The Democracy of Objects

4.3. Autopoietic and Allopoietic Objects

Returning to the themes of the last chapter, we can now situate the functioning of systems with respect to how they produce and respond to information in terms of virtual proper being and local manifestation. As I observed in 4.1, Maturana and Varela distinguish between autopoietic and allopoietic machines. Autopoietic machines are machines or objects that produce their own elements and “strive” to maintain their organization across time. Our bodies, for example, heal when they are cut. The key feature of autopoietic machines is that they produce themselves. Not only do autopoietic machines constitute their own elements, but they paradoxically constitute their own elements through interactions among their elements. By contrast, allopoietic machines are machines produced by something else. Generally the domain of allopoietic machines refers to inanimate objects. Here it's worth noting that the distinction between autopoietic objects and allopoietic objects is not a hard and fast or absolute distinction, but is probably a distinction that involves a variety of gradations or intermediaries.

Despite the differences between allopoietic machines and autopoietic machines, I want to argue that both undergo actualizations through information and both involve system/environment distinctions that constitute their relations to other objects. Here a major difference between autopoietic machines and allopoietic machines would be that allopoietic machines can only undergo actualization through information, whereas autopoietic machines can both be actualized in a particular way through information and can actualize themselves in particular ways through ongoing operations internal to their being. Here it might appear strange to speak of information in relation to allopoietic or inanimate objects. However, we must recall that information is neither meaning, nor is information a message exchanged between objects. Rather, as we have seen, information is a difference that makes the difference or an event that selects a system state. In this regard, there is no reason to restrict information to autopoietic objects, for such events take place within allopoietic objects as well.

Before proceeding to discuss the differences between how these two types of objects relate to information, it is important to make some points regarding the system/environment distinction as it is deployed in autopoietic theory. Maturana, Varela, and Luhmann tend to speak of the distinction between system and environment as a distinction that systems draw such that this distinction allows systems to observe their environment. In my view, these are conventions that should be abandoned, or rather, that should be evoked in highly system-specific contexts. Rather than claiming that systems draw distinctions between themselves and their environment—implying that there's a homunculus that does the drawing—we should instead say that systems are their distinction or form. Here it will be recalled that “form”, as Spencer-Brown understands it, is the unity of the marked and unmarked space produced by a distinction. The distinction that generates the marked and unmarked space is, of course, self-referential in the sense that it belongs to one side of the distinction: the system. Insofar as objects are autonomous and independent, they are necessarily self-referential in that their separation from the environment is produced by the object itself. It is the distinction between system and environment that both constitutes the closure of objects and their particular form of openness to other objects. In the case of more “advanced” systems like cognitive systems, social systems, and perhaps some computers, we get the ability to actively draw distinctions and follow through their consequences or what subsequent operations they generate, but in many other instances it's unlikely that systems have any real freedom in how the distinction between system and environment is constituted.

Likewise, rather than claiming that systems observe their environment through their distinctions, we should instead claim that objects interact with other systems through their distinctions. The emphasis on observation, in my view, is one of the greatest drawbacks of various strains of autopoietic theory. Observation implies a distinction between self-reference or reference to internal states of the system and other-reference or references to the environment. The distinction between self-reference and other-reference, in its turn, requires a doubling of the distinction between system and environment within the system itself. That is, systems that distinguish between self-reference and other-reference are systems where the distinction between system and environment re-enters the system that draws this distinction so that the distinction between system and environment can itself be observed. In other words, self-reference and other-reference requires a self-referential operation whereby the system observes how it observes and thereby distinguishes between what arises from within the system itself and what comes from without. Rather than simply undergoing a perturbation, I now treat this perturbation as something that issues from the environment and register that this perturbation comes from the environment. This doubling of the system/environment distinction is a necessary condition for observation.

