Dusk Somewhere

What is human nature?

Posted at — Aug 24, 2024 by Izzy Meckler

In the ant community, each worker performs its proper functions. There may be a separate caste of sol­diers. Certain highly specialized individuals perform the functions of king and queen. If man were to adopt this community as a pattern, he would live in a fascist state, in which ideally each individual is conditioned from birth for his proper occupation: in which rulers are perpetually rulers, soldiers perpetually soldiers, the peasant is never more than a peasant, and the worker is doomed to be a worker.

It is a thesis of this chapter that this aspiration of the fascist for a human state based on the model of the ant results from a profound misapprehension both of the nature of the ant and of the nature of man. I wish to point out that the very physical development of the insect conditions it to be an essentially stupid and unlearning individual, cast in a mold which cannot be modified to any great extent. I also wish to show how these physiological conditions make it into a cheap mass-produced article, of no more individual value than a paper pie plate to be thrown away after it is once used. On the other hand, I wish to show that the human individual, capable of vast learning and study, which may occupy almost half of his life, is physically equipped, as the ant is not, for this capacity. Variety and possibility are inherent in the human sensorium­ and are indeed the key to man’s most noble flights – because variety and possibility belong to the very struc­ture of the human organism.

– Norber Wiener, The Human Use of Human Beings, pp 51-52. (Emphasis mine)

If you are interested in this post, you might be interested in this one as well.

At the end of this video, Søren Mau mentions he is working on a new book on the question: what are “the corporeal roots of history and freedom”? In other words, what is it about the configuration of the human organism that makes “history” and “freedom” possible?

This is a question which can be investigated using Turing’s famous argument for his model of computation, which is itself based on the structure of the human organism. It is clear from the above quote of Norbert Wiener, founder of cybernetics, that a relationship between computation and the variety of human behavior has been intuited since the inception of computer science.

We will get to that in a minute. To start approaching this question we first need rough definitions of “historicity” and “freedom”. Leaving aside the correspondence with the everyday notions of these words and whether these are good definitions, I believe what Mau has in mind is the fact that human beings exhibit a wide repertoire of ways they sustain themselves in a given environment.

In contrast to other creatures and plants, each of which appears to have a relatively fixed method of obtaining necessary material and energy from the rest of nature, different societies, though composed of biologically comparable human beings, have very different ways of obtain material and energy from the rest of nature. That we can have multiple modes of life-sustaining behavior compatible with our biology is a kind of freedom.

History is the name given to the process of the evolution of this mode of metabolizing nature. It occurs through a dialectical interaction between humanity and the non-human environment, driven by changes internal to each as well as resulting from the interaction between the two.

I would leave it to more philosophically well-versed authors to properly situate such definitions or argue for their correspondence to other definitions of history and freedom.

What about humans produces historicity?

Now that we have such a definition of historicity, we can really pose the question, is there something about the human organism that means it has this property whereas others appear not to? If so, what is it?

A negative answer to the first question is possible: for example, it may be mere historical contingency that humans exhibit this behavior while others do not. Perhaps many other species are capable of it in the right circumstances, and those circumstances have occurred for us but not for them. Or, maybe other species actually have history to some degree and we have not perceived it.

The fact that humanity exhibits “historicity” can be fruitfully understood from the perspective of computer science. A given species-metabolism approximately corresponds (by an appropriate Church–Turing thesis) to a computable process, evidently one which is computed in a distributed manner by the individuals of the species. The potential for a wide repertoire of species-metabolisms then simply corresponds to a wide range of individual-behaviors being implementable on the individuals of the species.

This implies that other species that have the requisite biological/psychological attributes will also exhibit history, albeit at a different rate (and with different implications for their environment) from human beings.

The rate of change of the species' history (i.e., of the dialectical interaction between that species and the environment, resulting in changes in the metabolism) will depend on the speed and extent to which an individual of the species can physically intervene in the environment, the rate at which individuals reproduce, and the computational capacities of an individual of the species.

The corporeal requisites of complex behavior

The behavior of a biological entity (on the level of the individual) can be understood partially in terms of

  1. Its sensory apparatus: what information is it able to obtain from its environment?
  2. The range of physical actions which it may perform in any given environmental context (e.g., moving its limbs, vocalizing, secreting compounds It has produced).
  3. The range of internal states it can have which are distinguishable to its cognitive apparatus. States that are not distinguishable by its own cognitive apparatus (i.e., which result in the same behavior given the same stimulus) are to be considered identical.
  4. Its cognitive apparatus. I.e., the biological mechanism which produces a particular physical action given sensory information and the organism’s internal state.

If you are not familiar with the theory of computation and its history, consider reading this post before continuing.

