Task Sensitivity in L2 English Speakers’ Syntactic Processing: Evidence for Good-Enough Processing in Self-Paced Reading

Abstract

Native (L1) and second-language (L2) sentence processing can sometimes be shallow. A Good-Enough approach suggests that speakers may engage in shallow processing if the task permits. This study tests English native speakers and native Chinese L2 learners of English to explore whether different task demands affect their sentence processing. In a self-paced reading task, participants read globally or temporarily ambiguous sentences with relative clauses preceded by a matrix clause containing two noun phrases (NPs). Comprehension questions modulated task difficulty: Half the participants received comprehension questions that probed their interpretation of the relative clauses whereas the remaining half received superficial questions unrelated to the relative clauses. Task difficulty affected reading times at the point of disambiguation for both L1 and L2 participants. Additionally, participants' attachment choices for globally ambiguous sentences were consistent with reading times of the disambiguating region in both L1 and L2 readers. The results suggest that both L1 and L2 syntactic processing is modulated by the task at hand. We argue for a similar treatment of shallowness for L1 and L2 speakers in models of sentence processing, along the lines of the Good-Enough approach to language processing.

Music to Your Ears

“…The lack of semantic processing has been discussed as one of the main reasons why jumbling of segments in looped nonspeech excerpts does not block the transformation to music in a similar way to how jumbling of syllables in looped sentences blocks STS (Simchy-Gross & Margulis, 2018). Assuming that lexico-syntactic processing in a listener's non-native language is shallow and taskdependent (Favreau & Segalowitz, 1983;Kilborn, 1992;Segalowitz et al, 1998;Tan & Foltz 2020), the L2-effects observed in Experiment 2 corroborate the idea that meaningfulness of the acoustic signal mediates the transformation. The ''repetition-to-music'' effect is therefore likely to be stronger in non-speech than in speech, as results of a previous study suggest (Rowland et al, 2019), though pertinent evidence for the meaning hypothesis is yet to be provided.…”

“…Given that Experiment 2 did not include assessment of L2-comprehension, there is a possibility that listeners (particularly those with low proficiency) did not attempt to process L2-phrases linguistically and experienced them as if they were spoken in a completely unfamiliar language, thus demonstrating the previously observed foreign-language effect (Jaisin et al, 2016;Margulis et al, 2015). Task demands are known to affect processing and comprehension of L2-speech (Kilborn, 1992;Tan & Foltz, 2020) as well as the perception and action more generally (Memelink & Hommel, 2007), and might have led listeners of this study to exclusively attend to the sound structure of the stimuli, bypassing other sources of linguistic information in the acoustic signal or engaging in a shallow encoding of L2 sentences. Experiment 2 will thus benefit from a replication design to include tests of lexicosyntactic integration and access in L2-listeners.…”

Native and second language processing of quantifier scope ambiguity

“…The results showed that L2 learners have control over morphosyntactic information like the native speakers and that L2 processing can be viewed as GE processing that is qualitatively similar to native processing. Similarly, a number of recent studies that have examined L2 comprehension using the GE approach have found their results for L1 and L2 processing to be largely compatible with the predictions of GE processing (Fujita and Cunnings, 2020;Lee and Shin, 2016;Tan and Foltz, 2020). Although L2 speakers exhibited greater difficulty recovering from the initial misinterpretation and displayed lower accuracy scores for comprehension questions than L1 speakers overall (Jacob and Felser, 2016;Pozzan and Trueswell, 2016), they did not lack detailed syntactic representations of ambiguous sentences and were able to reanalyse them successfully with increasing proficiency (Hopp, 2006).…”

When more is more: redundant modifiers can facilitate visual search

“…Given that the modifiers in Experiment 2 were noisier—or at least were less beneficial to subjects—it is possible that subjects used a shallow processing strategy when reading the search instruction (e.g., used good-enough processing, Ferreira 2003 ). Reading studies have shown that temporarily ambiguous sentences were read faster when they were followed by superficial questions about the sentences as opposed to when the questions probed how the ambiguity was interpreted (Swets et al 2008 ; Tan and Foltz 2020 ). In other words, task difficulty modulated how carefully subjects read the sentences, because reading the sentence carefully would improve performance on difficult comprehension questions, suggesting that subjects use shallow, good-enough reading strategies when reading more carefully offers them no clear benefit.…”