In a speech in January 2006 to the Oxford Media Convention, the then Secretrary of State for Culture, Media and Sport, Tessa Jowell, making the case for media literacy in an age of media convergence, argued:

I do not exaggerate when I say that media literacy in its widest sense is as important to our development as was universal literacy in the 19th century. Then, the written word was the only passport to knowledge. 1 Now, there are many more.  And the most insidious digital divide is between those equipped to understand that and those who aren't.

Even accounting for political rhetoric, Jowell’s remarks reveal a common assumption: that the nineteenth century was dominated by the word. However, a glance at the front pages (not to mention the wrappers) of any of the titles in ncse reveals this assumption to be largely false. This page provides some brief accounts of both the importance of visual material in nineteenth-century print culture, and our treatment of it in ncse.

Importance of the visual

Although Jowell is trying to claim that there is something distinctly twenty-first century about media literacy, this is clearly untrue. Nineteenth-century proprietors and editors were adept at using combinations of words, blank space and images to create distinct identities for their publications that engaged with social codes and attempted to interpellate groups of readers. Even an early, relatively austere unillustrated periodical such as the Monthly Repository employed a visual repertoire through its typography, graphic devices to separate departments and items, icons to indicate notes, and the way words were displayed on the page. For instance, issue no. 1 of the Monthly Repository, from January 1806, has a variety of different sizes and formats of type; uses a range of icons to indicate footnotes, and breaks up its single block of text on page 49 into two columns. Although it never brought illustrations as such into its letterpress over its 32 year run, the Monthly Repository did issue high quality portraits and other illustrations as annual supplements, often with instructions that they be bound in at the front of the volume (for instance see the ‘Memoir of the late Dr Cogan’ in vol. 14 (1819) here). Although the Monthly Repository, which began publication in 1806, predates the expansion of the illustrated press in the 1830s, its use of visual material demonstrates the popularity of prints for its readers. The high quality illustrations celebrate notable people and buildings within the periodical’s culture, while the way they were issued links them to its serial production.

The other five periodicals in ncse all employ visual material in different ways. This is partly related to the changing economics of print production, which reconfigured the practical demands of getting images into letterpress. The economics of different modes of reproducing images are demonstrated in the Northern Star. Although not cheap because it paid the newspaper stamp tax, the Northern Star was aimed at a working class readership and produced as cheaply as possible. The most visually striking pages in its letterpress are those containing advertisements. For instance, the front pages of the first 60 issues of the Northern Star (roughly up until March 1839 with the preparation of the first Chartist petition) are devoted to six columns of advertisements. Each advertisement uses a different combination of type and space in order to distinguish itself visually from the others that surround it. Many, such as those for Whitehead’s Tea and Coffee shop on Briggate in Leeds (just down the road from the Northern Star’s offices) are high quality and used for a number of issues. Others, such as the small woodcut for Batty’s Circus Royal appeared for only a handful of issues, before the circus moved on.

Although the Northern Star did include some images to accompany its articles, these tended to be quite crude woodcuts. In the early nineteenth century it was difficult to commission, design and execute a woodcut within weekly publishing rhythms. A good example of this is on the front page of the Northern Star vol. 5, issue no. 249 (1842). The first two editions of this issue feature a blank space where a woodcut was supposed to appear. The subsequent three editions all have the woodcut. In the white space there is an explanation as to why the woodcut does not appear. The editors had been promised it would arrive from London that morning, but it was not on the expected 6 am train. The editors claimed that readers of their first edition would be ‘naturally anxious to hear the news’ so would prefer an incomplete paper than a delayed but fully-illustrated one. Of course there may be other reasons to do with distribution and printing schedules that would make such a delay unaffordable, but that the space was set, and then kept available until the last moment, gives the front page a timeliness that emphasizes its news value. The space itself is particularly suggestive of the absent image: the retention of the title and the caption allows readers to appreciate what should have been there, and the space itself indicates its size. By signalling the absent image, the editors of the Northern Star demonstrate the expense they have gone to in order to produce the engraving (even thought it isn’t there), while sacrificing the space by not filling it with letterpress or advertisements. The space invites readers to imagine the absent image, especially as the other relevant components on the page – the tribute to Henry Hunt printed in the space usually reserved for the proprietor Fergus O’Connor in the first column, and the other cut showing Hunt in full flow with the Peterloo Massacre raging around him at the bottom – are present. Yet the white space reminds readers that this is an incomplete version of the Northern Star and that a different, more up-to-date version exists elsewhere. This white space, therefore, does not just represent an absent image, but also reminds us of the presence of its other editions.

