A connectivity model for semantic processing

Summary A connectivity model for semantic processing, based on two major assumptions, is proposed and tested. The representation assumption states that semantically related features are represented by an interconnected structure of features. A network is termed interconnected if each node is linked to at least two other nodes. According to the processing assumption, indirect activation becomes increasingly effective as the number of interconnected nodes increases. The connectivity model thus predicts that more complex concepts will be processed faster than less complex concepts. Experiment 1 was performed to determine the number of features for a set of 72 words and to show that subordinate concepts share common features with their natural superordinate concept. In Experiments 2 and 3 subjects had to judge whether or not a word belonged to a previously defined natural superordinate concept. The results confirm the connectivity model and show that concepts with many features can be classified faster than concepts with only a few features.This research was supported by a Grant of the Max Kade Foundation     

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。     

The neural consequences of semantic richness: when more comes to mind, less activation is observed

Some concepts have richer semantic representations than others. That is, when considering the meaning of concepts, subjects generate more information (more features, more associates) for some concepts than for others. This variability in semantic richness influences responses in speeded tasks that involve semantic processing, such as lexical decision and semantic categorization tasks. It has been suggested that concepts with richer semantic representations build stronger attractors in semantic space, allowing faster settling of activation patterns and thus faster responding. Using event-related functional magnetic resonance imaging, we examined the neural activation associated with semantic richness by contrasting activation for words with high and low numbers of associates in a semantic categorization task. Results were consistent with faster semantic settling for words with richer representations: Words with a low number of semantic associates produced more activation than words with a high number of semantic associates in a number of regions, including left inferior frontal and inferior temporal gyri.     

Orthographic and phonological activation in recognizing Chinese characters

This paper is a contribution to the study of whether visual contact with lexical units in any writing system necessarily arouses their corresponding phonological information. In two experiments it was investigated whether phonological information is automatically activated during the semantic processing of Chinese characters. In these experiments, both using a semantic-categorization task, subjects produced the same proportion of false positive categorization errors and showed the same decision latencies on homophone foils and their non-homophonic controls, thus indicating that phonological information does not seem to affect the semantic task. Experiment 2 further revealed that subjects made more errors and produced longer response times on graphemically similar foils than on the corresponding controls. The absence of phonological effects and the presence of clear effects of visual similarity for Chinese characters in semantic tasks can be taken to indicate that phonological information may not be automatically activated during the processing of meanings of Chinese characters. The present results also cast serious doubts on the hypothesis that phonological activation is a universal principle of lexical processing.     

Understanding misunderstanding: a study of sex differences in meaning attribution

There are biologically based sex differences in verbal abilities and in neuropsychological systems of verbal processing. Measurement of observable behaviour, however, does not say much about sex differences in the internal, semantic processing of verbal material. The present study, which was conducted in Canada, China and Russia, investigated sex differences in connotative meaning attribution to the most common concepts using an object scale symmetry in the choice of the nouns and bipolar adjectives (projective semantic method). The results showed that males had a tendency to estimate reality- and work-related concepts more negatively and social- and physical attractors more positively than women. The paper hypothesizes that at the level of the most fundamental semantic processing men favour more exceptional objects than women, and women favour more predictable objects, including rules and routines.     

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.     

Lexical but not semantic priming in Alzheimer's disease.

The hypothesis that patients with Alzheimer's disease (AD) have a disturbance in semantic processing was tested using a new lexical-priming task, threshold oral reading. Healthy elderly controls showed significant effects of priming for word pairs that are associatively related (words that reliably co-occur in word association tests) and for word pairs that are semantically related (high-frequency exemplars that belong to the same superordinate category but are not high-frequency associates). AD patients showed effects of priming for associatively related words but not for word pairs that are related only by shared semantic features. These results are consistent with the hypothesis that semantic processing is impaired in AD and suggest that independent networks of relationships among words and among concepts in semantic memory may be differentially disrupted with various forms of brain damage.     

Performance in picture naming and word comprehension: evidence for common neuronal substrates from intraoperative language mapping

Correlations of naming ability and performance in the Token Test are known from aphasia; however, the mechanisms underlying these correlations are unclear. Naming tasks are commonly used in intraoperative mapping for identification of cortical areas involved in language processing. In the present study, we measured performance in an elementary Token Test task, i.e., single word comprehension, during electrical stimulation of cortical sites at which this stimulation previously had disturbed the naming process. It was found that at about half of the sites at which naming could be disturbed electrical stimulation also led to disturbances in Token Test performance, indicating that there are common neural structures critical for performance in both tasks. These findings are discussed in terms of a multilayered semantic network in which the level of the simultaneous binding of features into concepts and the level of units that are organized with respect to semantic relations may be disturbed separately.     

