To put it another way, there is no average drift towards larger brains, but individual lineages do have trajectories up or down. Encephalization is one adaptation suite out of many possible ones, and depends on context.
And even if there were no unusual evolutionary pressures on individual lineages toward greater intelligence—then it would be a random walk, starting from zero. The expected distance from the origin still grows over time.
Mammals evolve fast (I use them as the reference class here because they have encephalization about an order of magnitude higher than most nonmammals). Recent work seems to indicate that the entire diversity of mammals on Earth has a common ancestor after the KT event of 65 megayears ago. In an early fraction of that time, the diversity of mammals seems to have filled up a vast range of the possible body sizes and habitats. It would not surprise me if lineages also quickly explored most of the available encephalization ranges (ignoring weird feedback loops like those that lead to us) and the initial increase as those lineages were founded flattened on average.
Increased encephalization has evolutionary costs—high energy expenditure and (in humans) difficult birth of underdeveloped infants due to head size. Natural selection will not allow encephalization beyond a certain size unless it has offsetting benefits, which depend on the species and its environment. So it’s not a random walk (=neutral genetic drift) once you get far enough from the average, like we have.
To put it another way, there is no average drift towards larger brains, but individual lineages do have trajectories up or down. Encephalization is one adaptation suite out of many possible ones, and depends on context.
And even if there were no unusual evolutionary pressures on individual lineages toward greater intelligence—then it would be a random walk, starting from zero. The expected distance from the origin still grows over time.
Mammals evolve fast (I use them as the reference class here because they have encephalization about an order of magnitude higher than most nonmammals). Recent work seems to indicate that the entire diversity of mammals on Earth has a common ancestor after the KT event of 65 megayears ago. In an early fraction of that time, the diversity of mammals seems to have filled up a vast range of the possible body sizes and habitats. It would not surprise me if lineages also quickly explored most of the available encephalization ranges (ignoring weird feedback loops like those that lead to us) and the initial increase as those lineages were founded flattened on average.
Nitpick: that link talks about placentals; mammals also include monotremes and marsupials.
Increased encephalization has evolutionary costs—high energy expenditure and (in humans) difficult birth of underdeveloped infants due to head size. Natural selection will not allow encephalization beyond a certain size unless it has offsetting benefits, which depend on the species and its environment. So it’s not a random walk (=neutral genetic drift) once you get far enough from the average, like we have.