A programme for bidirectional phonology and phonetics and their acquisition and evolution

This paper summarizes an existing bidirectional six-level model of phonology and phonetics (and a bit of morphology). Bidirectionality in this case refers to the modelling of both the speaking process (production) and the listening process (comprehension). The elements of the grammar (the constraints) are bidirectional in the sense that the speaker and listener use the same sets of constraints, with the same rankings. In contrast with Blutner’s and Mattasusch’s bidirectional OT models, the evaluation is the simplest possible, i.e. it is performed unidirectionally in both directions of processing; still, listener-oriented effects tend to emerge from having learning algorithms for the comprehension direction alone. This paper describes a great number of learning algorithms in both directions of processing, and their typical results across one or multiple generations.