Electronic Discovery Search-Strategy
Litigants by their nature are in search of advantage over their adversaries. In this computer age, new tools for advantage emerge every day.
An intriguing discipline that could yield new tools is linguistic analysis, especially sentiment analytics. A company named Jodange searches vast quantities of text in social media (Twitter, Facebook, LinkedIn, Yelp, online dating sites) to ascertain the opinion (the sentiment) of thought leaders or interest groups. It employs computer-driven algorithms to analyze, for instance, the compositional semantics within text (e.g., tweets on Twitter) to glean what, say, US consumers think of brand X.
For sentiment analytics, I see a promising application in what we lawyers call "eDiscovery." eDiscovery is a rapidly growing field, where much money is at stake. eDiscovery includes the collection and review of vast quantities of data (especially email) in a lawsuit or investigation. A business or government lawsuit can involve tens of millions of emails -- too many to read and understand one-by-one. Linguistic analysis, and sentiment analytics, could be valuable to a litigant seeking e-Discovery (EDD) advantage.
Intent Influences Legal Results
Why do I think sentiment analytics could be especially relevant to a lawsuit? The reason is that very often legal liability or outcome depends on the “intent” of a person or an enterprise. For example, to commit a crime a defendant must usually intend to do something wrong. No intent, no crime.
In complex criminal cases, like one against a corporate defendant, the defendant strives to show that it had no intent to behave badly or break the law.
To deduce the intent of people or entities from large numbers of communications such as e-mails is not easy for a jury or a judge. Sentiment analytics could help. Just as it can help interpret the opinion of 20,000 consumers, it could help answer questions like, what was the corporation thinking, what did it mean to do, what was its intention?
Example in a Lawsuit
I imagine this kind of argument being made by an antitrust defendant in court: “Sentiment analytics of four million emails within ABC Corporation reveals that neither the company, its staff nor its leadership had a plan, desire or intention to fix prices, to conspire with competitors or to otherwise violate the antitrust laws.” Presented correctly, that kind of argument could be powerful, and could open a new chapter in American jurisprudence.
The cases to which sentiment analytics applies need not necessarily involve extremely large numbers of messages. Sentiment analytics might be useful for interpreting, say, a few hundred messages in a child custody battle or the augmented reality advertisements broadcast to the patrons of a restaurant during a particular week.
Investigators -- whether they be courts, auditors or regulatory authorities -- need new tools. Linguistic analysis tools could help any kind of investigator make sense of voluminous electronic evidence.
Mr. Wright teaches e-discovery and e-investigations law at the SANS Institute.