Similarity-based Interference Research Articles

A Matter of Memory? Age-Invariant Relative Clause Disambiguation and Memory Interference in Older Adults.

Past research suggests that Working Memory plays a role in determining relative clause attachment bias. Disambiguation preferences may further depend on Processing Speed and explicit memory demands in linguistic tasks. Given that Working Memory and Processing Speed decline with age, older adults offer a way of investigating the factors underlying disambiguation preferences. Additionally, older adults might be subject to more severe similarity-based memory interference given their larger vocabularies and slower lexical access. Nevertheless, memory interference and sentence disambiguation have not been combined in studies on older adults before. We used a self-paced reading paradigm under memory load interference conditions. Older (n=30) and Younger (n=35) readers took part in the study online; reading times were recorded and measures of comprehension accuracy and load recall were collected. This setup allowed for the implicit measurement of attachment biases and memory interference effects interactively. Results show that similarity-based interference affected both age groups equally, but was more pronounced in NP2-biased structures, which took participants generally longer to read. Attachment preferences did not differ by group and were unaffected by Working Memory span. However, accuracy on recall prompts was predicted by Working Memory span in both groups. Findings of greater interference in syntactically dispreferred structures support unified processing models where parsing constraints naturally interact. The lack of age differences on our measures further aligns with research finding age-invariant implicit language processing.

Memory Retrieval in Online Sentence Parsing: Empirical Evidence, Computational Modelling, and Simulations

This paper reports two experiments (Experiments 1 and 2) and computational simulations designed to investigate and model memory retrieval processes during real-time sentence processing. Central to this study is the hypothesis that linguistic information serves as a cue to retrieve target representations from memory during dependency formation. The basis for this cue-based memory retrieval stems from research showing that non-target representations that match a set of retrieval cues interfere with target retrieval. The susceptibility to this similarity-based interference has been debated in the sentence processing literature, and various hypotheses and models have been formulated and developed. This issue is addressed empirically in Experiments 1 and 2, which investigated similarity-based interference in sentences with a floating quantifier. Bayesian linear mixed models and Bayes factor analyses suggested similarity-based interference. However, the patterns of interference were not consistent with existing theories and models. To reconcile these findings within the framework of cue-based memory retrieval, this paper implements the Revision Integrated Cue-Based (RICB) model based on the ACT–R architecture. This model assumes that structural information is heavily weighted and incorporates the notions of initial retrieval and revision. The results of the simulations indicate that the RICB model successfully predicts the observed data, highlighting the central role of structural information and revision in memory retrieval during real-time sentence processing.

Agreement attraction in grammatical sentences and the role of the task

This study evaluates two broad classes of language processing accounts that make predictions for sentences like “The admirer of the singer(s) apparently thinks...”. Feature distortion accounts predict increased processing difficulty at the verb in sentences with a plural distractor noun (singers) while similarity-based interference accounts predict the opposite: increased difficulty in sentences with a singular distractor noun (singer). Neither of these effects was reliably observed in earlier research, and the Bayesian meta-analysis of 31 published studies reported here is almost perfectly inconclusive. An explanation may be that both effects occur simultaneously and therefore mask each other. To test this idea, we conducted three single-trial self-paced reading experiments (N1=4,296, N2=3,920, N3=3,559) which orthogonally manipulated agreement attraction and inhibitory interference. Surprisingly, all three experiments produced evidence for agreement attraction but none for inhibitory interference, which supports feature distortion but not similarity-based interference accounts. Experiment 4 (N4=3,535) tested the role of the expected task by preparing participants for a comprehension question (vs. acceptability judgment in Experiments 1–3). It showed neither agreement attraction nor inhibitory interference effects. Our findings demonstrate that agreement attraction effects can arise in grammatical sentences – contra earlier research – but also that these effects crucially depend on the task. This explains inconsistent results in prior research and supports feature distortion as the driving force behind attraction effects in grammatical sentences.

Online reflexive resolution and interference

ABSTRACT A reflexive referentially depends on another NP. Long-standing inquiries in sentence processing literature revolve around whether and how forming this coreference relation during online sentence processing is susceptible to similarity-based interference. To address these questions, we conducted four experiments using L-maze tasks. The results revealed (delayed) interference in ungrammatical sentences, indicating facilitatory interference. We also found that structural information associated with a distractor and related to clause finiteness may influence memory retrieval. Interference effects were not observed in grammatical sentences across all experiments. These results suggest that online reflexive resolution may be susceptible to interference only when the reflexive mismatches its structurally accessible antecedent in features, and that structural information may influence how a distractor interferes with memory retrieval during online sentence processing.

