Distributed Consciousness Theory (DCT)
Distributed Consciousness Theory (DCT) is a new theory of consciousness, which primarily addresses the substrate problem: what type of physical substrate best supports the phenomenology of consciousness? The theory describes mutual information non-locally ‘distributed’ in physical systems. This is formally described as closed systems having reduced degrees of freedom (entropy) due to mutual-information with the environment. This thermodynamic principle is generalised to a field with distance metric equivalent to the inverse of shared information. The theory describes how conscious objects may be represented in this distributed field and details information flows associated with conscious perception as well as processes in the brain which may interact with distributed information.
The theory is able to solve key problems in consciousness theory, including The Binding Problem (Bayne 2010) and a novel solution to The Hard Problem (Chalmers 1995). Consequences and remaining unsolved problems of the theory will be discussed including the problem of causal emergence.
consciousness, theory
Supplemental Material
A presentation from Models of Consciousness (Sapporo 2025) covering the key concepts of the theory is available online at https://r.greenant.net/published/DCT_presentation.
Introduction
This paper provides an outline of the key concepts of a new theory of consciousness, the Distributed Consciousness Theory. This theory has been developed on the basis of communications with a number of theorists, see Acknowledgements.
The Distributed Consciousness Theory is primarily motivated by the hypothesis that consciousness is a distributed phenomenon. We will formalise this hypothesis but for the moment it is enough to state that, for the purposes of the theory, ‘distributed’ means that the processes and contents of consciousness are reliant on non-local phenomena. The theory focuses largely on the question: “what physical processes can allow the types of conscious experience we experience to operate?”. Therefore, DCT is largely a theory of the mechanism of consciousness, with the contents and states of consciousness being secondary considerations.
DCT attempts to address some of the key metaphysical problems inherent in other theories of consciousness. This paper will briefly contrast DCT’s perspectives with alternative theories. DCT is grounded in modern physics, particularly information theory and quantum theory. However, it strenuously avoids new physics. While some physical interpretations may be novel, the theory does not rest on creating new forces or particles, and should be interpretable with well-established physical concepts and mathematics.
Background
Before getting into the specific elements and predictions of the theory, it is worthwhile identifying some of the key influential ideas which we will reference:
John Archibald Wheeler: “It from Bit”
In Wheeler’s 1989 paper, “Information, Physics, Quantum: The Search for Links”, he outlined a set of questions related primarily to the primacy of quantum information as a substrate for physics, aptly summarised in the term ‘It from Bit’ (Wheeler 1989). Wheeler’s arguments included some cursory discussion of consciousness, particularly from the perspective of binary information – “yes-or-no indications” – and the role of “observer-participants”. Importantly, he implored an understanding of “all of physics in the language of information”, which he conceived as fundamentally based on binary conclusions about observables.
Many interpretations of this concept followed, including Anton Zeilinger’s statement:
“My interpretation [of ”it from bit”] is that in order to define reality, one has to take into account the role of information: mainly the fact that whatever we do in science is based on information which we receive by whatever means.”(“It from Bit? Plus.maths.org” 2015)
Generally, the ‘It from Bit’ concept has been interpreted with reference to the dominant Copenhagen Interpretation of quantum physics, in which the probability of events evolve deterministically (under the Schrödinger equation (Equation 1), sometimes called Unitary Evolution) until the act of ‘observation’ collapses the wave-function indeterministically (the projection postulate)(Lombardi and Dieks 2014).
\[ i \hbar \frac{d}{d t}|\Psi(t)\rangle=\hat{H}|\Psi(t)\rangle \]
However, this interpretation raises a number of metaphysical challenges, not least of which is the Observer Problem; which types of observation lead to the collapse of unitary evolution, and what constitutes a measurement device.
It is important to be clear that the Distributed Consciousness Theory does not rely on this collapse mechanism. Instead, we ascribe to a ‘no-collapse’ interpretation, in which quantum information persists without being collapsed to classic observables (Passon 2018). These no-collapse interpretations share the common characteristic of formulating quantum theory without the projection postulate.
In Relational Quantum Mechanics, systems are described by the interaction between systems, rather than an absolute state of the system. This view does not establish primacy of a given, underlying state, rather a measurement is only relevant to the relationship between systems. This has been described as a “net of relationships” (Laudisa and Rovelli 2014). In Rovelli’s formulation (Rovelli 1996, 1997), an event or state is only validly defined with respect to a system (for instance an observer system).
(Rovelli 1996): “Quantum mechanics is a theory about the physical description of physical systems relative to other systems, and this is a complete description of the world”.
Importantly, even though the events and state of an observer are relative to the system they observe; and different observers may have different observables, the rules of quantum mechanics ensure consistency when different accounts are reconciled by physical interaction.
