There is certainly a coming together of many related ideas which is very exciting, but there are some implicit assumptions in that “convergence” that blur some details that may not actually be right in any of the three fields.
This post is to record a position. The linked paper ….
“CONVERGENCE of Neuroscience, Biogenetics and Computing
– a convergence whose time has come.”
by Dr Michael Brooks
… is part of a series linking the work of Dan Dennett on the computational aspects of evolution, Craig Venter on the digital informational aspects of genetics and David Deutsch on the fundamental nature of information in physics. I’m a fan of all three, and have referenced their works multiple times in this blog, but I believe there are a couple of traps to avoid in the rush to converge:
Information & computation – the manipulation of information with other patterns of information, in real or virtual “machines” – is a very fundamental process. Information is simply “significant difference”. Possibly more fundamental than physics itself as currently understood in the standard particle model(s).
Mind & brain – cognitive sciences generally are right to see Mind & Brain as a “computer” – that is as a “machine” that does computation, but clearly it’s important not to fall into the trap of thinking of machine here as a physio-mechanical device. Computation is a many layered process, and when it comes to the computer itself, distinctions between hardware and software need not map simply to the brain and the mind. Information and computation processes are fundamentally independent of any physical substrate in which they may be represented. Independent of the substrate notice, not just independent of their representation.
At that level, avoiding the trap of over-simplifying the hardware-software view, there is lots of scope for careful work to bring these ideas together. But there is a second trap to be aware of before looking at the convergence with Genetics. That trap is accidentally assuming the digital nature of what is being considered. And there are two sides to this trap, both to do with digital objectification – one that genetics is necessarily digital, two that the computation is necessarily digital.
Genetics – is real, and it really is about information encoded in molecular patterns of bases in DNA. However, the objectification of those significant patterns as “genes” with distinct boundaries and clear definitions is part of the ontology of bio-genetic science. Useful to the science but not fundamental to the information patterns – there are a lot of fuzzy edges and apparent trash in between. We have a useful digital model of genes, but the genetics – the significance of and manipulation of the information – are not necessarily digital.
Furthermore, this same trap also exists in the Mind-Brain convergence too. There is nothing above that says either of these concern digital information. We tend to think of physical world computers as familiar digital computers, and whilst there is excitement about potential growing realisation of quantum computing, non-digital computing is actually as old as analogue computation – I know, many years ago I used to do it for a living.
In the famous Registry Assembly Programming case, the exercise is indeed fundamentally digital, and yes, it does illustrate the fundamental nature of computation. How can computation not be fundamentally digital?
What that exercise does show is that basic computation steps lead to complex processing – any unlimited sophistication – only by their combination. The underlying processes remain very simple, even when higher level languages and tools are used. The integer registries in the RAP case are themselves a representation of the information, which further represent (human) semantics. The model – the ontology – is a digital abstraction, but the information need not be.
You might argue that even in analogue computing, there are still digital particles involved – individual electrons in the electrical currents and voltages, or water molecules in the physical flows and levels – but as already noted above the information is (may be) more fundamental than even the particle physics.
Food for thought and a fascinating topic.