Cumulative semantic interference for associative relations in language production

Similarity-induced interference or facilitation in language production reflects representation, not selection

Researchers have long interpreted the presence or absence of semantic interference in picture naming latencies as confirming or refuting theoretical claims regarding competitive lexical selection. But inconsistent empirical results challenge any mechanistic interpretation. A behavioral experiment first verified an apparent boundary condition in a blocked picture naming task: when orthogonally manipulating association type, taxonomic associations consistently elicit interference, while thematic associations do not. A plausible representational difference is that thematic feature activations depend more on supporting contexts. Simulations show that context-sensitivity emerges from the distributional statistics that are often used to measure thematic associations: residual semantic activation facilitates the retrieval of words that share semantic features, counteracting learning-based interference, and training a production model with greater sequential cooccurrence for thematically related words causes it to acquire stronger residual activation for thematic features. Modulating residual activation, either directly or through training, allows the model to capture gradient values of interference and facilitation, and in every simulation competitive and noncompetitive selection algorithms produce qualitatively equivalent results.

Why is a flamingo named as pelican and asparagus as celery? Understanding the relationship between targets and errors in a speeded picture naming task

Speakers sometimes make word production errors, such as mistakenly saying pelican instead of flamingo. This study explored which properties of an error influence the likelihood of its selection over the target word. Analysing real-word errors in speeded picture naming, we investigated whether, relative to the target, naming errors were more typical representatives of the semantic category, were associated with more semantic features, and/or were semantically more closely related to the target than its near semantic neighbours were on average. Results indicated that naming errors tended to be more typical category representatives and possess more semantic features than the targets. Moreover, while not being the closest semantic neighbours, errors were largely near semantic neighbours of the targets. These findings suggest that typicality, number of semantic features, and semantic similarity govern activation levels in the production system, and we discuss possible mechanisms underlying these effects in the context of word production theories.

The unseen, the seen, and the spoken: Latent and overt priming in cyclic picture naming

Language production entails selecting words in the face of internally, and sometimes externally, driven competition that influences the long-term accessibility of both selected words and competitors. Because both endogenous and externally presented competitors usually result in semantic interference, it is often assumed that they engage the same underlying processes. We question this assumption. Specifically, we propose that latent primes may more naturally commingle with endogenous lexical activation whereas overt primes elicit strong control processes. Two experiments examined the effects of latent (masked) and overt (picture-word) priming in combination with cyclic picture naming of small sets of taxonomically or thematically related or unrelated pictures. A subsequent continuous picture naming phase was designed to assess enduring effects of service as a prime or target. Only taxonomic relations showed substantial interference in cyclic naming. Latent priming tended to increase interference for taxonomic relations, but it produced facilitation for thematic relations. In contrast, overt priming induced interference for both types of relation, indicating exertion of cognitive control. In the continuous picture naming phases, accessibility was hindered for previously presented latent primes (Experiment 1) but enhanced for previously overt primes (Experiment 2). Surprisingly, interference in cyclic naming did not carry forward to the continuous phase. These findings suggest that masked priming may be a viable intervention in the internal dynamics of lexical selection. In contrast, they also add to the body of evidence questioning the validity of using picture-word procedures to study inherent semantic interference.

Taxonomic and thematic semantic relationships in picture naming as revealed by Laplacian-transformed event-related potentials

Semantically related concepts co-activate when we speak. Prior research reported both behavioral interference and facilitation due to co-activation during picture naming. Different word relationships may account for some of this discrepancy. Taxonomically related words (e.g., WOLF-DOG) have been associated with semantic interference; thematically related words (e.g., BONE-DOG) have been associated with facilitation. Although these different semantic relationships have been associated with opposite behavioral outcomes, electrophysiological studies have found inconsistent effects on event-related potentials. We conducted a picture-word interference electroencephalography experiment to examine word retrieval dynamics in these different semantic relationships. Importantly, we used traditional monopolar analysis as well as Laplacian transformation allowing us to examine spatially deblurred event-related components. Both analyses revealed greater negativity (150-250 ms) for unrelated than related taxonomic pairs, though more restricted in space for thematic pairs. Critically, Laplacian analyses revealed a larger negative-going component in the 300 to 500 ms time window in taxonomically related versus unrelated pairs which were restricted to a left frontal recording site. In parallel, an opposite effect was found in the same time window but localized to a left parietal site. Finding these opposite effects in the same time window was feasible thanks to the use of the Laplacian transformation and suggests that frontal control processes are concurrently engaged with cascading effects of the spread of activation through semantically related representations.

