Representation at different levels in a conceptual hierarchy

The present study examines the influence of hierarchical level on category representation. Three computational models of representation – an exemplar model, a prototype model and an ideal representation model – were evaluated in their ability to account for the typicality gradient of categories at two hierarchical levels in the conceptual domain of clothes. The domain contains 20 subordinate categories (e.g., trousers, stockings and underwear) and an encompassing superordinate category (CLOTHES). The models were evaluated both in terms of their ability to fit the empirical data and their generalizability through marginal likelihood. The hierarchical level was found to clearly influence the type of representation: For concepts at the subordinate level, exemplar representations were supported. At the superordinate level, however, an ideal representation was overwhelmingly preferred over exemplar and prototype representations. This finding contributes to the increasingly dominant view that the human conceptual apparatus adopts both exemplar representations and more abstract representations, contradicting unitary approaches to categorization.     

Semantic significance: a new measure of feature salience

According to the feature-based model of semantic memory, concepts are described by a set of semantic features that contribute, with different weights, to the meaning of a concept. Interestingly, this theoretical framework has introduced numerous dimensions to describe semantic features. Recently, we proposed a new parameter to measure the importance of a semantic feature for the conceptual representation—that is, semantic significance. Here, with speeded verification tasks, we tested the predictive value of our index and investigated the relative roles of conceptual and featural dimensions on the participants’ performance. The results showed that semantic significance is a good predictor of participants’ verification latencies and suggested that it efficiently captures the salience of a feature for the computation of the meaning of a given concept. Therefore, we suggest that semantic significance can be considered an effective index of the importance of a feature in a given conceptual representation. Moreover, we propose that it may have straightforward implications for feature-based models of semantic memory, as an important additional factor for understanding conceptual representation.     

Generalizing Deontic Action Logic

We introduce a multimodal framework of deontic action logic which encodes the interaction between two fundamental procedures in normative reasoning: conceptual classification and deontic classification. The expressive power of the framework is noteworthy, since it combines insights from agency logic and dynamic logic, allowing for a representation of many kinds of normative conflicts. We provide a semantic characterization for three axiomatic systems of increasing strength, showing how our approach can be modularly extended in order to get different levels of analysis of normative reasoning. Finally, we discuss ways in which the framework can be used to capture other formalisms proposed in the literature, as well as to model searching problems in Artificial Intelligence.

     

Dev E-R: A computational model of early cognitive development as a creative process

This paper describes a computational model named Dev E-R (Developmental Engagement-Reflection) that, inspired by Piaget’s theory, simulates the assimilation-accommodation adaptation process. It is implemented with a new extended version of the computational model of creativity known as Engagement-Reflection. That is, this model simulates adaptation as a creative activity. We introduce here the implementation of our model on an agent that is initialized with basic reflex conducts and that through the interaction with a 3D virtual world, it is able to build new behaviors autonomously. The new acquired skills, according to Piaget’s theory, are typically observed in children that have reached the second substage of the sensorimotor period.     

Informative cues can slow search: The cost of matching a specific template

During visual search, observers hold in mind a search template, which they match against the stimulus. To characterize the content of this template, we trained observers to discriminate a set of artificial objects at an individual level and at a category level. The observers then searched for the objects on backgrounds that camouflaged the features that defined either the object’s identity or the object’s category. Each search stimulus was preceded by the target’s individual name, its category name, or an uninformative cue. The observers’ task was to locate the target, which was always present and always the only figure in the stimulus. The results showed that name cues slowed search when the features associated with the name were camouflaged. Apparently, the observers required a match between their mental representation of the target and the stimulus, even though this was unnecessary for the task. Moreover, this match involved all distinctive features of the target, not just the features necessary for a definitive identification. We conclude that visual search for a specific target involves a verification process that is performed automatically on all of the target’s distinctive features.     

