# Epistemological Foundations

## Table of Contents

1. [What Are First Principles?](#what-are-first-principles)
2. [Philosophical Roots](#philosophical-roots)
3. [The Three Firsts](#the-three-firsts)
4. [Distinguishing First-Principles from Other Methods](#distinguishing-first-principles-from-other-methods)
5. [Epistemic Hierarchy](#epistemic-hierarchy)

---

## What Are First Principles?

First principles are the most basic propositions or assumptions that cannot be deduced from any other proposition. They are the foundational truths from which all other knowledge in a domain is derived.

**Key characteristics**:
- **Irreducible**: Cannot be broken down into simpler components
- **Universal**: Hold across the entire domain of applicability
- **Independent**: Do not depend on other propositions
- **Necessary**: Could not be otherwise without logical or physical contradiction

**In practice**: First principles are the propositions that must be accepted as true because questioning them leads to infinite regress or contradiction. They are the stopping points of reasoning.

### Mathematical Example

In Euclidean geometry, the first principles (axioms) include:
- A straight line can be drawn between any two points
- All right angles are equal
- Given a line and a point not on it, exactly one parallel line exists through the point

From these axioms, all geometric theorems are derived.

### Physical Example

In physics, conservation laws serve as first principles:
- Energy is conserved in closed systems
- Momentum is conserved in the absence of external forces
- Electric charge is conserved

These cannot be derived from anything more fundamental; they are empirical first principles.

---

## Philosophical Roots

### Aristotle (384–322 BC)

Aristotle introduced the concept of "aporia" (ἀπορία) — a state of puzzlement or impasse that marks the beginning of philosophical inquiry. He argued that to understand any subject, one must first identify its archai (ἀρχαί) — the origins, principles, or primary causes.

**Method**: Start with common beliefs (endoxa), subject them to dialectical scrutiny, identify contradictions, and arrive at first principles.

### René Descartes (1596–1650)

Descartes developed "methodic doubt" — systematically doubting all beliefs until reaching an indubitable foundation.

**Famous principle**: "Cogito, ergo sum" (I think, therefore I am) — the first principle derived from radical skepticism.

**Method**: Reject everything that can be doubted; accept only what cannot be doubted as first principles.

### Karl Popper (1902–1994)

Popper introduced falsifiability as the criterion for scientific theories.

**Key insight**: A theory is scientific not because it can be proven true, but because it can be proven false. First principles in science are those propositions that have withstood repeated attempts at falsification.

### Ludwig Wittgenstein (1889–1951)

Wittgenstein distinguished between "showing" and "saying" — some truths can be shown but not articulated as propositions.

**Relevance**: Some first principles may be tacit, embedded in practices rather than explicitly stated.

---

## The Three Firsts

### 1. Logical First-Principles

All reasoning must be logically consistent. Violations indicate hidden assumptions or flawed premises.

**The Three Laws of Logic**:
1. **Law of Non-Contradiction**: A proposition cannot be both true and false at the same time and in the same sense. (A ≠ ¬A)
2. **Law of Excluded Middle**: Every proposition is either true or false; there is no third option. (A ∨ ¬A)
3. **Law of Sufficient Reason**: Nothing happens without a reason. Every true proposition has a sufficient explanation for why it is true.

**Application**:
- When encountering a contradiction, one or more premises must be false
- When a conclusion is uncertain, check the logical chain
- When a proposition seems unsupported, seek the sufficient reason

**Example**: If you claim "This strategy will succeed" but also admit "It violates basic economic incentives," you have violated the law of non-contradiction. One claim must be false.

### 2. Physical First-Principles

All models must align with physical reality. Distinguish between laws of nature and human conventions.

**Key Physical Principles**:
- **Conservation Laws**: Energy, momentum, charge, mass cannot be created or destroyed (only transformed)
- **Thermodynamics**: Entropy increases in closed systems; no perpetual motion
- **Causality**: Effects cannot precede causes; no action at a distance without mediator
- **Speed Limits**: Information cannot exceed the speed of light; no instantaneous communication
- **Scalability Laws**: Physical constraints on scaling (square-cube law, metabolic scaling)

**Application**:
- Test models against conservation: If something seems to come from nothing, the model is wrong
- Check for thermodynamic violations: Free energy claims are impossible
- Verify causality chains: Ensure cause precedes effect
- Respect scaling limits: Small-scale experiments may not scale linearly

**Example**: A business model claiming exponential growth without resource limits violates physical constraints. Real growth faces diminishing returns, competition, and finite markets.

### 3. Epistemic First-Principles

All claims must be verifiable or falsifiable. Distinguish knowledge from belief, evidence from authority.

