Archive | November, 2012

Bossy pants: How do superiors and subordinates talk?

5 Nov

There are a fair number of corpora the involve friends/family or strangers and it’s likely that there are all sorts of power-plays happening in these. But it’s a lot easier to see what’s happening in terms of power if you start with a corpus where there are very clear power differentials. Here are some thoughts about where to look:

  • The Enron Corpus–there are a few places to find this. The great thing about this corpus is that you can see how people change language based on whether they are emailing subordinates, bosses, or peers. Anecdotal from years working in industry: it is interesting to see how certain phrases circulate. At Microsoft the word “super” (e.g., “I’m super-excited”) seemed to move around mostly at the executive level. In the IT consulting universe, consultants seemed to introduce (and then overuse) colorful phrases like “open the kimono”, “come to Jesus moment”, maybe as a way to stand out/be perceived as valuable.
  • Part of my dissertation is about understanding how the word “little” is used (as a hedge, “Can I have a little water”, for affection, “Look at your little toes!”, and for insulting, “How’s your little project?”). Pages 145-158 use two different corpora to this end: CHILDES, between parents and small children and the ICSI Meeting Corpus between academic computer scientists.
  • Power and Supreme Court Justices (plus Wikipedia editors, too): https://corplinguistics.wordpress.com/2012/09/19/power-supremes-wiki/
    • I think it’ll be hard to prove who has more power in terms of the justices, but there is a clear hierarchy where Justice X > Lawyer Y.
  • You may be able to use the Cornell Movie Dialogs Corpus if you did some extra annotation.
  • Russell Lee-Goodman also suggests this work on small talk and workplace communication: http://www.victoria.ac.nz/lals/lwp/research/small-talk.aspx

If you’re curious about “power” more generally, there’s lit review stuff in my dissertation.

And here’s a short synthesis of how various people in various fields have defined “power” (and an attempt to make it operationalizable for linguists):

You might also check out Jim Scott’s Domination and the Arts of Resistance, which has the idea of “hidden transcripts” (what the powerless say behind the backs of the powerful). But if you get into that, then definitely see Gal’s response for a critique. Gaventa and Lukes work on three dimensions of power is probably useful. And you might have fun with Butler, Muñoz, and maybe Valentine (search these class notes for the names).

Some other related posts on this blog:

I/me/my and presidential candidates

1 Nov

As a follow up to my work on “strength” and weakness” in presidential debates and the Language Log’s Obama Pronouns Again post. The question here: how many first person pronouns do various candidates use? Here I’m defining first person pronouns as I/me/my/myself/mine.

There are 16,076 first person pronouns used by all the debaters and questioners since 1960. But users have very different numbers of words total. Obama has 43,863 words across all debates while Kennedy only has 18,264 words.

Since Obama has 7.73% of all the words in the corpus, we’d expect him to use 7.73% * 16,076 = ~1,242 first person pronouns (if everything were even). But he actually just uses 1,022 (about 82% of what we’d expect).

In the table below, I’m going to report all of the people who have at least 7,500 words. In it, you see that the big users are the Bushes, Dole, and Clinton. The big underuser is Jim Lehrer (he has a lot of words because he’s moderated multiple times–and as a moderator we *would* expect him not to be talking about himself all that much; notice that vice presidential candidates also tend to use fewer first person pronouns, as we’d probably expect).

For thoughts about what first person pronoun uses *mean*, I strongly encourage you to check out James Pennebaker’s guest post on the Language Log. Just because you use a lot of “I/etc” doesn’t mean you’re a narcissist/dominant/arrogant.

Speaker Observed FPP Total words Percent of corpus Expected FPP Obs/Exp
BUSH-GHW 1444 32190 5.67% 911.51 1.58
DOLE 597 15184 2.67% 429.96 1.39
CLINTON 1116 29419 5.18% 833.04 1.34
BUSH-GW 1482 40153 7.07% 1136.99 1.30
KERRY 754 20985 3.70% 594.22 1.27
PEROT 523 15027 2.65% 425.51 1.23
ROMNEY 778 24421 4.30% 691.52 1.13
MCCAIN 639 20205 3.56% 572.14 1.12
GORE 972 31523 5.55% 892.62 1.09
REAGAN 669 21795 3.84% 617.16 1.08
DUKAKIS 418 14007 2.47% 396.63 1.05
QUAYLE 313 10605 1.87% 300.30 1.04
KENNEDY 501 18264 3.22% 517.17 0.97
BIDEN 385 14217 2.50% 402.58 0.96
FORD 393 15245 2.69% 431.69 0.91
NIXON 459 18597 3.28% 526.60 0.87
MONDALE 263 10793 1.90% 305.62 0.86
CARTER 587 24280 4.28% 687.53 0.85
OBAMA 1023 43863 7.73% 1242.05 0.82
PALIN 175 7714 1.36% 218.43 0.80
CHENEY 249 13380 2.36% 378.88 0.66
EDWARDS 129 8196 1.44% 232.08 0.56
LEHRER 205 16574 2.92% 469.32 0.44