Consciousness as the only true emergence|
Aug 9, 2013, 9:41a - Consciousness
As a neuroscientist obsessed with consciousness, I've spent the past 6 years in grad school grappling with what my philosophical point-of-view should be on the topic. I just read a new favorite, William Seager's Natural Fabrications (2012), and it has motivated me to write my current thoughts on a whole host of philosophical topics related to consciousness.
Since "consciousness" means different things to different people, I'll start with my definition. By "consciousness" I mean the experiences and feelings that are the basic components of our waking existence. As an example, let's take vision. Processing light to produce environment-dependent behavior is certainly a critical function of the brain, but the brain does something more: it generates a perceptual world of seeing, and it's this experienced sight that I call "consciousness". This can be contrasted with function in the absence of consciousness, of which blindsight is an interesting example. Philosophers also use the word "qualia" to refer to conscious experience. I've written more on my definition of consciousness in a previous blog post.
Based on the current scientific worldview, it's quite unexpected that the brain produces conscious experience at all. We have no "laws" for how consciousness can be generated from the biochemistry of physical matter, yet this is clearly what seems to be happening in the brain. So how consciousness arises from matter is an unanswered philosophical and scientific question. I find it to be the most interesting question in the world.
In the final chapter of his book, Seager articulates 4 views that one can adopt in trying to explain how physical matter generates consciousness. I'm going to go through each of these briefly, and though Seager is too shy to commit to any of them, I am not.
But before I do this, I need to explain the concept of emergence. I think most people find emergence intuitive. The idea is that in a sufficiently complex system high-level properties can "emerge" from the intricate low-level interactions that occur. Some people also refer to this as "chaos". A canonical example involves the weather, where macroscopic objects, such as a hurricane, emerge from the interactions of microscale volumes of air of varying pressures and temperatures. Another common example is that of an ant hill, where the entire colony seems to organize itself as a superorganism to accomplish feats that individual ants cannot.
But lets dig a little deeper into emergence. Seager distinguishes between 2 kinds of emergence. In conservative emergence, the macroscale properties that appear follow completely from the microscale laws of physics. Another way to think about this is that if we had unlimited computing power and ran a simulation of the system with just the location and momentum of every microscale particle, and set the simulation in motion following the basic laws of physics, then the macroscale properties would emerge without having to add any additional macroscale rules. A hurricane is an example of conservative emergence. I also believe that many current scientists would agree that the macroscale phenomena they study, from the economy to neuroscience, can be thought of as conservative emergents.
This is not to say that the best way for a person to understand a hurricane is by doing an atomic simulation of one. It is often better to rely on concepts that are much higher than the microstates of particles in order to get a feel for what is happening, since our brains are good at conceptual hierarchies and bad at keeping track of lots of little things at once. But I would argue that those intermediate levels of explanation are still the result of microstate interactions, nothing more.
Now, conservative emergence stands in stark contrast to radical emergence, which holds that certain macroscale properties do not follow from the basic laws of physics but require new macroscale laws to explain their existence. Another way to think about this is if we had unlimited computing power and we ran a simulation of a system incorporating just microscale features and laws, emergent properties of the real system would not emerge in the simulation. Additional rules at the macroscale would need to be included to simulate the radically emergent properties.
At first glance, I find the view of radical emergence absurd, because the reductionism that is modern science has been ridiculously successful in explaining all sorts of phenomena, from black holes to life. Current science does not have any macroscale rules of the radically emergent sort, except where the macroscale rules assist in understanding phenomena. But in principle, these assisting rules conservatively emerge from basic physics, even if the path is circuitous and inconvenient to calculate.
Now, I think something funny happens with the macroscale phenomenon of consciousness. I've already argued that modern science cannot account for how consciousness conservatively emerges from matter. Seager proposes 4 alternatives for how we can account for consciousness:
(1) The first view he calls Watchful Waiting. This view holds that although we currently don't have a clue as to how consciousness is generated from matter, this doesn't mean we won't figure it out eventually, using standard reductionist science. After all, in the early 20th century the phenomenon of life seemed impenetrable by science, so we invoked the theory of vitalism, an extra force, to account for life. But after a century of work science has now shown that life can be reduced to biochemistry in cells, nothing more. This view holds that something similar will happen with consciousness, and consciousness will be shown to be conservatively emergent from the brain. We are just missing key concepts that connect physical matter to experience, which we'll learn through the indefatigable scientific hunt.
