The story of lactose tolerance evolving multiple times has blown up a bit, thanks to Nick Wade at The New York Times. Some people are making analogies to light skin evolving via different genetic architectures (remember, skin color is a polygenic trait, albeit dispersed over ~4 loci of large effect). But there is a difference, light skin color emerges via loss of functionality or expression on the loci which result in pigment production. There are many ways to lose function, but it generally is considered more difficult to gain function. And yet this is what lactose tolerance is. Or is it? Remember, infants and toddlers can digest lactose fine, the issue is that tolerance persists in some populations. In other words, whatever pathways shift so that metabolic shunts are closed which drive lactose breakdown, all you need to do is short circuit this abolishment of gene expression on the LCT locus. The data implies that in Europeans a cis-acting element continues to induce transcription from LCT in adults, but other data implies that African groups may have stumbled upon a strategy utilizing the epistatic interaction across loci via a trans-acting factor to retain function on this locus. Fundamentally this isn't as if various human groups have evolved ways to breakdown something evolutionarily rare such as cellulose, rather, it is simply utilizing the standing genetic variation to tweak a mundane physiological process which is found across young mammals.
But we are always told how humans lactose tolerance is strange among mammals, and yet it occurs in multiple flavors in humanity. Why is this? I think the lesson is that there was never a great physiological or biochemical constraint, but the gentle force of selection operating against adult persistence has resulted in a lack of this trait across mammals. We know that even in humans there is some selection against it, I am to understand that in East Asians it seems there has been selection for alleles which are particular trenchant in shutting off this metabolic pathway in adults. So the difference was positive selection, not a change in the genetic or biochemical structural constraints. And, that selection was driven by a process particularly powerful in humans: gene-culture co-evolution. Just as selective forces uniformly constrained this trait across mammals and human populations before 10,000 B.P., so the rise of farming and animal husbandry probably made the emergence of this trait inevitable in a culture-bearing species. This might give us clues as to what makes humans different from other species: it isn't that our "essence" is somehow peculiar or special, rather, we have drifted onto a path of gene-culture co-evolution which is very rare, but inevitably necessitates many parallelisms in culture and genetics across our species.1
1 - In terms of culture, note the multiple hearths of farming across the world. Farming induces changes in the optimal manner in which one metabolizes starches, etc. The convergences in cultures will result convergences in phenotype, perhaps derived from alternative genetic architectures.