Aristotle taught that all knowledge comes through the senses.
Part I summarized Chomsky's theory of internal language and said that an alternate theory is possible. In this part I present that alternative.
Alternate Theory: Community
First, we can say there are several fundamental differences between humans and other primates, not just one.
- Humans can have parents from culture A but, when raised by members of culture B, they act like members of culture B. Chomsky often mentions this point, yet he never puzzles over what a strange fact this is. If you try to raise a wolf as a dog, you get a wolf. Birds who try to raise a cuckoo as, say, a robin, get a cuckoo. Yet families in Paris who raise babies born to natives of New Guinea get Frenchmen. This suggests that while most animals are born with a set of instincts to make them like their ancestors, humans are born with instincts to acquire the ways of their neighbors.
- Human communication is triangular and conversational while animal communication is a monologue. If a rhesus monkey sees a leopard, it lets out a distinct call. When other monkeys see the leopard, they join in making the same call. When birds sing a territorial song, other birds of the same species may avoid the area or sing the same territorial song. When fire flies signal their presence, potential mates approach. In none of these cases does anything like a discussion take place. The basic human arrangement of speakers and listeners discussing news about topics is nowhere to be found. Yet it is universal among humans.
- Human males take on social responsibilities. I do not wish to get too excited about the patriarchy, but, compared with most species, human males are not so terrible. Even an extremely social species like the African elephant never found a solution to the problem of males beyond driving them out of the herd. Somehow human males became willing to take some responsibilities for some women and children.
- Humans have moral codes which set rules on how to behave toward one another.
A theme is apparent in this list: humans are more than social, they are members of mutually dependent communities. We are, by nature, members of a culture that tells us who we are and how to act. Thus, the reason for suspecting that language's primary function is communicative is that we cultured animals need a tool like language to survive. We needn't ever have an original thought in our lifetime, but we must learn the ways of our neighbors.
This observation rejects one of Chomsky's most controversial conclusions:
- The primary function of language is thought, not communication.
Alternate Theory: Perceptions
A second part of an alternate theory might be that elementary concepts come from perceptions, not the other way around. Chomsky likes to argue that we have an innate concept of a river, and that idea allows us to recognize a real-word instance such as the Charles that flows by MIT or the Hudson that flows a few blocks west of my apartment. Thus, the person who lives an entire life in a Saharan oasis may never need to use the concept, but a descendant raised in Paris will find the concept ready to take off the rack and give meaning to the Seine. An alternate account might be that we get our concept of river from our experiences with them and that the oasis-born Parisian knows what the Seine is because it is there to be perceived.
Should we flip a coin to decide which theory is correct? Better might be an experiment. Oh, wait. Such experiments have already been made. The psychology world is full of experiments in which rats and pigeons are trained to use things their lineage has never before encountered, things like light switches, electronic locks, pulleys, etc. Humans might have gained these concepts by sticking more basic concepts together, but Chomsky tells us that animals cannot do such Merging. So we need some other explanation than innate prejudice for how animals come to make use of lab equipment. One alternative is that they have somehow learned from their experience with the gizmos.
Perceptions do not completely abolish the notion of inborn prejudices, but the biases do differ in their content. Instead of having concepts like rivers, perceptions assume relationships.
Imagine a zebra and a lion. The zebra's survival requires it be able to tell when a lion is charging it. The zebra's efficient survival requires that it be able to tell when a lion is moving but not coming. This subtlety of distinction means that zebras do not run every time they see a lion. Zebras seem to be able to perceive not just objects but a relation between the object and themselves. It would be most charming to conduct a series of experiments to determine which relationships animals can perceive.
As it happens, a number of perceivable relationships (called case relationships) are universal in languages. All languages, for example, distinguish between subjects and objects. English marks the distinction mainly by word order; many other languages change the noun to indicate case. Old English used case markers and a few such markers survive: e.g., he (nominative), his (genitive), him (accusative).
Although the experimental evidence is not complete, it would not surprise me to learn that chimpanzees can distinguish between subject and object, direct and indirect object, thing and part of a thing. These are the elementary case relationships and they are universally translatable. So, until the experimental evidence shows otherwise, I will assume that no special evolution was required to acquire the case relationships.
Apes have been taught to make sign language words, and the apes themselves have spontaneously stuck a few words together. Most notably, a bonobo once coined the phrase water bird to indicate a swan. So it seems that as soon as humans had the motivation to share perceptions with one another, they had the ability to draw a companion's attention to something. Words and short phrases appear to have been immediately available to them.
What we still have no evidence of is perceptual Merge in apes. Chomsky's Merge sticks elementary concepts together, but to believe in that process you have to believe in innate concepts. Perceptual merge unites different points of attention. A sentence like Jeter caught the ball focuses on two separate perceptions: (1) somebody named Jeter, and (2) a ball. They are in a subject-object relationship and are held together by a binding word, caught. I call it a binding word because it ties the two perceptions into a complete thought. The binding word can go with either point of attention: Jeter caught … shows Jeter doing something; … caught the ball shows something being done to the ball. So there is some kind of perceptual Merge capacity in humans and, apparently, humans alone.
This part of the alternate theory rejects a few more of Chomsky's conclusions:
- Thoughts are applied to perceptions and not vice-versa.
- All uniquely human thought is symbolic.
- The concepts never refer to actual things.
Alternative Theory: Meaning
Chomsky’s Merge is a computational process, which is to say it is automatic, determined and reflexive. Merge sticks the concepts together and comes up with Jeter caught the ball. Perceptual Merge is different, relying on at least a crude recognition between separate points of attention. A complete thought identifies the topic (Jeter) and news (the ball), but it is up to the perceiver to determine what’s the topic and what the news. Every topic could be the news and vice-versa. For example, a person could say The ball flew straight to Jeter to describe exactly the same scene in which the ball was the news rather than the topic.
While switching subject and object around requires a simple inversion of case indicators, the change to the binder word requires some understanding of what the word means. Language in this alternative theory works partly at the conscious level. There are unconscious processes at the level of phonetics, word choice, and grammar, but organizing an utterance to express a complete idea calls for a conscious awareness of the topic, news, and binders’ meanings.
So we can also scratch out:
- · Meaning is less important than syntactic structure in sentence construction.
Alternative Theory: Multi-Step Evolution
This alternative theory implies a multi-step evolution, First came a motive for sharing perceptions, and this motivation required a change to a group dependence more profound than anything currently known in the ape world. The ability to generate words and phrases drawing attention to specific perceptions was immediately at hand. Other steps included the ability to express case relationships and use metaphors. Another step was the development of a working memory that lets us keep track of topics in an utterance like The man my mother saw left town. Thus, we can also cross out:
- · Language is an all or nothing phenomenon.
I'll finish this essay in a future post.
Comments