摘要：In syntactic satiation, a linguist initially judges a sentence type to be unacceptable but begins to accept it after judging multiple examples over time. When William Snyder first brought this phenomenon to the attention of linguists, he proposed satiation as a data source for linguistic theory and showed it can be induced experimentally. Here, three new studies indicate (i) satiation is restricted to a small, stable set of sentence types; (ii) after satiation on one sentence type (e.g., wh-movement across … wonder whether … or … believe the claim …), acceptability sometimes increases for distinct but syntactically related sentence types (… wonder why …; … accept the idea …); (iii) for sentence types susceptible to satiation, the difficulty of inducing it (e.g., number of exposures required) varies systematically; and (iv) much as satiation in linguists persists over time, experimentally induced satiation can persist for at least four weeks. These findings suggest a role for satiation in determining whether the perceived unacceptability of two sentence types has a common source.