Cumulative semantic interference is blind to morphological complexity and originates at the conceptual level

When naming a sequence of pictures of the same semantic category (e.g., furniture), response latencies systematically increase with each named category member. This cumulative semantic interference effect has become a popular tool to investigate the cognitive architecture of language production. However, not all processes underlying the effect itself are fully understood, including the question where the effect originates from. While some researchers assume the interface of the conceptual and lexical level as its origin, others suggest the conceptual-semantic level. The latter assumption follows from the observation that cumulative effects, namely cumulative facilitation, can also be observed in purely conceptual-semantic tasks. Another unanswered question is whether cumulative interference is affected by the morphological complexity of the experimental targets. In two experiments with the same participants and the same material, we investigated both of these issues. Experiment 1, a continuous picture naming task, investigated whether morphologically complex nouns (e.g., kitchen table) elicit identical levels of cumulative interference to morphologically simple nouns (e.g., table). Our results show this to be the case, indicating that cumulative interference is unaffected by lexical information such as morphological complexity. In Experiment 2, participants classified the same target objects as either man-made or natural. As expected, we observed cumulative facilitation. A separate analysis showed that this facilitation effect can be predicted by the individuals’ effect sizes of cumulative interference, suggesting a strong functional link between the two effects. Our results thus point to a conceptual-semantic origin of cumulative semantic interference.

Mental chronometry of speaking in dialogue: Semantic interference turns into facilitation

Numerous studies demonstrate that the production of words is delayed when speakers process in temporal proximity semantically related words. Yet the experimental settings underlying this effect are different from those under which we typically speak. This study demonstrates that semantic interference disappears, and can even turn into facilitation, when semantically related words are embedded in a meaningful communicative exchange. Experiment 1 and 3 (each N = 32 university students) implemented a picture-word interference task in a game played between two participants: one named the distractor word and, after a stimulus-onset-asynchrony of -150 ms or -650 ms, the other named a semantically related or unrelated target picture. Semantic interference reappeared with identical experimental parameters in a single-person picture-word interference setting (Experiment 2, N = 32). We conclude that the inhibitory context effects leading to semantic interference in single-subject settings are attenuated whereas facilitatory effects are enhanced in communicative settings.

The Effect of Lexical Cohort Size Is Independent of Semantic Context Effects in a Picture-Word Interference Task: A Combined ERP and sLORETA Study

Lexical cohort size is known to play an important role in the magnitude of semantic interference during picture naming in continuous and blocking naming tasks. Nevertheless, whether and how lexical cohort size influences semantic context effects in a picture-word interference (PWI) task remains unclear. To address this issue, participants were required to name pictures, which were paired with both semantically related and unrelated distractors, from both large and small lexical cohorts while electroencephalogram (EEG) signals were recorded. Behavior results showed a semantic interference effect but no interaction between semantic relatedness and lexical cohort size in naming latencies. ERPs and correlation analyses revealed that semantic interference effects occurred at the lexical level in the time windows of 200-400 and 400-600 ms, and lexical cohort size effects occurred at the conceptual level in the time window of 100-200 ms and at the lexical level in the time windows of 200-400 ms. Critically, no interaction between two variables was found, reflecting that lexical cohort size is independent of semantic interference for categorical relations in the PWI. sLORETA results found stronger brain activations for large lexical cohorts at the left superior temporal gyrus and inferior frontal gyrus in the time interval of 250-300 ms, which may relate to lexical selection and self-monitoring. Our findings provide evidence for the swinging lexical network rather than the response exclusion hypothesis in spoken production.

Word naming slows picture naming but does not affect cumulative semantic interference

Two experiments are reported which investigate the effect of processing words prior to naming target pictures. In Experiment 1, participants named (read aloud) sequences of five printed prime words and five target pictures from the same semantic category, and also sequences of five prime words from a different unrelated semantic category to the five related target pictures. Picture and words were interleaved, with two unrelated filler stimuli in between prime and target stimuli (i.e. a lag of 3 between primes and targets). Results showed that across the five target picture naming trials (i.e. across ordinal position of picture), picture naming times increased linearly, replicating the cumulative semantic interference (CSI) effect (e.g., Howard, Nickels, Coltheart, & Cole-Virtue, 2006). Related prime words slowed picture naming, replicating the effects found in paired word prime and picture target studies (e.g., Tree & Hirsh, 2003). However, the naming of the five related prime words did not modify the picture naming CSI effect, with this null result converging with findings from a different word and picture design (e.g., Navarrete, Mahon, & Caramazza, 2010). In Experiment 2, participants categorised the prime word stimuli as manmade versus natural, so that words were more fully processed at a conceptual level. The interaction between word prime relatedness and ordinal position of the named target picture was significant. These results are consistent with adjustments at the conceptual level (Belke, 2013; Roelofs, 2018) which last over several trials at least. By contrast, we conclude that the distinct word-to-picture naming interference effect from Experiment 1 must originate outside of the conceptual level and outside of the mappings between semantics and lexical representations. We discuss the results with reference to recent theoretical accounts of the CSI picture naming effect and word naming models.

