Every so often a new development in digital hardware or software is heralded as the beginning of artificial intelligence. And every time, I look at it and see a machine - a machine designed to store and manipulate symbols (data) according to a set of rules - that works a little faster and more efficiently than its predecessors. What I don’t see is intelligence. The desk-top on which I am writing this article can’t do anything that the Apple II I used in the 70s couldn’t do. It just does it a whole lot faster and more effectively.
The problem, I think, is that the word “intelligence” is tossed around without much thought into what it means. Partly that is because we can’t decide what it means. The definition of intelligence includes aspects like the ability to learn, be creative, apply logic, think abstractly, communicate efficiently, remember information, or merely to be educated.
When we talk about human beings, intelligence usually in practice simply means academic ability. When applied more broadly in Zoology, it tends to mean complexity or flexibility of behaviour, particularly in response to novel stimuli. But intelligence also implies something else, something more fundamental, an awareness of self and consciousness. All vertebrates seem to have a well-developed sense of self, other organisms less so, computers never. It may be possible to simulate consciousness using computer software. Equally one can simulate rain—but no one gets wet.
So in summary, there appears to be qualitative differences between vertebrates and complex machines, notably self-awareness and consciousness.
Broadly speaking, suggestions about the mechanisms that might cause consciousness can be grouped into three: (i) the vital spark of life, (ii) an emergent property from basic control functions of the central nervous system and (iii) a special relationship between spooky quantum interactions and life.
Let’s start with category (i). Nowadays, “vital spark” concepts of consciousness tend to fall into the supernatural realm and so are outside the scope of this article—but that wasn’t always true. The mechanistic “clockwork” view of the universe devised by scientists such as Galileo and Sir Isaac Newton, and engineers like Jacques de Vaucanson with ingenious automata like the Flute Player, gave rise to a mechanistic model for living organisms. After all, had not William Harvey shown that the heart was simply an organic version of a mechanical pump?
However, there still seemed to be some qualitative difference between the animate and the inanimate, not matter how cleverly the latter was fashioned, some vital essence that only life possessed that was forever denied to objects constructed solely from the mineral.
Science soon provided a candidate for this vital spark. The concept of energy was one of the great triumphs of classical physics. Energy clearly interacted with material objects, doing things like keeping the planets in their orbits and powering steam trains, so it was not too great a jump to envisage some sort of bioenergy powering living organisms.
In 1845, Baron Carl von Reichenbach, no relation to the famous Falls as far as I know, introduced the concept of Odic Force, named after the Norse God, which was supposed to be the biological equivalent to energy. It still lives on in various mystical and pseudo-science philosophies such as “auras”—and in a couple of stories I wrote for Baen, “Storming Hell” and “Storming Venus.”
Electricity itself came close to replacing Odic Force as the vital spark of life, and it at least had the advantage of being demonstrably real. Back in the eighteenth century, Luigi Galvani discovered he could make frogs’ legs twitch by applying an electric current to the muscles. His nephew, Giovanni Aldini, applied the same technique to an executed criminal at London’s Newgate, appearing to at least partly reinstall the corpse with life. Horrified accounts from the time refer to quivering jaws, twitching legs, clenching fists and, most horrible of all, the corpse opening one of its eyes.
(As an aside, I always seem to hear clashing atonal chords when I write something like that.)
This was a period where medical techniques were apparently raising people from the dead. The drowned could be reanimated by resuscitation and skilled doctoring could bring someone back from a condition at least indistinguishable from death. Dictionaries of the period divided death into two types, incomplete and absolute, and the nervous began to be buried in graves equipped with little bells that could be rung from inside the coffin.
All of which brings us to science fiction and Mary Shelley’s famous book, Frankenstein, published in 1818. We all know the story. Igor winds up the rod from the castle tower, lightning flashes all about while the mad baron cackles dementedly.
(More clashing atonal chords.)
