How Do We Segment Text? Two-Stage Chunking Operation in Reading

Abstract

Chunking in language comprehension is a process that segments continuous linguistic input into smaller chunks that are in the reader's mental lexicon. Effective chunking during reading facilitates disambiguation and enhances efficiency for comprehension. However, the chunking mechanisms remain elusive, especially in reading, given that information arrives simultaneously yet the written systems may not have explicit cues for labeling boundaries such as Chinese. What are the mechanisms of chunking that mediates the reading of the text that contains hierarchical information? We investigated this question by manipulating the lexical status of the chunks at distinct levels in four-character Chinese strings, including the two-character local chunk and four-character global chunk. Male and female human participants were asked to make lexical decisions on these strings in a behavioral experiment, followed by a passive reading task when their electroencephalography (EEG) was recorded. The behavioral results showed that the lexical decision time of lexicalized two-character local chunks was influenced by the lexical status of the four-character global chunk, but not vice versa, which indicated the processing of global chunks possessed priority over the local chunks. The EEG results revealed that familiar lexical chunks were detected simultaneously at both levels and further processed in a different temporal order, the onset of lexical access for the global chunks was earlier than that of local chunks. These consistent results suggest a two-stage operation for chunking in reading, the simultaneous detection of familiar lexical chunks at multiple levels around 100 ms followed by recognition of chunks with global precedence.

Shared neural computations for syntactic and morphological structures: evidence from Mandarin Chinese

“…Topographical similarity analysis (i.e., representational similarity analysis, spatial similarity analysis) is an analysis for comparing the topographical distribution of ERP responses that is free from bias in electrode selection (Tian & Huber, 2008; Murray, Brunet & Michel, 2008; Tian, Poeppel & Huber, 2010; Wang, Zhu & Tian, 2019). Quantifying topographical similarity with this method has been used to hallmark similarity in the underlying neural processes in many language comprehension studies (Yang, Cai & Tian, 2020; Wang et al, 2020; Wang & Kuperberg, 2023; Hubbard & Federmeier, 2021; Wei et al, 2023; Zhao et al, 2023; Huang, Feng & Qu, 2023; Ding, ten Oever & Martin, 2023).…”

Evaluation of L2 Listening Comprehension: In Pursuit of New Measurement Units

“…Hence the issue of appropriate measurement units is of particular methodological importance and interest. In this regard, there are also gamut of approaches that are used throughout the L2 studies and adjacent research areas for measurement purposes -words (Yang et al, 2020), sentences (Bardovi-Harlig, 1992), propositions (Sato, 1988), clauses (Taboada & Zabala, 2008), T-units (Sotillo, 2000)b, C-units (Pica et al, 1989)but that opportunities for nonnative speakers (NNSs, AS-units (Foster et al, 2000) there must be agreement on the nature of the unit, and it must be possible to apply this unit reliably to a range of different types of speech data. There are a number of different units in use, the various merits of which have been discussed by Crookes (1990, prosodic units (Izre'el et al, 2020, idea units (Richards et al, 2016), discourse units (Houtkoop & Mazeland, 1985), turn-constructional units (Liddicoat, 2004) etc.…”

Unsupervised text segmentation predicts eye fixations during reading

“…In a recent review, Leminen et al (2019) analyzed more than 100 neuroimaging studies of inflected words (e.g., walk-ed), derived words (e.g., dark-ness), and compounds (e.g., walk-man). As they summarized, most studies of the processing of derivational/inflectional morphology agree that such complex words are decomposed during processing; but studies of the processing of compound words show inconsistent results: some support the access of constituent morpheme units (Fiorentino et al, 2014;Koester and Schiller, 2011), some support the access of whole-word units (Stites et al, 2016), and some support the mixed access of both (Kaczer et al, 2015;MacGregor and Shtyrov, 2013;Yang et al, 2020a) In addition to subword units, the cognitive system can also make use of supra-word units. Some studies provide indications that supra-words such as frequent phrases and idioms (e.g., "I don't know") are stored in our long-term mental lexicon (Arnon and Snider, 2010;Bannard and Matthews, 2008;Jackendoff, 2002), implying that supra-words can be processed directly.…”

“…Baayen (2007) has argued that the mental lexicon involves storage (of the wholes) and computation (of the combinatorial rules), and that they counterbalance each other. Yang et al (2020a) also considered the counterbalancing, arguing that storing more supra-words in our mental lexicon could reduce the cognitive load of computation since larger units (e.g., "I am /going to" vs. "I /am /going /to") imply fewer processing steps (e.g., two retrievals + one combination vs. four retrievals + three combinations). Taken together, this diverse evidence shows that cognitive units exist at various linguistic levels, and that cognitive units have a wide range of possible lengths.…”