Hey everyone, a brief note from your author before we get started. I’ve been paying attention to the responses the essays have gotten from around the community, and I am elated with the level of interest we’re seeing in the project here! I have to send a shout-out to Andrew and The Quest of Evolution team, especially to the excellent design and development team, for their total support of the Worldview Ethics project.
I’m going to be changing a few things about the Worldview Ethics project for this essay and the final one. Each will begin with a brief note like this to break the ice and ensure that people feel secure contributing to the project in any way that adds value, big or small. I want to be sure that people understand that I regard this series of essays as sandbox, not as a showcase. These works are often drafted in a hurry under the influence of a muse, and they contain some experimental ideas that may or may not stand the test of time and make it into the final book.
There can be peer-reviewable, formal, APA-style responses, or there can be single word responses that more or less serve the function of enabling a collector to write “I was here” on the token before flipping it for a profit. Anything in-between works as well. The more eyes and pens end up in this sandbox with mine, the more successful the project will be. I’m most excited about The Quest of Evolution format because it enables a large open community to have a stake in the creation of a larger work.
Without further ado, let’s jump into this month’s installment: Essay 5 - What Are Enactive Agents? Defining Agency
What is the differentiating factor of agency that enables us to best get a handle on what consciousness is and where it comes from in the enactivist model?
Suitable differentiators should be present in living conscious organisms but should be abstract enough from life itself to denote any process in which a decision is made and to refer to any sort of thing that turns out to be capable of taking action on the basis of decisions made. The definition of agency is simple enough: it is the capacity of an object for the action of animus
, the driving force behind the changes in the world caused by life and living bodies, in the broad sense.
In the narrower philosophical sense, agency is a faculty people have and can train. For agents as complex as people, sometimes referred to as moral agents, agency is a spectrum - prisoners and slaves and children sit at one end, and grown adult people who have the capacity to navigate their society spread out toward the other end, thinning until only the best decisionmakers who see the furthest remain because they have the most agency.
More concretely, agency scales with the organism. It begins with an impulse in a single cell, but collectives of approximately thirty-seven trillion cells in individual human bodies learn to work together so effectively that each benefits by surrendering a bit of personal autonomy to produce a process that manages all in the service of goals at the largest scale the aggregate intellect is able to comprehend. Among humans, there are some who can analyze and stress about goals that are relevant to hundreds of millions or even billions of individual people, so the scope of the definitions of the relevant terms here is massive - agency, life, consciousness, intelligence. Each must be useful at the smallest and largest scales as well as the area in between.
It seems there are a few examples of major ideas about how to handle the initial approach to an understanding of the relationship between agency and consciousness. These are just starting to emerge, so space is still available for additional ideas if anyone has a good one. Because enactivism is assumed as a prerequisite for agentic behavior in the emerging framework of cognitive behavior, we can start fresh with a few ideas that assume the enactive view’s truth.
This is the view I have taken for the Worldview Ethics project because it is the most precise differentiator between a living object and a nonliving object. All that needs to happen in this view is that a collection of matter must demonstrate its ability to function as a first mover, causing observable change in its environment.
Also known as Gödel-completeness in this body of work, self-justification is the ability to act in the world, as evidenced by action that is visible as change in a system that includes the actor. If a bacterium is able to move to where it can find sugar, or if a dog can pant, self-justification is occurring.
a. Self-justification is the basis from which spontaneous action emerges. A cell or many cells does something, the environment is changed by the action perceptibly.
b. Problems with this view include the potential for things that don’t have minds to be considered conscious or otherwise, that it is perhaps better described as expression, and that there is a relatively heavy lift needed for a layperson to understand it.
i. The first problem with this view can perhaps be sidestepped by switching our default framework for mind from the historic precedent of Cartesian Dualism and instead focusing upon mind and environment as enactive complex.2. Goal-directed behavior.
The goal-directed view is the brainchild of Michael Levin. Hear his presentation on this podcast
(Jarow, 2022). Levin is an equal part co-author of this paper
(Doctor, 2022). Levin views goals as a primary building block of consciousness because goals supply the foundation for decisions, which themselves supply the foundation for action. We might say that DNA encodes for certain goals that can then be related to by conscious processes at varying levels of organizational complexity.
