Ternarity through binarity
Ternary stress patterns—the lapse of more than one syllable between stresses—have been challenging for metrical stress theory. The existence of a ternary primitive predicts the existence of trisyllabic word minima, trisyllabic reduplicants, and most directly, stress on every third syllable. The alternative to a ternary primitive has its own predictions, such as the systematic skipping of one light syllable between metrical feet. As such, this, too, would predict the systematic occurrence of stress on every third syllable. Most of the worlds languages that have stress are binary and show strict alternating rhythm on stressbearing units, and so these ternary languages are somewhat problematic. This research shows that the predictions borne of a ternary primitive or its binary alternative do not hold. More specifically, that systems widely believed to be ternary are more uniformly identified as binary, with sporadic ternary effects under duress. This is particularly important, because we predict that the existence of those requirements that yield a ternary effect should apply across-the-board in some languages, and this prediction is shown not to hold. Evidence from Finnish, Estonian, and Koniag Alutiiq suggest that both the quantity of the individual syllable in specific contexts as well as the morphological status of certain syllables can be stress-attracting such that a binary rhythmic alternation can be interrupted to the extent that it appears ternary. Additional evidence from Old English and Bani-Hassan Bedouin Arabic, originally thought to be evidence of the binary equivalent of a ternary primitive, are shown to be well-formed binary systems whose sporadic ternary effects are predictable on the basis of requirements we know to hold in other languages. Finally, an examination of Sentani reveals that remaining instances of ternarity are predicted on the basis of the unusual application of requirements known to hold in other languages.