The Origin of Speech is the title of an article published in Scientific American in 1960 by Charles F. Hockett. It is often cited, even today, for its suggestion that there are 13 "design-features" shared by all languages.
The 13 design-features of language:
- Vocal auditory channel – While it is true that we have cultural adaptations that don’t require vocalizing communication – sign language, writing – nonetheless evolution has emphasized our vocal capabilities.
- Rapid fading - see next point.
- Broadcast transmission and directional reception – these two are inevitable corollaries of language's vocal component and the physics of sound. They refer to the fact that sound is non-permanent, and has spatial limitations and identifiers.
- Interchangability – if a person can understand an incoming message then they can also communicate it outwards. Compare the communication between a mother and infant, or the asymmetry of many mating rituals.
- Total feedback – the speaker is aware of what they're saying. This contrasts, for example, with insects that do not see what they are presenting in their colour displays, but nonetheless have the instinct to display them.
- Specialization – the communication is for its own sake, and not a by-product. A panting dog is communicating that it is hot, but that is only a by-product of expelling heat.
- Semanticity – the effect of language depends on its linking words to representations.
- Arbitrariness – the relationship between the word and its meaning is arbitrary.
- Discreteness – this is easiest to understand with an example: if I pronounce "bin" similar to "pin" I am not saying a word which lies between the two (as if language were continuous), but rather merely mispronouncing "pin". Hence words are discrete.
- Displacement – capacity to talk about things that are non-present, whether temporally or spatially.
- Productivity – capacity to create an infinite number of phrases from a limited pool. By comparison, the author notes, gibbon communication is limited and so does not display productivity.
- Transitional transmissions – transmission that is non-genetic, even if the capacity is genetic. The author notes that with other animals it is not clear how much their language is genetic and how much learned from their environment. Today, for instance, we know that some birds undergo genetically-induced periods of neurogenesis during which they take up (viz. "culturally") new songs.
- Duality of patterning – the limited sounds that make up language are meaningless though they combine to form meaningful words. Consider the words "cat", "act" and "tack", which are each made of the same three sounds.
The author suggests that the last four are the most important for understanding human language. The paper also contains a comparison of human language to other species' communicative capacities.
This is an excellently fun read. First of all because it’s from 1960, and is so assured – as we are today – of the progress made by the scientific community in its day. Hackett talks about recent progress, especially in human evolution. But what did they actually know? You’ll forgive me my temporal myopicism if I answer "basically nothing". The genetic code was yet to be cracked, the human genome project wouldn't be drafted for another fifty years, and if you’re going to talk about making strides in understanding how we got here, well it was only 2010 when we sequenced the Neanderthal genome and found out that we’d spent a good portion of our pre-history in bed with them. Molecular biology as we know it didn’t exist, let alone exciting new fields which are just beginning to explode like systems biology and synthetic biology. It's all happening now.
Patronizing aside – it's worth reading the first two pages just to grasp the problem of studying the origin of language. As Edward Sapir had already noted, we're not always sure if a society has art or religion, but we've yet to find one without language. There isn't much information out there on the origin of language, and it's not clear where we should be looking.
The thirteen design-features posited by Hockett don’t seem to justify what he says in his introduction. In his introduction he says that now, more than ever before, we're in a position to ask about the origin of language. But in what sense are his design-features modern? At most, the "modernity" of his approach is limited to his incorporating the fact of evolution into his model, placing importance on the understanding of the communication of proto-humans, other primates, and other animals, for understanding modern human language.
Do I agree? No idea, but I do have some opinions. It feels like Hockett's approach is viable – look for commonalities necessary for all languages. The most obvious danger that should be avoided is unfalsifyability – it's tempting to define X as necessary for language, and then when challenged with a language which appears to lack X to simply claim that the language has been misunderstood. The actual terms used by Hockett, and their definitions, are sufficiently vague to be prone to this fallacy - so while I like his approach I think that the details need some major reworking.
Another approach, and one which Hockett was in no position to appreciate, is the modern neuroscientific approach – attributing functional capacities to neuronal processes – in conjunction with powerful new molecular tools such as optogenetics (creating transgenic animals that have selected neural pathways that can be turned on/off with light). Guess we’ll see.