Abstract
<jats:p>The article presents an approach to defining the category of scientific knowledge that integrates the principles of scientific realism and constructivism. The authors propose a coherent model designed to overcome the limitations of the traditional definition of knowledge as “justified true belief” (JTB) and to account for the complexity of actual scientific inquiry, a process in which it is not always possible to evaluate current concepts and theories strictly in terms of truth. A central tenet of the paper is that knowledge should be understood as a reflection of reality in an ideal (mental) form. Within this framework, knowledge is categorized into two types: truth (which corresponds entirely to an object's established properties) and error (which corresponds only partially). The paper demonstrates that even refuted scientific concepts, despite being formally false, frequently contain objective elements that drive scientific progress. The authors analyze the correspondence between knowledge and the object of cognition, offering a focused critique of radical constructivism. They argue that the constructive nature of cognition does not contradict its reflective essence, but rather complements it; constructivist principles prove most productive when grounded in scientific realism. At both the sensory-empirical and rational-theoretical levels, this constructivity involves creating mental models that closely approximate reality. Furthermore, the study outlines a framework of criteria for evaluating scientific cognition. This includes standards for verifying truth, establishing scientific validity, ensuring disciplinary autonomy, and fostering a rigorous research culture. The article also explores the ontological status of knowledge. By critically examining Karl Popper’s “World 3” framework, the authors argue that knowledge exists exclusively as a mental phenomenon. In this view, texts and other semiotic systems act merely as transmission codes rather than direct containers of knowledge. Ultimately, the proposed coherent model establishes two boundaries: an external boundary demarcating science from non-science, and an internal boundary distinguishing verified true knowledge from knowledge yet to be validated. This approach maintains a firm commitment to the correspondence theory of truth (reflection theory) while successfully accommodating the constructivist dimensions of human cognition.</jats:p>