Where the Northern Star was more successful was in issuing one-off engravings as supplements. Almost from the outset it issued fine steel engravings of notable figures to encourage a particular form of accessing the paper, subscriptions, and to create a sense of unity amongst the developing Chartist movement. For more on the portraits, and to see them, click here. As steel engravings the portraits were difficult to produce in the necessary numbers. Although the Northern Star aggressively promoted the portraits, it was often forced to issue them on a regional basis over a period of weeks in order to meet demand. As the steel plates deteriorated, so did the quality of the portraits.

By the 1860s illustrated weekly journalism was well-established. The Illustrated London News and Punch appeared in 1841 and 1842 respectively, demonstrating the possibility of high-quality, timely illustrated journalism. These respectable titles were joined, in 1845, by the London Journal, a more down-market illustrated weekly that obtained large sales combining illustration with serial fiction. Twenty years later Tomahawk was set up as a direct rival to Punch and featured a large cut (either one or two pages) in every issue. These were engraved onto separate sheets and then bound into the issue. As the reverse side of the image was left blank, they could take up to four pages of a ten to twelve page issue. The images themselves were engraved by Matt Morgan to produce combinations of black and white line against a coloured ink wash. Each cut had a caption and was also often accompanied by a satirical article in the letterpress, creating a complex dialogue between the text and image across its pages.

Aside from the tomahawks that graced its masthead and leader, the large cuts were the only illustrative material to appear within Tomahawk. However, the journal was issued in an advertising wrapper that was marked by a striking title page. As you can see here both the wrapper and the advertisements it contains are packed with images, creating a visually dense page in stark contrast to the sparse paginated letterpres. As a trade journal, the Publishers’ Circular’s entire complement of images is located in its advertisements. At the end of the century a mature market for the production, distribution, and preserving of images coupled with new technologies such as photo-mechanical reproduction made it much easier to print visual material. However, as the Christmas numbers of the Publishers’s Circular demonstrate, there was still a substantial demand for illustration. The Christmas number was published in the first week of December in addition to the two usual fortnightly issues that month. It was intended to showcase the illustrations that were to be printed in the Christmas numbers of the illustrated press, and the abundance of visual material that this brought into the issue also made it more attractive to advertisers. These issues were usually four or five times the length of a regular issue, and included separately paginated inserts of illustrations from the periodical press that was interleaved with the regular advertisement section.

As reading necessarily entails looking, the visual is a key component of any text. As the periodical and newspapers in ncse demonstrate, visual material is not restricted to pictures. Although they did play a part in serials throughout the nineteenth century, other aspects such as mastheads, icons, typography and layout constitute the bulk of visual material in the press. The visual is not only a decorative addition to the textual content, but is the way in which that content is presented. As such, visual material is part of the content, structuring the ways in which it is encountered. It is through the combination of visual features, as well as what verbal text says, that individual issues posit an identity that can transcend their specific, dated material, linking them to the title as a whole.

Images in ncse

The predominant model for digitizing printed texts is to combine facsimile page images with a textual transcript, usually generated from optical character recognition (OCR). It is the OCR transcript that provides the searchable index to a digital resource, although this is usually complemented by other searches of the metadata. As search is essentially a textual procedure – searches work by looking for text strings that are similar to that proposed by a user – it is notoriously hard to include visual material within search engines. The most straightforward way of doing this is to include some sort of textual information along with the image that can be interrogated by a search engine. As this information is interpretive – it attempts to render the picture into words – it has to be entered by a human. Not only is this time-consuming but, as most digital resources tend to be top-down models, the process has to be carefully managed in order to standardize the judgements and control the vocabulary.

