Temporal order and individual differences in language comprehension

2009-09 till 2012-09
Research Areas: 

I examine how temporal order information affects language comprehension, because I believe it can contribute to our understanding of how the abstract notion of time is embodied through our experience of the linear procession of events. A second motivation is understanding how this process relates to the skills individuals use to organize verbal and non-verbal visual information.


Methods and Research Questions: 

Research on the comprehension of temporal order statements such as biclausal Bevor/Nachdem (‘Before’/‘After’) sentences is sparse and sometimes conflicting. Solidifying its base is important for examining both event structure and event ordering in cognition (e.g., with non-linguistic stimuli, Zacks, et al., 2007; Raisig et al. 2010) and language comprehension.

Analyses of event-related brain potentials (ERPs) suggest event order cues (e.g., bevor vs. nachdem) are processed immediately, although Bevor sentences cue greater working memory demands which HWM (but not LWM) readers are immediately sensitive to (Münte et al., 1998). We revisited Münte et al.’s findings using eye tracking, allowing sentence (re-)reading (unlike ERPs with rapid serial presentation). Longer reading time is interpreted as indexing greater working memory demands. First-pass time analyses revealed temporal cue effects (elevated times for before than after). As in Münte et al., temporal cue effects appeared shortly after HWM (but not LWM) participants read Bevor versus Nachdem. Surprisingly, both HWM and LWM readers showed longer re-reading (second pass) post-conjunction in Bevor versus Nachdem sentences. Thus, on the second (but not first) sentence reading, both HWM and LWM readers were influenced by increased working memory demands, suggesting these two groups differ only initially in event order comprehension.

Participants sat in a chair with their chins in a chinrest approximately 70 centimeters away from a Viewsonic 22 inch monitor, and read sentences presented in black Courier New font on a white background. At the beginning of each trial the participant saw a drift correct dot in the middle of the screen. Next, a dot appeared at the middle-left point on the display monitor for 1000 ms, and then the sentence. Participants indicated via a spacebar press that they understood the sentence. After that, they answered a comprehension question by pressing “Yes” and “No” marked keys.

After the eye-tracking part of the experiment, participants’ working memory was tested using the German version of the Automated Reading Span Test (AutoRST) created by Engle and colleagues, and translated by Dr. Helene Kreysa. For the AutoRST, participants read sets from two to seven sentences and judged the sentences as sensible or nonsense. After each sentence decision, a letter appeared. This letter had to be recognized in a 3 X 4 grid of letters at the end of the final sentence in a set. The participants were asked to mark the letters in the same order that they had seen them appear.



We tested reading time differences for the word region immediately following the conjunction Before compared to After for both HWM and LWM readers. A 2 X 2 mixed measures ANOVA resulted in a significant interaction between temporal conjunction and working memory. Paired-sample t-tests on mean differences in first-pass reading times for Before versus After sentences revealed a significant difference between Before (M = 476.06 ms) versus After (M = 442.29 ms) sentences shortly after HWM readers read Before. For LWM this difference was not significant.

We conducted the same 2 X 2 analysis for re-reading times. We found a significant main effect for temporal conjunction, indicating that participants’ re-reading of the word region following the conjunction (before, after) was longer for Before (M = 354.81 ms) than After (M = 307.51 ms). Unlike for initial sentence reading, high and low working memory readers, however, didn’t differ in their re-reading times. Thus, people rapidly process event order during comprehension, albeit with subtle differences as a function of their working memory capacity.