Full House: The Spread of Excellence from Plato to Darwin by Stephen Jay Gould
The central problem that Full House addresses is what Gould refers to as the “fallacy of reified variation”, where an abstraction like a mean or an extreme value is treated as a concrete thing and its progress tracked. This creates a bias in our perception, where instead Gould shows that we should consider what he calls “the full house of variation”.
The two principle case studies in the book are the disappearance of 0.400 hitting in baseball and the myth of evolutionary progress in biological complexity. What needs to be considered are how the normal distributions change over time and how they react with left or right walls, being minimum or maximally allowed parameters respectively. If the variation were arranged on an up-down axis, we’d call them floors and ceilings instead. In baseball (and other athletic activities), there is a right wall of maximum human performance that modern professional athletes are very near. In the evolution of life, “primitive” organisms necessarily began at the left wall of complexity, since only the simplest structures could have started out (you won’t find a giraffe randomly walking out of the primordial ooze).
A century ago, most baseball players (compared to those today) were not very good so the peak of the distribution, the mode, was far from the right wall, but there was a lot of variation in the individual skill of the players. In terms of batting average (BA), you have four possibilities:
- Skilled batter versus unskilled pitcher: high BA
- Skilled batter versus skilled pitcher: medium BA
- Unskilled batter versus skilled pitcher: low BA
- Unskilled batter versus unskilled pitcher: medium BA
In the early days, with wide variation it happened more often that a skilled batter would be against comparatively unskilled pitchers and so would secure a high batting average, even passing the vaunted (and arbitrary) 0.400 mark.
However, as the general performance of players increased (larger pool of potential players, better training, better coaching, better nutrition, better technology, etc), the mode moved toward the right wall of human ability and the variation shrunk, so nowadays most play is like #2 from the above list, where great batters and great pitchers are continually matched so there’s much less opportunity for standout performances (though the league average has remained constant, meaning batting and pitching have improved in equal measure).
The other major case study (there are two smaller chapters on the nature of mean life expectancy after diagnosis with an illness and the evolution of horses) is on the myth of an intrinsic drive toward complexity in the evolution of species. This is encapsulated in the stories we tell and the pictures we see: just think of what is conjured up by the cascade of the Age of Bacteria, the Age of Fish, the Age of Reptiles, and the Age of Mammals. As Gould points out, the real truth is that since the beginning life on Earth has always been one long Age of Bacteria and that our tunnel vision view of the evolution of humans skews our thinking.
As stated, life necessarily began at a left wall of complexity with a peak in the bacterial realm (what Gould calls the modal bacter). The supposed drive toward increased complexity is really just a drunkard’s walk from there (where a drunkard’s walk is a random walk with a reflecting wall, so that eventually the drunkard ends up far away from the bar). Life couldn’t get much simpler and still be considered life, so the only way to go was up, but by pure random statistical chance as the spread of variation increased (though maintaining a bacterial peak). Far from being the top of the evolutionary ladder, humans are a random fluke in a bacterial world.
I think Full House is an important book, since statistical thinking is so inherently difficult for us and so it’s easy to focus on abstract measures like an average and think we’re getting the whole story. Gould makes his case convincingly with logic and evidence, and it’s a story that needs telling (even if you don’t particularly care about baseball). He ends with “An Epilog on Human Culture”, pointing out that these arguments don’t necessarily translate over to what is often called cultural evolution (Gould prefers the term “culture change”, I think rightly) since human knowledge and technology is Lamarckian rather than Mendelian.