基于语义网络的创造性认知机制研究现状与展望

语义网络是主要的知识表征方式之一,对创造性研究具有重要意义。近年来不断发展的网络科学方法为相关研究提供了强大的方法论框架及量化工具,研究者得以分别从语义网络结构及语义相关认知操作的角度探讨语义网络与创造性的内在联系。已有证据表明语义网络中所储存的概念越丰富、概念间的连接通路越多元化或越灵活,则个体创造性水平越高;而个体创造性水平亦受制于个体对概念的检索、抑制、选择、整合及扩展等认知能力。在此方向上,未来研究工作应当更多地思考如何在一个相对整合的认知框架中深入探讨语义网络结构及相关认知过程对创造性的交互作用,并结合神经影像等技术手段,将语义网络之于创造性的认知机制嵌套于神经科学体系中进行解释。     

Associative thinking at the core of creativity

Creativity has long been thought to involve associative processes in memory: connecting concepts to form ideas, inventions, and artworks. However, associative thinking has been difficult to study due to limitations in modeling memory structure and retrieval processes. Recent advances in computational models of semantic memory allow researchers to examine how people navigate a semantic space of concepts when forming associations, revealing key search strategies associated with creativity. Here, we synthesize cognitive, computational, and neuroscience research on creativity and associative thinking. This Review highlights distinctions between free- and goal-directed association, illustrates the role of associative thinking in the arts, and links associative thinking to brain systems supporting both semantic and episodic memory – offering a new perspective on a longstanding creativity theory.     

The roles of associative and executive processes in creative cognition

How does the mind produce creative ideas? Past research has pointed to important roles of both executive and associative processes in creative cognition. But such work has largely focused on the influence of one ability or the other—executive or associative—so the extent to which both abilities may jointly affect creative thought remains unclear. Using multivariate structural equation modeling, we conducted two studies to determine the relative influences of executive and associative processes in domain-general creative cognition (i.e., divergent thinking). Participants completed a series of verbal fluency tasks, and their responses were analyzed by means of latent semantic analysis (LSA) and scored for semantic distance as a measure of associative ability. Participants also completed several measures of executive function—including broad retrieval ability (Gr) and fluid intelligence (Gf). Across both studies, we found substantial effects of both associative and executive abilities: As the average semantic distance between verbal fluency responses and cues increased, so did the creative quality of divergent-thinking responses (Study 1 and Study 2). Moreover, the creative quality of divergent-thinking responses was predicted by the executive variables—Gr (Study 1) and Gf (Study 2). Importantly, the effects of semantic distance and the executive function variables remained robust in the same structural equation model predicting divergent thinking, suggesting unique contributions of both constructs. The present research extends recent applications of LSA in creativity research and provides support for the notion that both associative and executive processes underlie the production of novel ideas.     

The cognitive underpinnings of creative thought: A latent variable analysis exploring the roles of intelligence and working memory in three creative thinking processes

The field of creativity has largely focused on individual differences in divergent thinking abilities. Recently, contemporary creativity researchers have shown that intelligence and executive functions play an important role in divergent thought, opening new lines of research to examine how higher-order cognitive mechanisms may uniquely contribute to creative thinking. The present study extends previous research on the intelligence and divergent thinking link by systematically examining the relationships among intelligence, working memory, and three fundamental creative processes: associative fluency, divergent thinking, and convergent thinking. Two hundred and sixty five participants were recruited to complete a battery of tasks that assessed a range of elementary to higher-order cognitive processes related to intelligence and creativity. Results provide evidence for an associative basis in two distinct creative processes: divergent thinking and convergent thinking. Findings also supported recent work suggesting that intelligence significantly influences creative thinking. Finally, working memory played a significant role in creative thinking processes. Recasting creativity as a construct consisting of distinct higher-order cognitive processes has important implications for future approaches to studying creativity within an individual differences framework.     