Theorem 1 \[ H^{u n i v}=H^1 \otimes H^2 \otimes \ldots \otimes H^j \otimes \ldots \]
Decomposition of Hilbert Space
See Theorem 2.
The interpretation retains the mathematical frameworks of quantum mechanics, with the dynamical state of the system corresponding to the usual density matrix, evolving as defined by the Schrödinger equation. In contrast, the value states for a given system are constrained by the non-commutativity of observables, thus defining joint existence of properties, independent of knowledge of a system
An implication of this interpretation is that it is necessary to define a “privileged observable”, and by extension the Hilbert space of the universe must have a preferred factorisation (as opposed to all factorisations being equally plausible).
\[ H^{u n i v}=H^1 \otimes H^2 \otimes \ldots \otimes H^j \otimes \ldots \]
In the Atomic Modal Interpretation (AMI), this is usually interpreted as being synonymous with the standard model (the fundamental objects of the universe being elemental particles). However, for the purposes of DCT, we will take the fundamental objects to be composed of quantum informational objects, with particles ‘dancing around’ this information. In this way, we can extend the ‘It from Bit’ analogy into a formal description. In DCT, the properties of particles (position, momentum, etc.) are constrained by the more-fundamental quantum information of systems.
Importantly, this type of quantum information is not a derived quality of a system, it is fundamental to the dynamics of the universe’s state space, \(H^{u n i v}\).
Importantly, this interpretation allows to build a realist substrate for consciousness, expressed in the precise mathematics of quantum mechanics. Furthermore, it applies to all physical phenomena, no matter the measurement scale. Another interesting implication of the Atomic Modal Interpretation is that composite quantum systems can have collective properties that may not be explained by their atomic components (Dieks 1998). This will become important when we consider the causal efficacy components of the DCT.
Healey (1989) also considered how non-local quantum properties of entangled particles may possess a “holistic” property.
An extension of the modal interpretation is the Perspectival Modal Interpretation (PMI, Bene and Dieks 2002), in which the the quantum state of the whole universe with respect to itself is utilised as a “reference system”. Observers in this formulation have different relational descriptions from their particular perspectives, but the position of macroscopic objects shows good agreement. These relational states are fundamental, not being derived from more fundamental non-relational states.
Modal interpretations define a space of possible events, from which only one of these becomes actual. Therefore, the interpretations are probabilistic (in contrast to many-worlds interpretations, for instance).
Out-of-equilibrium behaviour of quantum systems under drive
While the presence of quantum effects in biological systems is well established, there remains significant skepticism regarding whether quantum effects can be coherent and meaningful in large systems like the brain. This section will discuss some proposals for large-scale coherence in noisy environments.
Viewed from a traditional localist paradigm, the brain is a complex organ with emergent dynamics built on the activity of neurons, generating local field potentials and broader coherent activity which we observe behaviourally, and with the tools of electrical and magnetic neuro-imaging. The localist approach assumes that the coding of neural activity arises from the summed activity of neural firing, perhaps encoded in the firing patterns and/or the local fields generated in neural aggregates. However, consciousness theories usually require this local activity to be converted into global states, whether these be for perception, awareness or cognition.
This mapping from (micro) neural activity to (macro) states of consciousness has plagued all theories of consciousness. Various proposals attempt to explain emergent states, including minimisation of free energy, or causal integrations, functional workspaces, or higher-order monitoring processes. These theories often suffer from mathematical complexity (or intractability), making their mechanisms implausible or inefficient. It is the assertion of this paper that the core deficiency of these approaches is that they attempt to turn local activity into brain states while adhering to a localist view.
Instead, we can take a different approach, in which we consider these micro-states as contributing to a global, non-local field instantiated in the brain. Our best theory for describing non-local phenomena is quantum theory, in fact it could be argued that non-locality is the key defining feature of quantum theory. Non-locality in this context, does not mean just a distribution of information in a system, it implies that states themselves (and their accompanying informational states) are distributed and not describable simply by the sum of microstates.
There are many examples in physical experiments of these non-local states.
Elements of the Theory
The following elements form the basis for DCT. DCT is a topological information theory with falsifiable predictions. The basis to measure preductions is thermodynamic. We know the brain utilises a high and stable metabolic rate compared to other organs. This significant [Vijay Balasubramanian, Brain Power, 2021 PNAS]https://www.pnas.org/doi/full/10.1073/pnas.2008173118) “Brain Power” is remarkably stable in different conscious states, restful vs activated, asleep vs awake, receptive or quiescent.
