Various responses to this problem include that consciousness therefore forms a logically separate domain (as my old classmate and friend David Chalmers would hold). Another response is that this shows consciousness in this sense cannot be approached scientifically. I’d be happy if Buena Vista Consciousness could contribute to understanding how one aspect of what we mean by consciousness evolved within animals.

@Fred – I’m not sure I understand your point. If it’s one about the logic of consciousness not being empirically accessible, I could see some part of what you mean about its non evolvability. But, at the very least, we can agree that the biological antecedents — such as, if I’m right, the contemplation of multiple futures — have an evolutionary history.

This is a very important piece of research IMO, because it gets beyond pure speculation based on observable behavior of whether other organisms are conscious.

The best argument I have seen for associating the planning aspect of behavior with actual awareness comes from studies of neurologist and author V.I. Ramachandran, who reported that the ability to experience qualia–which are the essence of the hard problem–is closely associated with the existence of multiple behavioral options. He contrasts responses associated with qualia in humans with the more-or-less stereotyped responses linked to invertebrate vision, which is presumed not to involve qualia. There are some possible criticisms of this conclusion I won’t get into here, but this is an important piece of work that does get at this issue in a novel way.

Yogi-one gets at this issue in noting that AI researchers assume awareness emerges with growing complexity. But this is a very controversial idea, and many philosophers would argue against it. Searle’s Chinese room argument is perhaps the most well known and discussed argument against it.

If one understands consciousness as the experience of what it is like to be something (Nagel)–the so-called hard problem of consciousness–then there is a complete disconnect between ability to plan or any other measure of what in effect is intelligence, and consciousness. Not only could be very complex and intelligent behavior be carried out by a machine that we humans would all describe as unconscious, but conversely, relatively simple organisms could be conscious.

For example a Venus flytrap: it has a simple open-shut mechanism that simply closes whenever there’s pressure on the trap. It cannot even distinguish whether the stimulus is prey or not.

So perhaps there is an even more primary level of awareness that can simply receive information and even react, but not plan for the future.

So then, you would have three levels of awareness: simple awareness, consciousness, and self-awareness.

It would be interesting to explore the boundary between simple chemical reactivity and primary awareness. An interesting question would be whether a single cell has awareness, or it is simply a chemical machine. I think the prevailing viewpoint is that processes at the chemical level are pretty much done with no awareness, they are simply driven by the laws of chemical attraction and repulsion.

But does a one-celled plant or animal have awareness?

And how would you define awareness so it can be distinguished from simple chemical reactivity?

The view for AI researchers seems to be that as a system grows in complexity, it develops awareness. In other words, with enough processing power, and enough memory, you could build a fully conscious, self-aware machine.

This would imply that the idea, beloved by all religions, that something fundamentally outside the known laws of physics is necessary to give rise to conscious living beings, is a fallacy.

In other words, humans have self-awareness and robots don’t simply because humans are orders of magnitude more complex than the robots we have managed to build so far.

Also, the way we process huge amounts of information so fast is due to a technology that has so far evaded robotics. The way human nerve cells actually transfer, store, and apply information has no analog in the computer world.

Perhaps the advent of quantum computing will lead to progress on that front. In other words, humans are quantum computers on a mega-scale, according to this line of theory. Then the line of development from reactivity to simple awareness, to consciousness, and to self-awareness becomes simply a trajectory of rising complexity.

Also, I think that folks need to distinguish between ‘consciousness’ and ‘self-awareness’. Using your definition of consciousness, an animal can be conscious without being self-aware. In fact a plant can be conscious, because it can sense and react to surrounding stimuli.

Awareness of self as a separate entity is a different trait, and probably requires some specialized development in the brain. In humans, this trait of self-awareness seems to develop quite late, like 1-year or more after birth.

In fact, human babies give a good example of ‘consciousness’ being distinct from ‘self-awareness’. Self-awareness, according to this observation, would then be a specialized type of consciousness, but not a prerequisite for consciousness.