In their discussions of autopoietic theory, Maturana and Varela often evoke cells as a prime example of autopoietic systems. However, this example, above all, indicates just why we should not talk about the self-referential distinction upon which any system or object is founded in terms of observation. Although cells cannot exist without a boundary between system and environment that is constituted self-referentially by the cell itself, it is misleading to suggest that there's any meaningful sense in which cells observe their environment or make other-references to the world independent of them. To be sure, cells interact with their environment and are, like any other system, perturbed by their environment, but there's no meaningful sense in which they refer to their environment. To suggest otherwise is to imply that entities like cells operate according to meaning. Rather than speaking in terms of observation and other-reference, both of which are far too epistemological and cognitive in their connotations, we should instead speak in terms of how systems are selectively open to their environment and how they interact with their environment. Other-reference and observation, rather, seems to be something that only emerges with more complicated systems such as tardigrades, frogs, and perhaps certain computer systems.

The term “information” is fortunate in that it contains within itself a certain productive polysemy that allows it to resonate in a variety of ways. In addition to treating information as an event that selects system states, we can also read the term “information” avant la lettre to play on the more literal connotations of the term. When we break information into its units, we can say that information refers to what is in formation. Here information refers to the genesis of local manifestations as ongoing processes rather than as fixed identities. The identity of objects is not fixed, but is rather a dynamic and ongoing identity that is in formation. While there is indeed an identity to the object, in the sense that it has a virtual endo-structure that persists across time, this identity is always manifesting itself in a variety of ways. Similarly, we can also read information as “in-form-ation”. Here information does not refer to the ongoing genesis and openness of objects—that which is “in formation”—but rather refers to the manner in which objects take on new form or come to embody new form with their actualizations in local manifestations. Returning to the distinction I drew between the topology of objects and the geometry of objects in the last chapter, information as in-form-ation here refers to the transition that takes place within an object from the domain of virtual proper being and the potentialities populating virtual proper being to the geometric actualization of a form or quality in an object. In other words, in-form-ation refers to the local manifestation of an object embodied in a specific quality.

In both allopoietic and autopoitic systems, information is an event that makes a difference by selecting a system-state. However, information functions in very different, yet related, ways in the case of allopoietic and autopoietic systems. In both cases, information is non-linear and system-specific, existing only for the system in question and as a function of the organization or endo-structure of the object. In saying that information is non-linear, my point is that it is an effect of the endo-structure of the object as it relates to its environment and how this endo-structure resonates within the field of differential relations that define that structure. Information is not in the environment, but is a product of the system perturbed by its environment. In the case of allopoietic systems, information functions to actualize a degree in the phase-space of the virtual proper being of the substance, leading to the actualization of a particular quality in a local manifestation.

Here the point I wish to make is so basic as to appear trivial. However, this point has important consequences for how we analyze allopoietic objects in the world. When an allopoietic object is perturbed in a particular way, it produces an actuality proper to the endo-structure of its being. One and the same perturbation can produce very different local manifestations in different allopoietic objects. Thus, for example, water behaves differently than rocks when hit by another object or heated up. When water is heated up, it locally manifests itself in the quality of boiling. When a rock is heated up, heat is distributed throughout the rock. When water is hit by another object, it produces waves. When a rock is hit by another object, it begins to roll and perhaps vibrates.

These are obvious and familiar points about the objects that populate our world. We all recognize, even if only implicitly, that different objects or different types of substances respond differently to one and the same perturbation. However, while this is an obvious point, it is nonetheless a point that needs to be accounted for. It is precisely this which the concepts of virtual proper being, local manifestation, and information attempt to account for. When an allopoietic object is perturbed in a particular way, information is produced as a consequence of how the object in question is organized. This information, in turn, selects a system-state which actualizes a potentiality in the virtual proper being of the object in the form of a particular quality or local manifestation.

