The traditional explanation for such a change is that it must provide an advantage in either survival or sexual selection. In this case, though, the finches were well fed, and there were no predators. Meanwhile, mate selection was done by breeders, who were influenced only by feather coloration.
Enter my friend Terry Deacon, a scientist who has made fundamental contributions in widely diverse areas of research. He is a professor of anthropology at the University of California at Berkeley and an expert on the evolution of the brain; he is also interested in the chemical origins of life and the mathematics behind the emergence of complicated structures like language.
Terry proposed an unconventional solution to the mystery of Bengalese finch musicality. What if there are certain traits, including song style, that naturally tend to become less constrained from generation to generation but are normally held in check by selection pressures? If the pressures go away, variation should increase rapidly. Terry suggested that the finches developed a wider song variety not because it provided an advantage but merely because in captivity it became possible.
In the wild, songs probably had to be rigid in order for mates to find each other. Birds born with a genetic predilection for musical innovation most likely would have had trouble mating. Once finches experienced the luxury of assured mating (provided they were visually attractive), their song variety exploded. Brian Ritchie and Simon Kirby of the University of Edinburgh worked with Terry to simulate bird evolution in a computer model, and the idea checked out.
Recent successes using computers to hunt for correlations in giant chunks of text offer a fresh hint that an explosion of variety in song might have been important in human evolution. To see why, compare two popular stories of the beginning of language.
In the first story, a protohuman says his first word for something—maybe ma for "mother"—and teaches it to the rest of the tribe. A few generations later, someone comes up with wa for "water." Eventually the tribe has enough words to have a language.
In the second story, protohumans have become successful enough that more of them are surviving, finding mates, and reproducing. They are making all kinds of weird sounds because evolution allows experimentation to run wild, so long as it doesn't have a negative effect on survival. Meanwhile, the protohumans are doing a lot of things in groups, and their brains start correlating certain distinctive social vocalizations with certain events. Gradually, a large number of approximate words come into use. There is no clear boundary at first between words, phrases, emotional inflection, or any other part of language.
The second story seems more likely to me. Protohumans would have been doing something like what big computers are starting to do now, but with the superior pattern-recognizing capabilities of a brain. While language became richer over time, it never became absolutely precise. The ambiguity continues to this day and allows language to grow and change. We are still living out the second story when we come up with new slang, like bling or LOL.
Even if the second story happened, and is still happening, language has not necessarily become more varied. Rules of speech may have eventually emerged that place restrictions on variety. Maybe those late-arriving rules help us communicate more precisely or just sound sexy and high status, or more likely a little of both. Variety doesn't always have to increase in every way.
Variety could even decrease over time. In fact, there may be a bizarre example of that happening right now in human song. We can easily explore the changing amount of variety in songs over the last hundred years because of an amazing data archive: audio recording. Since the beginning of recorded music, the sound of human song has changed with each new generation of people. There's no confusing a 1930s song with a 1940s song, or a 1950s song with a 1960s song. The pattern sticks until roughly the end of the 1980s. It's not easy to tell whether a song came from 1990 or 2000.
This might sound like an extraordinary claim, but you can test it yourself. Listen to random clips from the many sources of songs available on the Internet and don't peek at the year they were produced. You'll discover that it's harder to date songs from the last two decades than songs from previous decades. Terry and I are now considering this experiment on a more formal basis.
If you accept that there has been a recent decrease in stylistic variety in human song, the next question is "Why?" There are plenty of possibilities: Maybe the Internet makes too much information available, so everyone has the same influences to absorb—and songs lose flavor and take on a generic quality. To be more cynical, it could be a sign of cultural decline.
Another explanation, which is the one I suspect, is that the change since the mid-1980s corresponds with the appearance of digital editing tools for music. Digital tools are more suggestive about results than previous tools: If you deviate from the kind of music a digital tool was designed to make, the tool becomes difficult to use. For instance, it's far more common these days for music to have a clockwork regular beat. Some of the most widely used music software becomes awkward and can even produce glitches if you vary the tempo much while editing. In predigital days, tools also influenced music, but with not nearly such a sharp edge.
So this is an ironic moment in the history of computer science. We are beginning to succeed at using computers to analyze data without the constraints of rigid grammarlike systems. But when we use computers to create, we are confined to equally rigid 1960s models of how information should be structured. The hope that language would be like a computer program has died. Instead, music has changed to become more like a computer program.