AI summary by TwinMind
Scientists Discover a Law of Natural Laws
Sabine Hossenfelder
6m 56s
Mathematical Language vs Natural Language
- Mathematics entirely self-referential, doesn't refer to anything besides itself (pure math)
- Physics mathematics becomes language to describe observations
- Question: why this particular math used in physics?
Zipf's Law in Natural Languages
- Regular decrease in word frequency when words assigned numbers by frequency rank
- Most frequent word = #1, second most = #2, etc.
- Follows power law decay pattern
- Found in text bodies of all known languages
- Speculation: most efficient way of communicating information
- Similar power laws may explain knowledge increase slowdown in large language models
Research Methodology
- Authors studied syntax of equations, breaking into mathematical operations
- Addition, multiplication, second/third powers
- Multiplication with constants of nature
- Complex functions: exponential, trigonometric
- Counted frequencies across three samples:
- Feynman Lectures of Physics
- Wikipedia equations list
- Encyclopedia Inflationaris (inflation models)
Key Findings
- Mathematical expressions follow regular law but NOT same as natural languages
- NOT power law decay - follows exponential decay instead
- Surprisingly regular despite small sample sizes
- Very intriguing pattern discovered
Two Possible Explanations
- Tells us something about mathematics in laws of nature
- Tells us something about how we use mathematics
- Possibilities not mutually exclusive - likely mix of both
Human Contribution Factor
- Tendency to rewrite equations for easy visual deciphering
- Factorization done to make patterns clearer (computers wouldn't care)
- Powers of quantities mostly integers for practical reasons
- Could define unit for π-th root of energy but would cause practical problems
- Simplification to notation humans accustomed to
Deeper Physical Truths
- Rare fractional powers of quantities may reveal something about origin of dimensional units
- Physics rarely has more than two time derivatives
- No deeper theoretical reason except "doesn't describe what we see"
- Physicists so used to this they don't think about it
- Suggests missing underlying structure
- Alternative: "universe is just trolling us"
Sponsor Content
- Brilliant.org educational platform mentioned
- Interactive visualizations and follow-up questions
- Topics: science, computer science, mathematics
- Author has quantum mechanics course on platform
- Special offer: 30-day trial + 20% off annual premium via brilliant.org/sabine