In Levin’s view, any time an organism or other sort of agent acts, a goal is implied - this applies equally to the smallest single celled organism and to a society at scale.
a. The goal is the basis from which behavior emerges. Coordination to pursue a single goal can unite any number of individuals in a “higher more collective sense of selfhood,” (Jarow, 2022).
b. Problems with this view include the definition of the word “goal” and uncertainty that it is the correct focus here.
c. As above, the observable component here is only available to external observers during the action of the organism or other sort of agent. But we are layering goal on top of the observable change as a sort of inference after the fact.
In the self-justification view, observable behavior is forwarded as the ground of consciousness. In the goal-directed view, explaining the emergence of intelligence is the primary concern. Spontaneous behavior occasionally occurs in chaotic systems and in normal systems that experience unstable conditions, and in neither case is life or consciousness required in the traditional sense.
If we’re working our way toward a single factor that distinguishes consciousness from non-consciousness, it’s probably the case that the self-justification hypothesis is the stronger of the two. Under normal circumstances, the environment has a rather high degree of determinism, i.e., we’re able to predict to some extent the behavior of billiards balls in a game of pool. That’s what makes the game fun. If the balls had the ability to decide where to go and go there instead of where the forces the players exerted sent them, the game would be very different.
Perhaps the correct approach here is to synthesize the two approaches and say that the billiard balls lack self-justification, which living beings possess. This distinguishes living matter from non-living matter and then, as patterns emerge over time, goals become observable to the extent that the living system of matter is able to self-organize in its own interest of continuation and reproduction, whereas rocks and metals and salts and such are instead inert. Agency is a property associated with life in all cases of life that, at very small levels, consists mainly of the action of proteins in response to changing environmental circumstances and animated by behavior. Can nonliving things possess agency? Care Ethics & Enactivism
The idea that care is related to intelligence is profoundly supported by the contemporary literature, from the work of the existentialists up through Robert Pirsig and indeed Michael Levin & co. Care is, for Levin, closely related to intelligence. There are no conflicts between Worldview Ethics and care ethics in my estimation, because both are compatible with enactivism. In fact, the general relationship between care and attention has long held academic interest for me.
In Formal Dialectics
(Daniel, 2023), I argue that observation follows sensation precisely because sensation is not always accompanied by attention, but when it is, robust memories can be formed, attachments can be developed. The ground from which these possibilities emerge is an enactive one, in which an agent is able to develop care for objects in its environment. These ideas are well supported by the literature, and perhaps a ladder can one day form here, to enable scientists to measure different general levels of consciousness/intelligence/care/agency.
For Levin, the relationship between care and intelligence is simple enough; the organism’s goals determine the organism’s cares. The intelligence of a particular organism is manifest in the complexity (or lack thereof) of that organism’s goals. These goals are the foundation of the organism’s motives, which may (or may not) become conscious in sufficiently advanced organisms.
Agents could be a better word than organisms here, but it is difficult to move past seeing the base unit of consciousness as the same thing as the base unit of life and intelligence and even goal-directed behavior (whether said goal is conscious or not). We have a complex train of thought here, but it is worth the cognitive effort - if we puzzle this out and it works, we’ll have our first real cybernetic framework for ethical thinking, and Worldview Ethics will be that much closer to completion.
Care is clearly not a process which must be conscious or explicit to play out in the real world through acts by the organism, and it seems there is some level of overlap between this concept and the concept of Gödel-completeness that first appeared in Essay 3. As we unpack these questions, we’ll develop a more thorough understanding of the problems found at the cutting edge of contemporary philosophy. Pleasure & Pain: The Bedrock of Our Lived Experience
What’s real to us is the experience that we have of the world. Lived experience is ineffable, which means that its complexity transcends that of our language such that linguistic representations can never map 1:1 onto the lived experience they represent. Fortunately, the vast majority of our interaction with the world never needs to be verbalized. And yet we need to understand our immediate circumstances to feel comfortable being ourselves. Despite the sense in which our conscious thinking takes place within our bodies and is not directly available to others, our environment is nonetheless a powerful determinant of the thinking we do.
Our environment is the main source of the set of things that come into our awareness. By act of will we may modify our surroundings but not our histories, by speaking we can describe a limited portion of our experience with others, and by adding music, we can bring our words to life. But despite our best efforts to create a universally shared language that conveys all we hope to mean by using it, we continually fall short.
