Spreading Activation in an Attractor Network With Latching Dynamics: Automatic Semantic Priming Revisited

Localist models of spreading activation (SA) and models assuming distributed representations offer very different takes on semantic priming, a widely investigated paradigm in word recognition and semantic memory research. In this study, we implemented SA in an attractor neural network model with distributed representations and created a unified framework for the two approaches. Our models assume a synaptic depression mechanism leading to autonomous transitions between encoded memory patterns (latching dynamics), which account for the major characteristics of automatic semantic priming in humans. Using computer simulations, we demonstrated how findings that challenged attractor‐based networks in the past, such as mediated and asymmetric priming, are a natural consequence of our present model’s dynamics. Puzzling results regarding backward priming were also given a straightforward explanation. In addition, the current model addresses some of the differences between semantic and associative relatedness and explains how these differences interact with stimulus onset asynchrony in priming experiments.     

Lexical processing of ambiguous words: dominance or associative strength?

Four experiments examined the role of meaning frequency (dominance) and associative strength (measured by associative norms) in the processing of ambiguous words in isolation. Participants made lexical decisions to targets words that were associates of the more frequent (dominant) or less frequent (subordinate) meaning of a homograph prime. The first two experiments investigated the role of associative strength at long SOAs (Stimulus Onset Asynchrony) (750 ms.), showing that meaning is facilitated by the targets' associative strength and not by their dominance. The last two experiments traced the role associative strength at short SOAs (250 ms), showing that the manipulation of the associative strength has no effect in the semantic priming. The conclusions are: on the one hand, semantic priming for homographs is due to associative strength manipulations at long SOAs. On the other hand, the manipulation of the associative strength has no effect when automatic processes (short SOAs) are engaged for homographs.     

The effects of associative and semantic priming in the lexical decision task

Four lexical decision experiments were conducted to examine under which conditions automatic semantic priming effects can be obtained. Experiments 1 and 2 analyzed associative/semantic effects at several very short stimulus-onset asynchronies (SOAs), whereas Experiments 3 and 4 used a single-presentation paradigm at two response-stimulus intervals (RSIs). Experiment 1 tested associatively related pairs from three semantic categories (synonyms, antonyms, and category coordinates). The results showed reliable associative priming effects at all SOAs. In addition, the correlation between associative strength and magnitude of priming was significant only at the shortest SOA (66 ms). When prime-target pairs were semantically but not associatively related (Experiment 2), reliable priming effects were obtained at SOAs of 83 ms and longer. Using the single-presentation paradigm with a short RSI (200 ms, Experiment 3), the priming effect was equal in size for associative + semantic and for semantic-only pairs (a 21-ms effect). When the RSI was set much longer (1,750 ms, Experiment 4), only the associative + semantic pairs showed a reliable priming effect (23 ms). The results are interpreted in the context of models of semantic memory.     

Does<Emphasis Type="SmallCaps">LGHT</Emphasis> prime<Emphasis Type="SmallCaps">DARK</Emphasis>? Masked associative priming with addition neighbors

We present two masked priming lexical decision experiments in which we examined whether a nonword prime word would activate associative/semantic information from its corresponding addition neighbor (e.g.,lght-DARK via the addition neighborlight), producing associative/semantic priming. The rationale was the following: If a nonword prime with a missing letter produced a semantic/associative priming effect, this would clearly indicate that this nonword was activating the lexical/semantic representations of its base word, thereby reinforcing the models of visual-word recognition in which the orthographic representations produced bylght (orligt) andlight are quite similar (e.g., SOLAR, SERIOL, open-bigram, and overlap models). The results showed that the magnitude of the masked associative priming effect with subset primes was remarkably similar to that of the priming effect with the corresponding word prime. Furthermore, the magnitude of the associative priming effect was similar when the deleted letter was a vowel and when the deleted letter was a consonant.     

Associative priming in faces: Semantic relatedness or simple co-occurrence?

In two experiments, we explored the effects of co-occurrence and semantic relationships in the associative priming of faces. In Experiment 1, pairs of computer-generated human faces were presented simultaneously (i.e., they co-occurred) with no associated semantic information attached to them. A significant facilitation effect in the subsequent recognition of these paired faces (priming) was observed. Thus, repeatedly presenting faces together while keeping semantic information to a minimum appears to be enough to produce associative priming. In Experiment 2, the computer-generated faces were associated with semantic information and again presented ipairs. Priming effects arising from co-occurrence and semantic relatedness were observed. The results from these experiments show that semantic relatedness is not the sole cause of the association between faces; co-occurrence plays a crucial role too. This conclusion has significant implications for the current computational models of face processing.     

