The Legal Definition of, “Insight”

A designer’s notes from the field

The canonical image of lawyers as sleek, powerful people rushing from meeting to meeting is not without truth. They exist. But lawyers also do homework. Lots and lots and lots of homework. I’ve found this out up close. In fact legal teams spend so much time and energy conducting research that an eight billion dollar industry exists to serve their needs. A handful of companies have sprung up to modernize the tools of the legal workforce. One of them, Ravel, is where I work as a product designer. Having spent the last few months getting familiar with the practice of practice I’d like to share some of what I’ve learned.

What is Legal Research?

Lawyers look to prior decisions for examples that might help strengthen their case. They accomplish this with a variety of means, including books, Google and asking each other. There are also a range of industry-specific tools catering to the needs of the studious attorney. That’s where we fit in.

At it’s core legal research is an information retrieval task like any other. Subjects are assigned to or come up with a problem, form hypotheses about how to solve it and attempt to find relevant data to support, disprove or modify their assumptions. In this case, however, the body of information being addressed is all of U.S. case law — the record of every decision that has been made in a legal dispute as long as we’ve been keeping track. The laws of the United States are governed by a principle called stare decisis, or, the doctrine of precedent. The translation from Latin means, “to stand by things decided”. In practical terms this means that courts look to previous decisions to decide how they should act. If you’re trying to understand a piece of the law as it exists today you may be looking at law that was informed by centuries of preceding thought. The law isn’t a simple, static entity. When a judge makes a ruling in court they are consulting the past, and adding their opinion to the vast web of knowledge which forms our legal system.

The Roles at Play

The modern law firm is comprised of three primary roles: Partners — the most senior attorneys, Librarians — the information experts, and Associates — young lawyers honing their craft. The actual work of legal research may fairly be described as, “dry”. Some might go so far as to call it tedious. It’s the associates — the young and eager — that are responsible for most of it. Their job is to wade through a miasma of old books, half-remembered anecdotes and clunky websites, returning with a piece of history that perfectly supports the facts of the case at hand. And they shouldn’t take too long to do it. For questions of law, insight comes in the form of history repeating itself. Finding an old case that closely matches the present one is the holy grail. This is like trying to find a needle in a linguistic haystack.

The Information Space

As of October 2015 the Supreme Court counted 560 bound volumes in its opus, averaging between 800–1200 pages each. That’s anywhere from 448,000 to 960,000 pages of precedent, in very small font-sizes, using very dense language, just for the Supreme Court! The final decisions, published after a case concludes, can contain references to heaps of other decisions, which contain heaps of references themselves. There are multiple tiers of lower courts in every state, multiple published versions of many cases, drafts and revisions that never make it to the published work, opinions which are never officially published in the first place, collections of expert opinion and advisement for certain topics, various white papers published by law firms relating their experience and opinions … the list goes on. The average associate will spend roughly a third of their time conducting research in their first two years of work.

The Golden Sentence

In short, we have a massive, unstable, incomplete network of citations to wrestle with. What the associate trying to grapple with this mess needs is what’s called a, “golden sentence”. I’ve conducted a series of interviews with subjects ranging in experience and area of practice, and their processes all share a similar narrative: A partner comes to one of their staff with a request sounding something like, “find me a case that says X”. The associate is given a deadline and turned loose. Their first move, across the board, is to establish an intellectual beachhead — a toehold somewhere from which they can narrow their focus. Time and again words like, “Universe” and, “Ballpark”, have come up to describe this beginning phase of research. They seek to create bounds around some bulk of cases to filter through, and they approach this task in interesting ways.

Three Types of Behavior

This is not meant to be a formal analysis so much as notes from the field, so take it with a grain of salt, but I see three dominant patterns of behavior emerging from my interviews. I’ve given them the shorthand ofSurveyors, Scorched-Earthers and Specialists. These labels represent information-seeking behaviors common to groups of our participants, and I use them as I make design decisions in the Ravel product today. The terms are not perfect, nor are they meant to be, and not mutually exclusive.

Surveyors are so named because they often begin their inquest by scanning the field to see if anyone else has dealt with their particular problem. Algorithmically speaking we might consider this a form of depth-first search. Their sources may be colleagues or peers at other practices, experts in a given knowledge domain, or a variety of secondary sources available in print and online. They try to find treatments that can inform their own work so they don’t have to reinvent the wheel. They want to be pretty sure that they’re getting it right without having to boil the ocean.

Scorched-Earthers, by contrast, want everything. Breadth first. They want to lay the full span of human knowledge before them so that they may inspect it and pass judgment. This comes both from a desire to construct a thoughtful, deliberate argument, and to establish a strong position for their work product by consulting — and making it known that they consulted — every possible resource. They commonly express concerns about missing something. And they still might. They don’t necessarily spend more time conducting research based on what we’ve seen, but they are highly aware of the apparent completeness of the information they’re working with.

Specialists, as the name may imply, inhabit a niche. They operate in subdomains of law, either topically or geographically, which allow them to focus their attentions on a limited subset of source material. This group develops a deep expertise in a narrow focus of practice and can act as references for less familiar parties. They typically have a single resource or a small collection that acts as the known starting point for all other research efforts — their professional bible. Whatever field or region they’re in, they have their reliable go-to. If a tool doesn’t include their preferred source its utility to them is lacking. They need, however, the most help when starting projects in unfamiliar areas.

Traditional Search

Upon entering any law firm ten years ago you would be greeted by walls and rooms filled with heavy, thick books. In the face of rapid technological advance and an industry-wide race to cut costs these libraries have withered. They were, however, the length and breadth of research for a long time. As trials wrapped up and judges issued their opinions they were gathered and published in collections called, “reporters”. Back in 1872 a pair of brothers started this practice in St. Paul Minnesota and built out a nationwide network of regional reporters. The duo, Horatio and John, shared the surname, “West”, and named their fledgling company after it. Over the last 150 odd years West Publishing became the de facto standard for published legal information, and with the advent of the web created Westlaw, an online tool for legal research. Westlaw and its peer LexisNexis offer boolean search tools (the complicated system of, “and”-s, “or”-s, slashes and dashes that form a logical query) which access a huge body of proprietary content. That content is edited manually by armies of real live humans. Imagine if Google bought the rights to web pages and hired people to read each one, taking notes by hand. It’s kind of like that.

A New Approach

Some smart people think there’s a better way.

Enter the powers of computation.

Machine learning, at its simplest, is the practice of feeding computers a handful of examples of what you want them to do, then letting them tear through huge quantities of data, following your examples, and looking at the output to evaluate how well those examples worked. It leverages the strengths of machines, in terms of power, speed and consistency when performing rote tasks, and the strengths of people, in terms of judgment, creativity and interpretation when defining those tasks and assessing the results. My company is applying this and other emerging technologies to legal research.

As a designer, I get to imagine ways to connect the people with the system.

It’s a really cool job.

I’ve made it my mission to understand just what insight means to attorneys. Does it mean that golden sentence? Or maybe several sentences? Or a single word? Could the voting patterns of judges when serving on a panel suggest a different trial strategy? Might the definition of a federal regulation have shifted ever so slightly in 1983, and might that shift have profound implications for a case being litigated today? How does a first year associate practicing in Wyoming learn about contract law in Delaware? Important progress has been made towards answering these questions, including identifying the behaviors mentioned above. These and other findings achieved through a human-centered design process inform everything from interface design to data science projects at Ravel. I don’t have all the answers yet, but I have the tools and the means to find them out. Stay tuned for more thoughts on design in the world of law.