QUESTIONING THE CRANIAL PARADIGM
CAROLINE JONES: I want us to think about the gut-brain axis and the powerful analog system of our immune brain, also thought of as a mobile brain. The cranial paradigm is what I’m here to question and offer you questions about. Mainframe is a kind of discourse that haunts the field that we’re talking about, and the cranium comes with that metaphor that we all live by.
What do we mean when we say the word "intelligence"? The immune system is the fascinating, distributed, mobile, circulating system that learns and teaches at the level of the cell. It has memory, some of which lasts our entire life, some of which has to be refreshed every twenty years, every twelve years, a booster shot every six years. This is a very fascinating component of our body’s intelligence that, as far as we know, is not conscious, but even that has to be questioned and studied.
As you go to lunch, you will be putting things in your mouth that are not yourself. Your body, hopefully, at this point in its existence, knows better than to reject these not-self proteins and not-self photosynthesizing cells and pitch you into an immunohistological response, and to say, "Oh, this is friend, this should be tolerated. I, this aggregated entity of self, will learn that these things are friend. These things are to be tolerated, these things are to be learned from and incorporated and not rejected."
Yet, if that same food were somehow injected into your lungs, you might have a violent asthmatic response. You might die from that. The immune system is using the mouth as a category to learn and to train. It took scientists a long time to figure out where this learning and training was happening, which it seems is in the lymph system. If things are introduced there by the injection at the doctor’s office, a whole different set of learning is instigated that reads it as not-self, not friend, something which needs to be expunged and eaten by the macrophasias and remembered as not-self, as enemy.
This is an extraordinarily powerful metaphor, and it’s one that is parallel with AI and computer sciences now in active transformation. In other words, most of our pharmacological economies are organized around antibodies. But the probiotic industry, which is completely unregulated by the US Food and Drug Administration, is expanding through folk medicine.
When my own immune system was quasi-destroyed and rebooted by chemotherapy, I was like okay, how do I rebuild this? What are the probiotics? What’s out there? It's in folk medicine and in proprietary corporate formula. I can know that it’s some kind of yeast, but I can’t know what the exact sub-species is, owned by this or that corporation, that I'm trying to reeducate my immune system with.
This is a moment of paradigm shift in multiple fields. I’m recommending that we think about the way Catherine Bateson describes some of her father’s work, that mind does not necessarily stop at the skin. We are completely symbiotic on these planetary systems that form and have formed our consciousness and our capacities to learn, and to navigate, and to remember. Through our lifetimes, we become hosts, dependent on xenobacteria that we invite into our bodies and cultivate and grow as part of a self that is not yet ourselves, that is a not-self that we cohabit with and are completely dependent upon.
I just throw this out as a complete provocation, which I’m supported by through the cultural evolutionists we call artists, who are making art forms out of biological materials, out of living materials, to help us think through our symbiotic dependence on other life forms and our interesting non-conscious negotiation with self and not-self every day. I can be very brief and leave it at that.
Frank’s comment about just being in the world, embedded in an environment, and letting that surfing and negotiating with inputs that are analog and need to be responded to in an adaptive and flexible way—this is what I would call intelligence. The body is an amazing model that goes way beyond mind of learning and memory and how we can craft our epigenetics through certain cultural acts and practices, how we can supplement them prosthetically, epigenetically.
This is my provocation to reboot AI on a certain model of what Gibson would have called "environmental and ecological perspective."
* * * *
STEPHEN WOLFRAM: I’m curious as to what the current immunologists’ high-level model of the immune system ends up being. For a long time there were these network models of the immune system where there are antibodies and anti-antibodies, and there was this notion of dynamic equilibrium in the immune system. If you ask a random immunologist what their high-level view of the immune system is, in my experience, they'll tell you a very low-level view of a very specific part of the immune system. There used to be these network theories of the immune system. It would be interesting perhaps to compare those with the current models for brains and neural nets. I’m just curious if people know what the current view is.
JONES: I’m not going to be able to answer that. Are you speaking of theoretical biology? Are you speaking about practical immunologists in a hospital setting?
WOLFRAM: There are 100 billion possible types of antibodies, and any particular person has some number of those antibodies in reasonable concentrations. You can do an assay for a particular antibody, but this question of how many of the 100 billion possibilities do we have in decent numbers, I don’t think that’s known. You start getting more and more antibodies, why does it not run away?
JONES: The obsession has been on the antibodies, and that is part of the systemic immune system that has received all the research. What hasn’t been researched is the mucosal immune system, which is the system that learns, the system that builds tolerance, the system that trains and takes in and negotiates the self/not-self.
Part of the example I'm about to share comes from a neuroscience boot camp I took at Penn, which was great. In the presentation somebody said, "Oh, then there are the glia." And I said, "What are the glia?" And the response was, "They’re not important. They’re the house cleaning staff." I said, "I’m a feminist. The house cleaning staff is important to me. And by the way, your model of mental illness is a serotonin reuptake inhibition model, so you’re dependent on the cleaning crew to manage this uptake."
