The Psychology of Communication


5.3 Language Origin And Structure - Robin Dunbar

The traditional argument for the importance of language is that we needed to communicate to hunt large animals. In his book, Grooming, Gossip, and the Evolution of Language, Robin Dunbar argued that we also needed to communicate because large animals hunted us [DUNBAR]. As relatively small, slow, and puny animals, we banded together for mutual protection. Clustered in groups we had many ears and many eyes to warn us of danger. Thus trust, essential for social animals, was established within a group by mutual grooming. He argued that language evolved as a sort of grooming-at-a-distance strategy when the optimal group size got too large for direct grooming. Gossip too served an evolutionary function because it can lead to a bad reputation and thus ostracizing of members of the group who can't be trusted.

Despite such attempts by evolutionary psychologists like Dunbar, the origin of language in our species is still shrouded in mystery. It happened very long ago and it has left no physical record. The absence of evidence does not, however, prevent us from having theories about it. Indeed, it seems that the fewer the facts, the more the theories. Speech evolved as imitation of sounds heard in nature (ding-dong theory), as imitation of sounds made by animals (bow-wow theory), out of interjections (oof-ouch theory), to accompany strenuous group activity (yo-he-ho theory) are a few of the candidates. Speculation about the origin of language was so rife and viewed as so futile that, in 1866, Société Linguistique de Paris banned any further discussion in their journals [DEACON, Page 14].

Perhaps it is more important to find out what is unique about human language than to find out how it started. An alternative approach is to identify the design features of language and determine which features distinguish human from animal communication. Charles Hockett lists three such design features of human language - displacement, productivity, and duality of patterning [HOCKETT]. While I was a graduate student at Cornell University, a campus debate developed between Charles Hockett and Karl von Frisch, who had conducted extensive research on the "language" of bees [VON FRISCH 1950]. He had discovered that a bee could communicate the source of pollen to other bees in the hive by doing a dance in a figure-eight, in which the angle of orientation of the 8 indicated the direction and the number of wiggles in performing the figure-eight indicated distance from the hive. That is, it passed on the polar coordinates of the pollen source. Von Frisch argued that this "language of the bees" had the design feature of displacement - the bee could "talk" about things which are not here and now - and productivity - the bee could "say" things which have never been said before, when it gives precise polar coordinates never before used by its species.

Hockett argued that such communication between bees should not be described as "language" (hence the inverted commas around language in his title). It was pre-wired into the genetic code of bees - that is, it was genetic not extragenetic. This argument was vindicated by later work by von Frisch himself on dialects in bees. When North American bees were mated with European bees with a different "dialect", the sons of bees could not communicate with either parent, since their "language" was some compromise between the two dialects [VON FRISCH 1967].

In my introductory psychology textbook, I included a footnote to the third distinguishing design feature of human language - duality of patterning - which stated that I didn't understand this feature [GARDINER 1970]. In the second edition, I added a footnote to this footnote, in which I stated that I had talked to Charles Hockett about this feature and I still didn't understand it. You'll be happy to know that I now finally understand it. Language is a hierarchy of units plus rules for combining units at one level to create meaningful units at the next level. More precisely, language consists of phonemes (roughly equivalent to the letters of the alphabet), morphemes (roughly equivalent to the words in the dictionary), sentences, and discourses, plus the rules of vocabulary to combine phonemes into morphemes, of grammar to combine morphemes into sentences, and of logic to combine sentences into discourses (see Figure 5-1).3

Thus language has semantic rules, which link it to the objective world, plus those syntactical rules, which link its elements together. This duality of functioning is unique to human communication. Humberto Maturana makes this point more concrete by describing a dialogue with his cat [MATURANA & VARELA]. Normally the cat says "meow" and he feeds it. They communicate. He has forgotten to get the cat food. "Meow" no food. "Meow" no food. With it's third "Meow", the cat can't say "I've said 'Meow' two times before and still no food". They can't communicate about communication. They both know the semantic rules but only Maturana knows the syntactic rules. In Chapter 11, we will discuss the two distinct areas of the brain responsible, respectively, for the semantic and syntactic aspects of language.

3   I say "roughly" because there is not perfect phoneme-grapheme correspondence and there are some morphemes - the smallest linguistic unit which has meaning of its own - which are not words. For example "ed" at the end of a verb means past tense.