Improving our sophistication in the area of knowledge classification systems — a.k.a. taxonomies — is a key step in building more advanced legal knowledge systems. Taxonomies are synonymous with organized knowledge.
Do you organize legal articles or "precedent" contracts for future use? Go beyond a simple Web site favorites list and organize the resources you want to have at your disposal? You are working with taxonomies, whether or not you realize it!
Most taxonomies are subject lists or hierarchical outlines of subjects. The nested outline style of taxonomy is so pervasive that we may even think it is the natural order of things, rather than simply what has been feasible up to now. But there is no longer any reason to try to "fit" knowledge into the artificial rigidity of simple hierarchies.
Of course, an outline of broader and narrower topics — i.e., main topics, subtopics, sub-sub-topics — is a primary component of a thesaurus. But the thesaurus concept recognizes that this is but one of the many relationships that can exist between two pieces of information or concepts. There are also related terms, preferred terms, non-preferred terms, and other relationships.
If you have a little gray in your hair, you may remember what a change it was when we moved from black-and-white to color television. We knew all along the world came in Technicolor, but it took quite a while for our systems to be able to show what we already knew to be true. Thesaurus-based systems offer the same potential to show us, explicitly, knowledge we understand but have not been able to depict through our print and electronic systems, such as document managers and search engines.
In the legal profession, we love to kick around sophisticated abstract concepts and argue and debate about meaning and nuance. Particularly, we love analogy, which is a "related" sort of relationship between the current situation and some other.
Read any formal legal writing and you will find special signals or connectors: e.g., accord, see, see also, cf., contra, but see, but cf., see generally. Lawyers, judges, and legal scholars have always recognized that the intellectual aspects of law and law practice cannot be adequately expressed simply.
Some of these signals mirror classic thesaurus relationships. Cf. is just an obtuse way of saying "related."
SWITCHING TO A THESAURUS
We launched LawMoose in the fall of 2000 as a regional legal search engine. Initially, we covered Minnesota law sites, but we have expanded into Wisconsin. But most of our evolution has actually been about providing more organized knowledge and about making search secondary.
Our first step was to add a small-scale "Yahoo-style" directory of key law sites in Minnesota and Wisconsin so that people could navigate right to the key legal resources in our region without having to search for them. But it would not scale, mostly because we did not build a scalable architecture for it. It was conceived as a simple, quick directory. That it was, but its simplicity also prevented its evolution.
Meanwhile, we were doing some other work, organizing the areas of law practice, for another project we were working on. We just could not make the areas of law practice fit into a hierarchy. Reality was complex enough that it would not fit into any simple outline. We began to read about and then work with thesauri.
When we began to see logical and usability improvements with practice areas that stemmed from the use of thesauri, we began to think of building a replacement for our "Yahoo-style" LawMoose directory as well.
We unleashed our new thesaurus-based directory to the world in July 2003. At first glance, it was even simpler than our earlier directory had been. The entry point to the thesaurus of resources was just a two-column listing of a dozen topics corresponding to the "top terms" in our thesaurus of legal topics and resources. But underneath, if a visitor clicked any of those top terms, we were now publishing an actual collection of highly organized resources. Now, we had an equal partner for our existing search engine. Directories, built by real people and showing evidence of human insights, have strengths that search engines can never have.
ADVANTAGE OF FLEXIBILITY
We learned the value of having a highly flexible editing capability right away. When our old directory disappeared and our new thesaurus-powered one appeared instead, we heard from some folks who were fonder of that old directory than we were. They were used to it.
From our perspective, the whole idea was to take a risk, to experiment to see if we could advance the state of the art about how useful LawMoose could be. At the same time, we did not want to unnecessarily experiment on our visitors in real time.
So, temporarily, we brought back our "classic" directory as an alternative for those who still found it more familiar. Meanwhile, we reshaped the topics, particularly the top-level topics, in the thesaurus, and continued to add resources.
We felt we had turned the corner when we received this comment:
"Great new home page! ... It looks a lot cleaner. I didn't think you could have organized it any better but you, as usual, found new ways."
A few weeks later, we dropped the classic directory link. Since then, no one has ever inquired about it. Now, in our thesaurus-powered legal knowledge collection, resources appear under as many topics as make sense without having to duplicate them. Topics interconnect with other topics in as many ways as make sense. Sometimes we place a sub-topic under two broader topics instead of just one. And, most importantly, each topic is enriched and rendered more usable by a host of "related" links connecting a topic to others logically related to it. These connections would simply be "outside the grid" of an outline style hierarchical directory.
In the year that followed, we increased the number of topics we published by 400 percent from 250 to 1,000 and the number of resources by 1,300 percent, from 600 when we launched to 8,000 a year later.
Our collection of knowledge grew denser with interconnections. You can now locate tens of millions of relevant pages on the Web via our knowledge map — not because we have indexed the pages into a search engine, but because we have identified key intersections, entry points, and Web destinations. And created, over time, a detailed map of how to get to them, directly, or through one of the resources we list.
But what about search? We surely haven't abandoned search. Far from it. Our thesaurus comes with its own search capability. If it can associate a search term with any of the thesaurus topics we publish, the first thing one receives in search results will be the list of appropriate topics, followed by the specific resources matching the search term. That way, visitors can choose to approach their inquiry either topically or specifically, as they may prefer.
Based on the volume of information we all have to deal with these days, we believe that practical taxonomies and knowledge systems capable of organically evolving to handle growth and change will be increasingly critical to our ability to keep on top of everything we need to know.
Minneapolis' LaVern Pritchard is the founder of Pritchard Law Webs, and the publisher of LawMoose, www.lawmoose.com.