Learning in LIDA

LIDA is a systems-level, biologically-inspired cognitive architecture. More than a decade of research on LIDA has seen much conceptual work on its learning mechanisms, and resulted in a set of conceptual commitments that constrain those mechanisms; perhaps the most essential of these constraints is the Conscious Learning Hypothesis from Global Workspace Theory, which asserts that all significant learning requires consciousness. Despite these successes, many conceptual challenges remain, and bridging the divide between LIDA’s conceptual model and its implementations has been challenging.The contributions of this paper are threefold: We present a detailed survey of learning in LIDA, during which we clarify, elaborate on, and synthesize together ideas from numerous papers, using updated terminology that reflects the continuing evolution of LIDA. We explore foundational issues in learning, such as, “What must be innate or built-in?” versus “What can be learned?”, the nature of LIDA’s representations, and the relationship between the LIDA conceptual model and its computational realizations. Finally, we provide a roadmap for future work. We believe that this paper will direct and catalyze our research endeavors, and provide a thorough introduction to the conceptual foundations of LIDA’s learning mechanisms that will be useful to anyone that would like a deeper understanding of LIDA or for those that plan to implement LIDA-based agents.     

具身概念表征的研究及理论述评

具身认知强调认知在本质上是具身的, 身体在认知的实现中发挥着关键作用。传统的符号加工理论认为, 概念表征独立于主体的知觉运动系统并以抽象符号的形式储存于语言记忆中。概念表征的具身理论则认为, 概念表征与知觉运动系统具有共同的神经基础, 概念在本质上是主体经验客体时知觉与运动体验的神经记录, 而概念加工的基本形式则是身体经验的模拟与还原。关于该理论的实证研究主要集中于概念加工引发的知觉动作变化、身体动作对概念加工的影响、抽象概念加工的具身特征等领域。今后的研究应关注符号加工理论与具身理论的整合等。     

Linking somatic and symbolic representation in semantic memory: the dynamic multilevel reactivation framework

Biological plausibility is an essential constraint for any viable model of semantic memory. Yet, we have only the most rudimentary understanding of how the human brain conducts abstract symbolic transformations that underlie word and object meaning. Neuroscience has evolved a sophisticated arsenal of techniques for elucidating the architecture of conceptual representation. Nevertheless, theoretical convergence remains elusive. Here we describe several contrastive approaches to the organization of semantic knowledge, and in turn we offer our own perspective on two recurring questions in semantic memory research: (1) to what extent are conceptual representations mediated by sensorimotor knowledge (i.e., to what degree is semantic memory embodied)? (2) How might an embodied semantic system represent abstract concepts such as modularity, symbol, or proposition? To address these questions, we review the merits of sensorimotor (i.e., embodied) and amodal (i.e., disembodied) semantic theories and address the neurobiological constraints underlying each. We conclude that the shortcomings of both perspectives in their extreme forms necessitate a hybrid middle ground. We accordingly propose the Dynamic Multilevel Reactivation Framework—an integrative model predicated upon flexible interplay between sensorimotor and amodal symbolic representations mediated by multiple cortical hubs. We discuss applications of the dynamic multilevel reactivation framework to abstract and concrete concept representation and describe how a multidimensional conceptual topography based on emotion, sensation, and magnitude can successfully frame a semantic space containing meanings for both abstract and concrete words. The consideration of ‘abstract conceptual features’ does not diminish the role of logical and/or executive processing in activating, manipulating and using information stored in conceptual representations. Rather, it proposes that the materials upon which these processes operate necessarily combine pure sensorimotor information and higher-order cognitive dimensions involved in symbolic representation.     

Feature centrality: Naming versus imagining

Being white is central to whether we call an animal a "polar bear," but it is fairly peripheral to our concept of what a polar bear is. We propose that a feature is central to category naming in proportion to the feature's category validity--the probability of the feature, given the category. In contrast, a feature is conceptually central in a representation of the object to the extent that the feature is depended on by other features. Further, we propose that naming and conceptual centrality are more likely to disagree for features that hold at more specific levels (such as is white, which holds only for the specific category of polar bear) than for features that hold at intermediate levels of abstraction (such as has claws, which holds for all bears). In support of these hypotheses, we report evidence that increasing the abstractness of category features has a greater effect on judgments of conceptual centrality than on judgments of name centrality and that other category features depend more on intermediate-level category features than on specific ones.     

