Commentary: Tone languages and laryngeal precision

In this response I make three distinct points. First and most important, I identify a logical premise of Everett et al.’s (2016) explanation for the correlation between humidity and complex tone, and briefly present empirical evidence that the premise is questionable. Second (and ultimately equally important), I point out that their definition of ‘complex tone’ is equivocal between two quite distinct senses. Third, I offer a speculative alternative approach to explaining the correlation. Everett et al. emphasize that they have formulated a hypothesis and tested it empirically. That is, they predict on independent grounds that tone languages should be more common where humidity is high, and then provide evidence that this prediction is valid. The logic of their prediction is as follows: They provide plenty of evidence for the first premise, and (lacking relevant specialist knowledge in statistics) I am willing to accept their demonstration of the geographical correlation that supports the conclusion. However, there is good reason to doubt the second premise. Over the course of many years, I have conducted studies in which speakers read prepared sentences under studio conditions. These studies have generally aimed to test the effect of specific structural manipulations on fundamental frequency (F0) at specific points in the utterance. For example, in one such experiment (Ladd 1988), I compared English sentences of the form A and B but C and A but B and C (where A, B , and C are short clauses), looking for (and finding) F0 differences on the accented syllables of the clauses depending on the structure. A consistent finding in all these studies is that, for any given speaker, mean F0 …