Bad questions in evolutionary biology

Why questions of the format: "Why did 'X' evolve?" or "What is the evolutionary reason for 'X'" are almost always unanswerable questions.

Questions in this format are almost always completely unanswerable from a scientific standpoint (with some exceptions I will get into later). 

Most often the "X" in question happened hundreds of thousands to millions or billions of years ago (e.g. why did the eye evolve? what is the evolutionary advantage of 5 fingers and toes on each limb?).

At this timescale it is impossible to figure out what the selective forces were that were acting on the evolution of the trait, and for more ancient "X", which particular species the trait was evolving in and its physiological environment.

Most phylogenetic reconstruction to infer ancestral states use parsimony methods. These are methods that use data from various individuals or species to figure out the evolutionary history between them by constructing tree diagrams. These methods, while correct on average, have a significant likelihood of being wrong for any particular case due to convergent evolution, parallel evolution etc. E.g. on average the same mutation will not happen twice, but any particular mutation (particularly in species with huge population sizes) has a not insignificant probability of doing so. This fact further complicates the evolutionary inference by making it impossible to figure out what the ancestral state of a trait was. Knowing the ancestral state is absolutely necessary for figuring out how the trait evolved.

Basically, unless you can get a time machine to sample a large number of individuals across the entire range of time when the trait was evolving, figure out exactly which forces (all of the other species as well as the environmental conditions) were driving its evolution and what chance events were involved, it is essentially impossible to figure out why "x" evolved.

The exceptions are:

Recent evolutionary events, with fairly obvious selection pressures and known biochemical mechanisms- e.g. lactase, the enzyme that breaks down lactose in milk, is typically only present in infants and not in adults. There have been at least 3 different instances of humans evolving production of lactase into adulthood, which allows adults to consume milk. This has obvious selective benefits, and can be traced to time when humans started agriculture 10k years ago.

Sickle cell anemia and malaria resistance is another powerful example of knowing why something evolved.

We can do laboratory evolution experiments under controlled condiitons and use the power of modern genetics to dissect out why particular mutations are beneficial. for example, if yeast are grown in low glucose conditions, they typically produce more of the proteins that transport glucose into the cell, so they can maximize their energy intake.

Most of these exceptions are recent, and have well defined biochemical effects and obvious selective pressures that can sometimes be experimentally tested (can't really ethically give people malaria or deny people the use of milk to test their fitness).


The bottom line:

So for recent events, with known biochemical effects of mutations and good ideas of the external selective pressures, we can infer why a particular trait evolved. Outside of this, it is essentially impossible to figure out. We can speculate and make up what I like to call "just so stories" that sound plausible, but there is no hard evidence (short of a time machine and lots of genetic analysis) that we could ever generate to prove one mechanism over another.

This is also why questions about the origins of life are impossible to answer, as we will probably never know enough to give a plausible mechanism, much less find conclusive evidence proving one mechanism to be correct.