Knowledge is power: How conceptual knowledge transforms visual cognition

In this review, we synthesize the existing literature demonstrating the dynamic interplay between conceptual knowledge and visual perceptual processing. We consider two theoretical frameworks that demonstrate interactions between processes and brain areas traditionally considered perceptual or conceptual. Specifically, we discuss categorical perception, in which visual objects are represented according to category membership, and highlight studies showing that category knowledge can penetrate early stages of visual analysis. We next discuss the embodied account of conceptual knowledge, which holds that concepts are instantiated in the same neural regions required for specific types of perception and action, and discuss the limitations of this framework. We additionally consider studies showing that gaining abstract semantic knowledge about objects and faces leads to behavioral and electrophysiological changes that are indicative of more efficient stimulus processing. Finally, we consider the role that perceiver goals and motivation may play in shaping the interaction between conceptual and perceptual processing. We hope to demonstrate how pervasive such interactions between motivation, conceptual knowledge, and perceptual processing are in our understanding of the visual environment, and to demonstrate the need for future research aimed at understanding how such interactions arise in the brain.     

Why is short-term sentence recall verbatim? An evaluation of the role of lexical priming

By showing that short-term sentence recall can be significantly affected by words encountered in an intervening distractor task, Potter and Lombardi (1990,Journal of Memory and Language, 29, 633–654) argue that short-term sentence recall is often verbatim because of the availability of recently activated lexical entries during the regeneration of the sentence from its conceptual representation. We show that similar effects can be obtained even when bilinguals perform an intervening task in a different language from that of sentence recall, or when monolinguals perform an intervening task upon pictures. Furthermore, we show that the presentation of a word in P&L’s distractor task does not, in any case, affect subsequent retrieval of a semantically related word as measured in a picture-naming task. We suggest that the effects on recall reported here and by P&L should be explained in terms of conceptual level interference at the time of recall. We also discuss the implications of our suggestion for the issue of the verbatimness of short-term sentence recall.     

Continuity of Theory Structure: A Conceptual Spaces Approach

By understanding laws of nature as geometrical rather than linguistic entities, this paper addresses how to describe theory structures and how to evaluate their continuity. Relying on conceptual spaces as a modelling tool, we focus on the conceptual framework an empirical theory presupposes, thus obtain a geometrical representation of a theory’s structure. We stress the relevance of measurement procedures in separating conceptual from empirical structures. This lets our understanding of scientific laws come closer to scientific practice, and avoids a widely recognised deficit in current philosophy of science accounts, namely to risk a collapse of the physical into the mathematical.     

Linguistic Distributional Knowledge and Sensorimotor Grounding both Contribute to Semantic Category Production

The human conceptual system comprises simulated information of sensorimotor experience and linguistic distributional information of how words are used in language. Moreover, the linguistic shortcut hypothesis predicts that people will use computationally cheaper linguistic distributional information where it is sufficient to inform a task response. In a pre-registered category production study, we asked participants to verbally name members of concrete and abstract categories and tested whether performance could be predicted by a novel measure of sensorimotor similarity (based on an 11-dimensional representation of sensorimotor strength) and linguistic proximity (based on word co-occurrence derived from a large corpus). As predicted, both measures predicted the order and frequency of category production but, critically, linguistic proximity had an effect above and beyond sensorimotor similarity. A follow-up study using typicality ratings as an additional predictor found that typicality was often the strongest predictor of category production variables, but it did not subsume sensorimotor and linguistic effects. Finally, we created a novel, fully grounded computational model of conceptual activation during category production, which best approximated typical human performance when conceptual activation was allowed to spread indirectly between concepts, and when candidate category members came from both sensorimotor and linguistic distributional representations. Critically, model performance was indistinguishable from typical human performance. Results support the linguistic shortcut hypothesis in semantic processing and provide strong evidence that both linguistic and grounded representations are inherent to the functioning of the conceptual system. All materials, data, and code are available at https://osf.io/vaq56/ .     

Preservation of categorical knowledge in Alzheimer's disease: A computational account

Abstract

The distinction between knowledge of specific exemplars and knowledge of their general categories is central to much theorising on the nature of semantic memory. The dissociation between exemplar and category knowledge observed in Alzheimer's Disease (AD) would appear to support this distinction, and to suggest that different neural systems are involved in the representation of exemplar and category knowledge. We review the evidence for preserved category knowledge in the semantic memory impairment of AD, and propose an alternative interpretation, according to which category and exemplar knowledge are both represented in the same distributed neural substrate. The relative preservation of category knowledge is a consequence of the greater frequency, and hence greater robustness, of the representation of attributes shared by all or most members of a category, compared to exemplar-unique attributes. We test and confirm the computational adequacy of this hypothesis in two computer simulations.     

