The words 'yes' and 'no' may seem like two of the easiest expressions to understand in any language, but their actual behavior and interpretation are surprisingly difficult to pin down. In a paper published earlier today in the journal Language, two linguists examine the workings of 'yes' and 'no' and show that understanding them leads to new insights concerning the understanding of questions and statements more generally.

Floris Roelofsen of the University of Amsterdam and Donka F. Farkas of the University of California Santa Cruz have written what they call a comprehensive account of 'polarity particles', as these words are called, across languages, and explain the intricate pattern of their distribution. For example, "Yes, it is" and "No, it isn't" are acceptable answers to the question "Is the door open or is it not open?", but not to "Is the door open or is it closed?".

Furthermore, the intonation used when pronouncing a sentence can affect whether 'yes' or 'no' are appropriate responses to it.

The distribution of these particles, it turns out, is also affected by the polarity of the sentence they respond to. For example, both "No, he hasn't" and "Yes, he hasn't" are acceptable as agreeing responses to "Ben has not called today", but in an agreeing response to "Ben has called today", "Yes, he has' is acceptable but "No, he has" is not.

Roelofsen and Farkas build on previous insights from semantics and discourse models, as well as on quantitative surveys of how speakers judge various responses. The framework they create not only explains the distribution and interpretation of these particles in English, but also predicts what patterns one expects to find across languages.

These predictions are then checked and verified against data from French, German, Romanian, and Hungarian.