Computational Complexity and Sour-Grapes-Like Patterns

In this paper, we claim that attested sour-grapes-like patterns of featural and tonal spreading differ meaningfully from what we call the 'true' sour grapes spreading pathology. We propose that the 'false' sour grapes processes attested in some tonal systems are computationally less complex than the unattested true sour grapes pathology, due to the presence of what we refer to as 'zones of predictability' local to potential triggers of spreading. In particular, the false sour grapes spreading patterns can be shown to fall into the class of weakly deterministic mappings, while the true sour grapes spreading pattern does not.