Journal of Experimental Psychology: Learning, Memory, and Cognition, Mar 03, 2025, No Pagination Specified; doi:10.1037/xlm0001436

We provide a further examination of the influence of top-down sentence-level constraints on the transposed-word effect by manipulating a factor—phrase frequency—that directly implicates sentence-level representations. The focus was on ungrammatical transposed-word sequences, and under the assumption that top-down influences would play a role in driving transposed-word effects, we predicted that ungrammatical decisions would be harder (longer reaction times and higher error rates) when the ungrammatical transposed-word sequences were derived from high-frequency compared with low-frequency phrases. Five experiments were conducted in which participants performed a speeded grammatical decision task. The results are clear-cut. Although phrase frequency did influence grammatical decisions to grammatically correct phrases, with shorter reaction times and lower error rates for high-frequency phrases relative to low-frequency phrases (Experiment 4), ungrammatical decisions were not influenced by the frequency of the base sentences from which the transposed-word sequences were formed, neither in the auditory (Experiments 1 and 2) nor in the visual modality (Experiment 3). In Experiment 5, we show that a transposed-word effect is observed when comparing the transposed-word sequences of Experiment 3 with nontransposed control sequences. We conclude that frequency-sensitive sentence-level constraints, as measured as the frequency of occurrence of a sequence of words in corpora of spoken and written language, do not modulate transposed-word effects. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Mar 03, 2025, No Pagination Specified; doi:10.1037/xlm0001475

There is evidence that asking people to predict their own memory performance during learning (immediate judgments of learning, JOLs) can alter memory. Changes in the use of learning strategies have been proposed to contribute to these reactive effects of JOLs. This study addresses the impact of making JOLs on the use of learning strategies and the contribution of learning strategies to JOL reactivity. Across six experiments, participants studied related and unrelated word pairs and did or did not make JOLs during study, completed a cued-recall test, and reported the learning strategies they had used for each word pair. When we manipulated the requirement to make JOLs between participants, making JOLs enhanced memory for related pairs and impaired memory for unrelated pairs. Further, the learning strategies participants used differed across the JOL and no-JOL groups, and these differences mediated the detrimental effects of making JOLs on memory for unrelated pairs. In contrast, when we manipulated the requirement to make JOLs within participants, making JOLs enhanced recall performance for related pairs but did not impact recall for unrelated pairs or the use of learning strategies. Overall, our findings indicate that changes in the use of learning strategies underlie detrimental effects of making JOL on memory for unrelated pairs but only play a minor role in positive effects of making JOLs for related word pairs. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 27, 2025, No Pagination Specified; doi:10.1037/xlm0001442

The present study examined how Arabic–Hebrew–English trilinguals process double and triple cognate words in their third language (L3) across three different experiments. Utilizing the same set of critical cognate items, trilinguals completed a semantic relatedness task, a lexical decision task, or a sentence reading eye-tracking task. The results revealed a significant cognate facilitation effect in the semantic relatedness task, with no consistent differences in the magnitude of facilitation across double and triple cognates, suggesting that both L1 and L2 are activated during L3 processing. In contrast, no cognate facilitation was observed in the lexical decision or the sentence reading tasks. These results demonstrate that the cognate facilitation is task-dependent, varying with the degree to which meaning is activated, sentential context is available, and orthographic cues are involved. Critically, the study extends findings of phonologically mediated cross-language activation from bilinguals to trilinguals. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 27, 2025, No Pagination Specified; doi:10.1037/xlm0001459

Successful goal-directed behavior requires not only selecting the correct response to an object in our environment but also requires selecting the correct object in our environment upon which to act. While most task-switching studies investigate the selection and maintenance of mental representations of response options (so-called response sets), they often do not investigate the selection and maintenance of mental representations of object selection (so-called stimulus sets). In the present study, participants were exposed to a taskswitching paradigm with multiple stimuli in which the relevant stimulus set (i.e., which object to respond to) and response set (i.e., how to respond to that object) independently either repeated or switched on each trial. Of interest was the nature of the task set representation required, and whether response set and stimulus set could be updated independently. Guided by predictions from a computational model of dual-task control (executive control of the theory of visual attention; Logan & Gordon, 2001), seven experiments were conducted that evaluated the independence of task-set components. All experiments confirmed executive control of the theory of visual attention’s predictions of an underadditive interaction between response-set and stimulus-set sequence—diagnostic of independent and parallel reconfiguration of components. However, limitations to this independent updating were observed when participants were encouraged to selectively prioritize response-set or stimulus-set reconfiguration via component-specific preparation manipulations. The results are discussed in terms of various hypotheses on the structure of task-set representation. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 24, 2025, No Pagination Specified; doi:10.1037/xlm0001451

