Bringing out the Hidden emotions

Internet has ushered in a new age of communication, shrinking the whole world into the legendary Global Village. Email and instant messaging have changed the way people communicate with one another and that too in every sense; now an email can communicate within seconds which the erstwhile snail mail used to take 1-2 weeks, that too in a much more economical way. Even though the technical evangelists would love to uphold these as a triumph of their networks and protocols, social anthropologists are noting a grave shift in the way people now approach the interpersonal relations. Even though the internet has enabled us to be more connected, the same tool is causing people to be more isolated, cuddled within their comfort zones, so that there is less of a person to person interaction.

Along with the exit of face to face communication is the extinction of one of the most important gifts of human communications, the emotions. As any Star Trek fan would recollect, Vulcans void of emotions, charting out everything through pure logic were not an easy race to deal with. Their ambassador, Spock, half-Human half-Vulcan, still displayed the lack of emotion in many of his interactions, much to the annoyance of his crew members (even though his logic was much revered and which made him one of the iconic characters of all time). Given the bland communication, through emails and chats, are we headed for a Vulcan kind of society? Something to be pondered on!

Now coming to the real intent of the article, well, it is not much about the danger of humans going into an unemotional mode, but more about how we can incorporate or deduce emotions from this new means. Whereas the author lacks any formal training in linguistics and other communication means, the inspiration has been derived from my personal experience in the modern wired means of communication. The thrust of the article would be on the methods to deduce the emotional level of the author while authoring an email, SMS or chat.

The written forms of instant communication such as Email, SMS or chat are by virtue devoid of emotions and it would be very difficult to assess the emotional state during penning down of the same. For example, a heartbroken lover, when conversing through telephone may exhibit the effects of the break up through the variations in his/her voice, pitch, tone etc., all of them which are absent in an EMail, say written to a different audience group. It is in such situations where the Mail Analysis can come into play, in an effective deducing mode. This follows from the fact that each person has a writing style, i.e. in usage of words, structuring of sentences, breakup of paragraphs or even usage of smileys. Even though this remains more or less the same, there does occur certain microdeviations from this usual manner. It is by observing these microdeviations that the emotional state can be deduced by reading a simple mail.

Microdeviations in the written communication are very much analogous to the microexpressions in the non-verbal personal communication. Micro expressions (made popular by the popular TV series, Lie to Me) are brief, involuntary facial expression, which may last for about 1/25^th of a second. It usually occurs when one tries to conceal a true emotion, like when lying, out of courtesy, trying to be diplomatic etc. Similarly, micro deviation is a slight change in the usage of words, sentence structuring, usage of punctuation, smileys etc. It could vary from usage of a common synonym for a word to the most common deviation, i.e. excess usage of punctuation, usually the period. It has been commonly observed that people tend to use multiple periods when they are in a happy/energetic mood, with a wide array of synonyms and paragraphs often are shorter, where as in a sad or gloomy mood, a single word will suffice throughout the mail, with conservative punctuation and longer paragraphs.

All said and done, the effectiveness of this method depends on the relation and acquaintance with the sender and his/her writing style. How could this be made into a structured process, which can be scaled and replicated? In the system I suggest, a collection of mails authored while in a normal mood can be used as the control set and with these in hand, a basic structure of the author’s style can be structured in terms of

I. Usage, or rather repetition of punctuations at each punctuation node

II. No. of synonyms for each of the words used in each mail

III. Length of sentences, in terms of words

IV. Length of paragraphs, in terms of sentences

V. Usage of smileys

After this, the calibrating set can be formed by using the mails authored while in certain mood. By comparing with the control set, the deviations in each of the 5 parameters can be checked upon and the deviations be noted. By having calibrating sets for various emotions, such as happy/sad/gloomy/angry etc., the deviation set for each of the emotions can be formed. In this process, a special form of adaptive regression can be used to make the most accurate set of deviations. After this is formed, it’s only a matter of click of a button for the algorithm to analyse each of the 5 elements, find the deviation and deduce the emotional state of the author. So, Jeffie thinks bringing out the hidden emotions brought out through Word Analysis is possible!