Information objects

Information object

An information object, within the covolution framework, is an encapsulated entity whose identity, persistence, and causal role in the symvironment are constituted by the informational structure it holds and the switching operations it performs, rather than by the specific material substrate in which it is currently implemented. The defining feature of an information object is that it can in principle be re-instantiated in a different substrate without ceasing to be the same object, provided its informational structure and operational closure are preserved.

Information objects are the entities that covolve. Where switches are the elementary operations of realized distinguishability and encapsulations are the bounded units those operations constitute, information objects are the things to which switching and encapsulation give rise as enduring, identifiable, addressable entities in the symvironment. A horon is a specific kind of information object; not all information objects are horons. The genome of a stored seed, an institution, a memory representation, and a gene are all information objects; only some of them, at any given time, satisfy the further conditions for horonhood.

The concept draws on several established intellectual traditions while generalizing beyond any of them. In computer science, the object of object-oriented programming is exactly a substrate-independent entity defined by its state (held information) and its interface (operations on that state); the covolution framework's information object generalizes this beyond computational substrates. In information physics, the work of Landauer, Wheeler, and Lloyd treats information as physically embodied but not reducible to any specific physical implementation. In biology, the gene was the first widely accepted information object: a unit whose identity is given by its sequence and regulatory role, instantiated in DNA but in principle implementable in other substrates. The framework's information object concept generalizes the gene's logical structure to other biological and non-biological units without committing to gene-like inheritance machinery for each.

What an information object does

An information object performs four connected functions. The structure parallels the four-function tests for switches and encapsulations and is deliberately so, since information objects are the entities those functions operate on and through. A candidate that performs all four is an operationally complete information object. Candidates that perform only some, such as transient patterns, storage objects without active role, or operational units without internal informational structure, are not information objects in the framework's full sense.

It holds informational structure. An information object contains structure that distinguishes it from a generic substrate. The structure is informational in the sense that it can be characterized in bits, in distinguishable states, or in switching configurations, and is non-trivially different from a maximum-entropy distribution over the substrate. A gene encoding a specific sequence holds informational structure; an arbitrary stretch of nucleotides with random sequence matching the same chemical composition does not. The structure must be specifiable, not merely present.

It is encapsulated. The information object has an operational boundary in the sense developed in the Encapsulation page: closing, selecting, self-maintaining, and presenting. Without encapsulation, information may exist but does not belong to any particular object. A diffuse field of correlated states with no operational boundary is a pattern; it is not an information object.

It participates in switching. An information object is not a passive store. Its informational state participates in the four-function switching dynamics of its symvironment: it responds to specifiable inputs, distinguishes between states, holds them on relevant timescales, and couples to other information objects through directed informational flow. A book on a shelf has informational structure and a physical boundary but, unread, participates in no switching network; it is an information object only when embedded in a network in which its contents are read.

It is substrate-independent in principle. The object's identity is constituted by its informational structure and operational role, not by its material implementation. Two objects implemented in different substrates that share informational structure and operational role are, by definition, the same information object instantiated differently. This does not mean substrate is irrelevant in practice; substrate constrains which structures are realizable and on what timescales. The substrate-independence claim is in principle, not in practice, and the distinction matters when the concept is applied empirically.

A worked non-example

Negative examples discipline a concept more reliably than positive ones. Consider a cloud of gas in thermal equilibrium inside a sealed container. The gas occupies a definite volume, has measurable thermodynamic properties, and is bounded by the container walls.

It is not an information object. It fails the informational structure criterion: at thermal equilibrium, the gas occupies a maximum-entropy distribution and contains no specifiable structure that distinguishes it from a generic gas of the same composition. It fails the encapsulation criterion in the framework's strict sense: the container boundary is externally imposed, not produced by the gas itself, so the self-maintaining function is absent. It fails the participation criterion: the gas does not perform switching operations that couple to other information objects through specifiable input and output channels. It fails the substrate-independence criterion trivially, since there is no informational structure to be re-instantiated.

A rock satisfies the same set of failures. It has physical configuration but no distinguishable informational structure relative to its substrate, no operational boundary that performs the four encapsulation functions, and no role in any switching network. Its identity is exhausted by its material instantiation, which is the defining property of an entity that is not an information object.

By contrast, a transcription factor protein in a living cell satisfies all four criteria. Its three-dimensional fold and DNA-binding specificity constitute specifiable informational structure. The cell's membrane and the protein's own folding stability provide encapsulation. The protein participates in switching by binding regulatory sites and gating gene expression in response to upstream signals. And the same regulatory logic could in principle be implemented by a synthetic protein with different sequence but equivalent binding specificity, satisfying substrate independence.

