Distinguishing discrete and gradient category structure in language: Insights from verb-particle constructions.

Brehm, Laurel; Goldrick, Matthew

doi:10.1037/xlm0000390

Cited by 12 publications

(15 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This modification of the basic equations of ACT‐R would require thorough testing using data from properly powered experimental designs. Such data do not exist yet, so such a large‐scale evaluation is not currently possible; however, more and more researchers are attending to the power properties of their studies (Brehm & Goldrick, 2017; Stack, James, & Watson, 2018; Zormpa, Meyer, & Brehm, 2019), and we expect that such data will become available in the future.…”

Section: Discussionmentioning

confidence: 99%

A Principled Approach to Feature Selection in Models of Sentence Processing

Smith

Vasishth

2020

Cognitive Science

View full text Add to dashboard Cite

Among theories of human language comprehension, cue‐based memory retrieval has proven to be a useful framework for understanding when and how processing difficulty arises in the resolution of long‐distance dependencies. Most previous work in this area has assumed that very general retrieval cues like [+subject] or [+singular] do the work of identifying (and sometimes misidentifying) a retrieval target in order to establish a dependency between words. However, recent work suggests that general, handpicked retrieval cues like these may not be enough to explain illusions of plausibility (Cunnings & Sturt, 2018), which can arise in sentences like The letter next to the porcelain plate shattered. Capturing such retrieval interference effects requires lexically specific features and retrieval cues, but handpicking the features is hard to do in a principled way and greatly increases modeler degrees of freedom. To remedy this, we use well‐established word embedding methods for creating distributed lexical feature representations that encode information relevant for retrieval using distributed retrieval cue vectors. We show that the similarity between the feature and cue vectors (a measure of plausibility) predicts total reading times in Cunnings and Sturt’s eye‐tracking data. The features can easily be plugged into existing parsing models (including cue‐based retrieval and self‐organized parsing), putting very different models on more equal footing and facilitating future quantitative comparisons.

show abstract

Section: Discussionmentioning

confidence: 99%

A Principled Approach to Feature Selection in Models of Sentence Processing

Smith

Vasishth

2020

Cognitive Science

View full text Add to dashboard Cite

show abstract

“…Punske (2013) has even suggested the existence of three structures by proposing two different kinds of small clause structures. However, Brehm and Goldrick (2017) showed that VPCs may be associated with multiple syntactic structures along a gradient, rather than with only one of two discrete categories. The current study is not intended to differentiate among the various theoretical linguistic structures posited; therefore, we remain neutral toward these proposals and focus instead on processing aspects of different surface word orders.…”

Section: Introductionmentioning

confidence: 99%

How Do French–English Bilinguals Pull Verb Particle Constructions Off? Factors Influencing Second Language Processing of Unfamiliar Structures at the Syntax-Semantics Interface

2018

View full text Add to dashboard Cite

An important challenge in bilingualism research is to understand the mechanisms underlying sentence processing in a second language and whether they are comparable to those underlying native processing. Here, we focus on verb-particle constructions (VPCs) that are among the most difficult elements to acquire in L2 English. The verb and the particle form a unit, which often has a non-compositional meaning (e.g., look up or chew out), making the combined structure semantically opaque. However, bilinguals with higher levels of English proficiency can develop a good knowledge of the semantic properties of VPCs (Blais and Gonnerman, 2013). A second difficulty is that in a sentence context, the particle can be shifted after the direct object of the verb (e.g., The professor looked it up). The processing is more challenging when the object is long (e.g., The professor looked the student’s last name up). This shifted structure favors syntactic processing at the expense of VPC semantic processing. We sought to determine whether or not bilinguals’ reading time (RT) patterns would be similar to those observed for native monolinguals (Gonnerman and Hayes, 2005) when reading VPCs in sentential contexts. French–English bilinguals were tested for English language proficiency, working memory and explicit VPC semantic knowledge. During a self-paced reading task, participants read 78 sentences with VPCs that varied according to parameters that influence native speakers’ reading dynamics: verb-particle transparency, particle adjacency and length of the object noun phrase (NP; 2, 3, or 5 words). RTs in a critical region that included verbs, NPs and particles were measured. Results revealed that RTs were modulated by participants’ English proficiency, with higher proficiency associated with shorter RTs. Examining participants’ explicit semantic knowledge of VPCs and working memory, only readers with more native-like knowledge of VPCs and a high working memory presented RT patterns that were similar to those of monolinguals. Therefore, given the necessary lexical and computational resources, bilingual processing of novel structures at the syntax-semantics interface follows the principles influencing native processing. The findings are in keeping with theories that postulate similar representations and processing in L1 and L2 modulated by processing difficulty.

show abstract

“…This modification of the basic equations of ACT-R would require thorough testing using data from properly powered experimental designs. Such data do not exist yet, so such a large-scale evaluation is not currently possible; however, more and more researchers are attending to the power properties of their studies (Brehm & Goldrick, 2017;Stack et al, 2018;Zormpa et al, 2019), and we expect that such data will become available in the future.…”

Section: Distributed Features and Cues In Act-rmentioning

confidence: 99%

A principled approach to feature selection in models of sentence processing

Smith¹,

Vasishth²

2019

Preprint

View full text Add to dashboard Cite

Among theories of human language comprehension, cue-based memory retrieval has proven to be a useful framework for understanding when and how processing difficulty arises in the resolution of long-distance dependencies. Most previous work in this area has assumed that very general retrieval cues like [+subject] or [+singular] do the work of identifying (and sometimes misidentifying) a retrieval target in order to establish a dependency between words. However, recent work suggests that general, hand-picked retrieval cues like these may not be enough to explain illusions of plausibility (Cunnings & Sturt, 2018), which can arise in sentences like The letter next to the porcelain plate shattered. Capturing such retrieval interference effects requires lexically specific features and retrieval cues, but hand-picking the features is hard to do in a principled way and greatly increases modeler degrees of freedom. To remedy this, we use word embeddings, a well-established method for creating distributed feature representations, for lexical features and retrieval cues. We show that the similarity between the features and the cues (a measure of plausibility) predicts total reading times in Cunnings and Sturt’s eye-tracking data. The features can easily be plugged into existing parsing models (including cue-based retrieval and self-organized parsing), putting very different models on more equal footing and facilitating future quantitative comparisons. In addition to this methodological contribution, our results suggest that, contrary to Cunnings and Sturts’ original conclusions, focused words might be more prominent in memory, making them less susceptible to interference, as predicted by a recent extension to ACT-R (Engelmann, Jäger, & Vasishth, 2019).

show abstract

Distinguishing discrete and gradient category structure in language: Insights from verb-particle constructions.

Cited by 12 publications

References 52 publications

A Principled Approach to Feature Selection in Models of Sentence Processing

A Principled Approach to Feature Selection in Models of Sentence Processing

How Do French–English Bilinguals Pull Verb Particle Constructions Off? Factors Influencing Second Language Processing of Unfamiliar Structures at the Syntax-Semantics Interface

A principled approach to feature selection in models of sentence processing

Contact Info

Product

Resources

About