Why Hypertext for History?

History in various forms has been practiced for thousands of years without the benefit of computers. Many historians probably suspect this helped their craft considerably. Historians never had to ponder which font to use, and never had the opportunity of adjusting his margins after they'd written a paper to bring it in line with the length requirements of his day. As a result, they had to write about what happened and make sense of it rather than waste time with complicated foibles. To many historians, computers are number-crunchers, advanced calculators for physicists and chemists seeking microscopic truths of the universe, perhaps helping the occasional economist or political scientist too caught up in quantitative analysis, but of little use to academics studying the grand narrative of the world.

Many people still wonder how someone studying the humanities could find a use for a computer, except as word-processor or bibliography. In my own college career I never had the opportunity to make the computer an important participant in historical discussions. Today, however, I realize how much the computer and the way it deals with information have affected the way I think about history, and as a result I am trying to integrate the computer and my discipline. In my view, the computer offers history a chance to build on its oldest traditions and makes possible the linking of many kinds of disparate information that history has tried to accommodate. This means more than using the computer as word processor or database; it means that the computer must be used as word processor and database simultaneously.

This seemingly perverse combination of applications is called hypertext. In a hypertext system, the reader is free to navigate information by exploring the connections provided. There may be a broad map or overview, or the story may simply start, but it is up to the reader to navigate as they please. An author can pause to go into extreme depth on a subject, knowing that not every reader will find it interesting, without interrupting the flow of the work. Connections form a more intricate web of links than is permitted by the straight-ahead flow of a book, even a book with serious notes to allow the reader further research. Reading hypertext is a project of exploration. Readers have the opportunity to read in great depth, and will hopefully find well-written hypertext more interesting than even the most enthusiastic notes.

Allowing readers to roam as they please, to skip and jump between arguments and miss the point entirely, may seem dangerous to some. I hope to defuse that fear by presenting a vision of hypertext that illustrates its potential for making arguments but which also makes clear how different those arguments may be. Along the way I hope to demonstrate that hypertext seems in many ways to correspond to the projects of history, academic and otherwise, in many ways better than the printed book, and offers the chance for historians to reach a wider audience.

Why hypertext?

Hypertext is very different way of presenting information than the usual linear form. Text no longer flows in a straight line through a book. Instead, it is broken down into many smaller units (lexias, to borrow a term from literary criticism), each addressing a few issues. The fragmentation of what might have been a single huge document allows the author to experiment with a wider variety of links than the sequential "A happened then B happened and then C." The author gives (more) control to readers, allowing them to find their own paths through a jungle of information. If authors design links intelligently, readers shouldn't get lost, though they might have no idea of their precise location at any given moment. It's similar to reading an encyclopedia, jumping from keyword to entry to the next keyword. The computer takes care of page-flipping, allowing the reader to browse the offerings or move in for greater detail. Text is no longer only linear, as it is in books, for instance. Left-right-down-the-page still works (right now), but turning the page is a whole new experience. The straight lines are only fragments.

Many people think that this format is great - but only for encyclopedias and dictionaries. Students who already groan at the prospect of a 200-page book are probably less than thrilled about the idea of a 200-page book whose pages are in no particular order. Many of those books, however, already exist. Over the last half-century many authors have experimented with nonlinear forms, although they haven't always made sense to (or earned the approval of) everyone. Hypertext, as I like to think of it, is a continuation of this creative fragmentation. The author's and reader's streams of consciousness interact as the author provides multiple paths and the reader skips along. Does this cost the author control? To some degree it must, because the author can't force the reader to go in order ABC through a document. The author does, however, still retain control over whether the reader can jump from A to C, C to B, and B to H, and in what order. It's a different control, a different experience.

The encyclopedia metaphor (no high-flying streams of consciousness) connects well with the established uses of hypertext: providing easy access to large chunks of data that don't always fit together in a single linear narrative. Journalists are using hypertext to research background on stories without having to read years of clippings to find a comment on their topic. Scientists are using hypertext to coordinate lab reports and to keep track of ranges of widely divergent but still connected information. Doctors use hypertext and expert systems (which help navigate) to help diagnose patients, and, similarly, technical support workers use expert systems that are always growing with "experience" to diagnose and fix problems with electronic equipment. Although these information-retrieval uses are promising and important, I'd like to see hypertext have a more fundamental impact on the way we read and write, especially the way we read and write history.

What happened then?