How meaning shapes seeing

Inattentional blindness refers to the failure to see an unexpected object that one may be looking at directly when one's attention is elsewhere. We studied whether a stimulus whose meaning is relevant to the attentional goals of the observer will capture attention and escape inattentional blindness. The results showed that an unexpected stimulus belonging to the attended semantic category but not sharing physical features with the attended stimuli was detected more often than a semantically unrelated stimulus. This effect was larger when the unexpected stimuli were words than when they were pictures. The results imply that the semantic relation between the observer's attentional set and the unexpected stimulus plays a crucial role in inattentional blindness: An unexpected stimulus semantically related to the observer's current interests is likely to be seen, whereas unrelated unexpected stimuli are unseen. Attentional selection may thus be driven by purely semantic features: Meaning may determine whether or not one sees a stimulus.     

When senses meet functions: an amodal stage in conceptual processing

Research has shown that concepts are associated with modality-specific features widely distributed in the brain and that semantic hubs or convergence zones may exist to mediate or integrate distributed representations. Although much neuropsychological and neuroimaging literature has provided support for the existence of hubs/convergence zones from the spatial perspective, opinions differ as to whether the representations in the hubs/convergence zones are modality-specific or amodal. The current ERP study aimed to explore whether there is an amodal stage in conceptual processing from the temporal perspective. Participants were presented with prime-target pairs and were asked to judge whether the prime (a sensory or functional feature) could be an attribute of the target (a noun in a natural or artifactual category). The results revealed that a semantic hub or convergence zone may be reached no later than 350–450?ms in conceptual processing and that the representations in the hub may be amodal.     

Semantic processing of living and nonliving concepts across the cerebral hemispheres

Studies of patients with category-specific semantic deficits suggest that the right and left cerebral hemispheres may be differently involved in the processing of living and nonliving domains concepts. In this study, we investigate whether there are hemisphere differences in the semantic processing of these domains in healthy volunteers. Based on the neuropsychological findings, we predicted a disadvantage for nonliving compared to living concepts in the right hemisphere. Our prediction was supported, in that semantic decisions to nonliving concepts were significantly slower and more error-prone when presented to the right hemisphere. In contrast there were no hemisphere differences for living concepts. These findings are consistent with either differential representation or processing of concepts across right and left hemispheres. However, we also found a disadvantage for nonliving things compared to living things in the left hemisphere, which is not consistent with a simple representation account. We discuss these findings in terms of qualitatively different semantic processing in right and left hemispheres within the framework of a distributed model of conceptual representation.     

Red-hot: How colour and semantic temperature processing interact in a Stroop-like paradigm

Temperature concepts and colour are commonly associated (i.e., red is “hot” and blue is “cold”), although their direction of influence (unidirectional, bidirectional) is unknown. Semantic Stroop effects, whereby words like fire influence colour categorization, suggest automatic semantic processing influences colour processing. The experiential framework of language comprehension indicates abstract concepts like temperature words simulate concrete experiences in their representation, where expressions like “red-hot” suggest colour processing influences conceptual processing. Participants categorized both colour (Experiment 1: red, blue; Experiment 2: red, green, blue) and word-meaning with matched lists of hot and cold meaning words in each colour. In Experiments 1 and 2, semantic categorization showed congruency effects across hot and cold words, while colour categorization showed facilitation only with hot words in Experiment 2. This asymmetry reflects a more consistent influence of colour categorization on semantic categorization than the reverse, suggesting experiential grounding effects may be more robust than the effects of semantic processing on colour processing.     

汉语动宾结构惯用语加工的基本单元：来自词切分的证据

采用词切分范式研究汉语动宾结构惯用语的理解机制, 探讨在惯用语理解中加工的基本单元, 揭示汉语动宾结构惯用语的表征方式。研究发现, 词切分范式对语义分解性不同的惯用语有不同的影响。在空格切分的形式下, 语素切分和非语素切分抑制了对语义高分解的动宾结构惯用语的通达, 却未影响对语义低分解的动宾结构惯用语的理解; 语素切分、非语素切分和空格切分均增加了对语义高分解的惯用语的错误率。在阴影切分的形式下, 非语素切分抑制了对语义高分解的动宾结构惯用语的加工, 同样未影响对语义低分解的惯用语的理解。整个研究表明, 在汉语动宾结构惯用语的表征中, 既存在着语素单元, 也存在着整语单元, 但整语是汉语动宾结构惯用语加工的优势单元。     

语义表征具身理论:情绪在概念表征中的作用

概念如何被表征是认知科学的热点问题,其中抽象概念如何表征是当前具身认知最具争议性的话题之一。与前期概念隐喻理论(强调意象图式)和知觉符号理论(强调情境内省信息)的观点不同,语义表征具身理论强调情绪经验信息在抽象概念表征和加工中的作用。具体而言,具体概念的表征主要来自感觉运动信息,而抽象概念的表征主要来自情绪经验信息和语言信息。研究证明,抽象概念的高情绪经验信息能够促进词汇的加工,且这一促进作用受词汇情绪效价的调节。未来研究应进一步考虑影响情绪经验信息发挥作用的因素,比如语言信息的丰富性、情绪唤醒或个体的情绪状态等。     

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.     

