Loaded Questions

The Ptolemaic system of the universe explained many things and defined which questions were legitimate and which should not be asked. The system was destroyed by people who asked questions it had no place for.

Late in his book Adam’s Tongue, Derek Bickerton presents a list of conditions that any theory of language origins must satisfy. They are:

• The selection pressures had to be strong.
• The selection pressures had to be unique.
• The very first use of language had to be fully functional.
• The theory musn’t conflict with anything in the ecology of the ancestral species.
• The theory must explain why cheap signals should be believed.
• The theory must overcome primate selfishness. [pp. 165-166]

Mathematicians and metaphysicians can think in this absolutist way, but scientists should be more cautious.

Take that business about the very first use of language. Bickerton has his reasons for making the rule, but they are theoretical rather than empirical. It leaves him blind to a series of possibilities. He dismisses Robin Dunbar’s idea that language arose as a replacement for grooming when our ancestors lost their body hair. Bickerton suggests, probably correctly, that vocalizations without meaning would do just as well. He has similar objections to Dean Falk’s idea that language began with vocalizations to maintain contact when the mother had to put the baby down. (Apes don’t put their babies down because the infants hang on to the body hair.) But all this vocalizing without words can’t just be excluded because it isn’t fully functional language.

This blog is about the origins of speech, but Bickerton insists he is concerned with the origins of language and I’ve been going along with him out of a sense of fair play, but here we see where his approach leads. Language is an abstraction, and if you are going to look for its origins you are going to look for the rise of abstract properties. Speech is a behavior, something different from signing or writing, and if you are going to look for its origins you must look for the origins of that behavior. Obviously the rise of a series of ad hoc vocalizations that promote joint attention and emotional bonding is going to be of interest to anyone concerned with the origins of speech, while it will be overlooked by anyone focusing only on the abstract properties common to all forms of languages.

I believe Bickerton would say the vocalizations don’t matter because in the end there has to be a time when speakers begin to use words and combine them. The behavioral forerunners of that moment are irrelevant; a divide to meaning had to be crossed. The retort sounds reasonable, but vocalizations without semantic meaning may change the way our ancestors moved across the divide.

Bickerton says in another rule, that the selection pressure for crossing the divide had to be strong. That is probably true if the crossing came out of nowhere, but suppose emotional but meaningless vocalizations were going on and were a normal part of a population. It might not take much for an individual and a vocalization to become commonly associated. For example, an infant might associate the vocalization mama with its mother, and then a bit later the mother might notice that the infant uses that sound for her. In that case a vocalization would have become a kind of a word and other changes and responses to it become possible. So there we would see the human lineage being shaped by its speech even before it had anything Bickerton would recognize as functioning protolanguage and without requiring some strong natural selection pressure to get it going.

Did it happen that way? I don’t know, but Bickerton has no justification for ruling it out even before the motion is heard.

Bickerton’s confidence that he already knows what’s what explains why his book falls so flat, despite the zest and impishness of its prose. From the outset of his search Bickerton expected to learn nothing major about either language or humanity. Right up in the introduction he tells us he wants to persuade the readers that language is what makes humans human. He adds that he has been convinced of the truth of this lesson for years.

Frankly, it hardly matters what the take-away message is if the author has been convinced of it for years. If the scientist/storyteller’s greatest excitement is not over what he has learned on his quest for a bit of knowledge, how can he expect readers to be excited by what he learned? (A fine example of telling the story in the proper way is Bickerton’s own memoir Bastard Tongues  (see: From the Mouths of Babes) which describes Bickerton’s search for the origins of pidgin and creole langages. There he didn’t know what he was going to find, and was quite excited by his discoveries.)

Bickerton was driven to wonder about language origins by a technical paradox that emerges from the collision of generative grammar with the theory of evolution. Biology teaches that new things build on old things, but linguistics teaches there are no old things for language to build on. Even that puzzle was solved twenty years ago when Bickerton proposed that before there was a true language there must have been a protolanguage. Of course, that solution is a bit like finding out that below the tortoise supporting the world on its back stands a proto-tortoise. Where did it come from?

It is that level of problem—where did the protolanguage come from—that most engages Bickerton, and his most exciting moment is when he finds in niche-construction theory a possible way around the need for precursors. But then that idea too fails to charm because it is merely a way to outflank the question of how language could emerge from non-language. It doesn’t surprise because we learn nothing about the thing that ordinary readers care about. The nature of language and of humanity, remain unchanged.

I suspect Bickerton can shrug off that complaint with the thought, I’ve solved the puzzle, sorry if you wanted something more. And perhaps if he had presented this solution 20 years ago when the matter of evolution and language was first coming to a boil, he could have gotten away with it. But by now the inquiry has shown the insufficiency of both the generative view of language and the selfish-gene theory of humanity. It is time for a wowser.

Comments

[Yair Shimron]

You write "but linguistics teaches there are no old things for language to build on". It is the "revolution" of Chomskian linguistics that teaches this, while historical linguistics teaches quite the same principles as biology.

[uzza]

On this issue Bickerton has concluded “there must have been” a P-L, then made a career out of searching for real-world examples to support his conclusion. Every example he's proposed over the years—apes, children, deaf, pidgins (a structureless pidgin is an oxymoron)—has been shown to possess the structure he claims it lacks. It's difficult to take seriously his latest attempt to avoid saying “there must not have been”.

[tndal]

Lots of animals groom each other and many species "vocalize" to their young. If Dunbar or Falk were correct and Bickerton was wrong then birds and squirrels (indeed most animals) would have language just like us. The world would that of Doctor Dolittle.

We're looking for a unique situation that occurred in our species but not in others. A situation that allowed man to take that first step away from the here-and-now inside which other animals are trapped for their entire lives.

Note re uzza's comments:

Bickerton doesn't say that children and the deaf have language without syntax. He claims that children have syntactical ability built-in, that deaf children can learn syntax if they are exposed to signing sufficiently early, that pidgin languages are languages composed primarily by adults who must cooperate but whom do not share a language and that children, when taught a pidgin language will re-engineer that pidgin language to create a fully syntactical creole language.

Further, he states that any given pidgin language is actually a continuum of languages primarily based on historical time: the oldest version being spoken by the originators and later versions by those who learned the language later. Later versions of the pidgin are more complex, have larger vocabularies, more features and eventually over time (as children learn the pidgin) produce a creole language.