Analyzing Wrap-Up Effects through an Information-Theoretic Lens

TL;DR

The paper tackles wrap-up effects in reading by linking clause- and sentence-final reading times to information-theoretic properties of preceding text. Using surprisal-based predictors derived from multiple language models, it introduces inf^{(k)} as a family of measures that aggregate prior context information with tunable emphasis on high-surprisal moments. Across five English corpora with SPR and eye-tracking data, inf^{(k)} improves model fit for clause-final RTs (especially for larger k), while offering limited gains for sentence-medial RTs. These findings support theories that wrap-up processing involves resolving contextual ambiguities and information distribution, and they show that information content in prior context captures cognitive processes at clause boundaries with implications for psycholinguistic modeling and interpretability of reading dynamics.

Abstract

Numerous analyses of reading time (RT) data have been implemented -- all in an effort to better understand the cognitive processes driving reading comprehension. However, data measured on words at the end of a sentence -- or even at the end of a clause -- is often omitted due to the confounding factors introduced by so-called "wrap-up effects," which manifests as a skewed distribution of RTs for these words. Consequently, the understanding of the cognitive processes that might be involved in these wrap-up effects is limited. In this work, we attempt to learn more about these processes by examining the relationship between wrap-up effects and information-theoretic quantities, such as word and context surprisals. We find that the distribution of information in prior contexts is often predictive of sentence- and clause-final RTs (while not of sentence-medial RTs). This lends support to several prior hypotheses about the processes involved in wrap-up effects.

Figures (8)

• Figure 1: Distributions of residuals when predicting either clause-final or non-clause-final times using our baseline linear models. Models are fit to (the log-transform of) non-clause-final average RTs. Outlier times (according to log-normal distribution) are excluded. The top-level datasets contain eye-tracking data while the bottom contain SPR data. Full distributions of RTs are shown in \ref{['app:results']}, where we also show models fit to regression times, rather than full reading times.
• Figure 2: Mean ${\Delta\mathrm{LL}}$ as a function of the exponent $k$ in ${\textsc{inf}^{(k)}}$ for models of sentence and clause-final (top row) and sentence-medial (bottom row) RTs using surprisal estimates from different language models. The shaded region connects standard error estimates. Vertical intercepts at $k=0,1$ are for reference. We see that our information-theoretic predictors contribute much less modeling power to the prediction of sentence-medial RTs in comparison to sentence- and clause-final RTs.
• Figure 3: Distributions of average RTs for clause-final and non-clause-final words. Outlier times (according to log-normal distribution) are excluded from averages for both graphs. The top-level datasets contain eye-tracking data while the bottom contain SPR data.
• Figure 4: Version of \ref{['fig:densities']} where surprisal estimates do not include the surprisal assigned to punctuation, which is often a large contributor to clause-final surprisal estimates. We see very little qualitative difference with \ref{['fig:densities']}.
• Figure 5: Version of (a) \ref{['fig:densities_clause']} and (b) \ref{['fig:densities']} for regression times for clause-final and non-clause-final words. Only applicable for eye-tracking datasets