Categories in a Database of Eighteenth-Century Medical Recipes

By Katherine Allen

Creating a database is a valuable (though time-consuming!) methodological approach to the history of recipe collecting. The database that I constructed for my doctoral research catalogues over 5,000 medical recipes from 27 eighteenth-century manuscript collections. I used the data to help answer one of my thesis questions: in what ways did the recipe collecting tradition continue into the eighteenth century as part of domestic medicine in elite English households? The data suggests continuity and change in recipe collecting. Even as compilers maintained traditional methods of collecting, they were open-minded to innovations in medicine and other cultural and intellectual shifts.

I use a form-based system in Microsoft Access, with each entry receiving a unique searchable ID number for manuscript and recipe. The collector, when available, is noted, along with the date (for a specific recipe or the manuscript). By including the recipe title, I could also trace popular recipes and the ailments they treated.

Recipe Form in Database
Recipe Form in Database

‘Ingredients’ was an embedded form listing the ingredient item (modernised spelling), quantity (decimalised), unit of measurement, and a description of characteristics and preparations. I grouped like ingredients and categorised them as follows: plant, animal, mineral, prepared (compound), and unknown. I also found total counts for each ingredient, and how many individual (unique) ingredients comprised each category, enabling me to rank proportionally the ingredient categories.

One of the difficulties of categorising materia medica is finding balance between conveying diversity, while grouping like items for effective and efficient analysis. I used similar methodologies to those used by Jennifer Stine and Anne Stobart (see post on plant ingredients). Moreover, I endeavoured to categorise separately distinct components of plants and animals (for instance, seeds and petals, or, suet and blood) when each item was contemporarily viewed as having differing medicinal qualities and uses.

The ‘equipment’ form recorded household objects used to make medicine, the technique or instructions associated with tools, and descriptions. I used the data on distillation to determine how frequently stills were cited, and the types of distillation equipment cited in recipe collections. This data reveals that distillation remained an important component of recipe collecting up to the first quarter of the eighteenth century.

Data on the purpose and administration of remedies was compiled using a text box on dosage/administration and a drop-down menu on treatment. The dosage text box allowed me to define the most commonly used quantities of medicine, how they were administered, and frequency of dosage. This data shows that medicine was often tailored to an individual’s age, constitution, and tastes, and that regimens remained popular approaches to treatment in elite healthcare.

Drop-down menu for treatment categories in database
Drop-down menu for treatment categories in database

Categorising diseases in recipe books is challenging. These manuscripts present a long history of varied medical approaches and theories, and compilers frequently grouped remedies by the process of making the medicine, rather than intended use. We also need to contend with recipes being used to treat symptoms and/or diseases, and recipes that could treat several ailments, or were cure-alls.

I took an objective approach to categorising diseases, following how compilers described their remedies first, and later grouping the ailments thematically into categories such as: who they were treating, approach to treatment, and diseases. Using a drop-down menu for treatment allowed me to organise and amend my categories as I entered data. The data was entered according to the main ailment the remedies was intended to treat (if there was more than one), but recipes that cited many uses were often grouped as cure-alls. This approach to database categories attempts to provide a realistic portrayal of medicine in recipe books by following how eighteenth-century compilers understood medicine, but it also meets my needs as an historian to create manageable categories.

Treatment data allowed me to trace popular named medicines, like Peruvian bark and Gascoigne’s Powder, and this helps us understand how compilers categorised and understood diseases and approaches to treatment. I used Access reports to group the recipes by treatment, thus generating data to calculate percentages for each treatment category.

Finally, I had two text boxes for entering information on individuals attributed to recipes. Forty-percent of my recipes were attributed two one or more individuals. The queries used to run this data looked at recipes with no attribution, one attribution only, and multiple attributions. This data was used to understand direct and indirect citations referencing how medical knowledge was communicated and that recipes were often communicated in a multi-step transmission process. Compilers who referenced from whom (or what- with regards to print sources) they obtained recipes, from where, and when, demonstrate the significance of references in documenting a recipe’s credibility and the complexity of social networks used to send and receive medical knowledge.

The Colour ConText Database

By Sylvie Neven

neven fig 1
Fig. 1: screenshot of the starting page of Colour ConText database

Artisanal recipes are considered to be primary sources in the historical study of artistic practices and materials. Prominent examples of such documents include the De diversis artibus attributed to Theophilus and the Libro dell’arte by Cennino Cennini. However, hundreds of other such examples exist and were still largely unknown. In his 2001 publication The Art of All Colours, Mark Clarke compiled an inventory of 400 source documents, dating from the production of the first artists’ recipe collections up to 1500. Since then, dozens of other surviving writings containing artisanal recipes have been discovered. Many more recipes were written down in manuscript and print in the period after 1500.

