Full House

Stephen Jay Gould’s book Full House was a pretty good read. Taking conventional wisdom and turning it on its head, Gould critiques the common mystical pronouncement of evolutionary progress, and offers an intriguing paradigm in which to view the changes in complex systems. Gould does so by using the notion of “right” and “left” walls – limitations inherent in the system that restrict progression or regression. Gould first advances his model that progression is marked by the amount of variation in the system, not “a point moving somewhere,” and demonstrates its explanatory power by going to his lifetime fascination with Baseball, or more specifically, the disappearance of .400 batting averages.

He rips to shreds some of the common explanations for the disappearance of .400 hitting. He is particularly harsh on the belief that the absolute skill of modern MLB batters has declined. I honestly have no idea how someone could possibly believe such nonsense. The idea that hitters have gotten objectively worse is not even remotely credible, not with literally every sport getting more competitive.

He treats the hypothesis that Batters have simply gotten worse in a relative sense, not in an absolute sense, by getting oustripped by better pitchers and fielders, with a little more sympathy. But that explanation, too, does not accord with the verifiable facts (the mean batting average has held flat at .260, not what you would expect from increasing pitcher superiority). And while it sounds reasonable on the surface, its credibility quickly vanishes once one has caught a glimpse of Gould’s explanation of contracting variation.

Basically put, batters have run into the right wall of batting ability. Due to the limitations that bio-mechanics imposes on the human potential, batters can only be so coordinated, strong, etc, and the disappearance of .400 hitting has resulted from more players getting closer (if not reaching) that limit. Thus, the great players of old, who had less competition than that of modern players, could deviate so far from the norm to reach the magic .400 without being blocked by the right wall. As play continued to improve, the average batter had advanced ever so closer to the right wall. This shrunk the overall variation in play and even the best batters could not rise to .400 hitting without reaching the right wall of human limitation.

After the section on baseball demonstrating the effects of right walls, Gould proceeds to discuss the effects of left walls on evolution. Gould debunks the argument that evolution is characterized by ”progress” or “an increase in complexity.” It is not. Natural Selection does not have a bias towards increasing complexity, says Gould. Instead, some biological organisms have increased in complexity through random movements expanding from a left wall.

To illustrate this, Gould takes a drunkard walking out of a bar, who takes a left down the road. There is a gutter thirty feet to the right of the bar. So, even with completely random movements (left to right), the drunkard will fall down the cutter, every time, because the bar acted a wall that he would bounce off of until he eventually landed in the cutter. Even random movements can have a surefire destination.

In a similar manner, the minimal complexity that is required for a biological unit – a single cell bacteria – is the left wall, leaving all random variation only one direction to move: to the right. Most organism will still cling to their humble left origins, but some – ah, a miniscule few – will drift off significantly to the right. From an evolutionary perspective, we are not living in the “age of Man”, or before, the “age of mammals/reptiles/dinosaurs etc.” but have always lived in the “age of bacteria”, with expanding variation on the right tail.

Gould gives plenty of examples and re-phrases so as to soak the argument into the reader’s mind, certainly much more effectively and comprehensively than a medium-length review of a book can do here. So I urge you to read the book, if for nothing else than a refreshing perspective on evolution and statistical analysis.