I've been a post-secondary student for almost seven consecutive years now so I feel pretty comfortable saying that I'm familiar with academic research. I did my undergraduate in social psychology which means that most of my research involved various databases that linked to journals my school had subscriptions to and Google. I liked using Google mostly because I already knew how to use it. I would sometimes search for articles on Google, find something I wanted, hit a paywall, and then search for the same article through my university database. Of course, I also came across Google Scholar which was often a lifesaver.
As I entered legal education, I still used university databases and Google scholar for articles but now I had to look for cases and for that we were taught how to use Westlaw and LexisNexis. Searching for case law is a little different than searching for articles. With articles, I could tell from a brief glance at the abstract whether it would be useful or not. Not quite with cases ... I had to consider its authority in regards to jurisdiction and level of court, what the dissenting opinion (if any) was, what qualifications were made, what the specific fact scenario was, which area of law it covered, which other cases it cited and how, and most importantly, if the case had been overruled. And there could be hundreds and hundreds of potentially relevant cases depending on which keywords I put in. As you might have guessed, going on Westlaw and LexisNexis was not exactly my favourite thing to do.
But I wanted to get better at legal research. I like research! So I took the Intensive Legal Research and Writing course taught by John Davis, a former law librarian with a computer science background. The class was fantastic - I learned so much about legal research and had some opportunities to nerd out with Prof. Davis. However, one of the main things I learned was that legal research is kind of stuck in the past.
For instance, I was writing a copyright law paper and I wanted to look at earlier versions of the statute. Well, CanLII is great but its earlier versions only go back to 2003. There's also the official Department of Justice website that does slightly better, going back to 2002. What if I wanted something really old? Like 1985? Well, the Law Library Microform Consortium digitizes some revised statutes but they unfortunately haven't done 1985. So I'll look at the next oldest one available, 1952. (If you are an Osgoode student, you can click here to follow along if you'd like.) So everything is scanned in, which is great, but there's no text recognition so I can't CMD/CTRL+F. There's also no table of contents available so I can't jump to a page. No, instead I have to download PDFs almost randomly, hoping I land on the right table of contents page that will tell me which other pages I have to download to get the Copyright Act. This was incredibly tedious and frustrating.
Luckily, day to day research isn't quite so bad but the tools that are popular right now are far from ideal. For example, it's slightly ridiculous that I have to click at least three times to download a case from Westlaw when I only have to click once to buy something from Amazon. Westlaw Next is slightly more user-friendly, functioning much like a Google search but it's not yet available for Canadian case law.
Two days ago, Anurag Acharya, one-half the brainchild behind Google Scholar came in to speak at the Legal Technology & Informatics class I'm auditing at Stanford. The class is taught by Ron Dolin, who wrote his Ph.D dissertation on search and started the search evaluation team when he worked at Google. Ron later got his law degree for fun, something that I still can't get over.
Anurag recently delivered a presentation at Cornell on the Google Scholar approach to legal search that he shared with us again in part. Currently, like Westlaw Next, Google Scholar is only available for American legal documents, which is a shame. However, of what is available, Google Scholar is already better than everything I've tried. The best part of using a Google product is that I don't have to spend hours figuring out how to use it. Now, that's partially because I've been using Google since I was 10 years old but it's also because Google is designed to be intuitive.
Google Scholar emerged from a problem that engineers couldn't quite solve. They wanted to rank scholarly documents higher on search results but couldn't gauge whether the user was looking for scholarly results or just a quick answer. For instance, if I'm searching for "Miranda warning", maybe I'm a layperson who heard the phrase somewhere and wanted to know what it meant i.e. the right to remain silent or maybe I'm a legal student/professional and I want to find the 1966 case i.e. Miranda v. Arizona. Google couldn't tell which category I fell into and therefore couldn't optimize the search results. Eventually, it was decided that scholarly articles would be separated from the regular search engine and Google Scholar was born.
On the legal search end, the overall goal of Google Scholar was that everyone should be able to find and read the laws that govern them. The team is small (5 people!) and the task is complex so everything has to be automated. One challenge of legal indexing is working with citations. For example, ibids are difficult to deal with because sometimes it's unclear which source it's referring to. Similarly, determining the significance of a citation can be tricky. Judgments can cite another case as precedent, persuasive, or simply as obiter dicta. So how how can the process of determining which cases are cited and of those, which citations are significant, be automated? (I have no idea.)
Here's how Google Scholar for legal documents work right now ...
I can search for a doctrine/keywords - here, I chose "fair use".
Campbell v. Acuff-Rose Music, Inc. is the second result, great! (This is the one where 2 Live Crew parodies Roy Orbison's "Oh, Pretty Woman".) Now, I can view the case.
I can see where the case was cited. I believe there is some effort made here so that the most relevant citations appear on top.
I can also see how other cases cited it which is very helpful.
Like other legal search tools, I can restrict my search by jurisdiction.
Considering Westlaw and LexisNexis charge hundreds of dollars an hour, Google Scholar is not bad for a free service. It can't tell me whether a case was given positive or negative treatment and won't let me search through claims in the patents database but otherwise, I can't wait for them to build their documents and work their way to Canada, the UK, and Australia.
Google Scholar won't be rolling out anything intended for private markets but I have high hopes for Ravel Law, a startup that spun out of Stanford. Their search results aren't quite up to speed yet but the interface is beautiful.
Here's what I get when I put in "fair use".
If I put in "fair use" + copyright, it gets better.
Fingers crossed that between Google Scholar and Ravel Law, no one will ever have to spend hours (or have to take a class) to learn how to use legal databases again.