This suggests that a good toy-model for the behavior of an individual organism is given by

  1. A notion of a space In which that organism exists: for concreteness say a finite metric space $X$, and a set of states $S$ that any position $x \in X$ may be in. At a given time, the state of the world will be given by a function $S^X$. In addition there should be some subset $O \subset S$ consisting of states that the organism is in. If the state of position $x$ belongs to $O$, we imagine there is an organism at $x$.
  2. For some radius of perception and action $r$, A function $\tau \colon (x \in X) \times S^{B(x, r)} \to S$. In other words, $\tau$ is a transition function which is given a position $x$ and the states in the ball of radius $r$ centered at $x$, and produces a state for position $x$. $\tau$ represents the capacity of an organism to perceive and then act on its surroundings.
  3. A state $S^X$ evolves over time by applying $\tau$ to every position. This model has the environment itself acting in response to the presence of an organism in its vicinity; it could be modified so that the action of the organism is made more explicit at the cost of having to explain what happens when two organisms express conflicting actions or actions that are incompatible with the properties of the environment.

Now, this is simply a definition for a general cellular automaton, a model which is well-known to be equivalent (in its power to compute functions $\mathbb{N} \to \mathbb{N}$) to Turing machines, as well as to every other well-known model of computation. This is only a jumping off point for an investigation into human nature and not the end point.

To start, this is well-trodden ground for philosophy of mind and the modeling assumptions can of course be critiqued. Weaker criticisms include: the states of a human being are not finite (but per Turing’s topological argument in his original paper they probably are close enough), human beings do not have a single transition function (of course the model could take some variation into account), the model does not account for randomness (random input for the transition function at each step could easily be added).

A much stronger criticism is that it is easy to take this model too seriously: to imagine it is within our capacity to closely approximate a real human being via such a model. It is not. However, that does not mean such a model is useless.

Toy models of behavior can illustrate aspects of the dynamics of real world phenomena. In Marxism for example the computational model of human behavior explains for example the idea of a subject of history independent of human agency.

What is distinctive about humanity?

For the subject at hand, once we have this computational point-of-view on organism behavior, we may assert as a first attempt: humanity has a flexible relation to its natural environment because as a species it its transition function is universal1. That is, the behavior of any cellular automaton (or Turing machine) may be simulated by the behavior of the human social organism.

This puts the real questions into view though: what is the material nature of the simulation? It is not enough to say that any automaton can be simulated without describing the material nature of the simulation. A simulation by human beings could instantiate cells in physical space and their states as the presence or absence of a stone, or in silicon and the presence or absence of a charge, or as plants in a garden, etc., etc.

So to merely say its transition function generates Turing complete behavior is not enough. We must say something about the range of behavior and how it relates to the execution of a metabolism: the exchange of material with the environment.

It is not just that the species needs to be capable of exhibiting a wide range of behavior (for example by simulating an arbitrary computation): it needs to be able to exhibit a wide range of behavior which also results in it acting on the the environment in whatever way necessary to reproduce itself. E.g., we probably could program an ant colony to perform a computation by constructing an appropriate maze for it. But its execution of the computation would also have to result in the obtainment of food, the reproduction of the maze, the detection of changes in environmental conditions resulting in appropriate changes to the maze etc.

History beyond humanity

This is a hard question we won’t finish with now. In any case, it should be clear that other organisms that have the capacity for Turing complete behavior likely also possess the capacity for the kind of highly-varied metabolism, and thus history, that human beings have.

Crows for example, have a powerful cognitive capacity, a decent ability to manipulate their immediate environment using their beaks, and are omnivorous. On these grounds it is very likely they have had their own history, although perhaps one progressing at a slower rate given that beaks are generally worse manipulators than hands, that they cannot communicate vocally and manipulate the environment at the same time, and the cognitive capacity of an individual is likely lower than that of a human being.

Whales and dolphins probably have history as well, possessing enormous brains and the capacity for long-range communication. The difficulty in recording information in the ocean environment would limit the recording of history to an oral tradition, and probably a slower historical development as a result.

However, these independent historical traditions have been changed significantly by the human one and its impact on the species that are their bearers. To put it simply it may be difficult for their histories to develop as we are constantly shaking up their etch-a-sketch.

Deeper questions

One may ask why the creation of a species with a highly-varied metabolism seems to have required individuals which were themselves very intelligent, and wasn’t possible with say, insects or bacteria, despite it being possible to simulate a Turing machine or cellular automaton with extremely dumb components.

I will not do a deep exploration of this question here. I would guess it has something to do with the concrete efficiency improvement of instantiating a complex computation within a single organism rather than spread out between multiple simple biological individuals. This is because spreading it over multiple simple individuals involves a higher overhead cost per unit of computation (in terms of separate digestive systems, separate sets of motility organs, separate sensory organs etc.).

In effect, all of our neurons are simple biological individuals across which the cost of the rest of our bodies can be amortized.


  1. Some important background: there exist fixed Turing machines which are universal in the sense that they can simulate the behavior of any other Turing machine if their “tape” (or software) is configured appropriately. This is the fact that underlies the possibility of constructing general purpose computers (such as the one you are using to read this): a single piece of hardware that can run any conceivable compute program. ↩︎