The initial goal of the Colour ConText database is to facilitate the consultation and exploitation of a large corpus of recipes. The core data consists of medieval and early modern manuscripts and printed books.

The ‘Sources’ page

neven fig 2
Fig. 2: screenshot of the ‘Sources’ page in the database

To date, more than 500 sources (including manuscripts and printed texts) have been entered into the database, specifically located on the ‘Sources‘ page (fig. 2). In the ‘List view’, the entries are tabulated optionally by place of conservation or edition, or by title, author or date. Detailed information such as the source’s title, language, location, provenance and circulation of these manuscripts and books (place and date of origin/publication), scribes or authors, previous owners, and a description of their technical and/or general content can all be viewed on this first interface, on the ‘detailed view’.

From now, these sources can be searched by title, or by place of conservation or edition.The database also allows access to digital images of these sources via European Cultural Heritage Online (ECHO), or via digital collections made available by external institutes.

The ‘Recipes’ page

neven fig. 3
Fig. 3: screenshot of the ‘Recipes’ page in the database

The database also makes the content of the recipe collections accessible at the level of the individual recipes. To date, more than 6,500 recipes—some consisting of only a few lines, others covering several folios—have been transcribed and recorded on another specific page within the database. The ‘Recipes‘ page allows users to consult the transcription of a particular recipe, and sometimes also provides an English translation. This translation may either have been done in the framework of this project or be reproduced directly from existing edition. In such a case references to secondary sources, together with the related bibliographical data, are specified. Finally, this layout also gives access to the specific image of the original recipe text (fig. 3).

Users can search for a specific request by library, source, title or ID number—a consecutive and unique number assigned to each individual source. It is also possible to search for specific words that appear either in the transcription or the translation of the recipe.

Thanks to subject classification, keywords can be used when researching specific recipes, methods or materials. The general search button – situated on top of each page – allow users to combine an ingredient with a specific technique, mentioned in a limited geographical/chronological framework, when making their search.

neven figure 4
Fig. 4: screenshot of the results for combinated search

For example, the combined search for ‘alum’ (an potassium aluminium sulphate notably used in the art of dyeing and for the producing of lake pigment) and the colour ‘red’ shows a number of entries, which can be further filtered by glossary, artistic technique or content type. We can read that 305 recipes are concerned with the substance alum and with the production of the colour red. These recipes are more specifically related to painting, illuminating, writing, dyeing, gilding and metalwork (fig. 4).

The Colour Context database can also help to identify specific, datable practices and materials. For example, we have observed that a significant number of procedures involving anthocyanin colourants (obtained from the juice of flowers and berries, such as poppies, cornflowers or blueberries) are specifically described within a certain group of manuscripts. More precisely these texts were written in the south of Germany and the north of France between 1400 and 1560.

The ‘Glossary’ page

The database also includes a complete list of the ingredients and substances mentioned in the recipes, indexed both by their current scientific name (‘Current names’) and by the ‘historical’ terms precisely as they are written in the source texts (‘Historical names’). Objects and materials are linked by relational tables that allow the retrieval of all the different historical names used for one particular material—detailing the historical written context—as well as enabling the user to see the various materials that may be related to a specific name.

These lists notably shed light on the diversity of colour names and the complexity of the varied colour terminology used in artisanal recipes. For example, the puzzling denomination ‘red of Paris’ relates to several different substances. In the Illuminier Buch von Valentin Boltz von Ruffach (first edition dated of 1549), it is used to designate a red pigment obtained from brazil wood (Caesalpinia sappan Linn. or Caesalpinia echinata Lamarck). However, other sources make a distinction between Paris red and the red pigment obtained from brazil wood by recommending the use of either the former substance or the latter (‘Ein guot röselin oder pfirsÿg bluot Nu nim presilgen oder paris rot’, Colmarer Kunstbuch, pp. 124-125, recipe 35). In Heidelberg, Cod. Pal. Germ. 489 [252], ‘Paris rot’ is a colour made from brazil wood and an unspecified lake called ‘Lacta’. Within this same manuscript, recipe [391] describes the preparation of ‘Rotenn Paris’ from a lake. It has therefore been hypothesized that this appellation was used as a way to distinguish a specific hue.

For more information on the Colour ConText database, see my recent MPIWG feature story, Colours and Their Context.