Event-related potentials indicate different processes to mediate categorical and associative priming in person recognition

Whether representations of people are stored in associative networks based on co-occurrence or are stored in terms of more abstract semantic categories is a controversial question. In the present study, participants performed fame decisions to unfamiliar or famous target faces (Experiment 1) or names (Experiment 2), which were primed, either by highly associated celebrity names or by names from the same occupational category, or were unprimed. Reaction times and event-related potentials (ERPs) were recorded. Reaction times yielded significant priming effects for both associated and same category conditions. ERPs to targets in the associated condition were significantly more positive than were ERPs in all other conditions over central and parietal areas (300-600 ms; N400 priming effect). By contrast, a more posterior effect was found for categorical priming. These findings held for both cross-domain (Experiment 1) and within-domain conditions (Experiment 2). Results (a) demonstrate behavioral and ERP evidence for categorical priming in person recognition, consistent with the assumption that shared semantic information units can mediate semantic priming, and (b) suggest that associative and categorical priming are based on mechanisms that are at least partially different.     

Semantic priming effects for stimuli presented to the right and left visual fields

Studies that have addressed the possibility of hemispheric differences in semantic priming effects have yielded contradictory results. This paper reports the findings of two experiments intended to shed greater light on the issue of hemispheric differences in semantic priming. Experiment 1 used a hemiretinal paradigm and examined manual response latency and response accuracy to four types of word pairs; categorically related, syntactically related, unrelated, and pairs containing a nonword member. Experiment 2 examined the effects of unrecognized, disambiguating flank words on verbal responses to a centrally presented homograph. Experiment 1 yielded no significant visual field differences in magnitude of priming effects when response latency served as the dependent measure, although categorical relatedness facilitated response accuracy for left but not right visual field stimuli. In experiment 2, the disambiguating words had a significant effect on meaning interpretation of the homographs that was independent of visual field of presentation. Taken together, the results of these studies are interpreted as indicating that semantic aspects of linguistic input are automatically processed and can influence the content and latency of subsequent responses, whether presented to the left or right visual field.     

不熟练中–日双语者同形词和非同形词的隐蔽翻译启动效应

采用不熟练中–日双语者, 以中文与日文的同形词和非同形词为材料, 考察了翻译方向、词汇拼写形式和加工任务对隐蔽翻译启动效应的影响。结果表明：(1)不熟练中–日双语者的中文词和日文词之间存在着显著的隐蔽翻译启动效应。(2)字词拼写形式对不熟练中–日双语者两种语言的隐蔽翻译启动效应具有重要的影响：当中文词和日文词同形时, 中文词启动日文词和日文词启动中文词出现了对称的启动效应; 当中文词和日文词不同形时, 中文词启动日文词和日文词启动中文词出现了不对称的启动效应。这种不对称性与加工任务有关：在词汇决定任务中, 中文词对日文词的启动效应显著大于日文词对中文词的启动效应; 在语义决定任务中, 两个方向的启动效应没有显著的差异。(3)对中–日同形词, 任务类型对隐蔽翻译启动效应的影响不显著; 对中–日非同形词, 任务类型对隐蔽翻译启动效应具有重要的影响。根据研究的结果, 作者提出了不熟练中–日双语者语言表征的混合模型。     

Conceptual Hierarchies in a Flat Attractor Network: Dynamics of Learning and Computations

The structure of people's conceptual knowledge of concrete nouns has traditionally been viewed as hierarchical ( Collins & Quillian, 1969 ). For example, superordinate concepts ( vegetable ) are assumed to reside at a higher level than basic-level concepts ( carrot ). A feature-based attractor network with a single layer of semantic features developed representations of both basic-level and superordinate concepts. No hierarchical structure was built into the network. In Experiment and Simulation 1, the graded structure of categories (typicality ratings) is accounted for by the flat attractor network. Experiment and Simulation 2 show that, as with basic-level concepts, such a network predicts feature verification latencies for superordinate concepts ( vegetable ). In Experiment and Simulation 3, counterintuitive results regarding the temporal dynamics of similarity in semantic priming are explained by the model. By treating both types of concepts the same in terms of representation, learning, and computations, the model provides new insights into semantic memory.     

The structure of semantic person memory: evidence from semantic priming in person recognition

This paper reviews research on the structure of semantic person memory as examined with semantic priming. In this experimental paradigm, a familiarity decision on a target face or written name is usually faster when it is preceded by a related as compared to an unrelated prime. This effect has been shown to be relatively short lived and susceptible to interfering items. Moreover, semantic priming can cross stimulus domains, such that a written name can prime a target face and vice versa. However, it remains controversial whether representations of people are stored in associative networks based on co-occurrence, or in more abstract semantic categories. In line with prominent cognitive models of face recognition, which explain semantic priming by shared semantic information between prime and target, recent research demonstrated that priming could be obtained from purely categorically related, non-associated prime/target pairs. Although strategic processes, such as expectancy and retrospective matching likely contribute, there is also evidence for a non-strategic contribution to priming, presumably related to spreading activation. Finally, a semantic priming effect has been demonstrated in the N400 event-related potential (ERP) component, which may reflect facilitated access to semantic information. It is concluded that categorical relatedness is one organizing principle of semantic person memory.     

