Legal Business Blogs

Guest post: Big data and the linguistics of law – how to win cases and influence judgments

With the hue and cry about big data, artificial intelligence and law being accentuated by all the talk around Watson/Ross, I was interested to see this thoughtful, carefully carried out and documented piece of research into the qualities of brief writing before the US Supreme Court from US political scientist Adam Feldman

It uses corpus-based techniques to examine quantitatively the quality and sentiment of brief writing and its impact on judicial decisions. Its basic measure of quality is the extent to which the judges adopt phrases from the brief in their judgment, but also whether the party actually won the case. There are some interesting findings in support of the view that experienced lawyers do better.

Among the positive findings are:

‘Increased attorney experience positively affects the amount of language the justices share with merits briefs’ (they also win more cases)

positive brief sentiment positively affects the amount of language the justices’ share with briefs (though these elements are associated with losing more cases) suggesting, ‘that attorneys benefit from writing with a positive tone and from excluding negative and offensive language in briefs.’

‘…the justices’ tend to share less language with more complexly written briefs…’

‘Increased wordiness and passive verbs in briefs, both indicators of lower quality writ­ing, leads to less language shared between opinions and briefs.’

Increasing complexity as measured by reading age is beneficial but only, ‘until a post-high school grade level of around fourteen.’

Similarly, legal briefs should balance complexity (desirable up to a point) with readability.

The findings on more experienced attorneys is particularly interesting.  They do better in terms of the results they get and the extent to which the justices use their language, but there are, ‘no dramatic differences in the median values for any of the quality criteria between more and less experienced attorneys.’  Linguistically they are not much different to the inexperienced attorneys, although:

More experienced attorneys tend to write briefs that use more complex language. This is apparent from their higher readability and reading grade scores. Less experienced attorneys tend to write briefs that have more sentence complexity and as a consequence the justices may perceive their briefs as more rambling. More experienced attorneys write briefs with higher sentiment values, which may be a sign that they are attuned, with their additional experience, to the justices’ preferences towards such briefs.

…The overlap values indicate that the justices tend to adopt more language from briefs by more experienced attorneys when the quality levels of their briefs are the same as briefs from less experienced attorneys.

Of course, the study cannot explain why this happens: do experienced attorney’s pick better arguments; write more persuasively in ways not measured by the linguistic software; have clients with stronger cases; or, are they more trusted by the Supreme Court justices?

Adam Feldman has posted another paper on who wins in the Supreme Court here that I have not looked at yet.

Richard Moorhead is Professor of Law and Professional Ethics at UCL, you can read his blog here.