When we don’t get what we want, we have an experience that we call pain. Pain is not yet completely understood, but two main drivers of pain are the dopaminergic circuit in the brain and the nociception system throughout the body.
If you’ve ever had an experience that didn’t live up to your expectations, you know what it is like for a reward prediction error correction to take place inside your brain - not pleasant. Neurons change their firing patterns and metabolic behavior, resulting in measurable differences before and after the painful event.
If you’ve ever hurt yourself, say you fell down and skinned your knee, or broke your wrist falling out of a tree, you know what nociception feels like. That dull, throbbing pain that seems to keep time with your heartbeat is caused by damage to your blood vessels in the affected area. Endothelial damage causes the coagulation cascade to kick in to stop the bleeding, along with an accompanying hormone called endothelin-1 that gets released by the endothelial cells. Endothelin-1 then binds to the nerve receptors outside the blood vessel and interrupts the nerve signal transmission, itself a form of signal, which the brain interprets as ouch
In the mental sense of pain, the physical analogue is the neurons in the dopaminergic tract recognizing the poor outcome and downregulating their individual metabolic activity. If no new ideas are to be found to reroute the metabolic resources to, perhaps the downregulation in metabolism here can become persistent and lead to conditions associated with continual hypometabolism, like depression. Eventually, the system will adjust and depression symptoms will fade, but the cognitive pain that accompanies these events is an experience that in the past has seemed far more difficult to put into words. Mental pain is the loss of previously functioning circuitry just as physical pain implies damage to somatic circuitry that returns information that is inconsistent with prior maps.
Things are, of course, far more complicated the closer we choose to look. There are instances of pain which fit neither description, and there are a variety of well-documented exceptions to the general neurological rule that, in the dopaminergic tract, the energy always finds a new path to flow through. Despite the shortcomings in the analogies presented, however, a powerful mechanical-reductive explanation is taking shape. Fundamental human experiences are being phenomenologically related to documented observable phenomena in the neurosciences. This is par for the course in the neuroscientific field, which has developed rapidly after scientists figured out how to do experiments on nerve tissue in the mid-nineteenth century.
Ancient philosophers loved to discuss mankind’s two sovereign masters, pleasure and pain. Modern neuroscience would suggest that both take place in the brain, and insofar as the models we construct of our world turn out to be accurate, we tend to experience pleasure. Yet, inversely, when things go wrong for us, we experience pain - whether or not we end up with physical injuries in the process.
The idea of a situation model comes from Hermann von Helmholtz, who conceived of the brain as a “prediction machine.” Psycholinguists use the concept of situational modeling to explain the mental model a reader has of an abstract concept that is modeled upon the text being read and updated in real time as reading continues. Worldview Ethics postulates that consciousness involves something similar; a model of oneself and one’s immediate environment, a mapping of one’s beliefs extending into predictive space from there. This is the worldview of something conscious, no matter what conscious thing it might choose to focus its attention upon.Care is in the loop of cognition
In what is now a seminal text of cognitive neuroscience, Antonio Damasio broke new ground with his Somatic Marker Hypothesis. Descartes’ Error
(Damasio, 1994) was released in the mid-90s and took less than twenty years to penetrate the canon and enter circulation as a core textbook concept of cognitive science courses everywhere. The Somatic Marker Hypothesis, perhaps soon to be renamed as the Somatic Marker Law, maintains that feeling is a core component of reason - after all, we’ve all had days where we felt tired and couldn’t focus in class! Metabolism is required as an input for cognition to result, at least in our minds today.
In fact, it is possible that, if we pursue the Metabolic Theory of Consciousness detailed in Essay 2 to its logical conclusion, we might find ourselves in possession of something like a theory about cognitive computation as opposed to analog or digital or quantum computation we think about today. Metabolism gave us agency and language, the reasoning goes - why would we assume that the relevant processes could be achieved using a digital computational framework?