One of the most fundamental problems in writing history is simultaneity. Using the phrase "at the same time" over and over again jolts readers out of a smooth flow of narrative, but giving long narratives to each of several different simultaneous occurrences makes it hard for readers to keep track of what exactly was happening simultaneously and where. Timelines are sometimes a useful aid, but can obscure as much as they present, especially where details are critical. Hypertext offers several answers to the problem of explaining many things at once.

First, it's easier simply to have multiple windows available simultaneously. Books written in multiple simultaneous columns remain extremely rare except in simple timelines. A history of the Hundred Years' War could, for instance, include lexias on the battle of Agincourt that explained the French tactics and some on English tactics and others on the events of the battle itself. The reader could flip between lexias whenever they found it relevant or whenever a link was indicated, keeping all the relevant lexias on-screen for easy reference. Similar things could be done with articles on industrial productivity in Massachusetts textile mills in the 1800's and the lifestyles of their workers and owners, or with the politics of the Gilded Age and the economic cycles of the period, or with Latin American and African development. They can be on the screen simultaneously, but, more important, the author can demonstrate how the two articles fit together by providing appropriate links.

Another possibility for hypertext is the expansion of the timeline. Timelines have the advantage of a simple form that readers can easily relate to, though in isolation they generally prove meaningless to anyone but students cramming for a test. In a hypertext timeline, the points indicated on the timeline don't have to stand by themselves, and the reader doesn't even have to search a chapter or search through the index to find supporting information. The timeline becomes the trunk of a tree whose branches the reader can explore by clicking on particular entries. Lexias can connect to timeline entries and then to each other, providing the reader with a simple way to get into the material.

Other graphic frameworks might also prove adaptable to linking. Clicking on the labels of a graph, for instance, might provide more information about the kind of data being presented or its source, or might even lead to the raw data used to produce the graph. The computer is a 'natural' home for quantitative information. An advanced hypertext system might allow the reader to perform statistical analyses of data without having to export it to another system. If a reader has doubts about the author's quantitative methods, the data is there for the reader to examine and experiment on. Of course, this hardly begins to answer questions about the methods the author used to collected the data, but at least it creates an opening for more thorough examination.

Maps are another graphic form that could be used extensively, allowing the reader to get more information about a geographical area or an event at a particular site. Pictures can supplement textual descriptions, bringing, for instance, the horror of the trenches in World War I or the pomp of the Tsarist court to the reader. Digitization of pictures continues to improve, though in many cases the quality is equal to or worse than a newspaper halftone. Pictures can also be linked, allowing the reader to click on a picture of Churchill for more information about the man and his politics, perhaps even his own writings.

"Just the facts" - primary sources, history, and hypertext

In theory, all history derives from the judicious combination of primary source material into more complicated material that tells a story or makes an argument. Primary sources are important, but the job of historians is to interpret them, to make them speak with less distortion and more comprehensibility than they do in their raw form. Because primary sources don't always agree, much of historical writing is sorting out which sources we want to consider trustworthy and which we'll ignore, and the development of criteria for such decisions. Issues like relevance, reputation, and accuracy provide guides for most historians through the thicket of unexamined material.

While this has provided the backbone for history for centuries, it also means that almost no one except for historians ever sees what lies in the archives. Authors use source materials to support larger arguments, but all the reader has of the original sources by the time the writing process has finished is fragments: occasional quotes, a long listing of notes at the back of the book, and maybe even a picture or two if the publisher felt that the price of a photo insert was acceptable. The book is (supposedly) a seamless whole, but the pieces from which it was built are largely left behind. Indeed, a truly successful historical work obscures the primary sources on which it relies by making it seem unnecessary to trek to the archives, just as a successful translation makes it seem unnecessary for a reader to read the original.

Hypertext makes it possible to avoid this effect. Rather than just cite a work so that a reader can find it someplace else, the author can provide a copy of the original, whether it is text (in the author's primary language or in the original), graphics, or even sound or video. Technologies developed in the last decade, most notably the CD-ROM, have made it possible to provide incredible quantities of information in a lightweight and easily accessed format. A printed book in which the author wrote 300 pages and then had 1800 pages of supporting source materials and notes is almost absurd, but a CD-ROM with that ratio is easily possible. The cost, weight, and accessibility of a CD-ROM aren't dependent on the amount of information it holds the way a book is. The linking facilities provided by a hypertext environment allow the reader (if, of course, the hypertext is well-designed) to reference this information as easily as they would have previously looked up a footnote that told them virtually nothing without a trip to the library.

