Much like life, consciousness often eludes definition or even description. We have a tendency to create ambiguous definitions for consciousness that deify the category of whoever is doing the defining (ourselves) while at the same time locking down certain aspects of consciousness that lead us away from knowing anything useful about it.
Learning from our prior experience defining life, we can relax a bit and focus on a simple, functional definition that emerges from the prior subjects we have already covered (entropy ⇒ emergence ⇒ life). If we are even close to the mark then further structures will emerge that will align closely with observed reality, and if we are badly off course this will become evident because we will be obviously divergent from reality.
An Entropist definition of consciousness:
The capacity of a system to sustain a predictive model of itself and its environment.
We can think of consciousness as a major advancement in the ability for a living system to cope with the uncertainty that an entropic universe will constantly present.
Note here that we are not specifying anything about the manner in which the models exist, how they operate, whether they enable “imagination”, and so on. We are simply stating that two models must exist, one for the self and one for the environment, and that they must be predictive in nature.
To give a plot spoiler up front: consciousness exists on a spectrum. We will return to this point in a moment, but keep it in mind.
Returning to the rhetorical “Is a fire alive?” question posed in the discussion about life, we can think about how a conscious fire would be qualitatively different from an unconscious one. For example, a fire has no notion of what things might work as fuel, and therefore can’t make decisions about which direction it might be optimal to focus heat. If it were capable of distinguishing between, say, wood and concrete, we might suggest that a strong natural filter would exist that selected in favor of fires that are capable of deliberately directing heat toward wood over concrete. These mindful fires would propagate faster, survive longer, and generally out-compete their mindless cousins.
It may seem a bit outlandish to contemplate the survival properties of semi-conscious fires that can detect and react to better or worse fuels, but this is not idle speculation. If we generalize the idea to the ability to take an action based on even a single bit’s worth of information about the environment that had some heuristic predictive power, we would find that this is actually what occurs with many single-celled organisms.
Many bacteria, fungus and are capable of detecting a handful of chemical properties of their environment and react to them according to what is likely (though by no means certainly) to be beneficial vs what is likely to be detrimental.
More complex life forms can synthesize more information and improve their heuristic. The main advantage of life growing more complex rather than less over time is that it is required in order to enhance the predictive model. More specifically, enhanced modeling requires enhanced wetware, and enhances wetware equates to enhanced energy cost. As life grows more complex and more energy intensive, natural selection filters constantly push back, enforcing limitations on extra costs that do not result in enhanced survival properties.
The point is that consciousness is an emergent property of life.
This is distinct, however, from it being an inherent property of life, and that is why it is not found in the baseline definion of life. It manifests due to the meta-process of life dictated by the selection filters imposed by the laws of nature. That is, there are self-propagating phenomena in nature, and those with more robust properties are more likely to propagate the best. The ultimate propagators are the ones that can self-replicate, and the best self-replicators are capable of modeling themselves and their environment.
The reason any of this happens is the original trouble that the universe is entropic in nature, and that means the universe is, from the perspective of any living process with limited information (which is all of them), chaotic. Consciousness emerges out of necessity as a mechanism for coping with the chaos.
(We will avoid going on a tangent here about chaos, and how the space of known knowns is fundamentally smaller than the space of known unknowns which is absolutely dwarfed by the space of unknown unknowns — this idea is extremely important to understand, of course, but is best left to another essay. Or go read pre-capture Taleb.)
The universe being chaotic means that living processes are always critically short on information. The best they can do to cope with the chaos is to heuristically model the world and take actions based on their best guesses at whether something is beneficial or detrimental. More advanced consciousness even allows for modeling what the specific outcome of a given action may be. Even more advanced consciousness allows for exploring the model that another life form is likely to be following and incorporate the resulting behavioral pattern of that other actor into its own model of the world.
All of this begins with a heuristic categorization of the world.
A fungal colony identifies free phosphorous as beneficial, and this is hard-coded in its genetics. It has evolved a hard-coded preference for something beneficial. This is an adaptation at the base level of the organism, in the blueprints for the organism itself. The colony reacts by growing in the direction of the phosphorous, and strengthens internal communication and transit channels in the direction of it, reinforcing nodes that are beneifial and allowing nodes that are not working out to atrophy. Fungal colonies are hardly “conscious” by the human standard, but they actually have evolved non-trivial behaviors, merely by having a handful of guidelines for what to avoid and what to favor. As amazing as it sounds, some fungal species even have some concept of memory.
The reactive nature of even this very basic heuristic pattern manifests in complex behaviors in simple organisms, but in more complex organisms the effects of predictive modeling begins to change radically.
We may not consider a fungal colony to have much in the way of an IQ, or even most plants beyond that. But this picture begins to change as we move into the animal kingdom and evaluate the complex behaviors of, say, dragon flies or praying mantis. Even quite advanced insects pale in comparison to the capacity for clear self-modeling and social behaviors in birds, dogs, horses, and so on.
If we think through this carefully we start to get some idea that there consciousness occurs on a spectrum rather than as a simple “Conscious: Yes/No” boolean indicator.
It is very important to understand that consciousness occurs on a spectrum.