An Empirical Investigation of Featural Similarity in Wh-islands

According to the most recent version of Relativized Minimality, the ungrammaticality of weak islands is the result of featural similarity between elements in an intervention configuration. The theory posits that only features triggering movement have the potential to induce intervention effects leading to ungrammaticality. However, recent advancements in the theory have extended the set of features claimed to generate intervention effects to encompass lexical restriction. This theoretical move encounters several empirical challenges. In this paper, we address this question in 3 acceptability judgment experiments in French. We explore how featural similarity influences acceptability judgments across both wh-islands and minimally different grammatical structures, that-clauses extraction, focusing on three distinct features: (i) the feature associated with question operators, (ii) lexical restriction, both anticipated to show intervention effects according to the most recent version of Relativized Minimality, and (iii) animacy, which is not expected to show such effects. Results indicate that featural similarity in lexical restriction and animacy exerts a mild influence on acceptability ratings in both islands and grammatical structures, contrary to what predicted by Relativized Minimality, while similarity in the feature associated with question operators exerts a strong influence. We propose an empirically motivated account that restricts the set of features relevant to grammar-based effects à la Relativized Minimality to core syntactic features triggering movement, and groups together the milder effects arising from similarity in other linguistic features, like lexical restriction and animacy, as resulting from similarity-based interference in memory.

SEAM: An integrated activation-coupled model of sentence processing and eye movements in reading

Models of eye-movement control during reading, developed largely within psychology, usually focus on visual, attentional, lexical, and motor processes but neglect post-lexical language processing; by contrast, models of sentence comprehension processes, developed largely within psycholinguistics, generally focus only on post-lexical language processes. We present a model that combines these two research threads, by integrating eye-movement control and sentence processing. Developing such an integrated model is extremely challenging and computationally demanding, but such an integration is an important step toward complete mathematical models of natural language comprehension in reading. We combine the SWIFT model of eye-movement control (Seelig et al., 2023) with key components of the Lewis and Vasishth sentence processing model (Lewis and Vasishth, 2005). This integration becomes possible, for the first time, due in part to recent advances in successful parameter identification in dynamical models, which allows us to investigate profile log-likelihoods for individual model parameters. We present a fully implemented proof-of-concept model demonstrating how such an integrated model can be achieved; our approach includes Bayesian model inference with Markov Chain Monte Carlo (MCMC) sampling as a key computational tool. The integrated Sentence-Processing and Eye-Movement Activation-Coupled Model (SEAM) can successfully reproduce eye movement patterns that arise due to similarity-based interference in reading. To our knowledge, this is the first-ever integration of a complete process model of eye-movement control with linguistic dependency completion processes in sentence comprehension. In future work, this proof of concept model will need to be evaluated using a comprehensive set of benchmark data.

Similarity-Based Interference Effects in Reflexive Processing by Brazilian Speakers of English as L2

O objetivo deste artigo é investigar como ocorre o processamento anafórico por brasileiros falantes de inglês como L2, buscando descobrir se estes sujeitos se guiam pela sintaxe desde o momento inicial do processamento, ou se fazem uso de informações de outras naturezas, especialmente semântico-lexicais, para identificar o possível antecedente das anáforas. Para isto, realizamos um experimento de leitura automonitorada, que mostrou diferenças entre os sujeitos de nível de proficiência intermediário e os avançados. No primeiro grupo, houve uma facilitação da leitura quando os traços de gênero de ambos os antecedentes combinavam com o traço de gênero da anáfora. Já para os avançados, os resultados indicaram que este grupo tem a leitura facilitada quando ao menos um dos antecedentes combina com o traço de gênero da anáfora. Tais resultados são consistentes com modelos de processamento em L2 que defendem um acesso incompleto à sintaxe, dando prioridade à utilização de informações de natureza semântico-lexical e pragmática para a resolução anafórica (CLAHSEN & FELSER, 2006, 2017). Além disso, apesar de não se encaixar perfeitamente em nenhum modelo de interferência na memória de trabalho baseado em semelhanças, os resultados são indicativos de propostas que tratam sobre interferência na recuperação da memória de trabalho

Modulating Complex Sentence Processing in Aphasia Through Attention and Semantic Networks.

