Таблица 8
Kenneth Craik’s The Nature of Explanation (1943) argues that minds—biological and artificial—explain and predict by constructing internal, small-scale models of external reality. Craik proposes that explanations are model-based, that intelligence consists of manipulating these models to simulate outcomes, and that scientific progress is the refinement of such models. The book blends philosophy, psychology, and early cybernetic thinking; its core claim foreshadows later model-based and representational approaches in cognitive science, AI, and philosophy of science.
Kenneth Craik (1914–1945) was a Scottish polymath whose brief but influential work bridged psychology, philosophy, engineering, and early cybernetics. His book The Nature of Explanation (1943) presents a concise theory of mind and science centered on internal models: organisms and systems explain, predict, and control the world by constructing and testing simplified internal representations. Craik’s ideas anticipated later developments in cognitive science, control theory, and computational models of mind.
Before Craik, the question of "explanation" was largely philosophical or behaviorist. How does a human understand a falling apple? How does a soldier anticipate a bullet’s trajectory? The standard answer involved stimulus and response.
Craik rejected this. He argued that explanation is not just a linguistic act or a conditioned reflex; it is the internal process of modeling reality. He proposed that thought parallels external events. In his own iconic words: kenneth craik the nature of explanation pdf
"If the organism carries a 'small-scale model' of external reality and of its own possible actions within its head, it can try out various alternatives, conclude which is the best, and react before the external event has occurred."
This single passage is the reason scholars hunt for the "Kenneth Craik The Nature of Explanation PDF." It is the first explicit articulation of mental models, simulation, and prediction—the holy trinity of modern AI.
Without Craik, there is no Herbert Simon, no Allen Newell, and arguably no modern cognitive science. But his most direct heir was Philip Johnson-Laird, who expanded the "mental model" theory in the 1980s. Kenneth Craik’s The Nature of Explanation (1943) argues
More profoundly, Craik predicted Deep Learning and Generative AI. When ChatGPT generates a response, what is it doing? It is running a statistical "small-scale model" of human language. When AlphaGo defeats a grandmaster, it isn't just reacting; it simulates thousands of future moves internally before the opponent moves a single piece. That is pure Craik.
As the philosopher Daniel Dennett noted: "Craik saw that to be a predictor, you didn't need a perfect copy of the universe; you just needed a working model—a cheap surrogate that gets the job done."
In the annals of cognitive science, certain works appear so prescient that they seem to have been written decades ahead of their time. Kenneth Craik’s "The Nature of Explanation" (1943) is precisely such a text. Written during the turmoil of World War II by a brilliant Scottish psychologist and philosopher, this slim volume laid the cornerstone for what would later become cognitive psychology, artificial intelligence, and modern philosophy of mind. "If the organism carries a 'small-scale model' of
For decades, researchers, students, and AI enthusiasts have searched for the elusive "kenneth craik the nature of explanation pdf" — a digital key to one of the 20th century’s most foundational theoretical works. This article serves three purposes: first, to explain why Craik’s book remains essential reading; second, to summarize its revolutionary thesis on mental models; and third, to provide a legitimate roadmap for locating and understanding the PDF version of this classic text.
A large section of The Nature of Explanation is devoted to the nature of analogy. Craik points out that many scientific breakthroughs come from noticing structural similarities between different domains. For example, the flow of heat and the flow of electricity are analogous; explaining one via the other is powerful because you can literally build a physical model (e.g., a resistor-capacitor network) that mimics heat diffusion.
But Craik warns: analogies are not identities. A good explanation requires specifying the domain of isomorphism—the set of relations that hold true between the model and the world. This is precisely what modern computational models do: they capture certain relational structures while ignoring irrelevant details.