The role of variability in the property listing task

It is generally believed that concepts can be characterized by their properties (or features). When investigating concepts encoded in language, researchers often ask subjects to produce lists of properties that describe them (i.e., the Property Listing Task, PLT). These lists are accumulated to produce Conceptual Property Norms (CPNs). CPNs contain frequency distributions of properties for individual concepts. It is widely believed that these distributions represent the underlying semantic structure of those concepts. Here, instead of focusing on the underlying semantic structure, we aim at characterizing the PLT. An often disregarded aspect of the PLT is that individuals show intersubject variability (i.e., they produce only partially overlapping lists). In our study we use a mathematical analysis of this intersubject variability to guide our inquiry. To this end, we resort to a set of publicly available norms that contain information about the specific properties that were informed at the individual subject level. Our results suggest that when an individual is performing the PLT, he or she generates a list of properties that is a mixture of general and distinctive properties, such that there is a non-linear tendency to produce more general than distinctive properties. Furthermore, the low generality properties are precisely those that tend not to be repeated across lists, accounting in this manner for part of the intersubject variability. In consequence, any manipulation that may affect the mixture of general and distinctive properties in lists is bound to change intersubject variability. We discuss why these results are important for researchers using the PLT.     

On the learnability of frequent and infrequent word orders: An artificial language learning study

Most languages have a basic or “canonical” word order, which determines the relative positions of the subject (S), the verb (V), and the object (O) in a typical declarative sentence. The frequency of occurrence of the six possible word orders among world languages is not distributed uniformly. While SVO and SOV represent around 85% of world languages, orders like VSO (9%) or OSV (0.5%) are much less frequent or extremely rare. One possible explanation for this asymmetry is that biological and cognitive constraints for structured sequence processing make some word orders easier to be processed than others. Therefore, the high frequency of these word orders would be related to their higher learnability. The aim of the present study was to compare the learnability of different word orders between groups of adult subjects. Four artificial languages with different word orders were trained: two frequent (SVO, SOV) and two infrequent (VSO, OSV). In a test stage, subjects were asked to discriminate between new correct sentences and syntax or semantic violations. Higher performance rates and faster responses were observed for more frequent word orders. The results support the hypothesis that more frequent word orders are more easily learned.     

语言和情境对具体概念感知运动仿真的影响

具体概念加工中的感知运动仿真是概念表征的核心加工过程。本研究系统探讨了语言因素(语言类型：第一语言和第二语言)和情境因素(知觉情境：空间知觉情境和语义知觉情境)对具体概念加工中感知运动仿真的影响。实验1采用语义相关判断范式探讨了第二语言加工过程中是否存在感知运动仿真, 以及第二语言中的仿真和第一语言中的仿真是否存在差异。研究发现, 第二语言加工中依然存在感知运动仿真, 但是第一语言的感知运动仿真具有一定的优势。实验2采用语义相关判断范式及其变式, 通过2个分实验分别探讨了个体加工概念时空间知觉信息和语义知觉信息对于概念表征过程中感知运动仿真的影响。研究发现, 无论是在较弱的空间信息知觉条件下还是在较浅的语义信息知觉条件下, 具体概念加工过程中都产生了感知运动仿真。本研究的发现弥补了知觉符号理论未对第二语言中的感知运动仿真提出针对性预测的不足, 表明感知运动仿真具有一定的跨语言稳定性。同时, 具体概念加工中的感知运动仿真不受空间信息和语义信息的调节, 这表明感知运动仿真能在一定程度上自动化产生。     

On the learnability of frequent and infrequent word orders: An artificial language learning study

Most languages have a basic or "canonical" word order, which determines the relative positions of the subject (S), the verb (V), and the object (O) in a typical declarative sentence. The frequency of occurrence of the six possible word orders among world languages is not distributed uniformly. While SVO and SOV represent around 85% of world languages, orders like VSO (9%) or OSV (0.5%) are much less frequent or extremely rare. One possible explanation for this asymmetry is that biological and cognitive constraints for structured sequence processing make some word orders easier to be processed than others. Therefore, the high frequency of these word orders would be related to their higher learnability. The aim of the present study was to compare the learnability of different word orders between groups of adult subjects. Four artificial languages with different word orders were trained: two frequent (SVO, SOV) and two infrequent (VSO, OSV). In a test stage, subjects were asked to discriminate between new correct sentences and syntax or semantic violations. Higher performance rates and faster responses were observed for more frequent word orders. The results support the hypothesis that more frequent word orders are more easily learned.