I thought I would re-read a classic article titled “The Origin of Speech” by Charles Hockett and published in the Scientific American’s September 1960 issue, to see what has changed and what survived of the old way of thinking. Much has changed. Hockett was part of the age when behaviorism dominated psychology, when the difference between animal communication and language was still hard to define, and when evolutionary explanations tended to focus on the benefit to the species as a whole.
Yet the article is still worth learning from because it is based on what Hockett called the “design features” of language. His list of 13 features is still good, although some people might ask so what. The biggest change, perhaps, is that the list is no longer taken as much of a guide to the origins of speech.
Hockett said some of the features were common to most or perhaps even all land mammals: (1) use of the vocal-auditory channel; (2) rapid fading of signal; (3) total feedback (i.e., signal sender is aware of the signal, compared with say a fish that cannot see its own signaling body markings); (4) interchangeability (signaler and receiver can change places); (5) broadcast transmission and directional reception (signal travels in all directions but receiver can tell where the signal comes from).
Then come some features that are perhaps limited to primates: (1) semanticity (signal means something, i.e. danger); (2) arbitrariness (meaning not implicit in signal itself); (3) specialization (the signal is only a signal and not some side effect of other behavior).
Next came some features that Hockett suspected were available to pre-humans: (1) discreteness (meaning distinguished by a series of separate (discrete) sounds, e.g. with the sounds p-, t-, -in, -un we can create a set of different words: pin, tin, pun and tun); (2) traditional transmission (grammar and vocabulary is culturally rather than biologically based).
Finally, he listed three features he thought only available to humans speaking modern languages: (1) displacement (the ability to speak of something absent); (2) productivity (the ability to say and understand things never said before); and (3) duality of patterning (the overlaying of a code on top of meaningless carriers).
The example Hockett gave of duality of patterning was Morse code. The carrier is a series of dots and dashes, meaningless in themselves, but specific arrangements of them indicate letters. A few years later he probably would have chosen the genetic code as his example. A DNA molecule consists of a long strand of four seemingly randomly placed nucleotides that carry instructions for the fabrication of proteins. Duality of patterning is the secret for producing an unbounded series of intelligible messages, or as Hockett calls it “productivity.”
Surprisingly, however, Hockett believed duality of patterning was the last feature introduced, mainly because he saw no reason to evolve such a thing unless language was already fairly complicated. He got productivity of language before duality of patterning by saying new remarks took pieces from old utterances and assembled them according to familiar patterns. So Chomsky’s time has not been entirely wasted. He has produced many sentences that seem to match patterns, but not meanings. For example, somehow we know the verb has a different meaning despite the matching of sentence patterns: I ran the hotel and I ran the race.
Hockett’s article appeared late in the Age of Behaviorism, just as Chomsky was publishing his review that scuttled the ship. The premise of all behaviorists was that everything an organism does (including speaking or writing) depends on previous inputs. The organism can contribute nothing truly original as an output. That idea was generally attacked by humanists who believed people could be creative and was soon joined by cognitivists (like Chomsky) who believed in the possibility of some internal computation that did not rely purely on a stimulus. Living as we do in a world of computers, the behaviorist premise that denies internal processing seems untenable.
Hockett also believed that language was an extension of the communicative powers of animals. This point is still accepted by many. It is common for texts to begin their study of language by going over animal communications and different monkey cries. Chomskyites don’t accept that because they do not believe language’s primary function is communication. On this blog I take as a given a remark by David Leavens et al, “A defining characteristic of the human species is our capacity to rapidly establish topics of mutual contemplation.” Language can be used for animal communication, but it has unique features because it has a unique function: the sharing of topics for mutual contemplation.
The shocking nature of this function would not have been readily recognized by Hockett in 1960 because evolutionary theory had not yet been radicalized by William D. Hamilton and his theory of kin selection and inclusive fitness (i.e., “the selfish gene”). In those days it was common to read of behaviors that supposedly benefitted the species, even at a cost to the individual actor. Gazelles, for instance, jump up and down when they espy a potential danger. This activity, known as stotting, was said to be a signal of danger to other gazelles, and was considered a sacrifice, since the safest thing to do would be just run away without bothering to alert one’s neighbors. That sort of explanation went out the window when Hamilton came along. Suddenly, altruism became the mystery. There has been some effort to show that altruism does not really exist among humans, but it has failed. Plainly, human communities demand both trust and trustworthiness, whereas ape societies, for all their closeness, do not. The story of human language origins, therefore, is not just a tale of the rise of distinctive communicative features. It is about the rise of a communal species dependent for survival on sharing perceptions with neighbors.
What then is the value of reading Hockett 56 years on? Besides the fact that it is always refreshing to encounter an active, creative intelligence, there is also Hockett’s list of design features. He lists five features that are universal amongst languages and exclusive to humans. During the whole of the period since Hockett’s article, Chomsky et al have been trying to assemble a universal grammar and they still have not produced a list that is as clear and universal as Hockett’s. If you want to know what is technically distinctive about language, read through that fine old piece.