Language Log

« previous post |

Ian Prasad Philbrick and Ashley Wu, "The 9 Elements of a Trump Rally", NYT 10/8/2024:

The energy for Mr. Trump’s third White House campaign comes from his rallies. Since President Biden dropped out of the race and Vice President Kamala Harris took the helm, Mr. Trump has held nearly 20 of them, speaking for about 90 minutes at each.

Like most politicians, he repeats things at every speech. Unlike most politicians, he offers a grim view of the country, makes up nicknames for his opponents and pledges to use the power of the government to punish his rivals.

To help readers experience what a Trump rally is like, we used video to break down the nine themes he consistently returns to.

Those nine "themes" are:

1. Savior and protector
2. Insults
3. Deceits
4. The hits
5. Political violence
6. The Trump agenda
7. Digressions
8. Anti-democratic statements
9. Stumbles

Read the article for the details, as well as clever tableaux of video clips and a cool carpet plot of the theme-weave in his 9/29 Erie PA rally:

I like the general idea, Ashley Wu's graphics are impressive, and it's tempting to use modern topic and sentiment analysis techniques to derive similar things automatically.

But I wonder how good the inter-annotator agreement for the human version of this analysis would be?

One obvious problem is that the "themes" are potentially overlapping — Trump's digressions are often also insults or deceits or threats of violence, etc.; his stumbles can occur in any of the other segments; some of his greatest "hits" are also insults, deceits, or savior/protector assertions; and so on. Furthermore, some of the themes are matters of content or tone (e.g. insults), while others are a question of discourse structure (e.g. digressions) or overall topic statistics (e.g. hits).

Those overlaps and ambiguities will make it easier for an automatic analysis to produce plausible results, but they'll also make the overall results less informative. And while I agree that Donald Trump's rhetorical style is in some ways special, it would be better to demonstrate that with an analysis that positions him in the same space as other speakers.

The cited 9 "themes" can certainly be applied to speeches from other politicians (or other people in general) — but when we start analyzing others, we're going to want additional "themes", and the whole system will need to do a better job of engaging the general problem of discourse analysis.

Still, the article makes sense, and the graphics are great.

October 9, 2024 @ 10:42 am · Filed by Mark Liberman under Language and politics, Pragmatics

Permalink