A trip to Harvard: Connecting Language Data to Personal Impact
Diverse thinkers came together at the Ways with Words conference to find commonality in gender, language, and data.
Running a language data business means that you uncover interesting linguistic traits of groups of people every day. It is, after all, what our clients use Relative Insight to do. Understanding the small but significant differences in language between different consumer groups can reveal critical insights into those groups of people.
For example, understanding that 20 year olds tend to say they ‘wear’ makeup and 50 year olds are more likely to say they ‘apply’ makeup means you can connect with these different groups more effectively and understanding the difference between the words is important. Why do small differences in language make such a big difference? A few reasons, but mainly language use has a deep and personal use so small changes in words can have a large emotional impact.
It’s something you can lose sight of when you process language as data for a living. However, recently I had an experience which brought back the individual impact language can have. I was invited to speak at the Harvard University Radcliffe Institute conference ‘Ways with Words’ in early March. The two-day long conference explored the themes of gender and language, and how language affects gender perceptions.
It was the most diverse set of thinkers that I’ve been in the presence of: physicists, social scientists, advertising executives, political strategists, gender rights groups, trans-gender rights groups and data scientists were all represented. During the day I reconnected with the impact small changes in language can have on people’s personal lives.
One of the most interesting facets of the discussion of the day was language and gender in politics.
We’ve developed our analytical techniques far beyond simple ‘male’ and ‘female’ as we have found that the world simply doesn’t exist in a simple binary way.
Rather, we see language across a spectrum. While it is true that men use certain language attributes more than women and vice versa, it doesn’t mean that these traits are exclusive to either gender.
At the event, I presented some of this in relation to US politicians. Donald Trump has a linguistic signature which shows up irrespective of whether you compare him to Clinton or to Sanders. However, even more tellingly, Hilary Clinton has a language style which marks her out as Hilary Clinton, irrespective of whether you are comparing her to male or female politicians.
In other words, Hilary Clinton doesn’t speak like a female presidential candidate; she speaks like Hilary Clinton, and gender has nothing to do with it.
Sometimes society is desperate to label people into mutually exclusive groups but if you use language as data you find the world seldom works like that.
If you want to learn more about the insights I presented alongside Lyle Ungar, processor at the University of Pennsylvania and Alice E. Marwick, assistant professor at Fordham University, you can watch the entire presentation here:http://www.radcliffe.harvard.edu/video/ways-words-big-data
If you’re interested in all of sessions you can see them in the links to the left of the page. They are well worth watching, especially Janet Mock’s interview — http://www.radcliffe.harvard.edu/video/ways-words-janet-mock