As mentioned above, there are lots of different types of visual material in ncse. Although we recognized the importance of marking-up pictures, we also wanted to signal the other types of visual material and how they interact with each other. In the early stages of the project we attempted to delineate all the possible data structures in the edition, before analyzing them to work out which could be achieved with the resources available. Olive Software’s segmentation allowed us to decide what constituted an item, and our editorial policy, which recognized that issues are more than a series of articles, prompted us to identify any textual component on a page as a single item (for more on our editorial policy click here). This means that items could be mastheads, datelines or running heads as well as whole articles. Olive Software permits items to be one of three types: articles, advertisements or pictures. Where an item is a picture, we have marked it as such; where an item contains a picture (for instance a picture within an article), we have marked the picture as a picture and then embedded it within the item. This allows users to open the item as a whole in the component viewer or, if they only want to see the image, open the image without the accompanying text. The one place where this causes difficulties is in advertisements. As noted above, advertisements are often the most visually diverse portions of an issue. However, Olive’s software does not permit pictures to be embedded within advertisements. As we did not want to exclude this important source of visual material, we ensured that images within advertisements were marked up with image metadata. Although a search made from within the edition that is delimited to images will therefore exclude any images that appear in advertisements, browsing the image metadata will produce results from all the images in the edition, whether included in advertisements or articles. You can read more about our image metadata below.

Although our segmentation policy allowed us to recognize a more diverse repertoire of content than articles and images, we were restricted in the extent to which we marked them up. ncse contains things like handwritten notes, newspaper stamps, library accession numbers, rules, ornaments, fingerposts, mastheads – all of which constitute visual material of one kind or another. These are segmented within the edition as items, formally recognizing that they are as much part of the issue as verbal contents and conventional images. We could have also chosen to label them with metadata, distinguishing them from each other and allowing them to be included within searches. However, because of the size of the edition we were restricted in the extent to which we could add metadata at item level and, after carefully considering the benefits to users, decided to treat these types of content in the same way as other items: i.e., with metadata that is inherited from higher levels in the ncse structural hierarchy. For more about our metadata, click here.

Image metadata in ncse

All of our images, whether part of articles, inserts into front matter, or appearing in advertisements, are labelled with image metadata and so can be searched and browsed. It is notoriously difficult to come up with a classificatory scheme for images. We considered adapting well-known schema such as Iconclass as a way of labelling images in a way that would be compatible with other projects. However, because Iconclass is designed to label works of art (predominantly from a Western tradition), there are large domains that it does not cover that are crucial for labelling the types of images in ncse. For instance, a large number of our images are within advertisements and are therefore usually representations of the goods that are advertised. As Iconclass tends towards the thematic (it is, after all intended to be an iconographic schema) it cannot provide terms to label, for instance, types of bookcase.

Many projects that work with images evaluate Iconclass as a possible schema. The Database of Mid-Victorian Woodcut Illustration (DMVI) considered a range of possible iconographic schema, before opting to design their own. As their resultant iconographic schema was a bespoke system designed to label illustrations in nineteenth-century periodicals, it was very close to our needs. The Project Director of DMVI, Julia Thomas, kindly allowed us to use their schema for ncse. We adapted it by selecting its more descriptive categories , and then tested them on different types of content in order to find any areas in which we needed to supplement them. It is a testament to their schema that even when a group of editorial assistants all worked on the same content, they came up with largely the same terms. You can read more about the DMVI schema here.

As any markup of images is necessarily interpretive, it is a process that involves a great deal of human input. There are over 5000 individual images scattered throughout ncse: although we knew where most were concentrated (for instance Tomahawk’s central cut; advertisements etc.) and could use Olive’s image search to find many of them, we still needed to scan every page and mark up each as it was found. As images do not produce legible OCR transcripts, it was possible to speed up this process by scanning the lists of item titles that populated the tables of components in the Olive Administrator Tool. As this was the tool though which we added most of the metadata, it was a particularly convenient method for finding images. In order to ensure as much consistency as possible through the edition, we carried out extensive checking of image metadata, comparing different editorial assistants’ work and keeping a log of any decisions that were made.