Explore the vast range of titles available.

Korean visual abductive reasoning: AI Language Model’s ability to understand plausibility. Linguistic Research 41(2): 283-310. Visual abductive reasoning is the logical process of drawing the most plausible hypothesis based on given observations. This ability is fundamental to artificial intelligence because it enables inference from incomplete information. However, little research has been conducted on Korean visual abductive reasoning. To examine the capability of a multimodal language model’s Korean visual abductive reasoning, we set a simple baseline model and analyzed how it numerically estimated the plausibility for all Korean hypothesis sentences through a multiple-choice task. This task was implemented using a simple dual encoder model and the Korean Story Cloze dataset. After fine-tuning with the binary-choice task discriminating the plausible hypothesis from the implausible one, our baseline model shows an accuracy of 79.81%. In multiple-choice task designed to check for the influence of overfitting or annotation artifacts, the model estimated the plausibilities of four options in the order of Groundtruth≃ Plausible>Implausible≫Random. We also conducted experiments to analyze how the model performed Korean visual abductive reasoning. It was observed that the model made little use of the observation before the hypothesis but demonstrated a similar tendency to humans, struggling with data samples which humans also struggle with when evaluating the plausibility of given sentences. Our study sets a research foundation for numerically analyzing and understanding the language models’ visual abductive reasoning ability in the Korean context. It also shows both the potential and limitations of the language model’s Korean visual abductive reasoning ability and provides clues for future research directions.

In this paper, we explore the problem of explaining observations starting from a classically inconsistent theory by adopting a paraconsistent framework. More precisely, we consider theories formulated in the well-known Belnap–Dunn paraconsistent four-valued logic BD and its implicative expansion BD⊃.  Abductive solutions are then given in one of the two further expansions of BD: BD∘, which introduces formulas of the form ∘φ (‘the information on φ is reliable’), and BD△, which augments the language with formulas of the form △φ (‘there is information that φ is true’). We show that explanations in BD∘ and BD△ are not reducible to one another. We analyse the complexity of standard abductive reasoning tasks (solution recognition, solution existence, and relevance/necessity of hypotheses) depending on the language of the solution (BD∘ or BD△) and on the language of the theory (BD or BD⊃). In addition, we consider the complexity of abductive reasoning in the Horn fragment of BD⊃. By showing how to reduce abduction in BD and its expansions to abduction in classical propositional logic, we enable the reuse of existing abductive reasoning procedures.

In this paper, we want to shift the attention of our scholarly community to the living condition of doubt and its underappreciated significance for the theorizing process. Drawing on Peirce's notion of abduction, we articulate the relationship between doubt and belief in the everyday imaginative work central to theorizing, and establish the role played by doubt as abduction's engine in these efforts. We propose three strategic principles for engaging and using doubt in the research process. In concluding, we explore our field's overemphasis on validation to the exclusion of discovery processes and to the detriment of excellence in theorizing. We call for a broadening of our notions of \"methodology\" to incorporate discovery processes and to begin their explication.

PurposeThe aim of the paper is to address the widening theory-praxis gap in marketing. The authors propose that one viable solution to this challenge is involving practitioners in research processes as active, reflective and empowered participants. Most extant discussions addressing the inclusion of managers as partners in theorizing restrain themselves to an “if” question, arguing whether or not it is possible to create sufficiently rigorous knowledge in collaboration with practitioners. This leaves the “how” question unanswered, i.e. how should such gap-bridging research be conducted in practice.Design/methodology/approachBased on a literature review of collaborative theorizing processes, the authors develop a conceptual framework highlighting the main research design decisions when theorizing with managers. The use of the framework is illustrated with four research program examples.FindingsMost accounts of theorizing with managers use – explicitly or implicitly – abduction as the main mode of inference. In addition to this philosophical commonality, our literature review identified 12 themes that should be considered when designing collaborative research processes. The four illustrative examples indicate that theorizing with managers is an effective way of producing and socializing both academically sound and managerially relevant knowledge. On the other hand, collaborative theorizing processes are time-consuming and studies using abductive reasoning may be more challenging to publish in top-tier journals.Originality/valueThis paper makes two contributions. First, the authors go beyond the extensive academic literature which provides a plethora of explanations and ideas for potential remedies for bridging the theory-praxis gap by offering a detailed description how one particular solution, theorizing with managers, unfolds in practice. Second, the authors ground collaborative theorizing processes in the philosophy of science and put abduction forward as a common nominator for such studies.

