Evolution of language
Ed Yong beat me again, this time discussing an interesting paper in Nature on the evolution of language, but I'm going to comment anyway. Actually, there are two papers in the same issue, both showing that frequently used words change more slowly. For example, irregular past tenses of rarely-used verbs (bide, delve, etc.) have tended to disappear, but we still say "came and saw" not "goed and seed." Lieberman et al. (Nature 449:713) note that:
It is much rarer for regular verbs to become irregular: for every ‘sneak’ that ‘snuck’ in there are many more ‘flews’ that ‘flied’ out.
I bet this is also true of less-used definitions of words and of collective nouns: a group of football players will still be a "team" long after a "bouquet" of pheasants has become a "flock" and then a "group."
Languages evolve, but analogies with evolution of genes may be misleading. Genes pass only from parent(s) to offspring and their frequency changes over generations. Words and ideas can spread much more rapidly, including from child to parent. Cultural evolution seems more analogous to the spread of viruses, only some of which come from our parents.
Some ideas (e.g., religions) come packaged with explicit instructions to proselytize, like the rabies virus making its victims bite others. But the desire to spread our tastes in music, for example, seems intrinsic to us, not to the music. So cultural evolution seems similar enough to epidemiology that analogies will sometimes be useful, but only sometimes.