Opening a conceptual space for metamemory experience

The experiences associated with remembering, including metamemory feelings about the act of remembering and attempts at remembering, are not often integrated into general accounts of memory. For example, David Rubin (2022) proposes a unified, three-dimensional conceptual space for mapping memory states, a map that does not systematically specify metamemory feelings. Drawing on Rubin's model, we define a distinct role for metamemory in relation to first-order memory content. We propose a fourth dimension for the model and support the proposal with conceptual, neurocognitive, and clinical lines of reasoning. We use the modified model to illustrate several cases, and show how it helps to conceptualize a new category of memory state: autonoetic knowing, exemplified by déjà vu. We also caution not to assume that memory experience is directly correlated with or caused by memory content, an assumption Tulving (1989) labeled the doctrine of concordance.     

Graphics in psychology: Pictures,data, and especially concepts

Over the last 50 years, psychologists have included numerous graphics in their journals and textbooks. The textbooks have contained mostly pictures, whereas the journals have contained mostly data graphs. Surprisingly, conceptual graphs (such as a network chart of the relations among types of memory) have been very infrequent. There are two main reasons for this absence of conceptual graphs: a lack of awareness of the important role that they have played in the history of our discipline, and a lack of knowledge about how to make them. Four types of conceptual graphs are presented here: mathematical graphs (e.g., Fechner’s law), geometric models (e.g., Henning’s smell prism), networks (e.g., Atkinson and Shiffrin’s STM-LTM model), and EulerVenn graphs (e.g., Shiffrin and Schneider’s model of control and automatic processes). Suggestions are offered forheuristics that can be used to facilitate the perception of conceptual relations in these graphs. It is argued that appropriate conceptual graphs can help researchers and theorists communicate the “big picture” relevant to their ideas, can help educators more effectively communicate with students, and may help clinical psychologists communicate their ideas to clients better.     

The temporal locus of the categorical bias in spatial memories

People often use the region of a target's location to make judgements about its exact position. Although there is general agreement that this category-based effect is due to a blending of the specific coding of a target's location and the categorical coding of its region, there is little agreement on when the blending occurs. The current paper examines whether the systematic bias is the result of combining separate undistorted mental representations at the time of retrieval or a result of using a category-distorted mental representation. Participants studied a dot within a circle and reproduced its location from memory after a short or long delay, and new categorical information was introduced during target retrieval. The results showed a larger increase over time in the biases caused by the new category introduced at retrieval than in the biases caused by the originally encoded category. The findings are consistent with retrieval models, providing evidence that category biasing processes operate at the time of memory retrieval.     

Hybrid analogies in conceptual innovation in science

Analogies are ubiquitous in science, both in theory and experiments. Based on an ethnographic study of a research lab in neural engineering, we focus on a case of conceptual innovation where the cross-breeding of two types of analogies led to a breakthrough. In vivo phenomena were recreated in two analogical forms: one, as an in vitro physical model, and the other, as a computational model of the first physical model. The computational model also embodied constraints drawn from the neuroscience and engineering literature. Cross connections and linkages were then made between these two analogical models, over time, to solve problems. We describe how the development of the intermediary, hybrid computational model led to a conceptual innovation, and subsequent engineering innovations. Using this case study, we highlight some of the peculiar features of such hybrid analogies that are now used widely in the sciences and engineering sciences, and the significant questions they raise for current theories of analogy.     

Parsing surrounding space into regions

Surrounding space is not inherently organized, but we tend to treat it as though it consisted of regions (e.g., front, back, right, and left). The current studies show that these conceptual regions have characteristics that reflect our typical interactions with space. Three experiments examined the relative sizes and resolutions of front, back, left, and right around oneself. Front, argued to be the most important horizontal region, was found to be (a) largest, (b) recalled with the greatest precision, and (c) described with the greatest degree of detail. Our findings suggest that some of the characteristics of the category model proposed by Huttenlocher, Hedges, and Duncan (1991) regarding memory for pictured circular displays may be generalized to space around oneself. More broadly, our results support and extend thespatial framework analysis of representation of surrounding space (Franklin & Tversky, 1990).