Revisiting Mednick's Model on Creativity‐Related Differences in Associative Hierarchies. Evidence for a Common Path to Uncommon Thought

Fifty years ago, Mednick [Psychological Review, 69 (1962) 220] proposed an elaborate model that aimed to explain how creative ideas are generated and why creative people are more likely to have creative ideas. The model assumes that creative people have flatter associative hierarchies and as a consequence can more fluently retrieve remote associative elements, which can be combined to form creative ideas. This study aimed at revisiting Mednick's model and providing an extensive test of its hypotheses. A continuous free association task was employed and association performance was compared between groups high and low in creativity, as defined by divergent thinking ability and self‐report measures. We found that associative hierarchies do not differ between low and high creative people, but creative people showed higher associative fluency and more uncommon responses. This suggests that creativity may not be related to a special organization of associative memory, but rather to a more effective way of accessing its contents. The findings add to the evidence associating creativity with highly adaptive executive functioning.     

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.     

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.     

A graph theory model of the semantic structure of attitudes

The semantic structure underlying the attitudes of pretreatment and posttreatment drug addicts was modeled using a network analysis of free word associations. Measures of graph theoretic properties were used to assess structural differences in the associative networks of the two populations. These measures modeled the information processes of associative networks proposed in the spreading activation theory of semantic processing. As expected based on graph theory, the structure of the associative networks of post-treatment subjects was more dense, less constrained, and more hierarchically organized by the self concept. In a test of the network model, the subjects' evaluations of concepts in the associative network were found to be a function of their evaluations of semantically similar concepts. Although preliminary and limited, the results suggest that graph theory may provide a broad mathematical foundation for diverse models of cognitive systems.The research was funded by a grant from the National Institute for Drug Abuse to the Institute of Cultural Studies, Inc.Daytop Village.     

Integrating the Automatic and the Controlled: Strategies in Semantic Priming in an Attractor Network With Latching Dynamics

Semantic priming has long been recognized to reflect, along with automatic semantic mechanisms, the contribution of controlled strategies. However, previous theories of controlled priming were mostly qualitative, lacking common grounds with modern mathematical models of automatic priming based on neural networks. Recently, we introduced a novel attractor network model of automatic semantic priming with latching dynamics. Here, we extend this work to show how the same model can also account for important findings regarding controlled processes. Assuming the rate of semantic transitions in the network can be adapted using simple reinforcement learning, we show how basic findings attributed to controlled processes in priming can be achieved, including their dependency on stimulus onset asynchrony and relatedness proportion and their unique effect on associative, category‐exemplar, mediated and backward prime‐target relations. We discuss how our mechanism relates to the classic expectancy theory and how it can be further extended in future developments of the model.     

Looking for meaning: Eye movements are sensitive to overlapping semantic features, not association

Theories of semantic memory differ in the extent to which relationships among concepts are captured via associative or via semantic relatedness. We examined the contributions of these two factors, using a visual world paradigm in which participants selected the named object from a four-picture display. We controlled for semantic relatedness while manipulating associative strength by using the visual world paradigm’s analogue to presenting asymmetrically associated pairs in either their forward or backward associative direction (e.g., ham—eggs vs. eggs—ham). Semantically related objects were preferentially fixated regardless of the direction of presentation (and the effect size was unchanged by presentation direction). However, when pairs were associated but not semantically related (e.g., iceberg—lettuce), associated objects were not preferentially fixated in either direction. These findings lend support to theories in which semantic memory is organized according to semantic relatedness (e.g., distributed models) and suggest that association by itself has little effect on this organization.     

Structural Comparison of Cognitive Associative Networks in Two Populations1

The cognitive associative structure of 2 populations was studied using network analysis of free‐word associations. Structural differences in the associative networks were compared using measures of network centralization, size, density, clustering, and path length. These measures are closely aligned with cognitive theories describing the organization of knowledge and retrieval of concepts from memory. Size and centralization of semantic structures were larger for college students than for 7^th graders, while density, clustering, and mean path length were similar. Findings presented reveal that subpopulations might have very different cognitive associative networks. This study suggests that graph theory and network analysis methods are useful in mapping differences in associative structures across groups.     