Cognitive style modulates semantic interference effects: evidence from field dependency

The so-called semantic interference effect is a delay in selecting an appropriate target word in a context where semantic neighbours are strongly activated. Semantic interference effect has been described to vary from one individual to another. These differences in the susceptibility to semantic interference may be due to either differences in the ability to engage in lexical-specific selection mechanisms or to differences in the ability to engage more general, top-down inhibition mechanisms which suppress unwanted responses based on task-demands. However, semantic interference may also be modulated by an individual's disposition to separate relevant perceptual signals from noise, such as a field-independent (FI) or a field-dependent (FD) cognitive style. We investigated the relationship between semantic interference in picture naming and in an STM probe task and both the ability to inhibit responses top-down (measured through a Stroop task) and a FI/FD cognitive style measured through the embedded figures test (EFT). We found a significant relationship between semantic interference in picture naming and cognitive style-with semantic interference increasing as a function of the degree of field dependence-but no associations with the semantic probe and the Stroop task. Our results suggest that semantic interference can be modulated by cognitive style, but not by differences in the ability to engage top-down control mechanisms, at least as measured by the Stroop task.

Semantic interference in speech error production in a randomized continuous naming task: Evidence from aphasia

Naming pictures from the same semantic category hinders subsequent naming from that category (i.e., semantic interference), irrespective of the number of intervening different-category exemplars named. Persistent semantic interference has been well documented in chronometric studies, and has been attributed to experience-driven adjustments in the strength of connections between semantic and lexical representations. However, whether parallel effects exist in speech error data remains unclear. In the current study, people with aphasia, a speaker population prone to naming errors, provided naming responses to a large picture corpus presented in random order that comprised multiple exemplars drawn from several different categories. We found persistent semantic interference in the task in semantic error rates specifically, and that semantic similarity between consecutive related exemplars modulated the effect. The results provide further evidence for the presumed lexical-semantic locus and mechanism(s) underlying semantic interference.

Facilitation and interference in naming: A consequence of the same learning process?

Our success with naming depends on what we have named previously, a phenomenon thought to reflect learning processes. Repeatedly producing the same name facilitates language production (i.e., repetition priming), whereas producing semantically related names hinders subsequent performance (i.e., semantic interference). Semantic interference is found whether naming categorically related items once (continuous naming) or multiple times (blocked cyclic naming). A computational model suggests that the same learning mechanism responsible for facilitation in repetition creates semantic interference in categorical naming (Oppenheim, Dell, & Schwartz, 2010). Accordingly, we tested the predictions that variability in semantic interference is correlated across categorical naming tasks and is caused by learning, as measured by two repetition priming tasks (picture-picture repetition priming, Exp. 1; definition-picture repetition priming, Exp. 2, e.g., Wheeldon & Monsell, 1992). In Experiment 1 (77 subjects) semantic interference and repetition priming effects were robust, but the results revealed no relationship between semantic interference effects across contexts. Critically, learning (picture-picture repetition priming) did not predict semantic interference effects in either task. We replicated these results in Experiment 2 (81 subjects), finding no relationship between semantic interference effects across tasks or between semantic interference effects and learning (definition-picture repetition priming). We conclude that the changes underlying facilitatory and interfering effects inherent to lexical access are the result of distinct learning processes where multiple mechanisms contribute to semantic interference in naming.

Taxonomic and thematic semantic systems

Object concepts are critical for nearly all aspects of human cognition, from perception tasks like object recognition, to understanding and producing language, to making meaningful actions. Concepts can have 2 very different kinds of relations: similarity relations based on shared features (e.g., dog-bear), which are called "taxonomic" relations, and contiguity relations based on co-occurrence in events or scenarios (e.g., dog-leash), which are called "thematic" relations. Here, we report a systematic review of experimental psychology and cognitive neuroscience evidence of this distinction in the structure of semantic memory. We propose 2 principles that may drive the development of distinct taxonomic and thematic semantic systems: differences between which features determine taxonomic versus thematic relations, and differences in the processing required to extract taxonomic versus thematic relations. This review brings together distinct threads of behavioral, computational, and neuroscience research on semantic memory in support of a functional and neural dissociation, and defines a framework for future studies of semantic memory. (PsycINFO Database Record

Language production in a shared task: Cumulative Semantic Interference from self- and other-produced context words

This study assessed the effects of semantic context in the form of self-produced and other-produced words on subsequent language production. Pairs of participants performed a joint picture naming task, taking turns while naming a continuous series of pictures. In the single-speaker version of this paradigm, naming latencies have been found to increase for successive presentations of exemplars from the same category, a phenomenon known as Cumulative Semantic Interference (CSI). As expected, the joint-naming task showed a within-speaker CSI effect, such that naming latencies increased as a function of the number of category exemplars named previously by the participant (self-produced items). Crucially, we also observed an across-speaker CSI effect, such that naming latencies slowed as a function of the number of category members named by the participant's task partner (other-produced items). The magnitude of the across-speaker CSI effect did not vary as a function of whether or not the listening participant could see the pictures their partner was naming. The observation of across-speaker CSI suggests that the effect originates at the conceptual level of the language system, as proposed by Belke's (2013) Conceptual Accumulation account. Whereas self-produced and other-produced words both resulted in a CSI effect on naming latencies, post-experiment free recall rates were higher for self-produced than other-produced items. Together, these results suggest that both speaking and listening result in implicit learning at the conceptual level of the language system but that these effects are independent of explicit learning as indicated by item recall.