Information objects and switching

Information objects and switches are related but distinct primitives, and the framework needs to keep them clearly separated.

A switch is an elementary operation of realized distinguishability: an event by which one state becomes distinct from another, characterized by the four-function test (respond, distinguish, hold, couple). Switches are operations, not entities. They occur within and between information objects.

An information object is the kind of entity within which and between which switching occurs. It has informational structure, encapsulation, and operational role; the switches are the dynamics that operate on and through it.

The two concepts are mutually presupposing. Without switches, an information object would be a static structure with no operational role. Without information objects, switches would be elementary events with no entity to which they belong. The framework's commitment is that the world's coarse-graining into information objects, with switches as the operations among them, is the right way to describe the dynamics of covolving systems.

This distinction also clarifies the storage versus operation question. A stored information object (a sequenced genome in a freezer, an archived institution, a memory not currently being retrieved) retains informational structure and encapsulation but is not currently participating in switching. The framework treats this as a degenerate state of information object existence: identity persists across storage, but operational reality requires embedding in a switching network. Storage objects are real information objects in latent form.

Information objects and encapsulation

Encapsulation is the operation that produces an information object from a network of components. A configuration of components without encapsulation can have informational properties but does not constitute an information object as such; the encapsulation event is what gives the configuration an addressable identity.

The relationship is logical, not merely descriptive. The four conditions for being an information object (informational structure, encapsulation, participation in switching, substrate independence) presuppose encapsulation as the second condition; encapsulation produces the bounded unit within which informational structure becomes the property of one entity rather than a diffuse network property. Information objects are constituted by encapsulation events; horogenesis at any scale is, equivalently, the production of a new information object at that scale.

This connects information objects to the framework's account of hierarchical levels. The Markov-blanketed encapsulation criterion specifies when a network of lower-order information objects constitutes a higher-order information object. A cell is an information object whose constitutive components are themselves information objects (proteins, RNAs, organelles); the cell becomes a single information object at its level through the encapsulation event that produces its operational closure. This is the recursive hierarchy of information objects that mirrors and depends on the recursive hierarchy of switches.

Information objects and horons

The framework's term horon names a specific subset of information objects, those that additionally satisfy the conditions for predictive coupling and computation in the framework's technical sense. All horons are information objects, but not all information objects are horons.

A stored genome is an information object but not a horon: it lacks the active predictive coupling that horonhood requires. An institution in active operation is both an information object and a horon. A book on a shelf is an information object in storage; if the book becomes the focus of a research community's active engagement, the community plus the book becomes a horon while the book alone remains an information object.

The relation between the two concepts is therefore one of strict inclusion: horons are operationally active information objects whose state space includes a predictive model of their symvironment. This is a useful narrowing because it picks out the information objects that engage in covolution as such, rather than the information objects that merely persist.

Information objects and covolution

Covolution is the process by which information objects accumulate switching density and refine their informational structure through structured exploration of para-determined possibility-space. Information objects are what covolve; switches are the operations through which covolution operates; encapsulation is the mechanism by which new information objects come into being at successively higher levels.

A horon engaged in covolution is, equivalently, an information object whose internal switching architecture is increasing in density, whose encapsulation is becoming more discriminating and self-maintaining, and whose coupling to other information objects in its symvironment is becoming more specific. The framework's central empirical claim, that switching density accumulates directionally under covolution, can also be stated as: information objects in covolving systems become informationally richer over time, while information objects under random drift do not necessarily do so.

The framework's distinction from neo-Darwinian processes can be restated at the information-object level. Selectionism treats organisms as material configurations whose variation is generated externally to their informational structure and whose selection is performed by an environment that stands apart from them. The covolution framework treats organisms as information objects whose variation is generated by their own switching architecture and whose selection is one phase of a closed loop in which they participate as information objects. The conceptual shift is from organism-as-material-configuration to organism-as-information-object.

Information objects across substrates

The framework treats information objects as substrate-independent in principle. The four-function test applies in any substrate; what changes from substrate to substrate is the empirical difficulty of operationalizing each function.

Biological information objects include genes, proteins, regulatory networks, organelles, cells, tissues, organisms, and lineages. Each satisfies the four functions through substrate-specific mechanisms: genes through sequence-based encoding and replication, proteins through three-dimensional folds and binding specificity, cells through membranes and gene expression states, and so on.