Frankenstein, the kitsch gothic horror story, was actually cutting edge hardcore SF when it was written, tackling the science of the vital spark behind death and resuscitation. It was also a cautionary tale about the moral issues involved with artificial intelligence; issues that we still haven’t resolved in the modern world. The baron is doing way more than resuscitating a corpse; he is creating a new self-aware life where none had existed before.
Putting to one side moral considerations about A.I., which thankfully are a problem for another day, we are left with the fact that extensive research has failed to find anything resembling a unique bioenergy in biological organisms—notwithstanding any perceived qualitative difference between the living and the inert.
To turn to our second mechanism, the hypothesis that consciousness is an emergent property of the central nervous system has been a mainstream scientific concept since the vital spark fell out of favour. The idea of emergent properties goes all the way back to Aristotle. Essentially, it’s where a set of interacting processes produces a new property that is not found in any of the component bits. For example, life itself could be described as an emergent property of organic chemistry.
In this model, while the central nervous system evolved to coordinate internal body physiology (and the response of the organism as a whole to external stimuli), consciousness and self-awareness emerged as a by-product of the central nervous system becoming more complex as organisms became larger and more structured.
The processes producing a complex emergent property like life do not have to in themselves be particularly complex. All that is needed is a historical system (one that is not reversible—where the past governs possible outcomes in the future) and plenty of time. With enough time, a complex universe can be created by a few simple physical laws and a human being can arise from the limited chemical properties of carbon-based molecules.
Simple processes can also create complex emergent properties by replication—a sort of horizontal iteration through space rather than vertical integration through time. Plants, animals and fungi are made by replicating the basic building block of the single cell, large animals from smaller by simply replicating their bodies as repeated segments. The segmentation of ragworm bodies is obvious but actually insects and vertebrates also evolved larger bodies by similar replication—it’s not so visible but it’s still there.
The great Robert Heinlein explored the idea of a computer becoming self-aware in his 1966 novel, The Moon is a Harsh Mistress. The novel runs with an idea that was mainstream in computer cybernetics at the time which is that that more and more processes would be controlled by larger and larger centralised mainframe computers. Eventually whole cities, countries or even the world would be run by a single giant computer. The Soviet state, which with Japan and West Germany was one of the three fastest growing economies in the world after World War II, ran into severe problems when it tried to shift from a war-emergency economy that was mono-focussed onto defence and industrial production to one based on consumer industries that would provide the material rewards of the promised communist society.
Without the feedbacks and decentralisation associated with the consumer demand-led Western economies, the supply-led Soviet economy became increasingly out of kilter with demand. Shortages were followed by gluts in ever more extreme swings. The Soviet solution was to predict demand by cybernetics using “large” computers—actually IMB 360s—rather than introduce an ideologically unpalatable feedback system using market forces.
It didn’t work: the computers just weren’t capable of the massive calculations required within anything approximating to real time.
In Robert Heinlein’s novel the computer, Mike, becomes sentient because it has so many peripherals attached and it “dies,” becomes nonsentient again, when many of these peripherals are disconnected by damage during a war. But a giant computer is nothing like a vertebrate central nervous system. There are similarities in that both appear to reduce information to digital on/off (neuron fires or not) codes but there is no equivalent to a central processor in your head.
Nevertheless, the idea that a bigger brain is a more “intelligent” brain keeps resurfacing despite all the modern evidence to the contrary. It usually takes the form in the modern world that the more neurons or synapses a brain has the more intelligent it is. Human beings have a thousand times more neurons than a mouse so maybe that is why they are more intelligent than a mouse.
The problem is that brain size is largely a function of body mass and human brains are not statistically oversized for a mammal of our body mass. Also, accidental consciousness appears to be not a property of the central nervous system as a whole but of specific parts of the brain, notably the frontal lobes. Architecture rather than size appears to be the key factor.