In this study, we investigated human causal learning in a continuous time and space setting. We find participants to be capable active causal structure learners, and with the help of computational modeling explore how they mitigate the complexity of continuous dynamics data to achieve this. We propose that participants combine systematic interventions with a narrowed focus on causal dynamics that occur during and directly downstream of their interventions. This task decomposition approach achieves comparable accuracy to attending to all the dynamics, while discarding almost half of the data. We argue this strategy makes sense from a resource rationality perspective: Ignoring dynamics outside of interventions saves computational cost while the interventions naturally decompose the global learning problem into a series of more manageable subproblems. We also find that when the causal relata are given real-world labels, participants will use their domain-specific priors to guide their structure inferences. In particular, individuals with accurate prior expectations were less likely to make the common local computations error of mistaking an indirect for a direct relationship. Overall, our experiments reinforce the idea that humans are frugal and intuitive active learners who combine actions and inference to optimize learning while minimizing effort. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 24, 2025, No Pagination Specified; doi:10.1037/xlm0001463

How do people recognize objects they have encountered previously? Cognitive models of recognition memory aim to explain overt behavior using latent psychological processes, such as true recognition and pure guessing. Validation studies assess whether the mechanisms underlying cognitive models properly reflect the psychological processes they aim to explain. The present study provides such a validation study for models describing paired-word recognition—a paradigm in which participants have to categorize randomly constructed word pairs. Specifically, introducing a strength manipulation (Experiment 1), presenting certain words more often during study, a base-rate manipulation of response categories (Experiment 2), presenting certain pair types more often during test, a base-rate manipulation of overall frequencies of old and new words (Experiment 3), and a payoff manipulation, differentially incentivizing correct responses (Experiment 4), we assessed the validity of general recognition theory, a multidimensional signal detection theory model, and the paired two-high threshold model, a discrete-state model. Both models captured the strength manipulation as expected on mnemonic parameters describing memory sensitivity and detection probability. Unexpectedly, the base-rate and payoff manipulations affected (strategic) memory retrieval within the discrete-state model (Experiments 2–4) and both strategic retrieval (Experiment 2) and decision boundaries (Experiments 3 and 4) within the continuous model. Implications for model validity and the future use of these models for paired-word recognition are discussed. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 20, 2025, No Pagination Specified; doi:10.1037/xlm0001449

Planning the future relies on the ability to remember how long events last, yet how durations are stored in memory is unknown. Here, we developed a novel n-item delayed duration reproduction task to assess whether elapsed time is stored as a continuous feature or as a discrete item in memory. In three experiments (N = 58), participants were presented with nonisochronous sequences composed of empty time intervals delimited by brief tones. Time intervals varied in number and in duration. Participants had to reproduce as precisely as possible the duration of all time intervals in the sequence following a delay period. We manipulated the number of time intervals (n-item) and the sequence duration to separate their effects on recall precision. In all three experiments, the precision of recall decreased with the number of items in the sequence, showing that durations can be stored as discrete items in working memory. Our analyses emphasize the distinction between reproduction biases that are captured by relative reproduction and decreased precision which indexes working memory load. Future research is needed to spell out the conditions under which durations are fully abstracted in working memory. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 20, 2025, No Pagination Specified; doi:10.1037/xlm0001431

Numbers are highly relevant in our everyday lives. Besides intentionally processing number magnitude when necessary, we often automatically process it even when not required. The SNARC (Spatial–Numerical Association of Response Codes; Dehaene et al., 1993) effect, describing faster left-/right-sided responses to smaller/larger numbers, respectively, provides evidence for this automaticity. It arises in semantic number-processing tasks both, when number magnitude is task-relevant (e.g., magnitude classification) and task-irrelevant (e.g., parity judgment). However, findings on the SNARC effect in tasks requiring the processing of nonsemantic number features are mixed: While it has been observed in orientation judgment tasks, it was mostly absent in color judgment tasks. Importantly, previous studies were underpowered or did not control for confounding variables. In two highly powered online experiments, we found a small but significant SNARC effect in both nominal color judgment (cyan vs. yellow; slope = −1.71 ms) and color intensity judgment (light cyan vs. dark cyan; slope = −1.13 ms) of Arabic digits from 1 to 9 excluding 5, which did not significantly differ in size. Further, we found little evidence for the Linguistic Markedness of Response Codes (i.e., faster left-/right-sided responses to odd/even numbers, respectively; Nuerk et al., 2004) effect. Moreover, the odd effect (i.e., faster responses to even than to odd numbers; Hines, 1990) was detected. Taken together, both magnitude and parity are processed automatically even if participants respond to physical nonsemantic and nonspatial number features, but the spatial mapping seems more automatic for magnitude than for parity. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 17, 2025, No Pagination Specified; doi:10.1037/xlm0001467