Classicists have been among the first to take advantage of this raw bibliographic ability. Because Ancient Greek, for instance, has an extremely limited canon of works, enterprising classicists were able to fit the entire canon on a single CD-ROM disk and provide search capabilities, and the same is being done with Latin. In fields with a limited base of data, it's incredibly valuable to have the entire corpus of primary material available with a few keystrokes. In history, which uses millions and millions of documents, this kind of 'world on a disk' isn't (yet) possible. Historians still have the opportunity to build such databases for subsets of the field, possibly fitting an archive on a disk. Once digitized, the information is much more easily transferred than it is sitting in a folder in an archive. There are some who stand to lose from such a wholesale freeing of primary information, who argue that such anarchistic distribution of data to a readership unqualified to examine it will inevitably lead to a collapse of standards. My suspicion is that most of these kinds of opponents have something to hide: they fear widespread distribution of the information they "protect" in any but the most shrouded form. Other opponents complain of the great cost of digitizing information; even the best optical-character recognition systems can't yet read handwriting well, and even handwriting experts can't always determine what is being said. Twentieth-century works also present the difficulty of copyright, which may threaten this kind of project. These kinds of practical limits (and the money involved) make large-scale digitization of historical data presently impossible, but I hope that it can be done on a smaller scale (as in my proposal for much-expanded footnotes) or in fields where much of the same information is constantly in use.

Expanding the discussion -

dialogue and hypertext

Because hypertext rarely follows a single path for more than a few paragraphs, it becomes much easier for a group of authors to collect their work into a single unit. Keeping the entire unit organized and non-contradictory in the usual linear fashion is no longer the point. Many voices can tell a story, the links between them bringing them together for support or opposition. An editor no longer faces the particular problem of classifying different voices and determining an order for presentation in the way that an anthology provides. In hypertext, the actual sequence is no longer important, because the links between the articles are much wider-ranging. An article is no longer a twenty-page chunk; it becomes a collection of lexias, each with its own supporting documentation. The creator then links the lexias to each other and to other lexias in the compendium by subject or argument or simply because the lexias refer to each other.

These links might seem to require much more effort to produce than a simple table of contents, but the text itself provides tools for such connections. In the typical active link model, the editor must choose each link and create starting and ending points for each connection, allowing for tight control but also requiring massive effort and a thorough knowledge of the contents and relations of every entry. In a passive link model (which I hope to put into practice) the editor (or the author) has to assign each lexia a keyword or several keywords. Use of that keyword in other lexias produces a linking symbol of some kind (my prototype simply puts words or phrases into bold italic) that the reader clicks on to reach the connected lexia. Keywords offer the creator a chance to let the text link itself, though it also has some costs. Passive links are slower (because they require more processing), and run the risks of never connecting at all (because the keyword is never used) or connecting too many times (imagine a lexia with a keyword of 'the.') Again, much of the usefulness of hypertext is dependent upon the skills of its creators.

It is my hope that well-designed hypertext environments (and I'm trying to design one) will make it easy for such multi-author collaborations to take place, a type of electronic conference where multiple parties get to interact. It won't be 'live,' necessarily, but it will be more than a single viewpoint presented for a length of 200 pages followed by another viewpoint presented for 150 pages followed by another.

Empowering the historical reader

When the author provides complete documentation of sources by actually providing the source material, it becomes much easier for the reader to develop different opinions based on a different interpretation of the same source material. When multiple opinions are presented simultaneously for the reader's consideration, the reader is allowed (or perhaps forced) to make choices about the reading that they might not normally have noticed. Hypertext and related technologies make an immense amount of material available, so much that many readers will have to decide what is worthwhile and what isn't, a decision normally made by the author. Whether readers will want to make such choices is an open question, but it remains clear that hypertext has the potential to give the reader control over many decisions previously made by the author(s).

An author might still seek tight control over the access a reader has to material by strictly limiting their choices, or rewarding some choices with more interesting material and punishing other choices with piles of seemingly useless raw data. Hypertext doesn't yet obliterate issues of control, but moves them to a different level. The author no longer has direct control over the order in which the reader will approach the materials but maintains enough power to place limits on what materials can be read in what order. My worst fear for hypertext is that some authors will take advantage of this possibility to create subtler forms of control that may in fact lead to less possibility for the reader to question the text than would a book written in a less 'open' medium.

In the hypertext that I hope for, the author accepts the loss of control over his or her work and allows the reader the chance to join the dialogue. The reader will have real opportunities to explore a wide variety of views on a subject, with access to the original materials being discussed. The viewpoint of the author should be clearer, identified as the work of a particular person (especially in multi-author situations.) Because original materials are available, the reader can always verify claims with a minimum of effort, and develop views in conjunction with the author's work without necessarily being created by the author's particular work.