The development of digital technology in mathematics education continues to face challenges, particularly in settings with multidimensional resource limitations. This study was conducted in Indonesia, specifically in West Nusa Tenggara Province, Eastern Indonesia. I employed constructivist grounded theory methodology (CGTM) to examine how mathematics teachers manage digital technology in resource-constrained environments. Eight teachers were selected through theoretical sampling. Data were collected through interviews and participatory observations. I analyzed the data through CGTM using ATLAS.ti 24, including initial, focused, and axial coding, supported by memo writing. Abductive logic guided the interpretation. The analysis indicated that infrastructure limitations, internet access, and teacher competence affected teachers' autonomy in using educational technology. Digitally competent teachers demonstrated adaptive strategies and pedagogical independence, enabling them to navigate constraints and support students' computational and metacognitive development. I offer a conceptual framework for digital technology integration in low-resource contexts and provide practical insights for policy design, reflecting teachers' instrumental views on technology (WEBSTER, 2016). This builds on CGTM's adaptability to shifting social realities (CHARMAZ & KELLER, 2016), extending CHARMAZ's legacy through the application of CGTM in complex educational settings.

Aim/Purpose: The rise of modern artificial intelligence (AI), in particular, machine learning (ML), has provided new opportunities and directions for knowledge management (KM). A central question for the future of KM is whether it will be dominated by an automation strategy that replaces knowledge work or whether it will support a knowledge-enablement strategy that enhances knowledge work and uplifts knowledge workers. This paper addresses this question by re-examining and updating a critical argument against KM by the sociologist of science Steve Fuller (2002), who held that KM was extractive and exploitative from its origins. Background: This paper re-examines Fuller’s argument in light of current developments in artificial intelligence and knowledge management technologies. It reviews Fuller’s arguments in its original context wherein expert systems and knowledge engineering were influential paradigms in KM, and it then considers how the arguments put forward are given new life in light of current developments in AI and efforts to incorporate AI in the KM technical stack. The paper shows that conceptions of tacit knowledge play a key role in answering the question of whether an automating or enabling strategy will dominate. It shows that a better understanding of tacit knowledge, as reflected in more recent literature, supports an enabling vision. Methodology: The paper uses a conceptual analysis methodology grounded in epistemology and knowledge studies. It reviews a set of historically important works in the field of knowledge management and identifies and analyzes their core concepts and conceptual structure. Contribution: The paper shows that KM has had a faulty conception of tacit knowledge from its origins and that this conception lends credibility to an extractive vision supportive of replacement automation strategies. The paper then shows that recent scholarship on tacit knowledge and related forms of reasoning, in particular, abduction, provide a more theoretically robust conception of tacit knowledge that supports the centrality of human knowledge and knowledge workers against replacement automation strategies. The paper provides new insights into tacit knowledge and human reasoning vis-à-vis knowledge work. It lays the foundation for KM as a field with an independent, ethically defensible approach to technology-based business strategies that can leverage AI without becoming a merely supporting field for AI. Findings: Fuller’s argument is forceful when updated with examples from current AI technologies such as deep learning (DL) (e.g., image recognition algorithms) and large language models (LLMs) such as ChatGPT. Fuller’s view that KM presupposed a specific epistemology in which knowledge can be extracted into embodied (computerized) but disembedded (decontextualized) information applies to current forms of AI, such as machine learning, as much as it does to expert systems. Fuller’s concept of expertise is narrower than necessary for the context of KM but can be expanded to other forms of knowledge work. His account of the social dynamics of expertise as professionalism can be expanded as well and fits more plausibly in corporate contexts. The concept of tacit knowledge that has dominated the KM literature from its origins is overly simplistic and outdated. As such, it supports an extractive view of KM. More recent scholarship on tacit knowledge shows it is a complex and variegated concept. In particular, current work on tacit knowledge is developing a more theoretically robust and detailed conception of human knowledge that shows its centrality in organizations as a driver of innovation and higher-order thinking. These new understandings of tacit knowledge support a non-extractive, human enabling view of KM in relation to AI. Recommendations for Practitioners: Practitioners can use the findings of the paper to consider ways to implement KM technologies in ways that do not neglect the importance of tacit knowledge in automation projects (which neglect often leads to failure). They should also consider how to enhance and fully leverage tacit knowledge through AI technologies and augment human knowledge. Recommendation for Researchers: Researchers can use these findings as a conceptual framework in research concerning the impact of AI on knowledge work. In particular, the distinction between replacement and enabling technologies, and the analysis of tacit knowledge as a structural concept, can be used to categorize and analyze AI technologies relative to KM research objectives. Impact on Society: The potential of AI on employment in the knowledge economy is a major issue in the ethics of AI literature and is widely recognized in the popular press as one of the pressing societal risks created by AI and specific types such as generative AI. This paper shows that KM, as a field of research and practice, does not need to and should not add to the risks created by automation-replacement strategies. Rather, KM has the conceptual resources to pursue a (human) knowledge enablement approach that can stand as a viable alternative to the automation-replacement vision. Future Research: The findings of the paper suggest a number of research trajectories. They include: Further study of tacit knowledge and its underlying cognitive mechanisms and structures in relation to knowledge work and KM objectives. Research into different types of knowledge work and knowledge processes and the role that tacit and explicit knowledge play. Research into the relation between KM and automation in terms of KM’s history and current technical developments. Research into how AI arguments knowledge works and how KM can provide an enabling framework.

