Imagine being led down a garden path, only to realize you’ve taken a detour. That’s what happens when we encounter garden path sentences, which temporarily confuse us with syntax ambiguities. But just comparing reading times in garden path and non-garden path sentences isn’t enough to understand the true cost. There are other factors at play, like inattention and the brain’s tendency to reject and skip over these confusing sentences without fully reanalyzing them. To get more accurate estimates of the processing costs involved, scientists designed a computational model that replicates these cognitive processes. This innovative model considers trial-level contaminants as probabilistic latent processes, simulating how we mentally navigate garden path sentences. It takes into account different sentence types, such as noun phrase/zero complement (NP/Z) and reduced relative clause (RRC) garden path sentences. The model not only provides accurate predictions based on existing data but also sheds light on the effectiveness of disambiguation techniques like adding a comma or relying on semantic plausibility. Understanding these hidden costs helps uncover how our cognitive processes tackle complex language constructions.

Abstract
What is the processing cost of being garden-pathed by a temporary syntactic ambiguity? We argue that comparing average reading times in garden-path versus non-garden-path sentences is not enough to answer this question. Trial-level contaminants such as inattention, the fact that garden pathing may occur non-deterministically in the ambiguous condition, and “triage” (rejecting the sentence without reanalysis; Fodor & Inoue, 2000) lead to systematic underestimates of the true cost of garden pathing. Furthermore, the “pure” garden-path effect due to encountering an unexpected word needs to be separated from the additional cost of syntactic reanalysis. To get more realistic estimates for the individual processing costs of garden pathing and syntactic reanalysis, we implement a novel computational model that includes trial-level contaminants as probabilistically occurring latent cognitive processes. The model shows a good predictive fit to existing reading time and judgment data. Furthermore, the latent-process approach captures differences between noun phrase/zero complement (NP/Z) garden-path sentences and semantically biased reduced relative clause (RRC) garden-path sentences: The NP/Z garden path occurs nearly deterministically but can be mostly eliminated by adding a comma. By contrast, the RRC garden path occurs with a lower probability, but disambiguation via semantic plausibility is not always effective.

Read Full Article (External Site)