Does jugde activate COURT? Transposed-letter similarity effects in masked associative priming

Transposed-letter (TL) nonwords (e.g., jugde) can be easily misperceived as words, a fact that is somewhat inconsistent with the letter-position-coding schemes employed by most current models of visual word recognition. To examine this issue further, we conducted four masked semantic/associative priming experiments, using a lexical decision task. In Experiment 1, the related primes could be words, TL-internal nonwords, or replacement-letter (RL) nonwords (e.g., judge, jugde, or judpe, respectively; the target would be COURT). Relative to an unrelated condition, masked TL-internal primes produced a significant semantic/associative priming effect, an effect that was only slightly smaller than the priming effect for word primes. No effect, however, was observed for RL-nonword primes. In Experiment 2, the TL-nonword primes were created by switching the two final letters of the primes (e.g., judeg). The results again showed a semantic/associative priming effect for word primes, but not for TL-final nonword primes or for RL-nonword primes. Experiment 3 replicated the associative/semantic priming effect for TL-internal nonword primes, with, again, no effect for TL-final nonword primes. Finally, Experiment 4 again failed to yield a priming effect for TL-final nonword primes. The implications of these results for the choice of a letter-position-coding scheme in visual word recognition models are discussed.     

The kinds of information that support novel associative object priming and how these differ from those that support item priming

We investigated how the information that supports novel associative and item object priming differs under identical study/test conditions. In Experiments 1 and 2, participants rated the meaningfulness of sentences linking two object pictures at study. At test, they performed either a size judgement or an associative recognition memory task on intact, recombined and novel picture (Experiment 1) or word (Experiment 2) associations. Associative priming was modulated by subjective meaningfulness of the encoded links, and depended on study/test perceptual overlap. In contrast, item priming was neither affected by the meaningfulness of the sentences nor by study/test changes in the stimulus presentation format. Associative priming and recognition were behaviourally dissociated, and associative recognition was probably too slow to have seriously contaminated associative priming. In Experiment 3, participants performed a perceptually oriented task during both experimental phases, and both associative and item priming were observed. These results suggest that associative priming depends on stored associative semantic and perceptual information when the test task requires flexible retrieval of associative information. Under the same conditions, item priming may only require activation of items' semantic properties. When both study and test tasks stress perceptual processing, retrieval of perceptual information is sufficient to support both kinds of priming.     

Integrative and semantic relations equally alleviate age-related associative memory deficits

Two experiments compared effects of integrative and semantic relations between pairs of words on lexical and memory processes in old age. Integrative relations occur when two dissimilar and unassociated words are linked together to form a coherent phrase (e.g., horse-doctor). In Experiment 1, older adults completed a lexical-decision task where prime and target words were related either integratively or semantically. The two types of relation both facilitated responses compared to a baseline condition, demonstrating that priming can occur in older adults with minimal preexisting associations between primes and targets. In Experiment 2, young and older adults completed a cued recall task with integrative, semantic, and unrelated word pairs. Both integrative and semantic pairs showed significantly smaller age differences in associative memory compared to unrelated pairs. Integrative relations facilitated older adults' memory to a similar extent as semantic relations despite having few preexisting associations in memory. Integratability of stimuli is therefore a new factor that reduces associative deficits in older adults, most likely by supporting encoding and retrieval mechanisms.     

Graph‐Theoretic Properties of Networks Based on Word Association Norms: Implications for Models of Lexical Semantic Memory

We compared the ability of three different contextual models of lexical semantic memory (BEAGLE, Latent Semantic Analysis, and the Topic model) and of a simple associative model (POC) to predict the properties of semantic networks derived from word association norms. None of the semantic models were able to accurately predict all of the network properties. All three contextual models over‐predicted clustering in the norms, whereas the associative model under‐predicted clustering. Only a hybrid model that assumed that some of the responses were based on a contextual model and others on an associative network (POC) successfully predicted all of the network properties and predicted a word's top five associates as well as or better than the better of the two constituent models. The results suggest that participants switch between a contextual representation and an associative network when generating free associations. We discuss the role that each of these representations may play in lexical semantic memory. Concordant with recent multicomponent theories of semantic memory, the associative network may encode coordinate relations between concepts (e.g., the relation between pea and bean, or between sparrow and robin), and contextual representations may be used to process information about more abstract concepts.     