Some systems go further and allow the reader to participate as author, adding comments and even adding new lexias to existing works. This is impossible on a CD-ROM, but it seems like a logical extension of the hypertext medium in networked information systems. It carries some risks, of course. The random Nazi with a connection to the network could link an article praising Kristallnacht to a work on the Holocaust, or simply make threats. Without some controlling editor, such threats could overwhelm the original intent. This may be a danger, but I suggest that it's a danger worth accepting in the interest of giving these texts a 'life' of their own. Revisionist historians of any stripe have always been a part of the discipline. Outside 'kooks' may well provide something interesting or provocative, and, at worst, a reader can close a window on a computer screen as quickly as a book. The potential for readers to become authors seems to me the real goal of hypertext, allowing more people to make decisions about their reading and then allowing them to participate in the actual discussion.

Toward a hypertext historiography

At the moment, this entire paper is conjectural because hypertext has yet to become a significant force in any writing, and I have yet to see a single serious work of hypertext history. I know that there are, for instance, CD-ROMs about the Vietnam War, but have yet to see any of them. My own project of creating a hypertext system that would be appropriate for historical (among other) writing is in its infancy, and my hope of writing hypertext history myself is not one I expect to fulfill within the year. This makes it rather difficult to speculate on what impact hypertext might have on historical writing. Nonetheless, I have a few ideas, none of which I can be certain will come true.

I hope that hypertext's emphasis on the links between text rather than its content will bring new attention to the role of historian as the shaper of links between what are often widely disparate accounts. The 'facts,' if we may be certain enough of them to grant them that title, are only a small part of the historical enterprise. The job involves (re)searching through and selecting tiny bits to assemble a story about what might have happened and why. I can imagine an experimental history where the 'author' simply assembles primary sources and writes by placing links between them, allowing the reader to draw their own conclusions. This might not be an ideal model for all history, but its successful implementation could raise some good questions about the type of project history wants to be.

Hypertext's capacity for including multiple voices seems to me to offer a way out of the kinds of battles historians face, where book reviewers attack each other's works as 'orthodox' or 'revisionist.' Although it may be a compromise neither side wants to accept, hypertext offers both sides the chance to present their views, side-by-side, allowing the reader to make decisions between them. Outside reviews could become part of the fabric of a book. The politics of keywords might suddenly become important, or perhaps a hypertext system might allow for lexias that present multiple views on the same subject in multiple columns. The opening to multiple voices offered by hypertext suggests a serious heterodoxy - that all views are equal, worthy of being given equal time even though it's 'obvious' that this one is terribly biased by socialist leanings or that that one is just another neo-conservative blowing his nose on the downtrodden souls of the third world.

This is a heterodoxy I am willing to accept and even to endorse. History is, and always has been, about considerably more than 'rigor.' One person's truth is another person's opinion is another person's sin. Even after an author considers historical standards, the question of interpretation always remains. The gaps in the historical record are vast, and our ability to understand the motivations of others long past has always been limited. We need to realize and accept this fundamental problem, and move on. Hypertext, I hope, will allow us to realize the gaps in any fabric, and see the connections we use to fill the gaps as patches, illuminating us at the same time they demonstrate how little we really know.

What offers me the most hope, however, is the potential for bringing the reader into the discussion. We can build connections, provide as much information as we feel necessary, and make argument after argument upon that base, but the reader retains the right to navigate as they want. The author will have a much harder time excluding information, especially in large collaborative environments. As readers become used to searching for materials as they want them, they may become more inclined to question when they can't find something that seemed important. The potential of hypertext for amplification of material makes it harder to hide material that is deliberately kept silent. In seriously interactive systems, readers could even do their own research and add to the web of information, filling holes as they see them. Strong interactivity redefines the reader/author relationship, allowing the readers and authors to collaborate in the production of the work, rather than requiring the reader to simply absorb the production of the hard-working writer.

As little as historians may want to admit it, the type of history we have been writing is as dependent on the way we write it as it is on what we write about. Hypertext is a new medium, perhaps the first serious change in the format of writing since Aristotle and other classical writers laid out the basic models of rhetoric thousands of years ago. At present, it seems like only a dream, a vision of some crazed computer scientists that has yet to affect 'serious' historical writing. This brief essay is just a dream. Only the years ahead can tell if it will survive the waking.