Levin allows for this objection, but wants to keep the concept of care sufficiently abstracted from metabolic processes to avoid prejudice against non-traditional life forms being discovered to engage in cognition or consciousness or goal-making in the future. This is admirable, and the Metabolic Theory of Consciousness need not stand alone. It can detail its assumptions such that other scholars will possess its blueprint and be able to spin up competing theories to rapidly advance an interdisciplinary literature.Attentional Guidance
Attentional guidance is a core component of conscious thinking. Endogenous and exogenous guidance combine to direct our attention to the things it gets directed to.
Get distracted? Perhaps an exogenous stimulus prompted you to look over your shoulder. Maybe a notification appeared on your screen. Maybe you suddenly remembered to finish that shopping list you were working on earlier.
In each case, your lived experience of the world is being dominated by one thing, then you get interrupted, and perhaps afterward you go right back to the task at hand, no worse for having been distracted.
Attentional guidance of an endogenous nature tends to be influenced by the prefrontal cortex, whereas exogenous guidance of attentional networks seems to be involved with the limbic system. There is a lot to say about this but mostly it will need to wait.Taxonomy of Decisionmaking:
One project which seems increasingly important is the creation of a taxonomy of decisionmaking processes. Decisions have a massive number of moving parts, but a detailed map of drivers influencing a variety of decisions could yield considerable insight into the basis of cognitive behavior. The Worldview Ethics book should probably contain some amount of taxonomical analysis, and readers are encouraged to participate in the project by sharing experiences or insights on this front.Environmental feedback => Cognitive behavior
Environmental feedback yields cognitive behavior. Organism-level behavior is still too complex in cases such as human beings, but we can point to science to support the assertion that cognitive behavior even in human beings is directly related to the processing of environmental feedback at internal and external levels, relative to the self in question. It is important to note that Melanie Mitchell and Michael Levin and Antonio Damasio and Gualtiero Piccinini and Ines Hipolito likely all agree on this point; however, we must remember that identifying this emerging point of coherence in the literature surrounding the science of the self does not provide us with certainty about our conclusion. We must remember to question everything.
If the axiom “environmental feedback yields cognitive behavior in conscious organisms” turns out to be accurate in every observable circumstance, we will have created a bit of canon in the new interdisciplinary field Michael Levin and the other authors cited are founding. The goal of Worldview Ethics is to bridge the gap between cognitive neuroscience and ancient metaphysics by, in the tradition of Aristotle and Aquinas and Alasdair MacIntyre, going back through the foundations of ethical thinking while armed with the novel theories and discoveries of our age. Ongoing, complex, unrepeatable
Having reviewed a bit of detail around the essay’s title question, “What are enactive agents?” we have the opportunity to return to that core concept and venture a tentative answer. Enactive agents are ongoing, complex, unrepeatable processes that leverage powerful abstract representation, analog and possibly cognitive computation, and near-real-time decision-action capabilities that, in goal-directed or self-justified behavior, produce emergent and stochastic behaviors in and around living organisms acting alone or together.
The perspective can be shifted without losing anything since the system is an enactive one: an enactive agent is a member of an environment-individual complex with a body that enables self-justified or goal-directed action to play out within the greater environmental system. Highly sophisticated environment-individual complexes can emerge from sufficiently scaled superorganisms comprised of smaller individual units.
At any level of scale, the events that play out in enactive systems are best thought of as ongoing processes and not static states of affairs. We can also safely claim that these processes are complex (or they express sensitive dependence to initial starting conditions). And finally, no two states of an enactive system are ever identical. The states of the system can gain or lose complexity, but the arrow of time is irreversible and no previous state can ever be attained again. Even in a very simple organism, the environment changes.
1. Damasio, A. R. (1994). Descartes' Error: Emotion, Reason, and the Human Brain. New York, NY: G.P. Putnam's Sons.
2. Daniel, Thomas Dylan. (2023). Formal Dialectics, Serious Philosophy. https://opensea.io/assets/optimism/0xe7f967ed990c4db5262b592a3e5b70e29dd585c3/2
3. Doctor, et. al. (2022). “Biology, Buddhism, and AI: Care as the Driver of Intelligence,” Entropy 2022, 24(5), 710; https://doi.org/10.3390/e24050710
4. Jarow, Oshan. (2022). “Scaling Selfhood,” In Musing Mind. [Audio podcast episode]. Retrieved from https://www.musingmind.org/podcasts/collective-intelligence-cells-economies-cosmos-michael-levin