Uncertainty Reduction as a Measure of Cognitive Load in Sentence Comprehension

Frank, Stefan L.

doi:10.1111/tops.12025

Cited by 117 publications

(113 citation statements)

References 34 publications

(56 reference statements)

Supporting

Mentioning

111

Contrasting

Order By: Relevance

“…In related work, Frank (2013) tested a version of the entropy reduction hypothesis whereby entropy reduction was not bounded by 0 (was allowed to take negative values). A Simple Recurrent Network was used to predict the next four words in the sentence; the uncertainty following the current word was estimated as the entropy of this quadrigram distribution.…”

Section: Discussionmentioning

confidence: 99%

Investigating the role of entropy in sentence processing

Linzen¹,

Jaeger²

2014

Proceedings of the Fifth Workshop on Cognitive Modeling and Computational Linguistics

View full text Add to dashboard Cite

We outline four ways in which uncertainty might affect comprehension difficulty in human sentence processing. These four hypotheses motivate a self-paced reading experiment, in which we used verb subcategorization distributions to manipulate the uncertainty over the next step in the syntactic derivation (single step entropy) and the surprisal of the verb's complement. We additionally estimate wordby-word surprisal and total entropy over parses of the sentence using a probabilistic context-free grammar (PCFG). Surprisal and total entropy, but not single step entropy, were significant predictors of reading times in different parts of the sentence. This suggests that a complete model of sentence processing should incorporate both entropy and surprisal.

show abstract

Section: Discussionmentioning

confidence: 99%

Investigating the role of entropy in sentence processing

Linzen¹,

Jaeger²

2014

Proceedings of the Fifth Workshop on Cognitive Modeling and Computational Linguistics

View full text Add to dashboard Cite

show abstract

“…This same pattern is also part of the empirical support for the minimize domains (MiD) principle (Hawkins, 2004, Section 7.2 There is no necessary connection between Entropy Reduction as a complexity metric and Minimalist Grammars as a formalism. In fact, Frank (2013) recently applied Entropy Reduction to the analysis of British readers' eye fixation times using Simple Recurrent Nets as substitute for a grammar. In this study, Entropy Reduction emerged as a significant predictor of fixation duration.…”

Section: Entropy Reduction As a Complexity Metricmentioning

confidence: 99%

Uncertainty in processing relative clauses across East Asian languages

Yun

Chen

Hunter

et al. 2014

J East Asian Linguist

View full text Add to dashboard Cite

The processing difficulty profile for relative clauses in Chinese, Japanese and Korean represents a challenge for theories of human parsing. We address this challenge using a grammar-based complexity metric, one that reflects a minimalist analysis of relative clauses for all three languages as well as structure-dependent corpus distributions. Together, these define a comprehender's degree of uncertainty at each point in a sentence. We use this idea to quantify the intuition that people do comprehension work as they incrementally resolve ambiguity, word by word. We find that downward changes to this quantitative measure of uncertainty derive observed processing contrasts between Subject-and Object-extracted relative clauses. This demonstrates that the complexity metric, in conjunction with a minimalist grammar and corpus-based weights, accounts for the widely-observed Subject Advantage. Electronic supplementary material The online version of this article

show abstract

“…Two GAMMs were fitted: one for reading latencies on the TRY verb, and another for reading latencies on the following infinitive verb to test for spill-over effects (Just et al, 1982;Frank, 2013). For the TRY verb, the mean reading latency was 798.78 ms (Mdn = 713.63 ms, s = 334.71 ms), while for the infinitive verb the mean reading latency was 783.52 ms (Mdn = 660.96 ms, s = 396.03 ms).…”

Section: Resultsmentioning

confidence: 99%

A learning perspective on individual differences in skilled reading: Exploring and exploiting orthographic and semantic discrimination cues.

Milin

Divjak

Baayen

2017

Journal of Experimental Psychology: Learning, Memory, and Cogni

View full text Add to dashboard Cite

The goal of the present study is to understand the role orthographic and semantic information play in the behaviour of skilled readers. Reading latencies from a self-paced sentence reading experiment in which Russian near-synonymous verbs were manipulated appear well-predicted by a combination of bottom-up sub-lexical letter triplets (trigraphs) and top-down semantic generalizations, modelled using the Naive Discrimination Learner. The results reveal a complex interplay of bottom-up and top-down support from orthography and semantics to the target verbs, whereby activations from orthography only are modulated by individual differences. Using performance on a serial reaction time task for a novel operationalization of the mental speed hypothesis, we explain the observed individual differences in reading behaviour in terms of the exploration/exploitation hypothesis from Reinforcement Learning, where initially slower and more variable behaviour leads to better performance overall.

show abstract

Uncertainty Reduction as a Measure of Cognitive Load in Sentence Comprehension

Cited by 117 publications

References 34 publications

Investigating the role of entropy in sentence processing

Investigating the role of entropy in sentence processing

Uncertainty in processing relative clauses across East Asian languages

A learning perspective on individual differences in skilled reading: Exploring and exploiting orthographic and semantic discrimination cues.

Contact Info

Product

Resources

About