In a recognition memory task, correct “old” responses to previously studied target items (hits) trigger larger pupil dilation (PD) than correct “new” responses to previously not presented foil items (correct rejections). This pupil old/new effect reflects the specific processes involved in recognition decisions, with dilation being larger when decisions are based on recollection of contextual details rather than mere familiarity. However, previous research has been limited in determining the exact link between PD and recognition processes due to the categorical nature of tasks used to assess recollection/familiarity. To investigate this issue, we examined whether the precision of the recollected memory representation is related to PD during successful recognition. During encoding, target words were presented on the outline of an invisible circle, and during a subsequent recognition task, participants made old/new decisions. For “old” responses, participants had to indicate the exact location of the target on the outline of the invisible circle. We found that larger PD during the old/new decision was related to more precise subsequent localization decision. Furthermore, we also demonstrated that PD and memory precision are linearly related, and even hits followed by unprecise source localization trigger larger PD than correct rejections. Thus, increased PD is present for all recognition decisions, but its magnitude increases with increasing precision of source recollection. This pattern of results suggests that the pupil old/new effect might originate from two distinct components: The first is related to the mere recognition of a word, whereas the second reflects the quality of recollected source information. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 17, 2025, No Pagination Specified; doi:10.1037/xlm0001425

Semantic transparency refers to the degree to which the meaning of the whole word can be inferred from its constituents. For Chinese, semantic radicals generally carry information about the meanings of Chinese characters and, thus, can be used to reflect semantic transparency of Chinese characters. For those Chinese characters having the same semantic radicals (i.e., neighboring characters), their meanings are assumed to be semantically related to each other. However, to what extent those neighboring characters are close in their meanings remains unclear. A conventional crowdsource approach could provide a coarse measure of semantic relationships between semantic neighbors. However, those approaches are generally limited to a small sample size of characters. Here, we proposed a corpus-based measure of semantic transparency, termed semantic radical consistency (SRC). Specifically, we utilized the Word2Vec models to construct a Chinese semantic space and quantified the SRC for 3,423 characters. To evaluate the SRC, we first conducted linear mixed-effect modeling analyses to verify the explanatory power of SRC on a large-scale Chinese character naming reaction times. Second, we investigated the SRC effect by conducting a word naming task based on traditional factorial designs. Both the linear mixed-effect modeling and factorial naming results demonstrated that SRC was a unique and reliable variable to account for the variance in traditional Chinese character naming reaction times. The results indicated this innovative, corpus-derived SRC was able to effectively reflect the semantic transparency level by measuring semantic distances among characters in the same semantic radical category. Further investigations on the interaction between SRC and phonetic radical consistency demonstrated the cooperative nature between phonological and semantic reading pathways. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 17, 2025, No Pagination Specified; doi:10.1037/xlm0001437

Working memory (WM) is the cognitive system that allows the temporary holding of mental representations for use in thought and action. Long-term memory (LTM) refers to our ability to remember a potentially unlimited amount of information over longer time periods. Understanding how these two memory systems interact has important implications for theories of cognition, learning, and education. Here, we examined (a) whether a shared perceptual bottleneck accounts for the relation between WM and LTM accuracy, and (b) whether serial position effects in WM are mirrored in LTM. In two experiments, participants studied sequences of objects at varying set sizes and completed old/new recognition tests for some items immediately after encoding (WM tests) and for other items after all WM trials were completed (LTM tests). In Experiment 1 (N = 80), LTM performance was better for items presented in lower rather than higher set size sequences, indicating that limitations in WM capacity constrain LTM encoding, irrespective of perceptual bottlenecks. In Experiment 2 (N = 120), we observed WM and LTM recency effects, but primacy effects were only present in LTM and not in WM. Thus, serial position effects in WM did not consistently predict the relative rates at which items from different serial positions were preserved in LTM. These results reinforce accounts that view WM and LTM as having at least partially separate mechanisms, shedding light on the nature of these mechanisms. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 17, 2025, No Pagination Specified; doi:10.1037/xlm0001468

