Recently Adam Benton posted an article " Fear drives technological evolution" on the Evoanth blog which started with the above picture. It shows some of the fanciest stone age tools our ancestors made; all of which were invented after our brain growing. He then asked: "What caused their development?"
Your post misses an important point. If an individual is innovative all is lost unless the innovation can be passed on from generation to generation. So in evolutionary terms you only need a small number of innovators and a much larger number of efficient copycats. If a prehistoric inventor came up with a good tool he may have no personal descendants but the rest of the clan, who copied it, would have increases survival prospects. If fear is involved it is far more likely to be among members of the group benefiting from the invention who realise that if they don’t accurately copy the tool they will no longer be able to benefit from the advantages it gives. (The person who first discovered how to light a fire may have done so by accident – so fear was not involved. However many of his successors may have been terrified because their fire was going out and they had not learnt to copy how to relight it.)
The important thing to realise is that a human clan which is good at copying successful ideas (which may not always be their own) is going to be more successful, in evolutionary terms, than a clan which includes many innovators but which is not so good at passing inventions to the next generation. This means that populations that are good at copying useful inventions will grow and the bigger the population grow the smaller the percentage that needs to be inventors.
As long as there are a small number of innovators the evolutionary pressures will be to favour accurate copying. A major evolutionary pressure would be, for a juvenile, to correctly identify the most suitable adults to copy. This may well explain why now we so readily follow charismatic individuals – be they well-spoken politicians, leading sports stars, brilliant scientists, or spiritual leaders (both living and long dead).
I am working on a very different model to explain why, perhaps 150,000 years ago we suddenly became more innovative. I start with a deliberately very simple neural net model and ask how it might evolve, remembering that evolution does not design ahead – and is very good at converting unpromising starts into successful species.
The starting point assumes that (1) in any pattern recognition system the time taken in learning a pattern increases rapidly as the complexity of the pattern increases; (2) the patterns learnt are lost when the brain dies; (3) the size and “intelligence” of an animal brain is limited by what it can learn in a lifetime.
In effect this puts an upper limit to animal intelligence – except that humans appear to have broken through it.
The breakthrough is to realise that to make complex tools one needs a rule-based system to describe each step in the process – and the more steps the more complex the learning process becomes. The neural-net model I am proposing may seem simplistic but it can morph first into a concept based set processing system and then into a rule based system. Under normal circumstances (i.e. in animals) the barrier blocks progress because learning very complex tasks involving sequential rules would take more than a lifetime.
The key is the development of a simple proto-language. Because my model morphs to handling concepts it is possible to name objects. When an infant is told the name of an object it effectively “speed learns” the name to memory. As language grows it reaches a critical point where it can begin to describe rules for making tools and handle generalisations. At this point the following factors come into play:
(a) A simple language which the infant “speed learns” allows more cultural information to be transferred between generations faster and more accurately.
(b) Teaching generalisations means less time teaching and using less memory. (Learning the concept “mammal” and a list of exceptions is better that learning about hundreds of different mammals and only afterwards working out the commonalities.)
(c) The ability to teach simple tool-making rules makes it far easier to transfer toolmaking skills between generations, and far easier to invent new tools and extend existing ones.
(d) Language is itself a tool, which can be modified and extended generation by generation – and with each extension the processes described in (a), (b) and (c) become even more significant.
What happened circa 150,000 years ago was that language reached a tipping point where it suddenly broke through the “animal intelligence” barrier, and we switched from evolving a bigger brain to evolving transferrable cultural intelligence which does not require the brain to be any bigger.
Another change happened circa 10,000 years ago with the coming of civilization. As we settled into communities there was a division of labour – with specialists in various trades and children no longer needed to learn everything needed to live because someone else was there to cultivate the food, make the tools, or protect from dangers. At the same time teaching methods have improved to the point where we have time to explore areas such as the arts, literature, science and music which have little relevance to the future of our genes.
In fact the chances of an individual successfully breeding is now more dependent on the culture they live in than on genes that contribute to the size of their brain. In such circumstance one would expect factors such as genetic drift to come into play – which could explain why average brain sizes appear to be getting smaller since we started the process of urbanisation.
Interesting my model not only suggests reasons for the technological breakthrough circa 150,000 years ago, and our currently shrinking brains, but it also suggests reasons for confirmation bias and the unreliability of our long term memory.
The real problem I have is that the model is counterintuitive. Humans like to think we are important and in the past we thought we lived on a flat earth in the centre of the universe. My model suggests that, apart from size, the biggest difference between our brain and other vertebrates is that we are better copycats. I can find no other examples of research which have come up with an evolutionary model that suggests a predictive pathway between a simple neural net and human intelligence. Perhaps it is because everyone takes it for granted that human intelligence is so marvellous that it could not possibly depend on a mathematical model which is so crude that a typical 10 year old could find fault in its logic.
There are many cases where we are happy to point out Nature’s “design faults” – and marvel how millions of years of evolution have somehow found a work-round. Perhaps it is now time for us to climb down from our “supreme intelligence” pyramid and admit that at the genetic level our brain (although bigger) is no cleverer than our primate cousins. What we like to consider makes us different is due to our mountain of culture and the fact that, through culture, we have developed more effective methods of teaching.