“Sentiment Analysis”

From the Blog of Phyllis G. Pollack.

       Last week, I discussed a study focusing on the effect of e-mediation (that is, a software program) on negotiating behavior. The study sought to determine whether a computer could mediate as well as, if not better than, a person. While the computer did reach resolution more often, people still preferred using the “live” flesh and blood mediator.

       In my mediations, I have often found that a dispute arose due to a lack of communication or a miscommunication. That is, people did not mean what they said or did not say what they meant (or did not say it at all).

       Now, it seems that there is a computer program to help us accurately understand each other (and avoid disputes altogether.) In the Technology section of the October 6, 2009 edition of The Economist, the authors discuss using computers to analyze sentiments. In their article entitled, “An Emotional Response,” the authors discuss research (by Stephen Pulman of the University of Oxford and Karo Moilanen, one of his doctoral students), using “Sentiment Analysis” software, to assess the emotional meaning of text and then labeling the words as positive, negative or neutral:

      “The analysis is then broken into steps that progressively take into account larger and larger grammatical chunks, updating the sentiment score of each entity as it goes. . . .”

      “By applying and analyzing emotional labels, the software can construct sentiment scores for the concepts mentioned in the texts, as a combination of positive, negative or neutral results. . . .” (Id.)


       But the software does not simply do a tally; rather it applies a weighting to each word so that in the end, the software can determine whether a sentence has mainly a negative, positive or neutral meaning.

      So, you wonder – to what use can this software be put? It seems that it can be used by companies seeking to identify unhappy customers (in an effort to stave off litigation such as “lemon law” suits) and by intelligence agencies seeking to determine if that e-mailer really does pose a threat to national security.

       My question is whether such software should be part of my mediation toolbox so that when people say something to me, I can let the software tell me whether it was meant in a positive, negative or neutral context. It will let me know whether the party really said what she meant or meant what she said! If so, then so much for reading body language and knowing all of those other “tricks” in my toolbox. Now, computer software can do “it” all for me, including resolving the dispute! Technology . . .where is it taking us! Maybe, “back to the future!” 

       . . . Just something to think about.


Phyllis Pollack

Phyllis Pollack with PGP Mediation uses a facilitative, interest-based approach. Her preferred mediation style is facilitative in the belief that the best and most durable resolutions are those achieved by the parties themselves. The parties generally know the business issues and priorities, personalities and obstacles to a successful resolution as… MORE >

Featured Mediators

View all

Read these next


Mediating f2f Holographically

Ok, when you are truly a carbon neutral mediator in future years and refuse to travel big distances to your high stakes mediations, remember you saw world’s first live holographic...

By Geoff Sharp

Are They Crazy? No, Just Human

This article was published in the Detroit Legal News on March 2, 2021.In last month’s ADR Spotlight my PREMI colleague, Laura Athens, authored an incredibly insightful and important article on...

By Richard Hurford

Effective Alternatives Analysis In Mediation: “BATNA/WATNA” Analysis Demystified

BATNA: Best Alternative to a Negotiated Agreement WATNA: Worst Alternative to a Negotiated Agreement Popular Attribution to Fisher & Ury, Getting to Yes. Introduction: In most settlement negotiations, parties are...

By Jessica Notini

Find a Mediator