There is empirical evidence that people have some intuitive discomfort when they judge that a believable, but invalid response is logically valid. This has led to the hypothesis that there exists a form of “logical intuition” that is responsive to logical form. However, there is also clear evidence that when reasoning with identical forms of inference, responses are not uniform but are instead modulated by access to semantic information related to potential alternatives. In two preregistered studies, we examine the hypothesis that differential access to such information determines the extent to which intuitions signal discomfort. To examine this, we constructed syllogisms using the same logical form but having either few or many alternatives associated with the premises. In Study 1, we show that when accepting a believable conclusion as being valid, confidence was lower for syllogisms having many alternatives. In Study 2, we show that people “like” conclusions based on logically invalid syllogisms having few alternatives more than those based on syllogisms having many alternatives. These results provide clear evidence for a form of “logical intuition” that relies on access to semantic information rather than pure logical form. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 17, 2025, No Pagination Specified; doi:10.1037/xlm0001460

Repeatedly seen or heard statements are typically judged to be more likely true than statements not encountered before, a phenomenon referred to as truth effect. Similarly, statements judged to be old typically receive higher truth judgments than statements judged to be new. However, it is unclear whether and how this recognition-based truth effect depends on the latent memory states underlying observed recognition judgments. In order to investigate this question, we used a model-based approach to compare truth judgments as a function of recognition judgments (“old” vs. “new”) and their underlying memory states (state of memory certainty vs. state of uncertainty). In three experiments, we observed a recognition-based truth effect and found this effect to be larger in the state of memory certainty than in the state of uncertainty. This result also replicated for subjective instead of modeled memory states. Moreover, we found effects of recognition judgments on judged truth to be stronger than effects of factual repetition in all three experiments. Taken together, our research highlights the role of episodic memory processes in the truth effect and provides a methodological tool that takes underlying memory states into account. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 17, 2025, No Pagination Specified; doi:10.1037/xlm0001448

Everyday situations require us to face a trade-off to inform our decisions: exploit known information or explore for new information. Although both have risks, empirical research has not shown whether individuals prefer exploring or exploiting across contexts. In the present study, we examined the explore–exploit trade-off as a theoretical framework across two broad domains: decision making and spatial navigation. In this registered report, we applied computational modeling to human behavior on a novel version of the Iowa gambling task to predict behavior on a spatial navigation task in which the navigator must either exploit a learned, familiar route or explore a new shortcut. If the hypothesis that risk tolerance is a domain-general trait is correct, we predicted that explore–exploit patterns would correlate across these tasks. We also examine the predictive power of computational models for the Iowa gambling task on behavioral uncertainty and the role of confidence in spatial navigation strategy. Our findings suggest that, while there is some overlap in risk tolerance between spatial navigation and gambling, the influence of exploration and exploitation on navigational decision making is weaker than initially predicted. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 17, 2025, No Pagination Specified; doi:10.1037/xlm0001469

There is now robust evidence of priming effects during sentence processing for adjacent words that form collocations (statistically associated combinations). However, there is very limited evidence regarding how nonadjacent collocations might facilitate processing. Furthermore, no previous research has examined how nonadjacent collocations interplay with other (non)collocational material in the surrounding context. We employed “nested” collocations for the first time, in which more than one contextual element (verb, adjective) is a potential collocate for a noun. For example, in a verb–adjective–noun (V-A-N) phrase, two collocations may be “nested” (“express concerns” + “valid concerns” = “express valid concerns”) or only the verb (nonadjacent) or adjective (adjacent) might be collocational. In an eye-tracking experiment with L1 English speakers, we manipulated the collocational status of adjectives adjacent to the noun, (V)-A-N, and verbs nonadjacent to the noun, V-(A)-N. Our results replicated the basic adjacent effect and produced evidence of facilitation for nonadjacent collocations. Additionally, we find preliminary evidence for a syntactic primacy effect, whereby collocational links involving the verb prove more impactful than adjective–noun collocations, despite nonadjacency. Importantly, the results reveal cumulative facilitation in “nested collocations,” with a boost resulting from the simultaneous effects observed in adjacent and nonadjacent collocations. Altogether, the results extend our understanding of collocational priming effects beyond single collocations. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 17, 2025, No Pagination Specified; doi:10.1037/xlm0001454

Data visualizations are common in publications addressed to scientists and the general public. A common graph distortion effect can be obtained by changing the y-axis range. On bar graphs with lower truncated scales (the y-axis starting point is above the data origin), observers tend to perceive larger differences between the values depicted. Herein, we define anchors, information that can be perceived from a graph, to explain ratings of differences in bar graphs. Study 1 examined whether the upper y-axis truncation effect exists or not. We confirmed its existence even though the effect size is smaller compared to lower y-axis truncation effect. Study 2 examined lower and upper y-axis truncations and expansions. We found that, compared to graphs without distortions, observers perceive larger differences between values when there is truncation and smaller differences when there is expansion at either end of the y-axis. Study 3 examined whether the effects of lower and upper y-axis distortions are also present on reversed bar graphs. We found that the black bars biased observers more when they are truncated, as it reduces their area. Finally, Study 4 examined the impact of y-axis distortions on bar graphs, dot graphs, and line graphs. We found that a plot not showing bars results in less biased judgments in the presence of truncation and similar biases for lower and upper truncation. We discuss the results of other relevant research using these anchors and argue that characterizing graphs using the anchors proposed herein can be generalized to other data visualizations. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 13, 2025, No Pagination Specified; doi:10.1037/xlm0001450

