Institute of Formal and Applied Linguistics

Processing of garden-path sentences (i.e. The horse raced past the barn fell.) has been a prominent topic of the research in psycholinguistics since 1970s. Most of the researchers are interested in the way how full and correct processing of these structures works. About 2000, several authors came up with the “Good-Enough Approach” which claims that sentence processing is often only partial and that semantic representations are often incomplete. Under this view, we often use heuristics which enable us to spare computational resources while getting a “good enough” idea of what has been said. One of the main sources of data for this approach are experiments on garden-path processing. In their seminal paper, Christianson et al. (2001) presented sentences like While Anna dressed the baby that was small and cute spit up on the bed to the participants and asked them comprehension questions like Did Anna dress the baby? 65.6% participants answered the question incorrectly whereas only 12.5% participants gave the wrong answer on the same question after a control sentence with a different clause-order (The baby that was small and cute spit up on the bed while Anna dressed). The authors argue that the participants create a “good-enough” representation of these sentences.

I will present several experiments on the processing of Czech garden-path structures such as Kluci honili psa a kočku v podkroví znepokojovali šediví hlodavci. Altogether, these experiments show that the native speakers of Czech do indeed have problems in analyzing these sentences, but that it is not due to their use of heuristic or good-enough processing.