Topics in Cognitive Science最新文献

Semantic feature norms have been foundational in the study of human conceptual knowledge, yet traditional methods face trade-offs between concept/feature coverage and verifiability of quality due to the labor-intensive nature of norming studies. Here, we introduce a novel approach that augments a dataset of human-generated feature norms with responses from large language models (LLMs) while verifying the quality of norms against reliable human judgments. We find that our AI-enhanced feature norm dataset, NOVA: Norms Optimized Via AI, shows much higher feature density and overlap among concepts while outperforming a comparable human-only norm dataset and word-embedding models in predicting people's semantic similarity judgments. Taken together, we demonstrate that human conceptual knowledge is richer than captured in previous norm datasets and show that, with proper validation, LLMs can serve as powerful tools for cognitive science research.

Engraved ochres and ostrich eggshells from the South African Blombos Cave and Diepkloof Rock Shelter are among the earliest expressions of human symbolic behavior. They appear to document a continuous practice of mark-making across ∼40,000 years. During this time, the engraved markings change from simpler, unstructured patterns to more complex markings such as cross-hatchings. Previous work examining the cognitive implications of these changes concluded that the engravings were likely used as decorations and may have served as group identity markers, but not as denotational symbols. To inform discussions of the emergence of symbolic behavior, we conducted a two-part experimental study inspired by these engravings and based on the assumption that artifact use will motivate incremental adaptive refinements. Part 1 combined a delayed reproduction task with a transmission chain design to simulate an enduring mark-making practice. Eleven transmission chains were seeded with four drawings derived from the early Blombos and Diepkloof engravings and reproduced over eight generations. Transmission chain drawings showed a tendency to become increasingly regular, organized, and symmetric. Part 2 subjected a sample of the transmission chain drawings to a suite of psychophysical experiments to assess the cognitive implications of the accumulated structural changes. We found that the drawings became easier to discriminate, looked more like they had been intentionally made, and became easier to remember and reproduce, but there was no evidence of a systematic change in saliency or stylistic properties. Finally, we compared the results from the transmission chains with a similar analysis of the drawings derived from the original engravings. Although we observe interesting qualitative similarities between the original engravings and the experimental drawings, our findings suggest that cognitive biases and working memory constraints are not sufficient to generate the patterns observed in the archaeological record, highlighting the significance of social and functional contexts in shaping early symbolic artifacts. By integrating archaeological and experimental research, we can better inform inferences on sparse records of early symbolic behavior. Our study thus leads to a broader consideration of the role, strengths, and potential limitations of the transmission chain approach in analyzing trajectories of early symbolic behavior.

When and how do children come to understand various kinds of visual media (e.g., pictures, videos, scale models), and how does early experience contribute to variation in the development of visual media comprehension across global contexts? In this selective review, we show that while researchers have investigated how children from Western convenience samples understand visual media, less is known about how this comprehension varies across children in global contexts. Indeed, prior work investigating picture comprehension suggests that children in different contexts may understand pictures at different developmental time points, potentially due to variation in their early picture experiences. These findings demonstrate the need for more research investigating children's comprehension of additional kinds of visual media across contexts. The experience-dependence of visual media comprehension could provide important insight into these abilities' origins, as well as the appropriateness of cross-cultural use of visual media in early childhood measurement.

Cognitive technologies are socially shared and culturally evolved systems whose function is principally cognitive. Throughout human history and prehistory, they have aided in classifying, organizing, or managing information and knowledge, including ideas, language, and material culture. They range in scope from the highly artifactual (e.g., maps, scientific instruments, weights and measures) to the more abstract and conceptual (e.g., taxonomies, linguistic frameworks). Cognitive technologies thus scaffold many of the complex activities common to all human societies. Because they are both dynamic and culturally embedded, cognitive technologies, therefore, have histories, and are thus amenable not only to contemporary experimental methods, but also a range of historical and evolutionary approaches, including those from outside disciplines traditionally considered parts of cognitive science, such as classics and other humanistic disciplines. While the study of cognitive technologies is hardly new, many pre-existing studies can now be brought together under this framework in recognition that the field has been insufficiently integrated. This issue brings together a disciplinarily diverse range of scholars whose work employs the methods and concepts of specific disciplines while orienting itself around contemporary cognitive-scientific frameworks. The value of this integrative approach is to form a nexus around which a broader range of future interdisciplinary cognitive scholarship can coalesce, in which humanists and scientists have much to learn from one another through collaboration and shared concepts.

