Categorization is a dynamic cognitive process that organizes human knowledge. As such, categorization processes are integral of the linguistic system and are primarily manifested in the conceptual categories represented in semantic memory. Linguistic theory has proposed various models to describe the nature and structure of semantic categories. This paper reviews the traditional definitions of four types of linguistic categories (natural or taxonomic, ad hoc, radial and scripts) from the complex network paradigm. The semantic networks have been constructed from a semantic fluency task (Animals, Objects laid on the table, Games and Countryside) involving 680 native Spanish speakers. Both the structure of the network and its dynamics are analysed, remarking the process by which speakers create the network through their linguistic performance. For this purpose, traditional mathematical measures from network theory were employed, and new measures were proposed to more precisely distinguish between the four types of categories. The results support the principles of linguistic description, expanding and refining the properties of the different types of semantic categories. Furthermore, they highlight the importance of fine-grained modularity measures as key to interpreting differences in conceptual categories.

Age of first exposure (AoFE) is an important factor that influences the quality of L2 acquisition. This study aims to investigate the AoFE effect on the contextual learning of L2 novel words at the neural level, as measured by the N400 component from event-related potentials (ERPs). Eighty-eight participants were recruited for the experiment of L2 pseudoword learning, which includes a learning session and a testing session. The participants’ EEG data were recorded from the testing session, and the N400 effect was derived from target words that were either congruous or incongruous with the context. The linear mixed model and multiple regression model revealed a positive AoFE effect on the N400 effect in discourses that were designed for testing retrieval of episodic and semantic memory even after accounting for the variance contributed by several confounding factors. In addition to AoFE, the effects of total L2 exposure, L2 proficiency and personality on the L2 novel word learning performance indicated by the N400 effect were also confirmed in the statistical results.

Bio-inspired design (BID) involves generating innovative ideas for engineering design by drawing inspiration from natural biological phenomena and systems, using a form of design-by-analogy. Despite its many successes, BID approaches encounter research challenges including unstructured data and existing models that hinder comprehension and processing, limited focus on finding biological knowledge compared to defined problems, and insufficient guidance of the ideation process with algorithms. This paper proposes a knowledge-based approach to address the challenges. The approach involves transforming unstructured data into structured knowledge, including information about natural sources, their benefits, and applications. The structured knowledge is then used to construct a semantic network, enabling designers to retrieve information for BID in two ways. Furthermore, a three-step ideation method is developed to encourage divergent thinking and explore additional potential solutions by drawing inspiration and utilizing knowledge. The knowledge-based BID approach is implemented as a tool and design cases are conducted to illustrate the process of applying this tool for BID.

There have been growing uses of semantic networks in the past decade, such as leveraging large-scale pre-trained graph knowledge databases for various natural language processing (NLP) tasks in engineering design research. Therefore, the paper provides a survey of the research that has employed semantic networks in the engineering design research community. The survey reveals that engineering design researchers have primarily relied on WordNet, ConceptNet, and other common-sense semantic network databases trained on non-engineering data sources to develop methods or tools for engineering design. Meanwhile, there are emerging efforts to mine large scale technical publication and patent databases to construct engineering-contextualized semantic network databases, e.g., B-Link and TechNet, to support NLP in engineering design. On this basis, we recommend future research directions for the construction and applications of engineering-related semantic networks in engineering design research and practice.

Starting from the hypothesis that alcoholics have a specific semantic network which causes a perceptual-processing bias, we tested 30 male inpatients and 20 healthy male controls. Our modified card version of the Stroop color-naming task consisted of a neutral and critical word condition. The results revealed that alcoholic patients showed a small information processing bias under the critical experimental condition (alcohol-related words); although this was only a trend in the expected direction and statistically not significant. However, neuropsychological impairment of the patients was demonstrated with the “standard Stroop procedure”. The most significant deficits were found in the interference task, which requires cognitive flexibility.

Using a semantic priming experiment, the influence of lexical access and knowledge of semantic relations on reading comprehension was studied in Dutch monolingual and bilingual minority children. Both context-independent semantic relations in the form of category coordinates and context-dependent semantic relations involving concepts that co-occur in certain contexts were tested in an auditory animacy decision task, along with lexical access. Reading comprehension and the control variables vocabulary size, decoding skill, and mental processing speed were tested by means of standardized tasks. Mixed-effects modeling was used to obtain individual priming scores and to study the effect of individual differences in the various predictor variables on the reading scores. Semantic priming was observed for the coordinate pairs but not the context-dependently related pairs, and neither context-independent priming nor lexical access predicted reading comprehension. Only vocabulary size significantly contributed to the reading scores, emphasizing the importance of the number of words known for reading comprehension. Finally, the results show that the monolingual and bilingual children perform similarly on all measures, suggesting that in the current Dutch context, language status may not be highly predictive of vocabulary knowledge and reading comprehension skill.

Many attempts have been made to discover some systematicity in the semantics of phrasal verbs. However, most research has investigated the semantics of particles exclusively; no study has examined how the multiple meanings of the verb also contribute to the meanings of phrasal verbs. The current corpus-based (COCA) study advances the research on phrasal verbs by examining the interaction of the polysemy networks of both the verb and the particle in four phrasal verb constructions: get up, take up, get out, and take out. Following the Cognitive Linguistics (CL) based methodology set out by Tyler and Evans (2003) for analyzing the semantics of particles, in conjunction with Langacker’s (1991) analysis of the semantics of verbs, a replicable polysemy analysis of the semantics of get and take was established. The polysemy networks for both the verbs and the particles laid the foundation for investigating the multiple meanings of the phrasal verbs found in the corpus. The CL-based analysis of the semantics of the phrasal verb constructions provides evidence for the compositional nature of phrasal verbs, showing that the multiple meanings can be systematically accounted for through the interaction of the polysemy networks of the component verbs and particles.

Decreased productivity on verbal fluency tasks by persons with schizophrenia has been attributed to semantic system abnormalities. Semantic structure is often assessed using multidimensional scaling (MDS) to detect normal and aberrant semantic clustering. However, MDS has limitations that may be particularly problematic for such assessments. Here, we introduce a different clustering technique, singular value decomposition (SVD), to elucidate abnormalities of the semantic system in schizophrenia. We compared 102 treated outpatients with schizophrenia to 109 healthy adults on two category-cued word fluency tasks. Patients with schizophrenia showed semantic clustering patterns that differ markedly from those of healthy adults. However, SVD revealed more detailed and critical semantic system abnormalities than previously appreciated using MDS. Patients with schizophrenia showed less coherent semantic clustering of both low- and high-frequency category exemplars than healthy adults. These results suggest the intriguing possibility that impaired automatic activation of semantic information is a key deficit in schizophrenia. (JINS, 2012, 18, 565–575)

The present study examined the relationship between rate of cognitive decline in patients with Alzheimer's disease (AD) and the integrity of the network of associations that comprise their semantic knowledge. The integrity of the semantic network of 12 AD patients was determined by comparing their networks to a standard normal control network derived with Pathfinder analysis, a multidimensional graphic analysis technique. A simple linear regression analysis, comparing the degree of semantic network deterioration with rate of cognitive decline as measured by the difference between the Dementia Rating Scale (DRS) scores obtained at the time of the testing of semantic knowledge (Year 1) and one year later (Year 2), was highly significant (r2 = .84; p < .001). These results suggest that a sensitive measure of the structural deterioration of semantic knowledge may be useful for predicting the rate of progression of cognitive changes in patients with AD. (JINS, 1995, I, 297–303.)