In this video I discuss Price’s law and the problems of generalizing it beyond its initial scope.
To summarize this video:
Price’s law tells us that in a given scientific field, half of the papers in the field are produced by sqrt(N) people, where N is the total number of authors writing in the given field. This observation seems to be true, more or less.
There was a recent video by Prof. Jordan Peterson in which the concept of Price’s law is generalized to all fields, including music. He further claims that Price’s law is a stumbling block to companies as they grow, e.g., in a company of size 100 half the work is done by 10 people with 90 doing the other half, but in a company of 10000 half the work is done by 100 people and the other half is done by 9900 people.
The Peterson videos can be see here: https://youtu.be/fjtBDa4aSGM
There are two failings of this generalization.
First the scaling isn’t clear. The relationship between work per person per time isn’t obvious. The work performed per “productive” person and per “other” person is highly coupled. In addition some of the “other” people enable the success of the “productive” people. For example, a company may higher a communications specialist that writes press releases and advertisements. This person doesn’t contribute to “work,” whatever that is, but they act to multiply the effectiveness of the “productive” people’s work.
Second, the definition of “work” isn’t clear. Consider music. It might be true that in a given time frame, half the songs produced within a given genre, might be produced by sqrt(the artists producing music within that time frame and musical genre). But this doesn’t make the music good. Consider the band Nirvana. They only produced 2 studio albums, but their influence on modern music was huge. There are many such examples in many fiends, including science, where people outside the sqrt(N) sub-set make highly impactful contributions. On the other hand there are members of the sqrt(N) subset that are highly productive, but don’t make a meaningful impact. Consider, for example, the painter Thomas Kinkade; he was prolific, but did his work leave any artistic impact on the field? I don’t think so. (In a similar vein, consider many of those that are highly “productive” on Steemit. There are many so-called creators that if their work was eliminated there would be no loss?)
My contention is that Price’s law may be true and useful for quantifying the production of discretized items, such as scientific papers, or musical albums, but it cannot be generalized to quantify the impact that workers have, nor can it be used to discuss the scaling of endeavors.
In case DTube isn’t showing my video, it can be found on YouTube here:
