Abstract Computational linguistics is often construed as the enterprise of processing language with computers. But the field has much more to offer than just that. A computationally informed perspective of language offers profound scientific insights and can unearth new language universals. In this talk, I illustrate this point with a concrete case study.

Syntax and phonology are commonly considered radically different components of natural language. This view has been formally corroborated by proofs that the generative capacity of syntax vastly exceeds that of phonology. A very different picture emerges, however, if one follows standard linguistic practice and treats syntax as manipulating tree structures. I show that in this case, syntax and phonology are computational twins in that they use the same memory structures and inference patterns. The observed empirical differences are due to a difference in data structures: strings for phonology, trees for syntax.