These intuitions about structure can be reformulated as constraints. Let us propose that there is a structural-functional relationship in the brain (and other sufficiently structured systems) that is driven by two affordances:
- the interface with the patterns of nature
- resonance to structurally filter [wide bandwidth] coincidental structures in-vivo
It is hypothesized, that the origin of consciousness, and it’s evolution in organisms; is present in the capacity to modulate activity dependent on cyclical events. The neural matrix can become aware of the rising and setting of the Sun. This matrix can be very small indeed. It can be thought of as a temporal bit - light on… light off…
The fundamental functional requirement is then to somehow encode this bit in a temporal function. This has to occur in an environment which is thermodynamically driven. The aim is not to crystalize the information (that will come later), instead a relatively stable co-incident cyclical activation with temporal multiplexing achieves lower power requirement.
Signals can drift because the system is coherently driven. So the incident signal characteristics impart enough of a phase shift (with associated incident moment) to the base matrix. In this way cyclical processes can be encoded in spatial informational patterns.
note on non-locality here. This type of loosly-coupled amplifier would have limits on information transduction. However, the incident information is structured. It is structured by spatial distribution of von Neumann entropy. So a loosy-coupled amplifier can retain spatially distributed temporal structures that are coincident. The signal continues to have a non-local distribution in a sufficiently structured system. what do we mean by “sufficiently structured”? In the most general sense, there must be 1+ conduits for information to coincide on the “measuring system”. If sufficient coherence is maintained in the transformation to an electric field, then the field shape can be modulated by entangled information.
This coupling of internal state and external probability density field is able to develop complex re-organisation (over time). In every case with which we are aware, this process only occurs in living organisms, although the two could theoretically be decoupled (life and consciousness).
Now we have hopefully justified two further hypotheses:
- Transduction: there can be a conduit for statistical properties of nature to be encoded, first as temporally smeared temporal bits, then with spatially encoded tuning.
> we differ with the Spike-Timing Dependent Plasticity hypothesis here. While STDP encodes temporal information in spike density; this density matrix may not be of sufficient structural complexity to be suitable as the higher bandwidth encoding layer. Spike timing is more likely to be the spatially integrated field collapses. In DCT the descriptive level is the field modulations across the whole neural mass, including harmonics at ranges of approximately 2e at the first harmonic driver.
Theorem 2 \[ f(x)=2^e \]
A fundamental is subsegmented by harmonic inversion, waves can move as solitons
See Theorem 2.
This enables a spatial pattern to persist while streaming incident statistical properties. A multi-frequency binning filter. Perhaps like a complex plain of inverted tetrahedra, with tetrahedal fractalisation.
This fractal surface can act as a spatial amplifier. A method of frequency tuning is required, but temporally continuous frequency tuning is exactly what we need to learn.
- Modulated Spatial Encoding: that the spatial distribution of incident probabilities can be encoded via this transduction.
The next series of questions could be approached using a variety of mathematical and physical theorems. From the mathematical perspective, the descriptions in the rest of this chapter will be largely topological. From the physical perspective the theory is concerned with two main constraints.
- That information transduction is lossy, and generates heat.
- That a transductive network operating at the edge of coherence, requires a driven system. It is almost always far from equilibrium, apart from the sampling process.
We hypotheize that the interval of sampling and sampling width probably happens stochastically, but within an broad operating frequency range. From the information theoretical perspective, the receptive bands of this network can be peppered with a query…
What
Functionovertdescribes the temporal characteristics of this datagram
The most appropriate function is a temporal Fourier. In DCT this is a spatially distributed Fourier encoding. To visualize it, start with a view of a working brain, somehow the skull and brain matter has become transparent, and the electric fields of the neural aggregates are enlightened by their local coherence and coloured by their co-incident waveforms with other oscillators. So far this is a local description. Imagine watching a region or a slice of this view, the colours shift and some areas are brighter than others. Waves of colour spread out and coincide on other waves, the waves have different periods and at the backwash there is constructive interference. These waves slide over each other while peaks pop from the froth. The curious thing is that those peaks seen to dance to a different beat than the incident waves, their spatial distribution develops vortices.
In Bose Einstein condensates, a fluid can be rotated (driven). It slides over itself, until the moment of probabilistic excursion is large enough to require quantising. Two votices emerge from the possible configurations of the atoms in that condensate. Why not just one or 1.5 for that matter? The system has been driven to a probabilistic energy state where one is just as likely to switch to more than 2. And the energy level (in Hilbert information space - \(E^H\) ) of an indeterminate field driven by 1 vortex is greater than a stable two-vortex system. Similar constraints apply to the representation of nested informational structures.