Lexical processing impairments such as delayed and reduced activation of lexical-semantic information have been linked to syntactic processing disruptions and sentence comprehension deficits in individuals with aphasia (IWAs). Lexical-level deficits can also preclude successful lexical encoding during sentence processing and amplify the processing costs of similarity-based interference during syntactic retrieval. We investigate whether two manipulations to engage attention and pre-activate semantic features of a target (to-be-retrieved) noun will (a) boost lexical activation during initial lexical encoding and (b) facilitate syntactic dependency linking through improved resolution of interference in IWAs and neurologically unimpaired age-matched controls (AMCs). Eye-tracking-while-listening with a visual world paradigm was used to investigate whether semantic and attentional manipulations modulated initial lexical processing and downstream syntactic retrieval of the direct-object noun in object-relative sentences. In the attention and semantic manipulations, the AMC group showed no changes in initial lexical access levels; however, gaze patterns revealed clear facilitations in dependency linking and interference resolution. In the IWA group, the attentional cue increased and maintained activation of N1 with modest facilitations in dependency linking. In the semantic condition, IWA results showed a greater degree of facilitation during dependency linking. The results suggest that attention and semantic activation are parameters that may be manipulated to strengthen encoding of lexical representations to facilitate retrieval (i.e., dependency linking) and mitigate similarity-based interference. In IWAs, these manipulations may help to reduce lexical processing deficits that can preclude successful encoding.

Pre-verb reactivation of arguments in sentence processing

Major models of sentence comprehension assume that a verb triggers retrieval of preceding thematic arguments from memory to establish argument-verb dependencies. If so, longer argument-verb distance should lead to higher processing load at the verb (a locality effect),since the representation of the argument should suffer from decay and/or interference.However, verb-final languages have often failed to show the expected argument-verb locality effect. A possible account of the lack of the effect is that arguments and adjuncts before the verb reactivate each other, counteracting memory degradation. In a pair of self-paced reading experiments in Japanese, a verb-final language, we found evidence of such pre-verb reactivation.Specifically, there was a locality effect and a similarity-based interference effect at the head of the adverbial that follows the subject, both of which suggest the retrieval of the subject at that point. The results are difficult to accommodate with other accounts of the lack of locality effect, such as a confounding effect of expectation and the inherent locality-insensitivity of verb-final languages. It is further argued that the constructivist analysis of verbal argument structure, which has been developed in generative syntax, provides an explanation for why such pre-verb reactivation takes place. 

Syntactic and semantic interference in sentence comprehension: Support from English and German eye-tracking data

A long-standing debate in the sentence processing literature concerns the time course of syntactic and semantic information processing in online sentence comprehension. The default assumption in cue-based models of parsing is that syntactic and semantic retrieval cues simultaneously guide dependency resolution. When retrieval cues match multiple items in memory, this leads to similarity-based interference. Both semantic and syntactic interference have been shown to occur in English. However, the relative timing of syntactic vs. semantic interference remains unclear. In this cross-linguistic investigation of the time course of syntactic vs. semantic interference, the data from two eye-tracking during reading experiments (English and German) suggest that the two types of interference can in principle arise simultaneously during retrieval. However, the data also indicate that semantic cues are evaluated with a small timing lag in German compared to English. This cross-linguistic difference between English and German may be due to German having richer morphosyntactic marking than English, resulting in syntactic cues dominating over semantic cues during dependency resolution. More broadly, our cross-linguistic results pose a challenge for the cue-based retrieval model’s default assumption that syntactic and semantic cues are used simultaneously during long-distance dependency formation. Our work also highlights the importance of collecting cross-linguistic data on psycholinguistic phenomena which can potentially advance theory development.

Similarity-Based Interference in Sentence Comprehension in Aphasia: a Computational Evaluation of Two Models of Cue-Based Retrieval

Sentence comprehension requires the listener to link incoming words with short-term memory representations in order to build linguistic dependencies. The cue-based retrieval theory of sentence processing predicts that the retrieval of these memory representations is affected by similarity-based interference. We present the first large-scale computational evaluation of interference effects in two models of sentence processing — the activation-based model and a modification of the direct-access model — in individuals with aphasia (IWA) and control participants in German. The parameters of the models are linked to prominent theories of processing deficits in aphasia, and the models are tested against two linguistic constructions in German: pronoun resolution and relative clauses. The data come from a visual-world eye-tracking experiment combined with a sentence-picture matching task. The results show that both control participants and IWA are susceptible to retrieval interference, and that a combination of theoretical explanations (intermittent deficiencies, slow syntax, and resource reduction) can explain IWA’s deficits in sentence processing. Model comparisons reveal that both models have a similar predictive performance in pronoun resolution, but the activation-based model outperforms the direct-access model in relative clauses.