I call this the conservative emergence view of consciousness, and though I think most people would take this position, I don't. I believe taking this position means that you don't fully understand the difference between consciousness and the rest of physical reality, because if you did, you would see that there is conceptually no bridge for this gap. Perhaps this is too skeptical a view, but nonetheless it feels correct for me. Although science has been very successful in conquering previously insurmountable pinnacles, I think consciousness is in a different category. All previous pinnacles were about externally observable phenomena, yet consciousness is something that is only internally observable. Because of this, it is in a class that I believe is resistant to reductionist science, which only works with externally observable phenomena.
(2) The second view he calls Embracing Emergence. This view holds that unlike all other phenomena which are conservatively emergent, consciousness is the one good example of a radically emergent phenomenon. Although our microstate simulation of the brain will generate all the expected behaviors of a person, it will not properly simulate consciousness without adding additional rules that apply at this macroscale.
I call this the radical emergence view of consciousness. Although I have great misgivings about radical emergence applied to any other phenomenon, of the 4 possible views, this is the only one that I cannot make strong arguments against. It is simple and practical. So I'm left holding it, reluctant as I may be.
(3) The third view he calls Favoring Fundamentality. This view holds that conscious experience is a basic property of matter, in addition to the basic physical properties. Not only does matter have mass, charge and other primitive properties, it also has a class of proto-conscious properties that may be considered the underpinnings of consciousness. So the consciousness we know and love is a conservative emergent of this new class of properties.
I call this the panpsychism view of consciousness, and to me it is intuitively absurd. Why posit the existence of an entire class of properties that applies to some if not all matter? This seems to be a very complex theory to account for consciousness, and opens up a whole new set of questions. How do these proto-conscious properties assemble into consciousness? How do they account for the different qualia? This seems like a very difficult road to go down, given the absence of evidence and the additional work needed to clarify how it functions.
(4) Finally, the fourth view he calls Modifying Metaphysics. This view holds that science itself is not telling us about reality as it is, but is just giving us useful concepts for manipulating reality for our practical benefit. Since science is not showing us the Truth and we don't really understand anything, understanding consciousness is of course out-of-bounds as well.
I call this the mysterian view of consciousness, as it says that consciousness will always be mysterious. Like panpsychism, this view seems overly complex. Why assume that science is not showing us reality as it is? Of course it is some sort of approximation, but it seems to improve as it progresses, and the trend is towards Truth (though it may of course asymptote some distance away). Do we have any evidence for this pessimistic view of science? I don't think so, though I suppose someone who holds this view would hold up consciousness as the requisite evidence. But an example of an outstanding mystery does not call the whole system into question - it just gives the system a new target on which to work.
After considering these 4 options, I'm left thinking that consciousness might be the one true radical emergent that I see in the world. It's internal rather than external character make it resistant to reductionist explanation, be it traditional conservative emergence or the more speculative panpsychist conservative emergence. Like the basic laws of physics which at some point bottom out and just need to be accepted as primitive, perhaps there is a basic law of consciousness, though it differs from known laws in that it operates at the macroscopic scale.
Our work as scientists of consciousness is to identify those particle macroscopic phenomena that specifically generate consciousness. On first glance all neural activity appears very similar, but we know that some activity accounts for consciousness and most activity is not sufficient for consciousness. I believe our work is to identify those precise features of neural activity that distinguish consciousness-generating neural activity from the vast stream of regular neural activity devoid of consciousness.
Read comments (5) - Comment
« "Representation" in neuroscience
A time for feeling »
- Aug 12, 2013, 1:37p
This was interesting. Thanks. Could you explain further why you dismiss #1, conservative emergence? What is "the difference between consciousness and the rest of physical reality"? Is it really true that consciousness has no externally visible effects which could be definitively distinguished from non-consciousness?
- Aug 12, 2013, 5:47p
Thanks for the questions. Let me try to answer them, the second one first.
No one today has any external test for consciousness. Even if you walk and talk and do fancy things only humans do, you could still (theoretically) lack internal experience. Think about people who sleepwalk - they behave seemingly in the absence of experience, as proverbial "zombies". There are even extreme cases where people do complicated things "in their sleep", like driving many miles, going into someone's house and killing them. So behavior alone does not seem to be sufficient to show the presence of consciousness.
On the flipside, behavior also doesn't seem to be necessary for consciousness. There are examples of people with "locked-in" syndrome who are completely paralyzed but eventually figure out a way to communicate to a nurse by blinking or changing the pH in their mouth. So even in the absence of behavior, consciousness can persist.