**Key Epistemic Principles**:
- **Falsifiability**: A claim is scientific only if there exists a test that could prove it false (Popper)
- **Repeatability**: Results must be reproducible by independent observers
- **Occam's Razor**: Among competing explanations, prefer the simplest that explains the data
- **Bayesian Updating**: Beliefs should be updated in proportion to new evidence
- **Epistemic Humility**: Distinguish between what we know, what we believe, and what we don't know

**Application**:
- Ask: "What evidence would prove this claim false?" If none exists, the claim is not scientific
- Demand reproducibility: Single observations are not evidence
- Prefer parsimony: Extra entities require extra evidence
- Update beliefs: New information should change conclusions
- Acknowledge uncertainty: Unknowns are not gaps to be filled with speculation

**Example**: "This product will succeed because users will love it" is not a testable claim. Better: "If users report satisfaction > 4/5 in NPS surveys, the product is considered successful. The null hypothesis is that users report ≤ 4/5."

---

## Distinguishing First-Principles from Other Methods

### First-Principles vs. Analogy

| Aspect | First-Principles | Analogy |
|--------|------------------|---------|
| **Source** | Fundamental truths | Similar cases |
| **Method** | Deductive from axioms | Inductive from examples |
| **Validity** | Universally valid (within domain) | Probabilistic, may fail |
| **Risk** | Over-decomposition | False analogies |
| **Use case** | Novel problems, fundamental innovation | Routine optimization, pattern recognition |

**Example**:
- **Analogy**: "Electric cars will succeed like smartphones did" (assumes similarity between markets)
- **First-Principles**: "Electric cars need energy storage. Battery chemistry determines range and cost. Market adoption depends on total cost of ownership relative to alternatives."

### First-Principles vs. Induction

| Aspect | First-Principles | Induction |
|--------|------------------|-----------|
| **Direction** | Top-down (axioms → conclusions) | Bottom-up (observations → patterns) |
| **Basis** | Logical necessity | Empirical regularity |
| **Strength** | Certain (within axioms) | Probabilistic (past ≠ future) |
| **Limit** | Cannot discover new axioms | Cannot guarantee universal truth |

**Example**:
- **Induction**: "All swans we've seen are white, therefore all swans are white" (false; black swans exist)
- **First-Principles**: "Swans have genes. If a genetic mutation produces melanin in feathers, some swans will be black. The existence of black swans is biologically possible."

### First-Principles vs. Reductionism

Reductionism is breaking complex systems into simpler parts. First-principles goes further: it breaks all the way to irreducible truths.

| Aspect | Reductionism | First-Principles |
|--------|--------------|------------------|
| **Goal** | Understand components | Find foundational truths |
| **Stopping point** | Useful components | Irreducible axioms |
| **Reconstruction** | Often missing | Required |
| **Risk** | Missing emergence | Over-decomposition |

**Example**:
- **Reductionism**: "The brain is made of neurons. Neurons transmit signals." (useful but incomplete)
- **First-Principles**: "Information processing requires: (1) representation, (2) transformation, (3) memory. Neurons provide these through synapses, firing, and plasticity. Consciousness emerges from these computational principles." (goes deeper, attempts reconstruction)

---

## Epistemic Hierarchy

Not all truths are equal. Understanding the hierarchy helps prioritize verification efforts.

```
Level 0: Axioms (First Principles)
  └─ Cannot be derived, must be accepted or refuted

Level 1: Theorems (Logical Consequences)
  └─ Derived from axioms through valid reasoning

Level 2: Models (Approximations)
  └─ Simplified representations of reality based on axioms

Level 3: Theories (Unifying Frameworks)
  └─ Coherent sets of models explaining phenomena

Level 4: Hypotheses (Testable Claims)
  └─ Specific predictions derived from theories

Level 5: Observations (Empirical Data)
  └─ Raw measurements and facts

Level 6: Beliefs (Unverified Assumptions)
  └─ Propositions accepted without proof
```

**Application**:
- When challenging a belief, ask: What level is it at?
- If it's a Level 6 belief, challenge its axiomatic basis
- If it's a Level 4 hypothesis, design tests to falsify it
- If it's a Level 2 model, check its assumptions and approximations

**Example**: "Remote work destroys culture" is a Level 6 belief. To challenge it with first-principles, ask: What are the axioms of organizational culture? What does culture actually require? Can remote work provide those? This moves the discussion from belief to axiomatic analysis.

---

## When to Use This Reference

**Read before**:
- Starting any first-principles analysis
- Challenging a widely-held belief
- Designing a novel solution from scratch

**Consult during**:
- Encountering contradictions (apply logical first-principles)
- Verifying physical claims (apply physical first-principles)
- Evaluating evidence (apply epistemic first-principles)

**Use for**:
- Grounding the entire framework
- Distinguishing first-principles from other reasoning methods
- Understanding why certain claims are not first-principles