Case study #1: The history of "losers"

Read Paul Krugman, "You're all losers", NYT 1/13/2014, and Mark Liberman, "Losers", LLOG 1/14/2014.

Learn about the corpora that can be searched at Look through some random samples of "losers" and "loser" in at least two of the available collections (e.g. COCA and GloWbE). How would you divide the examples into "senses" or other usage sub-types? (Note that you will need to register at the site.)

Look up loser in the Oxford English Dictionary (OED), Merriam-Webster's online dictionary, and the various dictionaries available through Learn about Wordnet and look up loser in the online version. (Note that is freely available if you're accessing it through Penn's network; from outside the university, you'll need to go through the library's proxy service, starting with a search for OED in the "FindIt" box...)

Optional: consider the translation(s) of "loser" and "losers" into other languages known to you.

Come to class prepared to discuss these questions:

How do the senses or other sub-entries in those reference works compare to each other, and to your own analysis of the corpus-based text samples?

How do these analytic choices relate to the claim made in Paul Krugman's blog post?

I suggest that you take some notes on your investigations.

In the class discussion, we'll decide on a suitable classification scheme, and make a plan for applying it.

Against that background, we'll learn about algorithms for "word sense disambiguation", and explore how to evaluate them.