Factors that predict mind wandering in the laboratory and in daily life differ (Kane et al., 2007, 2017). However, it is unknown how these predictors may vary when considering two identified dimensions of mind wandering—intentionality and emotional valence. We examined this with a 1-week daily-life experience sampling study with laboratory-based measures of working memory, personality, anxiety, and dispositional mindfulness predicting mind wandering in daily life. Overall, our results suggest that predictors of mind wandering in daily life vary based on both the intentionality and emotional valence dimension of the off-task thought. Dispositional mindfulness was predictive of neutral, intentional, and overall rates of mind wandering. Interactions between working memory and concentration level were observed for some but not all dimensions of mind wandering. The current findings suggest that is it critical to consider both intentionality and emotional valence dimensions to understand individual differences in mind wandering in daily life. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 13, 2025, No Pagination Specified; doi:10.1037/xlm0001455

In two visual-world eyetracking experiments, we investigated how effectively addressees use information about a speaker’s perspective to resolve temporary ambiguities in spoken expressions containing prenominal scalar adjectives (e.g., the small candle). The experiments used a new “Display Change” task to create situations where an addressee’s perspective conflicted with that of a speaker, allowing the point of disambiguation (early vs. late) to be specified independently from each perspective. Contrary to existing perspective-taking theories, the only situation in which addressees resolved references early was when both perspectives afforded early disambiguation. When perspectives conflicted, addressees exhibited a lower rate of preferential looks to the target and slower response times. This disruption to contrastive inference reflects either the suspension of pragmatic inferencing or cognitive limitations on the simultaneous representation and use of incompatible perspectives. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 13, 2025, No Pagination Specified; doi:10.1037/xlm0001443

Spoken language understanding requires the integration of incoming speech with representations of the preceding context. How rich the information is that listeners maintain in these contextual representations has been a long-standing question. Under one view, subcategorical information about the preceding input—including any uncertainty about the underlying categories—is quickly discarded due to memory limitations. Alternative views hold that listeners maintain some subcategorical information far beyond word boundaries. This would facilitate more effective integration with subsequent context, under the assumption that subsequent context is informative about the preceding input. We thus ask whether listeners are sensitive to changes in the informativity of subsequent context by changing the expected utility of subcategorical information maintenance. In three experiments, we manipulate how informative subsequent context is about words that occur six to nine syllables earlier. We find that reduced informativity leads listeners to down-weight the importance of subsequent context. This suggests that listeners can adjust the degree to which they maintain subcategorical information. We do, however, also identify alternative interpretations that affect not only the present results but also the interpretation of previous work on subcategorical information maintenance. (PsycInfo Database Record (c) 2025 APA, all rights reserved)

Journal of Experimental Psychology: Learning, Memory, and Cognition, Feb 13, 2025, No Pagination Specified; doi:10.1037/xlm0001438

Reading relies on the incremental processes that occur across all words in a passage to build a global comprehension of the text. Factorial experimental designs are not well-suited to examine these incremental processes, which are influenced by multilevel factors in an overlapping manner. Exemplifying an alternative approach, we combined event-related potentials, probabilistic language models, authentic texts, and statistical methods to examine the time course of multilevel linguistic influences on the incremental processes which occur during reading each word. We found that indicators of the initial stages of word identification (N170 and P200) are sensitive to context-independent statistical information of a word, for example, word frequency. The later stages of word processing, involving processes related to meaning retrieval and integration (N400), heavily rely on the word’s context-dependent information measured by word surprisal. Syntactic processing, reflected by a word’s syntactic surprisal and the number of phrase structures it closes, was presented across multiple phases (an early negativity, N400, and a late positivity). Additionally, the effects of position factors at both the word and sentence levels emerged across multiple time windows (including N170, P200, and N400), suggesting their distinct influence beyond linguistic factors. These findings provide a theoretically coherent picture of incremental reading, partly convergent with conclusions from factorial studies but with novel results concerning the time courses and interactions of processing components. (PsycInfo Database Record (c) 2025 APA, all rights reserved)