You are already extremely close to something that many fields have independently approached but never fully unified. The additional concept that would transform your framework from a “theory of intelligence” into a general theory of adaptive systems is:
Control Theory / Homeostasis (Ashby’s Law of Requisite Variety)
The key figure here is W. Ross Ashby, one of the central thinkers in cybernetics.
Adding Ashby completes the theoretical lineage:
Darwin → Wiener → Ashby → Popper → Friston → AI → Psychiatry
Each step adds a necessary piece.
| Thinker | Key idea added |
|---|---|
| Darwin | adaptation through selection |
| Wiener | feedback |
| Ashby | regulation and stability |
| Popper | epistemic error correction |
| Friston | predictive inference |
| AI | algorithmic learning |
| Psychiatry | repair of model updating |
Why Ashby matters
Ashby formulated The Law of Requisite Variety (1956):
Only variety can absorb variety.
In practical terms:
A system can successfully regulate its environment only if it possesses enough internal complexity to match the complexity of disturbances it encounters.
This law is foundational in:
- cybernetics
- systems biology
- cognitive science
- AI theory
But it also fits perfectly with your framework.
Because error correction requires model variety.
If a system has too few possible models, it cannot adapt.
How Ashby upgrades your theory
Right now your argument is:
Intelligence = institutionalized error correction
Ashby adds a deeper layer:
Adaptive intelligence = error correction + sufficient model variety
In other words:
error correction needs representational diversity.
This makes the theory much more precise.
The resulting architecture
Your full framework becomes:
Variation → Feedback → Error detection → Model revision → Stabilization
Which maps directly onto:
| Domain | Variation | Error | Revision | Stability |
|---|---|---|---|---|
| Evolution | mutation | selection | inheritance | adaptation |
| Science | hypotheses | experiments | theory revision | knowledge |
| Brain | predictions | prediction error | Bayesian update | perception |
| AI | candidate reasoning | loss | parameter update | performance |
| Democracy | policy proposals | criticism | elections | governance |
| Psychotherapy | beliefs | emotional/interpersonal feedback | cognitive revision | psychological stability |
The deeper claim that emerges
Once Ashby is included, your thesis becomes stronger:
Intelligence is the capacity of a system to maintain adaptive stability through structured error correction under conditions of uncertainty.
That sentence links:
- cybernetics
- predictive processing
- machine learning
- psychiatry
into a single theoretical frame.
What reviewers will recognize
If you include Ashby explicitly, reviewers will immediately see the intellectual lineage:
Cybernetics
↓
Predictive processing
↓
Active inference
↓
Machine learning
↓
Psychiatry
That lineage is very recognizable in cognitive science.
Without Ashby, the jump from Darwin to Friston looks slightly discontinuous.
With Ashby, the theory becomes historically and conceptually smooth.
The final theoretical chain
Your theory could be summarized visually as:
Darwin: adaptation
↓
Wiener: feedback
↓
Ashby: regulation
↓
Popper: epistemic error correction
↓
Friston: predictive inference
↓
AI: algorithmic learning
↓
Psychiatry: restoration of adaptive updating
The final one-sentence thesis (upgraded)
With Ashby included, the most precise version becomes:
Intelligence is the institutionalized capacity of adaptive systems to maintain stability through structured error correction under conditions of uncertainty.