Cognitive information objects include perceptual representations, memories, beliefs, concepts, and self-models. Each is held in some neural substrate, encapsulated by attentional or representational boundaries, participates in cognitive switching, and is in principle substrate-independent in the sense that a sufficiently detailed simulation of the same representation would preserve its identity. The cognitive case is empirically harder than the biological one because the encapsulation and switching dynamics of cognitive objects are less well characterized.

Social and cultural information objects include institutions, organizations, communities, professions, languages, scientific theories, and bodies of law. Each is held in a network of practices and records, encapsulated by membership rules and identity markers, participates in social switching through decisions and operations, and is in principle substrate-independent in the sense that an institution can persist across complete turnover of its members.

Technological information objects include software objects, data structures, networked services, and engineered systems. Here the four-function test is least controversial because the substrate-independence and encapsulation are explicit by design.

In each case the framework's substantive claim is that the same concept of information object applies, with the same four-function test, despite the substrate variation. The cost of this claim is the difficulty of operationalizing the test uniformly across such diverse domains.

Limits

Several limitations of the concept should be acknowledged.

The substrate-independence claim is contested. Strong substrate-independence (an information object is fully specified by its informational structure regardless of substrate) is a philosophical commitment that not all biologists or physicists share. Some real biological information depends on substrate-specific properties such as timing, energy, stochasticity, and chemical context that may not be cleanly separable from the information content. The framework's commitment to substrate independence is in principle, not in practice, and this distinction should be honored. For most empirical work, substrate matters in ways that the in-principle claim does not fully capture.

The boundary between information object and pattern is not always sharp. A pattern (a correlation, a regularity, a structure) becomes an information object when it acquires encapsulation and a switching role. The transition is gradual in many real cases, and the framework's binary categorization (object or not object) is a simplification. A research program is an information object; a vague trend in research is a pattern; the transition between them is fuzzy and empirically difficult to localize.

The concept may be doing too much work. Information object is being used to cover entities ranging from individual genes to entire civilizations, spanning many orders of magnitude in substrate scale and temporal duration. Whether a single concept can do meaningful empirical work across this range, rather than serving as an abstract category that absorbs everything it encounters, is an open question. The framework's bet is that the four-function test is substantive enough to do the work; the honest acknowledgment is that this bet is a research wager, not a settled result.

The overlap with horon needs continual maintenance. Information object and horon are closely related concepts, and the difference (predictive coupling as an additional horonhood condition) is technical. The framework should resist the temptation to use the terms interchangeably. They are not interchangeable: every horon is an information object, but most information objects in the universe are not horons. Keeping the distinction sharp is part of the conceptual discipline the framework requires.

Operational counting remains unresolved. As with switches and encapsulations, counting information objects in a complex system requires choices that the framework does not yet fully specify. How many information objects are in a single cell? Does each protein count? Each regulatory network? Each compartment? The Markov-blanketed encapsulation criterion provides an in-principle answer (each operationally closed unit is one information object), but applying it empirically is substantial work. The level-individuation problem reappears at the information-object level just as it does at the switch level.

Why information objects matter for the framework

The concept of information object does several kinds of work that no other primitive in the framework does.

It provides the framework's ontological commitment. The framework is not committed to a world of material configurations on which information is superimposed; it is committed to a world of information objects, with material substrate as the implementation layer rather than the fundamental category. This is a substantive philosophical position and should be named as such. Calling the entities of covolution information objects rather than organisms, systems, or configurations makes the commitment explicit.

It generalizes the gene concept beyond biology. Genes were the first widely recognized information objects in biology because their substrate-independence is conceptually clean and their operational role is well characterized. The information object concept extends this clarity to other biological units (cells, organisms, lineages) and beyond biology (institutions, technologies, cognitive representations) using the same logical structure rather than treating each as a special case requiring its own concept.

It supplies the entity around which the framework's other primitives are organized. Switches are operations on and between information objects. Encapsulations produce information objects. Horons are operationally active information objects. Symvironments are networks of information objects. Without the information object as a unifying entity, these other concepts would lack a common referent.

It connects the framework to the broader literatures of information physics, computer science, and theoretical biology. The substrate-independence claim is the same one made by Wheeler, Lloyd, and Tegmark in physics, by object-oriented programming in computer science, and by computational biology in its treatment of cellular regulatory networks. The framework's information object concept is in conversation with these traditions rather than freestanding.