Semantic priming of person recognition: Categorial priming may be a weaker form of the associative priming effect

An interactive activation and competition account (Burton, Bruce, & Johnston, 1990) of the semantic priming effect in person recognition studies relies on the fact that primes and targets (people) have semantic information in common. However, recent investigations into the type of relationship needed to mediate the semantic priming effect have suggested that the prime and target must be close associates (e.g., Barry, Johnston, & Scanlan, 1998; Young, Flude, Hellawell, & Ellis, 1994). A review of these and similar papers suggests the possibility of a small but non-reliable effect based purely on categorial relationships. Experiment 1 provided evidence that when participants were asked to make a name familiarity decision it was possible to boost this small categorial effect when multiple (four) primes were presented prior to the target name. Results from Experiment 2 indicated that the categorial effect was not due to the particular presentation times of the primes. This boosted categorial effect was shown to cross domains (names to faces) in Experiment 3 and persist in Experiment 4 when the task involved naming the target face. The similarity of the pattern of results produced by the associative priming effect and this boosted categorial effect suggests that the two may be due to the same underlying mechanism in semantic memory.     

An Attractor Model of Lexical Conceptual Processing: Simulating Semantic Priming

An attractor network was trained to compute from word form to semantic representations that were based on subject-generated features. The model was driven largely by higher-order semantic structure. The network simulated two recent experiments that employed items included in its training set (McRae and Boisvert, 1998). In Simulation 1, short stimulus onset asynchrony priming was demonstrated for semantically similar items. Simulation 2 reproduced subtle effects obtained by varying degree of similarity. Two predictions from the model were then tested on human subjects. In Simulation 3 and Experiment 1, the items from Simulation 1 were reversed, and both the network and subjects showed minimally different priming effects in the two directions. In Experiment 2, consistent with attractor networks but contrary to a key aspect of hierarchical spreading activation accounts priming was determined by featural similarity rather than shared superordinate category. It is concluded that semantic-similarity priming is due to featural overlap that is a natural consequence of distributed representations of word meaning.     

Implicit memory for novel associations between pictures: effects of stimulus unitization and aging

Studies of implicit memory for novel associations have focused primarily on verbal materials and have highlighted the contribution of conceptually unitized representations to such priming. Using pictorial stimuli in a perceptual identification task, we examined whether new association priming can occur at a purely perceptual level. By manipulating the spatial contiguity of stimuli, we also evaluated whether such priming requires the creation of perceptually unitized representations. Finally, we examined the status of such priming in aging. In Experiment 1, we found that spatial contiguity of stimuli is not necessary for novel pictorial association priming to emerge, although such contiguity does enhance the magnitude of associative priming. In Experiment 2, we found that new association priming is age invariant, regardless of spatial contiguity. In Experiment 3, we provide additional evidence that pictorial association priming is perceptually based. These findings expand the scope and delineate the conditions of novel association priming and inform theories about the nature of implicit memory for new associations.     

韩-汉跨语言词义水平的重复知盲效应*

关于语义水平的重复知盲效应一直存在争论。本研究考察非平衡韩-汉双语者在语言内和语言间词义水平的RB效应。实验1采用韩语双字词为实验材料探究韩语语言内的RB效应,结果表明在韩语词汇水平存在RB效应。实验2采用两个关键词为韩语（R1）和汉语（R2）双字词,并通过韩-汉语间隔词W的语言类型变化探究韩、汉语言间的RB效应,结果表明：不同语言类型的间隔词的正确率有显著差异,语义重复词R2在重复条件的正确率显著高于非重复条件,出现语义重复启动效应。本研究结论是：（1）在韩语语言内词水平存在RB效应,具有语言类型普遍性;（2）韩、汉跨语言语义水平出现语义重复启动效应,而未发生RB效应,表明双语者在加工韩、汉两种语言时,采用分别形式表征和共享语义表征的方式。     

Semantic activation and letter search: Blocking or suppression?

Two experiments investigated associative priming with a letter-search prime task where either the prime and letter probe were presented simultaneously, or the letter probe appeared 200 ms (Experiment 1) or 300 ms (Experiment 2) after the prime. Weak associative priming was observed in both experiments, but unlike Stolz and Besner (1996) we found no evidence that priming was increased when the probe was delayed. However, strong associative priming was observed when a semantic decision had to be made on the prime (Experiment 3). Our results are consistent with an account where semantic activation of the prime occurs but its action on the target is suppressed by the prime task. The persistence of weak priming effects with the letter search task is explained in terms of the low-frequency items used.