Sources and mechanisms of modality-specific distraction in visual short-term memory

ABSTRACT Visual short-term and working memory can be disrupted by irrelevant, distracting input occurring after encoding. Distractors similar to the original memory are known to be interfering, but it is unclear whether dissimilar distractors have the same disruptive effect. The presence of dissimilar distraction would be problematic for views of similarity-based interference, hence the present study investigated modality-specific distraction using a procedure that required participants to compare single target and probe objects over a delay. An irrelevant distractor could be presented during the delay separating the target and probe, but it varied in its similarity to the target. In four experiments, recognition was disrupted by the presence of a distractor, even when the distractors were highly dissimilar to the target. Furthermore, the interference effect was not reduced when the same distractors were repeatedly used throughout the experiment, and interference from dissimilar distractors was only lessened when it was extremely predictable. These findings indicate that susceptibility to dissimilar distraction is a persistent limitation in visual short-term memory.

Irrelevant speech impairs serial recall of verbal but not spatial items in children and adults

Immediate serial recall of visually presented items is reliably impaired by task-irrelevant speech that the participants are instructed to ignore (“irrelevant speech effect,” ISE). The ISE is stronger with changing speech tokens (words or syllables) when compared to repetitions of single tokens (“changing-state effect,” CSE). These phenomena have been attributed to sound-induced diversions of attention away from the focal task (attention capture account), or to specific interference of obligatory, involuntary sound processing with either the integrity of phonological traces in a phonological short-term store (phonological loop account), or the efficiency of a domain-general rehearsal process employed for serial order retention (changing-state account). Aiming to further explore the role of attention, phonological coding, and serial order retention in the ISE, we analyzed the effects of steady-state and changing-state speech on serial order reconstruction of visually presented verbal and spatial items in children (n = 81) and adults (n = 80). In the verbal task, both age groups performed worse with changing-state speech (sequences of different syllables) when compared with steady-state speech (one syllable repeated) and silence. Children were more impaired than adults by both speech sounds. In the spatial task, no disruptive effect of irrelevant speech was found in either group. These results indicate that irrelevant speech evokes similarity-based interference, and thus pose difficulties for the attention-capture and the changing-state account of the ISE.

Interference and filler-gap dependency formation in native and non-native language comprehension.

The mechanisms underlying native (L1) and non-native (L2) sentence processing have been widely debated. One account of potential L1/L2 differences is that L2 sentence processing underuses syntactic information and relies heavily on semantic and surface cues. Recently, an alternative account has been proposed, which argues that the source of L1/L2 differences lies in how susceptible L1 and L2 speakers are to interference during memory retrieval operations. The present study tested these two accounts by investigating filler-gap dependency formation and susceptibility to similarity-based interference in L1 and L2 language comprehension. The results demonstrated that L1 and L2 speakers recover the information of the filler upon encountering a gap and are susceptible to similarity-based interference during filler-gap dependency formation. However, there was no significant evidence of L1/L2 differences. These findings suggest that L1 and L2 speakers similarly engage in cue-based memory retrieval operations during filler-gap dependency formation. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Subject Advantage in L1-English Learners’ Production of Chinese Relative Clauses

This study investigated whether L1-English Chinese learners show a subject preference in their oral production of Chinese relative clauses (RCs) and whether they show animacy effects. We conducted a picture-based elicited production experiment that compared subject and object RCs, varying the object animacy between animate and inanimate. The results from thirty learners showed more targetlike performance in subject RCs than in object RCs, both at group and individual levels, regardless of object animacy. Error analyses revealed that more object RCs were converted into subject RCs than vice versa. These results point toward a clear subject preference despite conflicted findings in previous research on RCs in Chinese as a foreign language. Animacy influenced subject and object RCs alike: both types were easier to produce when featuring an inanimate object. We suggested similarity-based interference or distribution-based effects to account for this finding.

The interaction of predictive processing and similarity-based retrieval interference: an ERP study

ABSTRACT Language processing requires memory retrieval to integrate current input with previous context and making predictions about upcoming input. We propose that prediction and retrieval are two sides of the same coin, i.e. functionally the same, as they both activate memory representations. Under this assumption, memory retrieval and prediction should interact: Retrieval interference can only occur at a word that triggers retrieval and a fully predicted word would not do that. The present study investigated the proposed interaction with event-related potentials (ERPs) during the processing of sentence pairs in German. Predictability was measured via cloze probability. Memory retrieval was manipulated via the position of a distractor inducing proactive or retroactive similarity-based interference. Linear mixed model analyses provided evidence for the hypothesised interaction in a broadly distributed negativity, which we discuss in relation to the interference ERP literature. Our finding supports the proposal that memory retrieval and prediction are functionally the same.

