4 ms·
》PLs was an indicator of randomness Exactly this! Meaning intelligent language has high entropy. And it has high entropy because it's compressed. We don't noti
by ffhhj 4y ago
》PLs was an indicator of randomness
Exactly this! Meaning intelligent language has high entropy. And it has high entropy because it's compressed. We don't notice this compression and seems counter intuitive that wordy natural language is actually compressed, but it has evolved to be optimal and compressed.
- flybrand 4y agoIs there a way to compare power-laws between two series and take something from the data? For example the slopes of public market returns are different than that of illiquid markets. Is a power law w more slope more random? I was really looking for a t-test or chi-squared type take away. My smarter, math under grad interns pointed out that we were at the edge of interpretive capabilities - but this was ~2000s.
- hammock 4y agoUnfortunately so many people encounter power laws in business and interpret the opposite
- doctor_eval 4y agoWould it be possible to test the law against a corpus of a creole or pidgin? Since such a language is by definition new and may be rapidly evolving, it might have less entropy, and therefore might not fit the law. That might provide some evidence for this idea.