"Like all science stories, this one begins with wonder." That's how Babel's Dawn the book starts. I''m quoting it because a dispute has popped up in the comments section on how science works and what sorts of questions are best asked. I was thinking one way, then another about my response and then I remembered how I began my manuscript and I thought, oh, yeah, wonder. There are lots of things to wonder about language: How did it begin? Is it really different from other forms of animal communication? How do words get meaning? What's a verb? Why do children learn to speak so easily while other animals don't learn it at all? Why can users ignore some rules and still be understood?
This much I know. Whatever question you begin with, be ready to reconsider it. In my own case I asked how language could have begun since it is passed down from generation to generation in an infinite regress like standing the earth on the back of a turtle. I was stuck at that level of insight for many years until one day I changed the question. It occurred to me that perhaps I was wrong about how language begins in children. Maybe it isn't passed on from generation to generation. Maybe each generation invents language for itself.
The great thing about that idea was that it was researchable. As soon as I thought of it, I sat up and wrote a list of things that would have to be true if my idea about language re-invention were true. I listed about five things, all of which turned out to be true. Eventually, I even wrote a book on the subject, a long-out-of-print work called So Much to Say. Naturally, it turned out that my trigger idea was too simple. There are two phases of children's language learning. The first typically lasts until age three, and during this period children do spontaneously make linguistic sounds, then words, then phrases. But even that is not purely instinct. Normally a toddler is raised among speakers and will mostly use their hand-me-down words. It is only the freak condition in which children are left to their own devices and create a personal lexicon.
When children turn three, however, they become much more communally-oriented and start learning a particular language with all its rich complications. This passage does not occur in the freak cases of children locked together in the basement. It's also the point where things can suddenly go wrong. Children who seemed to be doing fine sometimes don't make the leap to "real" language. In other cases, less common perhaps, a child is silent in the early period and then starts talking at age three. The story about Einstein is that he never spoke until he was three, and then he spoke a full sentence. Doubtful, but not impossible.
The debate in the comments section focuses chiefly on the issue of whether science builds more on facts or ideas. At least the dispute didn't focus on which helical strand is more important to DNA.
In my case, my wonder began with an idea about how language is learned. I then began looking for facts that would help explain my wonder. But science can run the other way. You can have a fact and look for an idea to explain it. The classic example is Einstein again, who noticed that light is never observed standing still. He then spent the next ten years looking for an idea that might explain tjat fact.
Whether you wonder at facts or ideas, the biggest danger in science is getting stuck on just one side of the street. You need to embrace both sides. Some people have a great ability to dig up facts, and as there is always another fact to find, they can make a career of it. Other scientiests like ideas and grind them out with a regularity that would astonish a sausage factory. But to escape the label of metaphysics, or mathematics, or blarney, someway or other you have to move from ideas to facts. When you've got both ideas and facts working together they twist into a cable that has some real strength.
Looking back at four and a half years of this blog I'm glad to see that I've covered both sides. Ideas tend to have a shorter half-life than facts, but the good ones have kept me occupied. They let me know what facts I need that can either support or cripple the idea. Meanwhile, facts come along and demand an explanation.
The main ideas that have shaped this blog are the speech triangle, the role of words as pilots of attention, the importance of linking two or more points of attention into a single idea, and multi-level selection. Each of those ideas is still standing because they are now anchored in facts.
The speech triangle (speaker, listener, topic) requires the support of two different kinds of facts. One set is data about cooperation that allows speaker and listener to work together honestly and trustingly. The other cata set concers attention to topic. Together these facts explain why apes and other primates do not have language. They don't trust each other enough to share their knowledge, and they are not curious enough about the world to examine a topic whose practical value isn't immediately apparent.
The idea of words piloting attention provides a way of looking at sentences and examining them. Do they really work by directing attention? It turns out that this combination of idea and data gets rid of mystical mumbo-jumbo about meaning, translation, and even the Whorfian hypothesis. It also ties into the speech triangle, indicating how speaker and listener relate to one another through language.
Multi-level selection is about evolution rather than language, and I came to it facts first. I kept coming across examples of language that seemed to be inexplicable in terms of maximizing one's own genetic strength. So when I read a paper by E.O. Wilson and David Sloan Wilson that rejected the selfish gene hypothesis, I was relieved. I was even more relieved when I found the statement that multi-level selection does not require a revision of the math supporting evolution. And once again this idea supports the speech triangle image, indicating that it is possible to evolve the necessary speaker-listener relationship.
Finally, the importance of linking multiple points of attention is an idea that came late to the blog, as a result of a paper James Hurford presented in Poland. This idea suggests on a grammatical level what happens at three years of age to change the speech from the original one (that even apes can be taught) into something that requires a culturally-defined sentence.
I have my scenario that hypothesizes how these pieces all came together, but the critical element is having the ideas and facts that make sense of the whole. In the back-and-forth over how science works, one commenter distinguished between information and understanding. The distinction is real enough, but I don't think we have to choose between them. Based on what I've read over the years working on this blog, I'd say we have plenty of both, and together they keep us moving along.