Basically, there was this tiny window into an under-researched entity in the brain that is entirely involved in the immune system. It used to be thought that the brain is somehow isolated in its beautiful ivory cranium, and it just doesn’t have to deal with the immune system, and it’s kept away from all those diseases. Well, no, the glia are there actively cleaning up, managing the garbage that is produced by the phasias that are eating the toxins and determining what is self and so on and so forth. I believe, not being a scientist in this world, that they are at the edge of shifting into some very different kinds of research not on the heroic actors with their shields and swords, but the clean-up crew that is determining how the body will go forward.
RODNEY BROOKS: You talked about the immune system as a separate system. We have the gut neurons, which are separate. Even C. Elegans (Nematode), which has 302 neurons, twenty of which are in a separate gut-brain than the central brain, and they have fifty-six glia cells. Even in that smallest thing, we see this structure. You were talking about the immune system in an interesting way, as learning, teaching, and remembering. When we look at plants and their capabilities, they don’t have neurons either, but the roots go out and search and the leaves do all sorts of things. There's a lot of activity that is underappreciated when compared to the neurons, which are seen as the only important cells in the brain. It’s even worse when you think about non-animals, because there’s stuff happening for which you don’t have computational models, which gets back to my earlier point. When do we have computation models and when don't we? Plants are obviously doing something very interesting in the way they adapt to their local environment and adapt to what’s happening and change themselves.
ALISON GOPNIK: This gets back to something that you were talking about, Rod, regarding adaptation. There’s one dimension, which is one of the things that Turing realized, and that's the idea of breaking up something complex into a process where you can describe it as parts of the process. That’s the big idea of computation. One kind of intelligence is being able to do that.
Another idea is being able to represent something that’s external to you, being able to take the external structure of the world and, in some way that we don’t quite understand, get a veridical account of what’s going on in the world around you or adapt to what’s going on in the world around you. There are interesting questions about what the relationship is between those two kinds of intelligence. And they might be orthogonal to one another in various ways.
Something like deep learning solves this problem of trying to adapt to the external world in a very simple way. It lets you take the statistical structure of the input, something like images on the Web, and incorporate those into a system that’s producing a particular kind of process. But that’s a primitive way of relating to the external world. I don’t think we have a very good theoretical account of how that process of adapting to the external world is related to the process of being able to compute. I don’t think we have a good story.
JONES: That’s true. In parallel to the neuron supremacists, we would have the representation fetishists. Part of the definition of intelligence is always this representation model. When we think about the immune brain, I don’t think we need to imagine the glia having a representation of the body or even a map of the body. It has pathways to circulate in. It might even respond to some vessel making components of the body to make new pathways, if it needs them. I don’t know those mechanisms.
The point is, it does not need a representation of the body; it needs to know where it needs to go, which is a different problem. I’m pushing this idea of distribution—homeostatic surfing on worldly engagements that the body is always not only a part of but enabled by and symbiotic on. Also, the idea of adaptation as not necessarily defined by the consciousness that we like to fetishize. Are there other forms of consciousness? Here’s where the gut-brain axis comes in. Are there forms that we describe as visceral gut feelings that are a form of human consciousness that we’re getting through this immune brain?
PETER GALISON: When the bio-artists look at microbiological forms and plant forms or animal forms, is there something suggested among them that might give us a different set of metaphors or conceptualizations of consciousness?
JONES: Through the artists, I’m coming up with this idea of symbiontics, "ontics" being that which is, and symbiosis being that which I wish we could be more completely aware of as we navigate this world. Many of them work with concepts and materials in the gallery that prompt me to think more robustly about our interdependencies.
GALISON: So, what are the artists doing?
JONES: Philippe Parreno, who we saw in Berlin, used bacterial motors to turn the lights of the gallery on and off and raise the window blinds. The bacterial motors are entrained with other forms of AI and digital computations that are responding to the presence and absence of humans and their movements through the space, as if we were invaders of a non-self that the gallery must then respond to as an immunological distributed system.
You could think of these metaphors in lots of different ways, but the artists are helping us evolve toward a more symbiotic understanding of our place.
DAVID CHALMERS: One angle on thinking about self and not-self in cognition and intelligence? We could bring in the literal immune system, as you’ve done, but we could also think about it in terms of having a separate cognitive immune system, drawing the self/non-self distinction at the cognitive level. A lot of that is done by things like trust. If you ask about self and non-self in cognition, my smartphone is totally self. It’s not non-self. It’s not something outside which is coming in.
JONES: You never get spam calls?
CHALMERS: It’s app sensitive or context sensitive. The phone numbers and Google maps, that’s just self, that’s my navigation system. I treat it as self and I trust it. It basically becomes an extension of my cognition.