Now, there are two important points worth making here. First, as in the case of autopoietic objects, allopoietic objects are only selectively open to their environments. Many events can occur in the environment of an object without all of these events being capable of perturbing the object and thereby being transformed into information. While rocks, for example, are certainly open to sound waves, they are not, as far as I know, open to signifiers. Uluru or Ayers Rock, for example, is in-different to its title as Uluru or any special legal status it is given. It does not get offended when a stranger that has never heard of it fails to refer to it by its proper name, it doesn't answer to its proper name, nor does it likely worry itself over any sacred or legal preferences it might gain through being Uluru. Here reference to Uluru's in-difference to its name should be taken quite literally as signifying that Uluru's name cannot select system-states within Uluru. Uluru is entirely closed with respect to its name.

Lest one conclude that this sort of closure to its name is merely a feature of the difference between culture and nature, I offer an example of (non)relations between completely natural beings as well. Neutrinos are extremely small elementary particles that travel close to the speed of light. Because neutrinos are electrically neutral, they pass through most matter completely undisturbed and without disturbing that matter. This causes, of course, massive problems in the detection of neutrinos as most detection devices we might use to detect them cannot be perturbed by them due to the electric neutrality of the neutrino. Here the neutrino is a perfect example of a strongly closed entity that cannot be perturbed by other entities and that cannot perturb many other entities. Between the indifference of Urulu to its proper name and the indifference of neutrinos to most other entities, there's a difference in degree rather than kind. While it is important to recognize that most inanimate objects cannot answer to their name (computers are quickly calling this generalization into question), there is no reason to treat culture as a special domain or distinct realm unlike material interactions. In both cases, the issue is one of how entities are selectively open to their environment.

The second consequence that follows from treating allopoietic objects in terms of self-referential system/environment distinctions that are only selectively open to their environment is that allopoietic objects cannot be treated as bundles of qualities. Qualities are results of how allopoietic objects are actualized by their perturbations. They are things that objects can do, but they do not define the proper being of objects which consists of powers. As I tried to show in my discussion of Bhaskar in the first chapter, objects can be “out of phase” with the events they're capable of producing. When situated in terms of qualities, this means that objects can exist, they can be there in the world, either in a dormant state where they produce no qualities of a particular sort, or in a state where, due to the intervention of other generative mechanisms or objects they produce exo-qualities that inhibit the production of particular qualities of which the object is capable.

The key point not to be missed is that the qualities of an object are variable. Every object, allopoietic or autopoietic, is capable of a variety of different local manifestations. And we can say that perhaps every object is capable of producing an infinite number of different properties. This is among the reasons that we cannot treat objects as bundles of qualities. Qualities are products of how allopoietic objects are perturbed, how those perturbations are transformed into information, and how that information selects system-states producing local manifestations.

The question that emerges here is that of why, if objects cannot be equated with their qualities, we have such a persistent tendency to reduce objects to their qualities. I think there are two basic reasons for this. The first has to do with the type of objects we are. Like all objects, we are operationally closed and relate to the world only through the distinctions that regulate our openness to the world. These distinctions, like all distinctions, have a marked and an unmarked space, such that the unmarked space becomes invisible or disappears. In the case of our perceptual world, one operative distinction seems to be the distinction between identity and change. Here identity functions as the marked state, while change functions as the unmarked state. If this schema plays such an important role in our experience of the world, then this is because, as Bergson observed long ago, our perception is geared towards action and our ability to act on other objects. Since action requires a more or less stable platform to take place, change and difference is thrown over into the unmarked side of the distinction governing our perception and cognition. When I go to grab my beloved coffee mug, I register it not as a series of variations or different local manifestations, but as a blue coffee mug. I register my mug in this way even when the lights are out and the mug is no longer blue. Here the blueness of the mug functions as a marker for returning to the mug. “Oh, there's my mug!”

However, while the manner in which we translate objects plays a role in our tendency to treat objects as bundles of qualities, there are object-centered reasons for this tendency as well. While objects are, in principle, independent of their relations, objects are only ever encountered in and among relations to other objects. Terrestrial existence is such that these relations are more or less stable and enduring. The consequence of this is that allopoietic objects tend to be perturbed by other objects in their environment in more or less constant ways. Insofar as objects are perturbed in more or less constant ways by other objects in their environment, they tend to have fairly stable and ongoing local manifestations. As a consequence, the volcanic powers objects have folded within them remain largely hidden from view.