This paper presents a novel model, called T om A bd , that endows autonomous agents with Theory of Mind capabilities. T om A bd agents are able to simulate the perspective of the world that their peers have and reason from their perspective. Furthermore, T om A bd agents can reason from the perspective of others down to an arbitrary level of recursion , using Theory of Mind of n th order. By combining the previous capability with abductive reasoning, T om A bd agents can infer the beliefs that others were relying upon to select their actions, hence putting them in a more informed position when it comes to their own decision-making. We have tested the T om A bd model in the challenging domain of Hanabi, a game characterised by cooperation and imperfect information. Our results show that the abilities granted by the T om A bd model boost the performance of the team along a variety of metrics, including final score, efficiency of communication, and uncertainty reduction.

We present an Information–Theoretic Model of Abduction for Detecting Hallucinations in Generative Models, a neuro-symbolic framework that combines entropy-based inference with abductive reasoning to identify unsupported or contradictory content in large language model outputs. Our approach treats hallucination detection as a dual optimization problem: minimizing the information gain between source-conditioned and response-conditioned belief distributions, while simultaneously selecting the minimal abductive hypothesis capable of explaining discourse-salient claims. By incorporating discourse structure through RST-derived EDU weighting, the model distinguishes legitimate abductive elaborations from claims that cannot be justified under any computationally plausible hypothesis. Experimental evaluation across medical, factual QA, and multi-hop reasoning datasets demonstrates that the proposed method outperforms state-of-the-art neural and symbolic baselines in both accuracy and interpretability. Qualitative analysis further shows that the framework successfully exposes plausible-sounding but abductively unsupported model errors, including real hallucinations generated by GPT-5.1. Together, these results indicate that integrating Information–Theoretic divergence and abductive explanation provides a principled and effective foundation for robust hallucination detection in generative systems.

Purpose Does future thinking enhance managers’ innovative behavior? This study aims to posit that the ability to project events while considering current/future variables and their development (i.e. future thinking) – inextricably linked with the knowledge creation process – may enhance the manager’s accuracy and the number of potentially successful innovative ideas for organizations. Design/methodology/approach The authors use a between-group experiment to examine the innovation choices of 47 subjects with experience in evaluating the market potential of new products when asked to support or otherwise reject real-life innovation-related ideas. The authors test the accuracy of decisions made by participants primed to apply future thinking, practically implemented through abductive reasoning, in their decision-making. Findings The authors found a significant change in managers’ innovative choices, with participants primed for future thinking making significantly more accurate decisions than the control group. Those participants both correctly chose innovation-related ideas with significant future potential and rejected ideas with limited potential that ultimately failed. Originality/value This study explores how future thinking enhances managers’ innovative behavior in organizations. It provides empirical evidence on how future thinking, practiced through abductive reasoning, can work to foster innovative behavior, which is an antecedent of knowledge creation. Organizations that foster future thinking concurrently create knowledge, increasing their competitive advantage in the long run.