Similarity-based interference and relative clauses in second language processing

Relative clauses have long been examined in research on first (L1) and second (L2) language acquisition and processing, and a large body of research has shown that object relative clauses (e.g. ‘The boy that the girl saw’) are more difficult to process than subject relative clauses (e.g. ‘The boy that saw the girl’). Although there are different accounts of this finding, memory-based factors have been argued to play a role in explaining the object relative disadvantage. Evidence of memory-based factors in relative clause processing comes from studies indicating that representational similarity influences the difficulty associated with object relatives as a result of a phenomenon known as similarity-based interference. Although similarity-based interference has been well studied in L1 processing, less is known about how it influences L2 processing. We report two studies – an eye-tracking experiment and a comprehension task – investigating interference in the comprehension of relative clauses in L1 and L2 readers. Our results indicated similarity-based interference in the processing of object relative clauses in both L1 and L2 readers, with no significant differences in the size of interference effects between the two groups. These results highlight the importance of considering memory-based factors when examining L2 processing.

An Information-Theoretic Account of Semantic Interference in Word Production.

I present a computational-level model of semantic interference effects in online word production within a rate–distortion framework. I consider a bounded-rational agent trying to produce words. The agent's action policy is determined by maximizing accuracy in production subject to computational constraints. These computational constraints are formalized using mutual information. I show that semantic similarity-based interference among words falls out naturally from this setup, and I present a series of simulations showing that the model captures some of the key empirical patterns observed in Stroop and Picture–Word Interference paradigms, including comparisons to human data from previous experiments.

Encoding interference effects support self-organized sentence processing

According to cue-based retrieval theories of sentence comprehension, establishing the syntactic dependency between a verb and the grammatical subject is susceptible to interference from other noun phrases in the sentence. At the verb, the subject must be retrieved from memory, but non-subject nouns that are similar on dimensions that are relevant to subject-verb agreement, like number marking, can make the retrieval more difficult. However, cue-based retrieval models fail to account for a class of interference effects, conventionally called “encoding interference,” that cannot be due to retrieval interference. In this paper, we implement a self-organized sentence processing model that provides a more parsimonious explanation of encoding interference effects than otherwise reasonable extensions that could be made to the cue-based retrieval approach. We first also present new behavioral evidence for encoding interference using a semantic similarity manipulation in two self-paced reading studies of subject-verb number agreement. The results of these experiments are more compatible with the self-organizing account. We argue that self-organization, which reduces all parsing to fallible feature match optimization and makes no a priori distinction between encoding and retrieval, can provide a unifying approach to similarity-based interference in sentence comprehension.

The Role of Animacy in Children's Interpretation of Relative Clauses in English: Evidence From Sentence-Picture Matching and Eye Movements.

Subject relative clauses (SRCs) are typically processed more easily than object relative clauses (ORCs), but this difference is diminished by an inanimate head-noun in semantically non-reversible ORCs ("The book that the boy is reading"). In two eye-tracking experiments, we investigated the influence of animacy on online processing of semantically reversible SRCs and ORCs using lexically inanimate items that were perceptually animate due to motion (e.g., "Where is the tractor that the cow is chasing"). In Experiment 1, 48 children (aged 4;5-6;4) and 32 adults listened to sentences that varied in the lexical animacy of the NP1 head-noun (Animate/Inanimate) and relative clause (RC) type (SRC/ORC) with an animate NP2 while viewing two images depicting opposite actions. As expected, inanimate head-nouns facilitated the correct interpretation of ORCs in children; however, online data revealed children were more likely to anticipate an SRC as the RC unfolded when an inanimate head-noun was used, suggesting processing was sensitive to perceptual animacy. In Experiment 2, we repeated our design with inanimate (rather than animate) NP2s (e.g., "where is the tractor that the car is following") to investigate whether our online findings were due to increased visual surprisal at an inanimate as agent, or to similarity-based interference. We again found greater anticipation for an SRC in the inanimate condition, supporting our surprisal hypothesis. Across the experiments, offline measures show that lexical animacy influenced children's interpretation of ORCs, whereas online measures reveal that as RCs unfolded, children were sensitive to the perceptual animacy of lexically inanimate NPs, which was not reflected in the offline data. Overall measures of syntactic comprehension, inhibitory control, and verbal short-term memory and working memory were not predictive of children's accuracy in RC interpretation, with the exception of a positive correlation with a standardized measure of syntactic comprehension in Experiment 1.