I refer to networks of exo-relations like this as “regimes of attraction”. Regimes of attraction are networks of fairly stable exo-relations among objects that tend to produce stable and repetitive local manifestations among the objects within the regime of attraction. Within a regime of attraction, causal relations can be bi-directional or symmetrical or uni-directional or asymmetrical. Bi-directional causation is a circular relation in which two or more entities reciprocally perturb one another in response to each other. Like fireflies signaling to one another, one lightning bug lights up and another lights up in response, leading the first to light up again. Similarly, one object perturbs another, producing an act in the second object that in turn perturbs the first object that started the sequence. As a consequence of these sorts of relations, we get constant local manifestations. The moon’s gravity affects the earth and the earth's gravity affects the moon. Likewise, we can have uni-lateral or asymmetrical relations of perturbation that bring about a largely constant state in an object.

Fire is a particularly good example for illustrating the idea of regimes of attraction. In its terrestrial manifestation, fire behaves in relatively predictable ways. It leaps up towards the sky and is characterized by pointed tongues of flame that dance and oscillate. As a consequence, we are led to think of this sort of behavior (these qualities) as constituting the essence of fire. However, in outer space, fire behaves more like water, rolling over things in waves, expanding everywhere like liquid on the surface of a table. In its terrestrial manifestation, fire behaves this way because of the gravity of the earth. Here fire exists within a particular regime of attraction that leads to very specific local manifestations. When situated in different regimes of attraction, fire behaves in a very different way.

The concept of regimes of attraction is of central importance to onticology and has profound implications for how we think about epistemology or inquiry. The concept of regimes of attraction entails that it is not enough for inquiry to merely gaze at objects to “know” them, but rather that we must vary the environments of objects or their exo-relations to discover the powers hidden within objects. Knowledge of an object does not reside in a list of qualities possessed by objects, but rather in a diagram of the powers hidden within objects. However, in order to form a diagram of an object we have to vary the exo-relations of an object to determine that of which it is capable. And here, of course, the point is that knowledge is gained not by representing, but, as Aristotle suggested in a different context in the Nicomachean Ethics, by doing. In the case of Aristotle, this doing consists of repeated actions so as to produce habits or dispositions of action. In the case of other forms of knowledge, by contrast, this doing consists in acting upon objects to see what they do under these conditions.

As should be obvious, the concept of regimes of attraction is crucial to our understanding of both allopoietic and autopoietic objects. In “A Developmental Psychobiological Systems View”, biologist Gilbert Gottlieb recounts his early doctoral research on the sensitive period for imprinting in ducklings. [200] Imprinting refers to any time-sensitive phase of learning that occurs very quickly and appears to be independent of behavior. As one might suspect, lurking in the background here is the issue of innateness or whether certain phases of imprinting are innate or learned. While Gottlieb did indeed discover a critical period of imprinting before and after which imprinting could not occur, he also discovered that the developmental age for imprinting could be moved around through a manipulation of the duckling's early environment. As Gottlieb puts it, “[t]he sensitive period for imprinting was not exclusively a function of maturation but depended also on the nature and extent of the bird's prenatal and postnatal experiences prior to entering into the imprinting situation”. [201] Here maturation, of course, refers to factors of innateness. The crucial experiences that played a role in the timing of the onset of imprinting (or, presumably, lack thereof) had to do with whether or not the duckling was reared with visual and social experiences with other ducklings, or whether it was raised in complete darkness and in complete social isolation.