JONES: So, it’s a prosthetic self.
CHALMERS: Yes, it becomes prosthetically part of the mind because I choose to trust it and identify it as self. And there’s other stuff out there. Spam, for example, that comes in over email, and who knows what on the Web—these things that I regard as not-self are no longer part of my cognition. This is the way that cognition gets distributed out from our brain into the environment.
JONES: Part of what I’m advocating, and what the art is helping me to do, is advocate for a much broader self. In other words, we know that humans have evolved clothing, and language, and heating, and H-vac, and architecture; if we take these things away, we’d survive, I don’t know, maybe two weeks.
In other words, it’s partly to acknowledge our existence as social animals, to acknowledge that the cranium is not where we do most of our thinking and being, and to figure out how to get our artificial systems, our prosthetic systems, to help us acknowledge our embeddedness. Partly I see this as a planetary dilemma. If we don’t feel our place in the planetary ecosystems, we deserve to go extinct, which we will.
ROBERT AXELROD: I was going to ask whether you think hormones provide another form of intelligence, adrenaline for example.
JONES: Oh, absolutely. Psychiatric diagnoses are being made on the basis of which drugs you respond to, which are influencing hormone cycles and their reuptake by the brain. If we look at how we’re practicing this medicine and dealing with this thing we call the brain, then the mind is totally distributed throughout the body. Hormones are very much a part of that.
IAN MCEWAN: Can you say something about the civil war that occurs when the body turns against itself—Crohn’s disease, arthritis?
JONES: There's an argument for ingesting in the oral tolerance portal that which our immune system is turning against. For example, collagen, in the form of certain autoimmune diseases. It doesn’t have to be human collagen. There’s enough molecular similarity between cows and chickens that if you ingest, rather than inject, this form of collagen, your body is like, "Oh, I don’t need to attack that. That’s an okay thing, collagen."
Why is the body attacking its own collagen in the first place? If mice are given neural sheets without adjuvants, without things that are alarming their immune system, they will stop having MS or something, their sclerota will stop being attacked by their body. So, there are incredibly promising therapies that are emerging from this.
MCEWAN: But the system can make mistakes.
JONES: The system absolutely can make mistakes. If you think about the form of the vaccine, you think you’re just getting polio, but you’re getting polio surrounded by a witch’s brew of cholera and diseased bacteria, and things that are saying to your body, "This is really bad! Turn against this!" It’s the adjuvant. And what adjuvants are we taking in? Let’s hope Monsanto isn’t on those fields out there. The pollution, the pesticides—these are adjuvants that are alerting us to attack certain things as toxic that then may be disrupting other parts of our immune system. Again, this is still in the realm mostly of folk medicine.
There's an idea that you can reduce asthma if you eat local honey, because what you’re eating is the bee’s concentrate of all the airborne proteins and pollens and dust, so you're eating all of your local airborne potential triggers. You are learning and training your body to tolerate them by eating the local honey.
Urban asthma is very bad, so should you be eating cockroach feces, or should you just get rid of the cockroach feces? This then becomes a social problem. I’m just recommending this as a model of intelligence which is quite distributed, not conscious. Does it have ethics? It certainly has goals, but they shift every day depending on what part of the system encounters what not-self element.
AXELROD: I would say the immune system has pretty stable goals, which is evolutionarily to protect the host.
JONES: That sounds like a reasonable thing.
AXELROD: The methods change every day, depending on the challenge, but the goal is pretty stable.
JONES: We assume cancer was always there, so why was that not evolutionarily eliminated? We’re looking at a system that didn’t completely eliminate these things that seem like they would have been evolutionarily problematic. So, is the goal wrong or is the system messed up? Or do we just need to see this homeostatic navigating as part of life?
AXELROD: Are you saying evolution is not finished?
JONES: Well, let’s hope it’s not. Do we want evolution to be finished?
WOLFRAM: Are you suggesting that the reason there’s a rise in autoimmune diseases is because there are more adjuvant-like things in the environment?
JONES: Well, that is the suggestion on the table. It’s not my suggestion.
WOLFRAM: What’s the leading suggestion for what the adjuvants are?
JONES: What we’ve described as pollutants is probably a pretty good category: Benzenes, BTEX chemicals, pesticides, things that we have produced either as by-products of energy or to kill lifeforms. Rachel Carson wanted to call these biocides, because they’re not just killing pests.
WOLFRAM: Epidemiologically, that would be a fairly easy question to test I would think. That’s an interesting theory.
JONES: They’re now beginning to test this. You have to understand that you’re dealing with industrial food and industrial medicine where those haven’t been the leading research questions. Monsanto doesn’t want us to ask about that. Monsanto wants us to buy Roundup, so it’s hard to get that research done into whether people living in blow fields near Monsanto are having more autoimmune diseases than people who don’t. It’s hard to find people who are not living near Monsanto drift. These are important questions, and they are starting to be tested.