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.     

Associative recognition processes are modulated by the semantic unitizability of memoranda

Although memory of episodic associations is generally considered to be recollective in nature, it has been suggested that when stimuli are experienced as a unit, familiarity processes might contribute to their subsequent associative recognition. To investigate the effect of semantic relatedness during episodic encoding on the processes of retrieval of associative information, we had participants interactively encode pairs of object pictures, vertically arranged so as to suggest a functional or configural relationship between them. Half the pairs were independently judged to be of related objects (e.g., a lamp over a table) and half of unrelated objects (e.g., a key-ring over an apple). At test, participants discriminated between intact, recombined, and new pairs while event related potentials (ERPs) were recorded. In an early ERP marker of retrieval success generally associated with familiarity processes, differences related to associative memory only emerged for related pairs, while differences associated with item memory emerged for both related and unrelated pairs. In contrast, in a later ERP effect associated with recollection, differences related to associative memory emerged for both related and unrelated pairs. These findings may indicate that retrieval of episodic associations formed between two semantically related visual stimuli can be supported by familiarity-related processes.     

On Being Found: How Habitual Patterns of Thought Influence Creative Interest,Behavior, and Ability

This study on 138 undergraduate students used path analysis to investigate the relationship between creativity (interest, measured by a creative activities survey; and ability: fluency, originality, and elaboration) and different aspects of thought patterns presumed to influence the preparation and illumination phase of the creative process: habitual patterns of thought (ruminative brooding, ruminative self-reflection), thought suppression, thought intrusion, mind wandering, and associative ability. Such relationship was hinted at in Wallas’s classical model of the creative sequence, but is rarely investigated. We found that creative behavior/interest was driven by self-reflection, thought intrusion, and the lack of a need for thought suppression; creative ability was fueled mainly by associative ability. The only variable that influenced both aspects of creativity was the lack of resistance to thought suppression; this distinguished the creativity variables from dysphoria, which was associated with a desire for thought suppression. The results suggest that what drives the need to create is not creative ability per se, but rather self-focused attention, as well as the feeling or experience of being found by thought, rather than finding it. That is, the need to create is associated with having thoughts that interrupt one’s ordinary stream of consciousness and that are seen as welcome rather than interfering.     

Semantic relations differentially impact associative recognition memory: Electrophysiological evidence

Though associative recognition memory is thought to rely primarily on recollection, recent research indicates that familiarity might also make a substantial contribution when to-be-learned items are integrated into a coherent structure by means of an existing semantic relation. It remains unclear how different types of semantic relations, such as categorical (e.g., dancer–singer) and thematic (e.g., dancer–stage) relations might affect associative recognition, however. Using event-related potentials (ERPs), we addressed this question by manipulating the type of semantic link between paired words in an associative recognition memory experiment. An early midfrontal old/new effect, typically linked to familiarity, was observed across the relation types. In contrast, a robust left parietal old/new effect was found in the categorical condition only, suggesting a clear contribution of recollection to associative recognition for this kind of pairs. One interpretation of this pattern is that familiarity was sufficiently diagnostic for associative recognition of thematic relations, which could result from the integrative nature of the thematic relatedness compared to the similarity-based nature of categorical pairs. The present study suggests that the extent to which recollection and familiarity are involved in associative recognition is at least in part determined by the properties of semantic relations between the paired associates.     

Dynamic Creative Interaction Networks and Team Creativity Evolution: A Longitudinal Study

To assess the dynamical effects of creative interaction networks on team creativity evolution, this paper elaborates a theoretical framework that links the key elements of creative interaction networks, including node, edge and network structure, to creativity in teams. The process of team creativity evolution is divided into four phases, including formation, growth, maturity and decline/restart. The importance of domain‐relevant knowledge, creativity‐relevant skill, interaction frequency, interaction length, network density and closeness centrality are emphasized in specific phases of team creativity evolution in a complex creative context. To test our assumptions, a longitudinal study of creative teams in a “Challenge Cup” Creative Business Plan Competition for university students is performed and the full networks of 17 creative entrepreneur teams are mapped. Both static comparison and dynamic analysis are conducted to analyze the relationship between creative interaction networks and team creativity evolution. For specific phases of team creativity evolution, we find confirmation of our predictions. The implications of dynamic creative interaction networks for all the phases of creative teams from formation to decline/restart are discussed.     