Nevertheless, The Moon is a Harsh Mistress was cutting edge hard SF for its time and a worthy successor to Frankenstein, albeit with a more positive theme.
Updating the idea, one can postulate that an interconnected net of digital systems might be the place to look for the accidental emergence of an artificial intelligence but we have had telephones and now the internet for some time and so far no ghost in the machine has been detected. Animal life didn’t jump from nothing to the sophistication of human consciousness. Consciousness gradually builds up through the vertebrates. Of course, the internet still has considerably fewer connections than a human brain but nothing appears to be emerging and we would have expected to see signs by now if this hypothesis was correct.
As an aside, perhaps this is just as well when we consider the end result of the emergence of self-awareness in the Terminator movies’ Skynet.
To conclude we have no evidence that consciousness and self-awareness are caused by the number of links in a digital system.
So with a shudder, and the image of Frankenstein’s monster as robotic human-extermination machines fresh in our minds, we turn to our third mechanism—quantum interactions as the source of consciousness. When we speak of quantum mechanics applications to biology we specifically mean a direct effect of those spooky parts of quantum theory such as superposition and quantum tunnelling.
The idea that quantum spookiness might explain some of the more puzzling properties of living organisms goes all the way back to the birth of quantum theory itself. Pascaul Jordan began exploring the idea in 1929, suggesting that life was different from most physical systems in that a very few molecules seem to have a wholly disproportionate control of the organism as a whole. Unfortunately, Jordan’s ideas became discredited by his claims that they proved the superiority of the Nazi Führerprinzip—not a wildly popular view by 1945.
Schrödinger, who had to flee Germany because his wife was considered a subhuman, published a book in 1944 called What Is Life? This work is probably the real start of quantum biology. He suggested that the units of heredity (their composition unknown at the time) replicated too efficiently to be obeying classical science and so must be molecular and subject to the laws of quantum mechanics. In doing so he predicted some of the characteristics of the theoretical units of heredity before DNA’s structure became known.
Schrödinger’s key insight is that living material displays one fundamental difference to nonliving material, namely that living materials are highly structured right down to the molecular level. The reason this is important is that spooky quantum effects only occur at the smallest level of organisation. The fact that living organisms are structured from the molecular level upwards means that quantum mechanics potentially can have impacts right up to the macro-level. DNA itself is an obvious example of how a highly structured molecule influences the whole organism.
However, there are good practical reasons why quantum processes should not apply to large, warm objects like a living organism. And these reasons are connected with superposition and decoherence.
I am not going to try to explain these terms in detail in this short essay but essentially superposition is the idea that a particle exists solely as an unlimited number of simultaneous probability waves. When our particle is touched by another particle it undergoes decoherence into an actual place where it has actual properties. The place the particle turns out to be could be on the other side of a barrier that it lacks the energy to surmount by normal classical physics.
This apparently “magical” leap through an obstacle is called quantum tunnelling—the particle tunnels through the barrier rather than jumping over it. Hydrogen fusion only occurs because of quantum tunnelling. There is a huge energy barrier to be surmounted for two positively charged hydrogen nuclei to get close enough to fuse.
Spooky quantum effects which depend on superposition do not apply to large, warm objects which have lots of particles buzzing around quickly because decoherence is very fast in such a situation. The only way around this is for the object to be very small, so have few particles, and/or be very cold, so they move slowly. Spooky effects therefore cannot apply to living organisms because such effects would be way too transient in such large, warm objects to be of any impact.
All efforts to explain the spooky properties of life by invoking spooky quantum effects foundered on this fact.
And then in 2007 the world changed.
Picture this scene. The prof’s office in a research lab where our science team leader has been rewarded for a lifetime’s achievement in research by making him sit in an office filling in grant applications and trying to decipher the latest insane communication from HR. In dashes a young post-doc declaring that they have worked out the efficiency of electron transfer in photosynthesis.
“Really,” says our hero, while struggling to grasp the implications of being allocated a zero-based budget.
