Syntactic and semantic interference in sentence comprehension: Support from English and German eye-tracking data

Abstract

A long-standing debate in the sentence processing literature concerns the time course of syntactic and semantic information processing in online sentence comprehension. The default assumption in cue-based models of parsing is that syntactic and semantic retrieval cues simultaneously guide dependency resolution. When retrieval cues match multiple items in memory, this leads to similarity-based interference. Both semantic and syntactic interference have been shown to occur in English. However, the relative timing of syntactic vs. semantic interference remains unclear. In this cross-linguistic investigation of the time course of syntactic vs. semantic interference, the data from two eye-tracking during reading experiments (English and German) suggest that the two types of interference can in principle arise simultaneously during retrieval. However, the data also indicate that semantic cues are evaluated with a small timing lag in German compared to English. This cross-linguistic difference between English and German may be due to German having richer morphosyntactic marking than English, resulting in syntactic cues dominating over semantic cues during dependency resolution. More broadly, our cross-linguistic results pose a challenge for the cue-based retrieval model’s default assumption that syntactic and semantic cues are used simultaneously during long-distance dependency formation. Our work also highlights the importance of collecting cross-linguistic data on psycholinguistic phenomena which can potentially advance theory development.

Do syntactic and semantic similarity lead to interference effects? Evidence from self-paced reading and event-related potentials using German

“…The reading times of the immediately following spill-over region and of the pre-critical region were also of interest. Previous work using the present design (Mertzen et al, 2023;Van Dyke, 2007) has consistently found effects in the pre-critical region (we return to this point in the discussion sections). A further issue with German embedded clauses (also see Mertzen et al, 2023) was that, without an appropriate intervening phrase, the clause embedding would lead to two verbs appearing one after another (erzählt hatte, trank, literally 'told, drank').…”

“…The studies had 120 items with four conditions in a Latin square design shown in (2). The items were constructed following Van Dyke's (2007) items and partially re-used Mertzen et al's (2023) The comprehension of the sentences in (2) requires a memory retrieval operation at the verb trank ('drank') to retrieve its non-adjacent subject Witwer ('widower'). During this retrieval, the distractor Einbrecher ('burglar') / Verlust ('loss') might cause interference and impede comprehension.…”

“…The priors assume a priori effect sizes for the main effects of syntactic and semantic interference, and their interaction, ranging from 0-8 ms, 0-40 ms, or 0-81 ms. These priors are based on the observed sign and sizes of effects and uncertainties in previous reading studies that use the present design (Mertzen et al, 2023;Van Dyke, 2007), and on meta-analyses relating to interference effects (Jäger et al, 2017). For the syntactic and semantic interference effects, we use directional priors so that any evidence for the effect of interest from Bayes factors can be interpreted as evidence of replication success (Verhagen & Wagenmakers, 2014).…”

“…This difficulty in identifying the correct dependent is called similarity-based interference. We present large-sample self-paced reading (N = 774) and event-related potentials experiments (N = 103) using a well-established design (Mertzen et al, 2023;Van Dyke, 2007), to investigate interference due to syntactic and semantic cues in German. A Bayes factor analysis showed that both the self-paced reading times as well as event-related potentials provided clear evidence for a semantic interference effect.…”

“…Another example is the recent heated debate (Nieuwland et al, 2018) on prediction in sentence comprehension, which was largely focused on the results of low-powered studies, leading to potentially misleading conclusions (Nicenboim et al, 2020). Second, although it has been established (e.g., Mertzen et al, 2023;Van Dyke, 2007) that syntactic and semantic cues are involved in sentence processing, it is not obvious which cues are relevant for a particular retrieval event. For example, Dillon et al (2013) argued, following Sturt (2003, that the gender cue was either downweighted or not used at all when an antecedent-reflexive dependency needs to be built.…”