Michael Polanyi
1891–1976
“We know more than we can tell.” Polanyi’s work on tacit knowledge explains why some skills resist documentation — and why apprenticeship remains powerful even in an age of abundant information.
The Tacit Dimension
The Tacit Dimension (1966) argued that all knowledge has a tacit component — a layer of understanding that can’t be fully articulated. You can describe how to ride a bicycle, but the description doesn’t enable riding. The knowledge lives in the body, in practice, in the relationship between knower and known.
This isn’t a deficiency in our descriptions. It’s a fundamental feature of knowledge. Some things can only be learned by doing, in the presence of someone who already knows.
Implications for AI
AI systems are explicit-knowledge machines. They operate on what can be written down, formalized, encoded. Polanyi’s insight suggests there’s a permanent boundary: some knowledge will always require human practice, human bodies, human attention.
An AI can describe how to knead bread. It can’t teach the feel of properly hydrated dough. A manual can document greenhouse management. It can’t teach the smell of soil that’s ready for planting.
This isn’t a limitation to be overcome. It’s a design parameter: know what AI can do (explicit knowledge) and what it can’t (tacit knowledge), and build systems accordingly.
Key Works
- Personal Knowledge (1958)
- The Tacit Dimension (1966)
Connections
- embodied-knowledge — Polanyi’s central concept
- Matthew Crawford — shop class as tacit knowledge in practice
- John Dewey — knowledge through experience
- Masanobu Fukuoka — decades of observation as tacit agricultural knowledge
- appropriate-technology — understanding the limits of formalization
“We know more than we can tell.”