The characteristic of the logarithm

https://arbital.com/p/log_characteristic

by Nate Soares Jun 13 2016 updated Oct 19 2016


[summary: Any time you find an output that adds whenever the input multiplies, you're probably looking at a (roughly) logarithmic relationship. For example, imagine storing a number using digit wheels. Every time the number goes up by a factor of 10, you need one additional digit wheel: It takes 3 wheels to store the number 500; 4 to store the number 5000; 5 to store the number 50000; and so on. Thus, the relationship between the magnitude of a number and the number of digits it takes to write down is logarithmic. This pattern is the key characteristic of the logarithm, and whenever you see an output adding when the input multiplies, you can measure the output using logarithms.]

Consider the interpretation of logarithms as the cost of communicating a message. Every time the number of possible messages to send doubles, your communication costs increase by the price of a coin, or whatever cheaper [-storage_medium] you have that can communicate one of two messages. It doesn't matter whether the number of possible messages goes from 4 to 8 or whether it goes from 4096 to 8192; in both cases, your costs go up by the price of a coin. It is the factor by which the set grew (or shrank) that affects the cost; not the absolute number of messages added (or removed) from the space of possibilities. If the space of possible messages halves, your costs go down by one coin, regardless of how many possibilities there were before the halving.

Algebraically, writing $~$f$~$ for the function that measures your costs, $~$c(x \cdot 2) =$~$ $~$c(x) + c(2),$~$ and, in general, $~$c(x \cdot y) =$~$ $~$c(x) + c(y),$~$ where we can interpret $~$x$~$ as the number of possible messages before the increase, $~$y$~$ as the factor by which the possibilities increased, and $~$x \cdot y$~$ as the number of possibilities after the increase.

This is the key characteristic of the logarithm: It says that, when the input goes up by a factor of $~$y$~$, the quantity measured goes up by a fixed amount (that depends on $~$y$~$). When you see this pattern, you can bet that $~$c$~$ is a logarithm function. Thus, whenever something you care about goes up by a fixed amount every time something else doubles, you can measure the thing you care about by taking the logarithm of the growing thing. For example:

Conversely, whenever you see a $~$\log_2$~$ in an equation, you can deduce that someone wants to measure some sort of thing by counting the number of doublings that another sort of thing has undergone. For example, let's say you see an equation where someone takes the $~$\log_2$~$ of a Relative likelihood. What should you make of this? Well, you should conclude that there is some quantity that someone wants to measure which can be measured in terms of the number of doublings in that likelihood ratio. And indeed there is! It is known as [bayesian_evidence (Bayesian) evidence], and the key idea is that the strength of evidence for a hypothesis $~$A$~$ over its negation $~$\lnot A$~$ can be measured in terms of $~$2 : 1$~$ updates in favor of $~$A$~$ over $~$\lnot A$~$. (For more on this idea, see [ What is evidence?]).

In fact, a given function $~$f$~$ such that $~$f(x \cdot y) = f(x) + f(y)$~$ is almost guaranteed to be a logarithm function — modulo a few technicalities.

[checkbox(proove_things): This puts us in a position where you can derive all the main properties of the logarithm (such as $~$\log_b(x^n) = n \log_b(x)$~$ for any $~$b$~$) yourself. Check this box if that's something you're interested in doing. y: path: Properties of the logarithm n: ]

[fixme: Conditional text depending what's next on the path.]