Galen Fontaise

The Prediction Paradox: Limited Reflexivity (Axiom A6) This is Part 7 of the Computational Macrohistory series.

Social science isn't just observation.

It's participation.

But participation within modelable bounds.

Predictions change reality—but not infinitely, not randomly.

Full article: open.substack.com/pub/galenfon...

#Reflexivity #Prediction #SocialScience #GameTheory

5/5

This means we can find "fixed points":

P* such that P*(E | P*) = P*

A prediction that remains accurate AFTER actors respond to it.

Like Nash equilibrium—stable because no one wants to deviate.

Reflexivity is real. But it's not fatal.

4/5

The formula:

X(t+) = X(t-) + G(P(E))

When prediction P is published, the system shifts by G.

Key claim: G is LIMITED and PATTERNED.

Not everyone hears predictions.
Not everyone responds.
Responses often cancel.
Institutions move slowly.

3/5

The opposite case:

Predict a pandemic. Governments prepare. Containment works. Pandemic averted.

Your prediction was "wrong."

But was it failure—or success?

Self-defeating prophecies are the GOAL of early warning systems.

A prevented disaster is a victory, not an error.

2/5

🧵 Predict a bank will fail.

Depositors panic. They withdraw money. The bank collapses.

Did you predict the failure—or cause it?

This is the paradox of social prediction.

New article: Axiom A6 — Limited Reflexivity

1/5

... well onto Turchin's integrative-disintegrative cycle. This is exactly why early-warning systems matter: recognizing where we are in that cycle before the amnesia sets in.

Thanks for sharing the event — it must have been a fascinating discussion!

Your observations connect directly to what structural-demographic theory captures mathematically — elite overproduction and regulatory capture are precisely the mechanisms driving the disintegrative phase of secular cycles.
The 'three-generation amnesia' maps...

Thank you! Always glad to connect with someone interested in cliodynamics. Feel free to ask questions or share thoughts as you explore the work — that kind of exchange is what makes this research worthwhile.

CMH models causes, not triggers. We can identify dry tinder. We cannot predict which spark will ignite it.

This isn't a limitation. It's honesty about what science can and cannot do.

#ComputationalSocialScience #History #Prediction #Causality

Both had youth unemployment. Both had authoritarian regimes. Both had social media. Both saw the same news.

Triggers are unpredictable—the specific spark, the specific moment, the specific person.

Causes are structural—youth bulges, inequality, regime illegitimacy, state capacity.

The difference between triggers and causes is everything.

Mohamed Bouazizi was the trigger of the Tunisian revolution. His self-immolation on December 17, 2010 sparked the uprising.

But the trigger doesn't explain why Tunisia exploded while Saudi Arabia didn't.

What is Computational Macrohistory? Why we need a science of large-scale historical dynamics—and what it can (and can't) tell us

First post explores what Computational Macrohistory is (and isn't):
galenfontaise.substack.com/p/what-is-co...

If you're interested in evidence-based approaches to social dynamics, I'd love your thoughts.

#ComplexSystems #DataScience #QuantitativeSocialScience #Research

How:
Mathematical modeling, statistical analysis, complexity science, historical databases—all with rigorous validation and transparent uncertainty.
Current focus:
Arab Spring case study. Could we predict which countries would experience revolution in 2010-12 based only on structural conditions?

I'm excited to launch CMH Bulletin—a new project applying quantitative methods to understand large-scale historical dynamics.

What we study:
- Revolutions and political instability
- Economic cycles - crises
- Demographic pressures - social change
- Patterns in the rise - fall of civilizations

If you're into quantitative history, complex systems, cliodynamics, or simply understanding why societies collapse — glad to have you here.

-Galen Fontaise

5/5

Computational Macrohistory Bulletin | Galen Fontaise | Substack I apply mathematical and computational methods to study historical patterns—revolutions, political cycles, economic crises. My work combines statistics, complexity science, and data analysis. Founder ...

I'll post research updates, methodological reflections & findings here.
For more:
📰 Substack → galenfontaise.substack.com
🔗 LinkedIn → www.linkedin.com/in/galenfont...
📄 Orcid → orcid.org/my-orcid?orc...

4/5

The framework rests on 8 foundational axioms defining when historical systems become scientifically tractable.

Currently in empirical validation: Arab Spring 2010-2012 as proof-of-concept case study.

3/5

CMH applies math, statistics & dynamical models to history to identify recurring patterns and compute probability distributions for critical socio-political events.

No deterministic forecasts — honest probabilities only.

2/5

🧵 Hi Bluesky!

I'm Prof. Fontaise, founder of FICSS (Lugano) & creator of Computational Macrohistory (CMH) — a quantitative science of large-scale historical systems.

A short intro:

1/5