Post-cognitivist approaches recognize cognition as a phenomenon that involves not just brains but all the sensorimotor apparatus of organisms. This means that brains are not always required for the emergence of cognition and that every organism can, in principle, be cognitive, unlocking a theoretical framework to explain the complex adaptive behavior of even non-neural organisms. This construct blurs the boundaries of cognitive processes, leading to what is known as the Extended Cognition thesis, where objects in the environment could become integral parts of an individual's cognitive system. Here, we explore the possibility that some species of entomopathogenic fungi in the families Cordycipitaceae, Ophiocordycipitaceae, and Clavicipitaceae could extend their cognition to the insects and other arthropods they infect. In this parasitism, the fungus takes possession of the sensorimotor apparatus of its host and coerces it to find the best places for fungal dispersal. We examine case studies where Ophiocordyceps spp. fungi induce ants to seek bright places in the forest. In this case, the fungus may be using the ant's sensorial and motoric apparatus to locate the places appropriate for its reproduction. This could be a remarkable example of extended cognition of a non-neural organism through a neural organism. However, empirical testing using a solid demarcating criterion is required to confirm this hypothesis.

The appearance of colorful earth pigments-primarily red ochre-in the archaeological record dating back to at least 300,000 years ago has long been held by researchers to signal the emergence of symbolic behavior in humanity. Specifically, the deliberate selection and use of this material is connected to its bright color and its ability to transfer that color onto human skin, rock surfaces, and many other materials. Despite this recognition that colorants-and color selection-represent some of the earliest evidence for emerging symbolic capacities in our lineage, clear description and discussion of the visual qualities of "symbolic material culture" recovered from Pleistocene-aged sites is surprisingly minimal. In this paper, we present results of a quantitative review of the archaeological literature surrounding early symbolic behavior in relation to how color and other visual aspects are reported. As well as highlighting trends and gaps in research to date, we outline how a more nuanced treatment of the visual qualities of raw materials and/or artifacts created by hominins may provide new insights into the origin and development of symbolism in our genus.

Humans can make moral inferences from multiple sources of input. In contrast, computational moral inference in artificial intelligence typically relies on language models with textual input. However, morality is conveyed through modalities beyond language. We present a computational framework that supports moral inference from natural images, demonstrated in two related tasks: (1) inferring human moral judgment toward visual images and (2) analyzing patterns in moral content communicated via images from public news. We find that models based on text alone cannot capture the fine-grained human moral judgment toward visual stimuli, but language-vision fusion models offer better precision in visual moral inference. Furthermore, applications of our framework to news data reveal implicit biases in news categories and geopolitical discussions. Our work creates avenues for automating visual moral inference and discovering patterns of visual moral communication in public media.

It is well known that context-dependent decisions incur mental costs. While previous research has sought to formalize these costs at various levels of analysis, we still lack basic insight into the nature of mental costs, including the underlying cognitive resources being consumed. Moreover, many computational models assume that mental costs scale linearly with the cognitive resource being used, an assumption of convenience that has yet to be systematically tested. To address these gaps, we build on rate-distortion theory by formalizing an information-theoretic notion of mental costs. Specifically, we define the cost of policies-the mappings from states to actions-as a function of the mutual information between states and actions, the policy complexity. Across four decision-making experiments featuring diverse task manipulations, we find that this mental cost formulation offers a parsimonious description of how humans adaptively adjust their policy complexity across different tasks. Notably, a quadratic mental cost formulation, where increases in policy complexity incur supralinear costs, provides the best fit. These findings highlight the meta-cognitive ability of humans to account for mental costs when forming decision strategies, and pave the way toward a domain-general quantification of mental effort.

Predictive Processing has been proposed as the single unifying computation underlying all of cognition, and proponents argue that all psychological phenomena can be explained as consequences of this principle. This theoretical framework has inspired many cognitive scientists and neuroscientists, but it currently has no developmental mechanism that would explain how infants begin to perceive and learn about the world. Rather, it treats human cognition as if it exists in a fully developed adult with a history of observations and world knowledge. In its current formulation, Predictive Processing only allows for perception of incoming stimuli given the existence of expectations based on previous experiences and as such does not allow for an infant to ever make a first observation. In this paper, we propose a possible starting point from which the infant can begin to develop predictive models, as well as a toolkit necessary to allow the infant to perform the range of cognitive operations on predictive models necessary for learning. The starting point we propose is a set of low-precision, low level-of-detail predictions with little or no hierarchical structure, which is very rapidly updated to reflect the infant's early environment. The toolkit contains a range of operations referred to collectively as structure learning, which are applied to models in order to allow for building adult-like hierarchical models. These modifications are necessary for developmental scientists to be able to adopt the Predictive Processing framework and benefit from its advantages, but also for Predictive Processing to be able to explain all human cognition, which inherently must include development.