创造性问题解决中的联想能力及其作用

长期以来,研究者一直认为联想能力与创造力之间存在着紧密的联系,但相关的实证研究并不能很好地支持这一论断,这可能是由于联想能力的界定不清所造成的。本研究梳理了以往研究中联想能力与创造力的关系,在此基础上辨析了联想能力测验所反映的实质,进一步提出了一个创造性问题解决中联想能力的作用模型：（1）联想能力主要作用于创造性问题解决中的重构阶段;（2）存在三种相对独立的联想能力,它们对空间类顿悟问题和言语类顿悟问题的影响不同;（3）联想内容中经验关系的联想能力与领域一般创造力关联最密切。     

Poster presentations

We analysed the organisation of semantic network using associative mechanisms between different types of information and studied the progression of the use of these associative relations during development. We aimed to verify the linkage of concepts with the use of semantic associative relations. The goal of this study was to analyse the cognitive ability to use associative relations between various items when describing old and/or new concepts. We examined the performance of 100 subjects between the ages of 4 and 7 years on an experimental task using five associative relations based on verbal encoding. The results showed that children are able to use the five semantic associative relations at age 4, but performance with each of the different associative relations improves at different times during development. Functional and part/whole relations develop at an early age, whereas the superordinate relations develop later. Our study clarified the characteristics of the progression of semantic associations during development as well as the roles that associative relations play in the structure and improvement of the semantic store.     

Are intelligence and creativity really so different? : Fluid intelligence, executive processes, and strategy use in divergent thinking

Contemporary creativity research views intelligence and creativity as essentially unrelated abilities, and many studies have found only modest correlations between them. The present research, based on improved approaches to creativity assessment and latent variable modeling, proposes that fluid and executive cognition is in fact central to creative thought. In Study 1, the substantial effect of fluid intelligence (Gf) on creativity was mediated by executive switching, the number of times people switched idea categories during the divergent thinking tasks. In Study 2, half the sample was given an effective strategy for an unusual uses task. The strategy condition interacted with Gf: people high in Gf did better when given the strategy, consistent with their higher ability to maintain access to it and use it despite interference. Taken together, the findings suggest that divergent thinking is more convergent than modern creativity theories presume.     

Adynamic Aphasia: A Transcortical Motor Aphasia with Defective Semantic Strategy Formation

Adynamic aphasia is a form of transcortical motor aphasia characterized by sparse but otherwise normal spontaneous speech that may improve when concepts are introduced by external stimuli. Akinesia, impaired concept formation, inertia of concept generation, a defective semantic network, damage or impaired access to the verbal output lexicon, and defective semantic strategy formation have been proposed to account for this disorder. We studied a patient with adynamic aphasia and frontal lobe systems dysfunction due to bilateral striatocapsular infarctions. The patient was not akinetic but did demonstrate inertia of concept generation that could be overcome with prompting. However, prompting did not improve the number of concepts generated. He demonstrated a generally intact verbal lexicon and semantic network and normal lexical priming. However, his ability to sort closely related items into different classes without prior cuing regarding the nature of the classes was defective. Although his verbal memory was normal, he appeared to use a serial rather than a semantic strategy to recall items. Finally, despite normal lexical priming, he was impaired on a letter fluency task. These results most clearly demonstrate a defect in semantic strategy formation but indicate an additional and possibly related deficit in concept formation and a partial deficit in lexical strategy formation. All of these deficits appear to reflect impairment in the hierarchical organization of knowledge specific to the task at hand. This appears to be a key component of executive functions supported by frontal lobe systems.