So there really isn't any definitive test for consciousness. If there was, we would hopefully be able to apply this test to other non-human creatures and assess their levels of consciousness. Then I could figure out if all this work with the worms had any chance of paying off, seeing as I don't even know if they're conscious or not!
Now on to the first question. The difference between consciousness and the rest of physical reality is that consciousness is an internal state (of experience), while the rest of physical reality is sufficiently encapsulated by external (observable) states (or so conventional science assumes). When matter affects matter, we can observe the change in various physical properties that occur. But somewhere in the chain, consciousness is produced, but this is different from the effect that matter was having on its surroundings in all the previous cases, because it can't be observed. So the hard question is how a physical state that lacks any internal states might build up into the complex internal state of consciousness. Science today has no way to bridge this gap between the physical and the internal (mental), and there's no clear path for how this can even be done.
Here's an analogy: making consciousness from matter is like making wine from water. If all you have is water and you can't add any other chemicals, you can never make wine. You can mix and cool and heat and beat it to your heart's content, but you'll never make wine. Similarly, I believe that no matter what you do to physical matter, you can never make consciousness in a conservatively emergent way.
Hopefully this explanation is clearer.
- Aug 30, 2013, 6:47p
Your blog post is super interesting. But faced with the four alternatives, I’m on the side of the reductionists. My main argument is simply that this is the most parsimonious explanation for how consciousness arises. It doesn’t invoke anything special. And I think one day we might be able to test for consciousness in simulated brains. Recent work in the group of Giulio Tononi developed an approach to measure the level of consciousness of patients in various states. By perturbing the brain and measuring its reaction patterns, they were able to discriminate patients that were awake, asleep, sedated or emerging from coma. Obviously this technique would only work with the human form of consciousness, but we could imagine creating a brain in a vat and looking for patterns of activity indicative of human consciousness. It wouldn’t be bullet proof because these patterns of activity could be associated with something else than consciousness, for example with something that always accompanies consciousness in humans, but not necessarily in a brain in a vat. But I think it’s a flaw to say that because we can’t currently test for consciousness, because it doesn’t have any external representation, it must be something special. They are presumably many phenomena that we can’t observe or test and yet it’s not a reason to invoke “radical” explanations.
Secondly, you write: “perhaps there is a basic law of consciousness, though it differs from known laws in that it operates at the macroscopic scale.” Is it possible that all laws exist at every scale, but that they only matter at a certain level? For example, from my understanding of physics, it’s not that the microscopic laws of physics don’t apply at the level of planets, but simply that they are not useful in understanding the global behavior of planets. Similarly, gravity exists at the microscopic scale, but is irrelevant. (I know that physicists are working on a unifying theory, yet one hasn't emerged yet.) Perhaps consciousness is similar in that it needs a complex enough system to become relevant. If this is true, then it’s not a fundamentally different property from all others.
- Oct 30, 2013, 8:13a
Thanks for the comment Alix.
To your second point: I think you're confusing the concepts of things that matter *in practicee* with things that matter *in principle*. You say that though gravity exists at the microscopic scale, it is irrelevant. By this you mean that it is irrelevant in practice - you can forget about gravity and just model other forces and you get very accurate simulations. So gravity is only irrelevant *in practice*, but in principle it is still there and likely has a miniscule effect, but an effect nonetheless.
So I take this to mean that gravity is a microscale force that also manifests at the macroscale, but merely as the sum or combination of the microscales (as a conservative emergent).
So when you say that laws only "matter" at a certain level, again I think you're speaking *in practice*. In principle the laws of physics can be viewed as being generated at the microscale and then affecting all higher scales. What I'm concerned with is what is *in principle* plausible with respect to the generation of consciousness.
The in principle/in practice dichotomy often also includes another source of linguistic difficulty: what we're comfortable understanding. Often what is practical is also what is easier for human minds to comprehend, but that is not what I'm interested in here. Although the limits of understanding are inescapable, what I'm trying to understand is not a feature of those limits, but rather the real Truth that lies out there in the world.
Also, your second point towards the end seems to represent a belief in panpsychism - consciousness is there at all scales, but then "needs a complex enough system to become relevant", as you say.
Make up your mind! Are you with the conservative emergents (#1 in the text) or the panpsychists (#3)? To put it another way: do you thing that mind emerges from basic objective features of matter, or do you think all forms of matter have a little bit of proto-consciousness as a basic property?
- Dec 11, 2013, 7:52a
Perhaps Consciousness isn't physical? Any first member in a given series of subsequent members can only pass on what it itself possesses. If this is so, unconscious matter can't pass on consciousness. Perhaps I have it backwards!