Words come in and out of fashion — and election year headlines provide exceptionally rich fodder. Some words and phrases will work their way into everyday speech while others are slated for the dustbin of history.

When new words emerge suddenly to express an opinion we’re tracking with our technology, we’re able to keep up. How?

Our technology “reads” language with an approach based on how humans are likely to express themselves. For example, if you say that a politician is a jerk, stupid, an idiot, etc. and then come up with a new word to describe him (say, fluborizer) that means the same thing , you’ll surely use that word as well as lots of other words already known to the algorithm. Human expression of opinion is often inefficient and repetitive — that’s what makes it recognizable and authentic. Humans’ predictable language patterns enable our technology to keep up with language shifts as they occur.

Our technology — which amplifies human intelligence and can understand new words — can be contrasted to tools which merely count words chosen ex ante. If you set a fixed lexicon and choose the wrong words, you can end up missing the opinion you’re targeting.