Please see Computing Services' blog post for more details
The Social Origin of Language
This page presents the form of an argument. It has links to references
and spaces for anyone to comment on the argument.
Human language can not have evolved by adaption of animal communication
1/ All instances of communication between animals can be situated somewhere on a spectrum between the two extremes of a) for self interest b) for the group interest
2/ In so far as a signal serves the self-interest of an individual animal, its existence must depend on its cost.
Evolution focuses especially hard on these signals
Because they have a direct link to reproductive success
If an animals can exploit the signal to its advantage, it will evolve to do so.
Because if it doesnt it will lose to others who do.
What if no animal tries to exploit the signal?
Its theoretically possible to imagine this but:
it doest match any animal observed in nature
such a species would be far too boring to evolve language
In addition non-human primates have been observed in deceptive behaviour
What is the consequence of such exploitation?
Other animals can not afford to be naive in responding to the signals
Because there is a danger that they will fall victim to exploitation
So what happens?
They require a guarantee of a signal's honesty before they will pay attention to it.
Cant a signal work without a guarantee?
No, the only possible outcome is somewhere on the spectrum between these two extremes:
1/ the signal is ignored by the listener, and thus falls out of use
2/ the signal is followed by the listener faithfully until there are no more such listeners in the population
either way the signal extinguishes itself.
So a guarantee is required and the only one available is the signal's cost.
Why is that the only guarantee, couldnt there be lots of others!
No, only cost will serve because the value is built in to the expression of the signal itself
so it cannot be faked.
What about choice of behaviour in aggression (Enquist), that is a signal which doesnt depend on cost!?
That is true, but this signal is guaranteed by the causal link between the choice of behaviour
and the animal that makes it. So once more, it is built-in to the use of the signal. If in an aggressive encounter an animal does something that suggests
confidence, this reliably communicates to the opponent. I expect that in many situations animals can
'read' an inference from another's behaviour, so strictly speaking, this is communication. However the
corroboration this requires is at least as restrictive as that in costly signalling
All other guarantees, as they are not built-in, are liable to the same exploitative intelligence described above
Well why is it a problem if the signals must use cost (or other corroboration)
Because this does not look like an adaptive route to human language, it looks like a development
in the opposite direction.
the range of expression open to a language governed by cost (or other corroboration) is severely limited in contrast to human language which is infinitely expressive
Why is it limited?
Because you can only express in cost what the signal can physically represent. (And you can only express via causal corroboration something about the physical nature of the animals situation) So you can express things like 'strength' 'agility' 'intelligence' "confidence"
But couldnt they have started from such a simple beginning and then evolve via adaption to a fuller language?
No because as soon as the signal extends beyond what corroborates it, it will be cancelled out.
So the range of expression is extremely limited and, as I said, the expressive power of human language is infinite.
Prove its infinite.
Language is combinatorial, recursive, and allows for the creation of new words = infinite
And anyway human language does not depend on cost (or other corroboration) for its communicative ability
To express something and then express its opposite requires no change in cost. This means
language cannot be employing the mechanism of costly signalling to communicate.
So are you saying language didnt evolve?
No, it did evolve, but not via adaption of animal communication, in which case it must have been via a discontinuity.,
Couldnt an increase in cognitive ability have overcome the problems described above?
No, and in what is one of the most devastating criticisms of the adaptive case; the opposite is the case! once a certain threshold of cognitive ability has
been passed, any further increase mitigates against the development of language, not in favour of it!
Because if an improvement in the ability to exploit leads to exploitation episodes occurring in areas of signalling which had previously been safe
the use of the guarantee of cost must correspondingly increase, so evolution moves further away from language not closer to it.
Couldnt cooperation have solved the problem?
It depends what you mean, if you mean cooperation in the widest sense of the word: including novel cultural order, then the answer is yes. This is
the discontinuity mentioned above.
But, if you mean 'cooperation' in the sense usually meant in cog sci, the answer is no
It is not the lack of cooperation was not an obstacle to the development of language, on the contrary
the encephalization which is presumably a pre requisite for language was driven by highly competitive
complex social reactions
What about proof that cheaters can be beneficial to the group? Craig McClean's yeast model
Craig Maclean's model shows that sometimes the group can accomplish a task better if it includes some cheaters.
But my assertion is not that cheaters damage group fitness, but the opposite: the level of cheating is maintained by its selective advantage for the group or individual
but while its adaptive for the species, its counter-adaptive for the emergence of language.