What we encounter here is the importance of regimes of attraction as they function in the development of allopoietic objects. The point here is that, if we don't attend to the regime of attraction in which the autopoietic system develops, we fall prey to a tendency to treat local manifestations as strictly resulting from innate factors in the system, rather than seeing them as results of an interaction between both system-specific properties of the system and perturbations from the environment that are translated into information which then selects system-states. Here the conclusion seems to be that development does not have any one particular attractor in the teleological sense. Rather, through entering into different exo-relations with other objects in the world, an allopoietic object can develop in a variety of different ways. This entails that a key component of inquiry consists in 1) mapping the exo-relations in which particular local manifestations take place, and 2) varying the exo-relations into which an object enters to determine the variations of which it is capable.

Nonetheless, there are significant differences between how autopoietic and allopoietic systems respond to information events. With allopoietic systems, the selection of a system-state is a terminal process. When the water is frozen in response to a change in temperature, it is frozen. There are no additional operations that take place within the system until it is once again perturbed in a new way. By contrast, in autopoietic systems, there are continuous operations that take place within the object even after the selection of a system-state through information. Taking an example from a social system, a news report that the value of the dollar has risen selects a system-state within the economic system. This information event, in its turn, kicks off a variety of subsequent operations within the social system. For example, people begin selling their stocks to maximize their profit. The point here is that, even in the absence of new information events, these subsequent operations continue apace. These system-states and operations are, of course, local manifestations of the autopoietic system in question.

Another, perhaps counter-intuitive, difference between autopoietic systems and allopoietic systems is that there's a way in which the local manifestations of allopoietic systems are more elastic than the local manifestations of autopoietic systems. As I noted in a previous chapter, it is necessary to distinguish between symmetrical and asymmetrical qualities. In the current context, the important nuance of this distinction is that symmetrical qualities are reversible qualities, while asymmetrical qualities are irreversible qualities. While there are certainly asymmetrical qualities that characterize a number of local manifestations for allopoietic objects (paper yellowed with age comes to mind), many qualities of allopoietic objects are symmetrical in character. I turn out the lights and my beloved coffee mug becomes black. I turn the lights on and the mug becomes a shade of blue once again.

In the case of autopoietic objects, by contrast, asymmetrical qualities seem to be the rule rather than the exception. Developmental processes, for example, appear to be largely irreversible, changing the structure of an autopoietic object's local manifestation irrevocably. In communication systems, a statement that is repeated is no longer the same statement, but has now taken on ever so slight new resonances. In the structural coupling of psychic systems and communications systems, I cannot read the same book twice because the very act of having read the book through once to the end already changes how the beginning of the book reads the second time when I begin it anew. As Bergson recognized at the beginning of the last century, the presence of memory as a dimension of all living, psychic, and social experience transforms each event, no matter how apparently repetitious in the brute sense, into a novelty.

However, where allopoietic systems often appear to have a greater degree of elasticity with respect to qualities, autopoietic systems seem to have a greater degree of elasticity with respect to distinctions or what we might refer to as “channels”. It will be recalled that distinctions play a key role in how closed systems are open to their environment or other objects in their environment. One of the crucial features of autopoietic systems is that they have the ability to develop new distinctions, thereby enhancing their capacity to be irritated or perturbed by other objects. This occurs in a variety of ways that are subject to very different degrees of freedom. Thus, for example, it is likely that many plants can only transform the distinctions through which it is possible for them to be irritated by their environment through evolutionary processes of random variation and natural selection. Throughout the animal world, we seem to get increasing degrees of freedom in forming new distinctions through developmental processes that take place through learning rather than innate structure. The same holds true of social systems. And finally, it appears that computers are slowly developing the ability to revise their own distinctions, broadening their ability to be irritated by their environment.

What is important here in these reflections on the difference between autopoietic and allopoietic objects is that both types of objects are organized around a system/environment distinction, both objects are operationally closed, both types of objects are only selectively related to their environment, and both objects transform perturbations into information that selects system-states presiding over local manifestations. In the case of both allopoietic and autopoietic systems, local manifestations are a product of actualizations of virtual proper being rather than fixed properties in substances.