Alonso-Almeida, "references to divinities are in decline after 1600" (2013: 79; see also de la Cruz Cabanillas 2017: 18), and in any case such references are more typical of lay than of professional texts (Sylwanowicz 2018: 83).In W8086, only 12 EPs refer to God, as shown in (20) above and also (21), the latter of which includes a specific EP: (21) and by Gods grace it shall breake it with out any tent and so heale it./(p.71)

Other mechanisms to convey the validity of the recipe
Another mechanism that confers validity to the remedies exposed in W8086 is the reference to other medical case reports or anecdotes, as shown in ( 22) and ( 23), or, following Taavitsainen regarding efficacy in ME recipes, "[a] precise time when the healing effects could be anticipated, or more precise conditions on achieving the desired effect may be given " (2001: 104).This is also the case of W8086, where information as to when or how a remedy is best prepared or ingredients collected, or when results should be expected, is occasionally used, as with ( 24) to (26).Time to expect results is linked to the presence of an EP, as shown in ( 26): (22) Also J Thomas Potter have found by often expe= rience that it is á singuler and speedy helpe for Bones out of Joynt.(p.99) (23) This hath helped them that hath had the Sea=water in their stomacks and bruised by forte of the Sea./ (p.149) (24) Jt must be made in the moneth of May/ (p.38) ( 25) and so lay it to the small of the back.where the kydneylyes.and weare it á weeke.refreshing it when it is drye./ Theis hearbs must be gathered in May. and stamped and strayned and kept in á Glasse./(p.106) (26) let him chafe á little of it betweene his hands, and smell thereto, and within 4 .dayes it will be whole./(p.117)

Storage and Expiry Date
These refer to "the manner of keeping the finished product for future use" and "how long the finished product can be preserved for use", respectively (Alonso-Almeida 2013: 80).Both Alonso-Almeida and de la Cruz Cabanillas (2017: 18) agree that the latter is more frequently attested than the former.
A substantial number of recipes mention storage and expiry dates, as exemplified in ( 27) to (30) below.As (27) reveals, dates of expiry are not always very precise, and can be reduced to the formula "keep it [for your use]" (Alonso-Almeida 2013: 81): (27) then strayne it into some cleane vessell and so keepe it for your vse./ (p.20) (28) and so straine it and it will keepe till that tyme twelve moneth./(p.156) (29) This being so vsed will continue good for twenty yeares./(p.163) (30) and soe keepe it in tynn potts./(p.211) As these examples show, vocabulary relating to containers and to the passage of time is recurrent to convey storage and date of expiry, respectively.

Virtues
This relates to "the uses of a particular product", hence the vocabulary related to diseases, adjectives presenting positive and/or negative qualities and sentences in the present tense featuring It is good for or This serves for, indicated by Alonso-Almeida (2013: 81).The same author further indicates that the main difference between virtues and EPs lies in the former's list format (2013: 81).
As mentioned in section 3, virtues sometimes appear separated from the rest of the recipe, under a different heading, as de la Cruz Cabanillas also found in her corpus (2017: 19).The list format is evident, for instance, in the recipes running on from p. 189 to p. 190, where the 24 additional virtues of a white ointment are listed and numbered.Nonetheless, virtues generally come at the end of the recipe, even after the efficacy section, as shown in ( 31) to (33).They may appear as a separate paragraph, as in (31), or be the only component of a recipe along with the ingredients, as in ( 32): (31) Jt is good for the Liver.for the milt.for a good appetite to meate.and very good to kill wormes in the bellye./(p.44)

Conclusions
This article has focused on the linguistic and structural features of the recipes contained in the EModE recipe collection in W8086.The analysis carried out has rendered interesting findings concerning not only these features but also the intended audience and authorship.
On the one hand, the examination of W8086 shows that it conforms on the whole to what is expected for recipes in the EModE period.This comprises the stages that may be found and even the typical ordering of such stages, as well as the linguistic traits surveyed (imperatives, lexical choices, limited use of Latin in efficacy phrases, and so on).More variety has been found, for instance, regarding types of recipe headings in W8086 than in the relevant literature, usually by elaborating on patterns already identified.Also, ingredients play a paramount role and become an essential component of recipes, although very rarely are they used in isolation, without other stages, despite the fact that this is the only element that has been considered to be obligatory for recipes, according to the literature reviewed.However, some ME traits are still attested, such as the occasional references to God in EPs or the frequency of certain evaluative adjectives in headings.It will be interesting to see if other features at other levels (e.g.linguistic) are in line with the ME touch of some recipe elements.
On the other hand, much of the evidence gathered about W8086 points to this being a lay text.This is supported, first of all, by the fact that the collection gathers various types of recipes (not only medical), and also by the lack of a structuring pattern or references to surgical procedures in medical recipes.Emphasis on health care within the household sphere is also strengthened by references to children not only as patients but also as distinct addressees of specific remedies.All these arguments are also consistent with considering a woman as the likely author of the text.Additionally, the repeated imprecision and vagueness that characterise references to authorities (most of whom are modern laypeople), measures and quantities of ingredients, dosage, and date of expiry are also noticeable and in line with lay texts.EPs also offer interesting information concerning the potential audience of the text.Despite their frequency (around a fifth of the recipes in W8086 contain an EP), which conforms more to learned that to lay texts for the period 1601-1651, their distribution into types fits the figures for lay texts in the said period.Nonetheless, Mäkinen's findings are more balanced than those for W8086, where the proportion of stock and specific EPs is 1.5 to 1.
This study could be complemented with a comprehensive analysis of the interpersonal strategies (overt references, authorities beyond those in recipe headings, etc.) used in order to assess the roles of author and reader/addressee.Likewise, other traits frequently attested in the recipes, such as the imprecise information in certain stages, could also be further explored.At the same time, more individualised studies of EModE recipe books should be undertaken so as to expand the evidence available on the features of recipes of the period and women's writings.This is particularly true of manuscript materials so that their results can be eventually contrasted to those obtained from printed materials.

( 32 )
Drink Vervin tempered with wyne. it drives away the stinck in the mouth.and makes á sweet Breath.(p.119) (33) This same is good to annoynt any swelling.or to be mixed with any other salve what you shall thinke good./(p.156)

Table 3 .
(Bator and Sylwanowicz 2018) headings in EMEMT(Bator and Sylwanowicz 2018)and W8086.Water of Calamus Aromaticus./" (p.202).Contrariwise, even if Bator and Sylwanowicz's classification of SP into structural types is fairly detailed (2018: 112-114), as shown in Table 4 below, it still does not cover all the patterns found in W8086.This has led us to add a subgroup called 'other' within each group (SP1, SP2 and SP3) which gathers the patterns that use PP, NP or infinitives, respectively, but present alterations with respect to Bator and Sylwanowicz's patterns: 13%Titles in W8086 conform to what is expected; i.e. the use of NPs, as in "A Strengthning meate."(p.25]), "A Purge."(p.173) or "

Table 4 .
Distribution of structural types of SP headings in W8086.

Table 5
below collects various examples under 'other' which reflect more elaborate patterns, hence increasing the complexity of the headings.This trait can be realised by a) coordinating two instances of the same pattern (examples under 1); b) coordinating or combining two different patterns either of the same group (cases under 2) or of different groups (examples under 3); c) modifying any one pattern (examples under 4): 13

Table 5 .
Examples of additional structural types for SP headings in W8086.