LING052

Notes on what words mean:
testing Paul Krugman's theory about loser(s)

In answer to the question "when did it become so common to disparage anyone who hasn't made it big, hasn't gotten rich, as a 'loser'?", Paul Krugman offered a plot of word frequencies from the Google Books Ngram Viewer:

Is this a valid answer? Well, that depends on whether the change in the frequency of "losers" is substantially due to an increase in disparaging people who haven't gotten rich -- as opposed to describing teams that have lost baseball games, or politicians who have lost elections, or companies who have been disadvantaged by changes in a law or a policy.

In itself, this is a simple question. We can answer it by looking at a suitable sample of losers examples from the past half-century or so; determining in each case whether Prof. Krugman's characterization ("[disparaging] [someone] who hasn't made it big, hasn't gotten rich") applies; and estimating how much of the overall increase in the word's frequency has been contributed by an increase in the usage Krugman is interested in.

This looks like a problem of word-sense disambiguation -- and word-sense disambiguation is one of the types of analysis that we'll study this semester. For an overview, see the Wikipedia article on Word-sense disambiguation. For a detailed review, see Roberto Navigli, "Word Sense Disambiguation: A Survey", ACM Computing Surveys, February 2009.

But there's a problem: Krugman's sense is not sharply distinguished from other cases in the senses for loser(s) given in dictionaries, or in resources like WordNet.

We could try making up a new sense-division and applying word-sense-disambiguation methods. But alternatively, we could note that our loser(s)-classification problem naturally divides into several sub-problems.

1. Is a use of losers is "disparaging" or not? (And maybe, how strongly disparaging is it?)

2. Are we talking about sports, gambling, politics, war, business, life in general, or what?

3. Who is doing the losing: a team, a company, a political party, a country, an individual? A class of teams, companies, parties, countries, individuals?

4. How is winning or losing being defined: points, games, money, popularity, or what?

5. Are we talking about a specific point of evaluation or a long-time pattern?

There are plenty of other relevant semantic issues, but these will do for the current experiment. Specifically, for each instance of losers, we want to answer five yes/no questions:

1. Does the writer or speaker intend to disparage the loser(s)?

2. Is the "game" life in general?

3. Are the losers individuals (as opposed to teams or countries or etc.)?

4. Is winning/losing defined in terms of money?

5. Is losing seen as a long-term pattern or essential characteristic?

If the answer to all five questions is "yes", then we have an example that fits Prof. Krugman's model.

We'll try asking these questions about six decades of samples, taken from the COHA database. I'll try to set up an interactive online method for us to cooperate on the project, but for a start, here are the texts:

1950s sample
1960s sample
1970s sample
1980s sample
1990s sample
2000s sample