Tagging the Infernal


In CB51: Making the Middle Ages, the teaching staff, consisting of Professor Dan Smail and TFs Rowan Dorin, Zoe Silverman, Joey McMullen, and Rena Lauer, used an online annotation tool to have students tag images and descriptions of hell and create a "tag cloud."  This activity engaged students in perceptions and interpretations of hell and the use of metadata.

The following description comes from digital history@harvard:   

In this project, students used a Collaborative Annotation Tool, developed by Harvard's ATG using the Confluence platform, to tag and annotate collectively a set of images and texts. The teaching staff constructed a database of fifty medieval hell visions and over a hundred depictions of hell [accessible only with Harvard ID], which the students then explored and analyzed in order to track changes in the representation of hell over the course of the Middle Ages.The goals of the project were:

  1. To track the development of a particular theme (in this case, representations of hell) across time, space, genre, and media
  2. To teach students to identify and express key elements of a text or image
  3. To have students engage actively with each others' analyses and interpretations of texts and images by recording their own responses
  4. To teach students to think critically about the elaboration, uses and limitations of metadata

First, students chose one-word tags to describe what they were seeing/reading as they explored the database, whether simple stylistic descriptions of the type of item, such as 'woodcut' or 'apocalypse'; identifying attributes of the persons, objects, or themes/activities embedded within, such as 'demon', 'sword', or 'burning'; relevant date or place markers; or any other notable elements of the image/text. They were also encouraged to review other students' tags in order to ensure consistency (and thus search functionality). As the students marked each image/text with tags, a 'tag cloud' was built up that visually reflected the relative frequency of particular tags. Students then wrote brief commentaries on the images or on short text selections (which they extracted themselves from the longer hell visions), noting features of interest or questions to pursue further. They were also asked to respond to several commentaries written by others. Some class time was devoted to discussing what made a 'good' tag, a good response, etc., and class time was also spent analyzing trends in the tag cloud (e.g. looking for sampling bias as opposed to robust patterns).

At the end of the project, the students were divided into groups and asked to produce a short, fictional hell vision that included elements appropriate to a particular time and place. These elements could be determined by starting from a particular tag (a century or a country, for example), and then tracking associated tags.  [This occurred in lecture and took 15 minutes.]  The results, which showed an impressive command of the material (and considerable creativity), were posted to the course Tumblr and read aloud in class.  It should be noted that other collaborative annotation platforms exist, some of which allow for comments to be attached to specific points on image files or PDFs (which the Confluence-supported CRT does not). The CRT does not allow individual user activity to be tracked at a class level (one must instead search for each student separately), and certain kinds of activity (tag creation, for example) go unrecorded on the summary pages for each user.


Here are the instructions from the course website:

This project is designed to give you some experience with tagging or marking-up data--in this case, data that takes the form of words or ideas found in texts and images. It will also enable us to think about the mental world of medieval people by considering how they pictured hell.

The wiki-supported database that we will be using, known as the Collaborative Research Tool (CRT), consists of over one hundred texts and images containing visions or representations of hell and purgatory. They date from the fourth through the fifteenth century and are drawn from across Europe and the Mediterranean. The texts vary in length from a few lines to dozens of pages; most have been translated into English, but those students who can read Latin, French, Italian or German are encouraged to tackle the handful of texts that are available only in those languages.

As you look at an image or text, you will be charged with three tasks. First, you will choose one-word labels to describe what you are seeing/reading. These can be simple stylistic descriptions of type of item it is: “woodcut” or “fresco” or “apocalypse”; and they can note a theme or activity (“burning”), object (“sword”), character (“demon”), or any other essential element of the item. You will start to see recurring images and ideas, and so some of your labels will begin to repeat – this is good! As you and your classmates build up labels for each item, a tag-cloud will be generated on the CRT’s homepage, so you can see what themes crisscross the whole hell database, and how common each theme is. In class we will discuss what makes a good label. (As a class, you will create a set of guidelines for consistency, such as using singular and not plural.) Second, you will be asked to reflect on a given image or a short passage in a comment. In these comments, you might want to think about how a given item describes hell, what it says about the mentality of the author or audience, or another aspect of the given item. Each comment should be the length of a short paragraph. Finally, you will be asked to interact with your classmates by replying to a number of their comments. Do you agree? Disagree? See their chosen passage from another angle? See a modern parallel? Offer a short paragraph in response.

During the week between Thursday, 4 October and Thursday, 11 October your at-home assignment is to mark up (label and comment) on 2-3 images, and 8-10 text segments. You should also reply to 3-5 comments written by your classmates.

Two important points:

  • A segment is NOT a whole written work; rather, it is a rich paragraph that you will (1) choose from the complete written work; (2) make a "daughter" page (a sub-page filed under the complete written work); (3) name according to the verses or page numbers;  and, (4) enter your labels and comments on that daughter page. The teaching staff will model for you how to do this. To see an example of how to break down a written work into text segments, look at the Apocalypse of Ezra.
  • When you click on a "hell text", you will notice that the text of the work will either be attached as a PDF or linked to an on-line source, or occasionally necessitate a trip to the library to get the hard copy. When you have chosen a text segment, create a daughter page, and transcribe or copy-and-paste the relevant paragraph into a new daughter page.