“Yes, to the nearest round figure it’s one hundred percent.”
And that is where the solids hit the climate control dispenser. Nothing in biology involving energy transfer is one hundred percent. Ten percent is good going and even one percent is usable. But repeated experiments have shown that electron transfer in photosynthesis has an efficiency greater than ninety-nine percent, probably because of superposition and electron tunnelling.
And evidence is gathering that many enzymes may use spookiness for electron transfer. Enzymes are the classic controlling agents for the body’s biochemistry: the hierarchy is that DNA makes RNA makes enzymes, which then go on to regulate everything else. Many other biological systems are currently being investigated for quantum influences such as vision, olfaction, magnetoreception (detecting magnetic fields) and Brownian motors (typically nano-scale engines in a cell that convert chemical into mechanical energy).
All of which brings us to the final conundrum. If life really is full of spooky quantum interactions that potentially affect the macro as well as the micro level then is the conscious mind to be found in superposition, tunnelling and quantum entanglement? Are quantum mechanics effects the missing hidden ingredient behind intelligence? Roger Penrose in his 1989 book, The Emperor's New Mind, certainly thought so.
The weakness in Penrose’s work was that this was purely theoretical without a known mechanism for the transfer of quantum effects from the world of molecules to the macro world of the brain. He suggested cellular tubules as the site of the quantum interactions but that is not convincing.
Apart from other problems such as size (too many particles involved), microtubules have no particular connection with the firing of the neuron’s action potential (the electrochemical wave that passes along the neuron to a synapse where it may or may not cause an action potential in another cell) which is the one thing we know is connected with thinking and brain function.
But there is another potential mechanism that might fit the bill. Neuron action potentials are mediated by the movement of sodium and potassium ions along voltage-gated ion channels in cell membranes. Ion channels are tiny, about 1-2 nanometres long and 0.6 nanometres wide so ions can only pass along them in single file. In addition, Austrian scientists have shown using simulations that such ions should show superposition and that the system works to slow decoherence. So all the ingredients exist for a quantum effect on a molecular mechanism that is directly connected with brain function.
The brain generates an electromagnetic field as it functions—this is what brain scanners detect—which can influence the voltage-gated ion channels in the brain, thus tying the whole system together. Recent research has shown that the brain’s electromagnetic field is strong enough to influence whether a neuron fires, presumably through an impact on the ion channels.
We therefore have the possibility—one could put it no higher—that we are looking for consciousness in the wrong place when we try to see it as the direct result of a digital web of interconnected neurons like an internet linking millions of classical computers.
Maybe, just maybe, consciousness is an emergent property of the positive and negative feedbacks between quantum mechanics effects at the level of single ions and an overarching electromagnetic field around the brain.
In some ways we are back where we started if this hypothesis proves to be correct. Consciousness is an emergent property, albeit in a rather strange place, and it does depend on quantum mechanics spookiness in very small but highly structured constructs in living tissue to give the “vital spark” that distinguishes the living from the inanimate.
This would certainly explain why no classical computer, no matter how powerful or cleverly programmed, has ever shown the slightest sign of self-awareness or consciousness.
But hang on a mo. Aren’t all those clever geeks currently in the process of designing quantum computers?
NOTE: Anyone wishing to follow this up could do a lot worse than looking at Life on the Edge, a popular science book by Jim Al-Khalili & Johnjow Mc Fadden.
Copyright © 2016 John Lambshead
Dr. John Lambshead is a retired senior research scientist in marine biodiversity at the Natural History Museum, London. He is also the Visiting Chair at Southampton University, Oceanography, and Regent’s Lecturer, University of California. He writes military history and designs computer and fantasy games. He is the author of swashbuckling fantasy Lucy’s Blade, contemporary urban fantasy Wolf in Shadow, and coauthor, with nationally best-selling author David Drake, of science fiction adventures Into the Hinterlands and Into the Maelstrom.