I'm looking at using Google's Ngram to pick out the word that is the closest to an opposite of a given word by looking at the frequency and correlation over time of the two words. For example, I predicted and found large and small have similar frequencies and appear positively correlated, and so do huge and tiny. Nevertheless, a careful reading of the descriptions of words from a dictionary is still likely to be the best way of finding an opposite.

What suggestions does everyone have for refining this approach? What concerns of potential pitfalls to this approach are there?


closed as not constructive by Matt E. Эллен, JSBձոգչ, Mitch, Daniel, tchrist Nov 13 '12 at 20:55

As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, visit the help center for guidance. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 1
    That would be the FAQ; I could see where some might regard this as a "chatty, open-ended" question. Of course a dictionary is a better place to learn about suitable antonyms than an Ngram. Your Ngram is interesting, and provides some measure of evidence that tiny & huge could be coupled, much like small & large are, but that won't always work. So, there's not much to say here except, "Hmm, that's interesting," and, "Yes, dictionaries & thesauri are the best resources." – J.R. Nov 13 '12 at 8:58
  • 2
    Choosing an antonym by frequency rather than context won't always work. Every word has connotations & denotations that have to be considered. Sometimes the obvious antonyms don't work because the style or register is wrong. Sometimes certain contexts demand certain pairs of words in opposition. I often decide that a standard antonym doesn't work in a specific sentence in the articles I edit every day. The four examples you've chosen don't need an Ngram to help you decide. The result was predictable to anyone who reads extensively. – user21497 Nov 13 '12 at 9:03
  • 1
    There's also a matter of polling. Considering your question: I'm wondering whether anyone has looked at the only direct answers would be "I did", "I didn't", and "I know someone who did". These are rather of dubious value. A question "Best method to find closest antonyms?" would be much better - with your idea given as your research outside of the common dictionary comparison, and one could expect other good ideas for this. – SF. Nov 13 '12 at 9:08
  • 1
    @BillFranke: I must agree. Direct antonym to "Short" would be "Long", but then when you're talking about a person, choosing the direct antonym would be plain wrong. The context dictates to use "Tall" and while strongly secondary, this is the antonym to choose here. – SF. Nov 13 '12 at 9:11
  • 1
    Interesting. You might want to run some tests with unusual pairs to see whether the method is actually helpful. Let us know. – user21497 Nov 13 '12 at 9:37

Nice method!. The only issue I keep finding is that as the graph comes near the present, the differences between the words seem really scarce. (use the examples I give) It seems like in the past there were more clear differences between: stature and height: or distance and length. Or maybe is just me?

Not the answer you're looking for? Browse other questions tagged or ask your own question.