This fundamental description of information fits well with the probability conservation rules in quantum theory
The bright spots in the spatial field also show entrainment of other oscillators. However, this is where non-locality is important. The temporal evolution of field densities can occur without neural tranduction delay. These long-range axons are important in seeding the driver signal but there is an overlay of more temporally dense information in the carried signal. And it is expressed in the temporally and spatially encoded filter function operating over the brain.
Now we can return to topology. We can think of this interface layer, whether it is the metaphor of coincident waves slapping against one another on wavy sandy beach; or the modulating electric fields of our (now imaginary transparent brain) as a hypersurface. However, we can also go further and hypothesize different holes in this surface. We also need to ask… what is outside this surface?
It is hypothesized that this topological object is a sliding knot. It is a pinch (also quite likely in the field energetic view) that gathers up the sheet of causal probabilities and slides across this surface. A point on the sheet will always be in the bulk of the sheet, but also pass inside the knot and then out again on a slightly different path, knots dissipate to the bulk - a process we’ll call crinkle.
Now, we have an informational space in the knot, it distorts incident probabilities, which creates a diff. Now we have the incident density field, the artefacts from transduction of this field, and the difference between the harmonics on that field last sample and this one.
The knot is permitted to roam over informational space. So it can re-sample, or even orbit a particularly rich informational environment. If the rhythms of nature can be understood, then phase difference can encode an orthogonal trace of this sample. The pattern has been learned, so now it can be quietened.
The neural field boundary over time is adapted to dampen the spatial pattern over shorter epochs. This allows higher density in the second-moment of integration. A learned pattern can be subtracted from the predicted incident. The better alignment (in time and space), the better is this dampening filter.
It is hypothesized that this filter can be activated to different states, and that these global states act as “memories”. An overlay looking for this particular arrangement of probabilities. If the incident is a good match to this memory state then further detail can be extracted from the higher frequency components. This occurs via spatial Fourier subtractive interference.
What does the evolution of sampling epoch look like in time and space? Let’s think again of just a sheet. The peaks and troughs are waving about, spatial nodes and antinodes of activity are becoming less coherent, but the sheet continues to wave and tremor. Then a clamp is applied, generated by a spatially distributed array of oscillators, field activity lowers, reaches a brief pinch as it changes state to a spatially distributed receptive field. The knot is sliding again, and the flood of new probabilities resonates through the spatial information field. Some neurons can’t help but be incident with a greater field density – this bit of information matters – and the neural aggregate responds by encoding that gradient within its own meso-scopic field coherence (neurotransmitter tuning is directed to better match the oscillatory signature (a type of load-matching to incident temporal signature) –I’ll see it again –. The evolution of this stochastic, multi-band illumination of the neural space is to carve it, with evolving field patterns. The longer-term adaptive neural changes are driven by the new “program” of that neural field generator complex. Of course this is also embedded in the broader field interactions of brain as a whole.
Back to our topological universe with one sheet and a single conscious sliding knot. The sheet is patterned with probability. Is this a wave, or a vortex or some pattern of those? The knot slides, slows and crinkles into the sheet. It becomes stiff, then relaxes and dissipates with a low glow.
In a longer temporal sample, it is ballooned out over a larger probability field. – Less resolution… but there’s some stuff I can see better out here –.
Over time this meandering with overlapping sample spaces reinforces statistical distribution of oscillator networks, imprinting a matching between [input] and [ongoing state] probability distribution. Such does the coupling of information temporal structure become encoded in the field generating capacities of the neural dampening field.
note that it is tempting to equate this state distribution to being similar to “resting state”. However resting state does not convey the information density relationships, and these states are not “resting” at all. They are representations of the temporal integration of many moments of receptivity. These receptive epochs are blips, – * move, crinkle, stop, unstick, remember * –. The signal in resting state fMRI data is likely to correlate most strongly with the energy required for the spatial electric field dampening “program” being ever-so briefly activated. Programs shift over time in the spatial nodes and antinodes.
The problem of complexity
Why should a system be organised in form \(A\) instead of form \(B\)? And how many forms are there? These forms can be though of as configurations of the probability density field. There is also a finite information density (the sum of forms -> weighted by their probability). Not all mathematically possible configurations are physically possible, there is a trade off in spatial coherence versus temporal density. This gradient defines the vortices of the bulk.
What this relationship implies is that the net of events over any probability sheet also has a trajectory. Informational structures, particularly non-local, temporal structures are limited in their complexity. The potential energy of many configurations is high.
which implies greater density of the sheet, but we must maintain a singular
Mutual Information
The Protoconscious Field
The Human Perspective: conscious processes
Implications of Theory
For the Hard Problem
The Binding Problem
- non-locality helps
Discussion
Testing the Theory
Speculations
- holography