Self-improvising Memory:

a perspective on memories as agential, dynamically-reinterpreting cognitive glue

Michael Levin1*
1
Department of Biology, and
Allen Discovery Center
Tufts University, Medford, MA

* Corresponding Author:
200 Boston Avenue, Suite 4600,
Medford, MA 02155-4243
Phone: +1 617 627 6161
Email: michael.levin@tufts.edu

Keywords: basal cognition, diverse intelligence, memory, learning,

morphogenesis

Running title: Self-improvising Memories

Abstract
Many studies of memory emphasize the material substrate and mechanisms by
which data can be stored and reliably read out. Here, I focus on complementary aspects:
the need for agents to dynamically re-interpret and modify memories to suit their ever-
changing selves and environment. Using examples from developmental biology,
evolution, and synthetic bioengineering, in addition to neuroscience, I propose that a
perspective on memory as preserving salience, not fidelity, is applicable to many
phenomena on scales from cells to societies. Continuous commitment to creative,
adaptive confabulation, from the molecular to the behavioral levels, is the answer to the
persistence paradox as it applies to individuals and whole lineages. I also speculate that
a substrate-independent, processual view of life and mind suggests that memories, as
patterns in the excitable medium of cognitive systems, could be seen as active agents in
the sense-making process. I explore a view of life and agency as a diverse set of
embodied Perspectives – nested agents who interpret each other’s and their own past
messages and actions as best as they can (polycomputation). This synthesis suggests
unifying symmetries across scales and disciplines, of relevance to research programs in
diverse intelligence and the engineering of novel embodied minds.

Introduction and overview

There is a paradox which points out that if a species fails to change, it will die out,
but if it changes, it likewise ceases to exist. The same issue faces all of us: if we do not
change, learning and growth is impossible. If we do change, doesn’t the current Self
cease to exist, in an important sense? This profound puzzle, which rests on pure logic –
not contingent facts of implementation or origin – highlights the deep symmetry of
existential concerns for agents existing at all scales, from subcellular organelles to
evolutionary lineages 1. Thus, it is also highly relevant to issues facing the engineering of
novel intelligences and fields from active matter research to AI and Artificial Life.
A solution to this problem has been suggested in the West as Process Philosophy
[10-12], and in the East as Buddhist and Indic conceptions of the no-Self or the non-duality
of Self vs. world [13]. The idea is to conceive of the Self2 as a process, not a thing, and to
consider snapshot Selves3 as low-dimensional projections of the deeper reality that both,
exists and constantly changes. Here, I discuss how evolution beat us to this solution, and
how it is implemented across spatiotemporal scales. Biology went all in, committing to the
fact that everything will change in ways that cannot be anticipated, and thus to the idea
that defending static Selves is futile. When it comes to information, biology commits to
optimizing salience and meaning [15-17], not fidelity – high level, agent-based
interpretation – the Gestalt, not the low-level details4. Crucially, this is not just about life
forms preparing for challenging external environments – it is even more about the
inevitable changes of our internal parts, which are subject to mutation, aging, cancer, and
hacking by other biota. More broadly, it’s about the passage of time.
The essential unreliability of the biological substrate was a key driver of an
architecture that is responsible for numerous fascinating capacities of the minds and
bodies of living form. Here, I discuss our nature as a hierarchy of essentially epistemically
vulnerable 5 agents, who must interpret the actions of our own parts and information
structures despite their being prone to error, decay, and autonomous behaviors. I argue

2
that this is not a limitation of living things but rather is the origin of life’s most unique
aspects, and of intelligence specifically. I provisionally call mnemonic improvisation the
dynamic ability to re-write and remap information (e.g., memories) onto new media and
new contexts, which occurs at many scales (behavioral, genetic, and physiological
memories). Thus, I propose that the ability to modify and interpret non-local memories as
one of the kinds of cognitive glue that reifies Selves (at all scales) [23,24].
Using the lens of memory, and specifically the transfer and remapping of memories
across space, time, and Selves in very diverse biological examples, I develop two key
ideas. First, that evolution makes problem-solving agents, not solutions, and that the
commitment to the reality of mutation and uncontrollable environment – to active
reinterpretation of information on-the-fly is the ratchet that gave rise to intelligence and
cognitive Selfhood6. Second, that the Self is a dynamical construct, not in the sense of a
misleading illusion to be eradicated, but in the sense in which all models are compact
perspectives created and continuously maintained 7 by agents under energy and time
constraints. Viewing the Self as an intelligent data pattern, which facilitates its own
transformation [35,36], is a helpful construct because it captures a core truth: while the
details of minds and bodies change, thought forms (salient cores abstracted from
experiences) remain and survive because they move across architectures that are able
to interpret them in new ways for sense-making as a gestalt [37-39], rather than committing
to their details. I trace back the origin of this property to the ubiquity of polycomputing8,
and propose the hypothesis that these dynamics enabled the evolution of minimal agency
in the thought-forms themselves: in other words, giving up the binary dichotomy between
the computational machinery and passive data, and exploring the idea that certain kinds
of information structures actively facilitate their transformation and remapping in ever-
changing cognitively-excitable media. As William James proposed, “thoughts are
thinkers”9.

“The material present in the form of

memory traces being subjected from time
to time to a rearrangement in accordance
with fresh circumstances – to a
retranscription.”
– Sigmund Freud writing to
Wilhelm Fliess on Nov 2nd 1896

Background: the shifting sands of Selves and memories

Confabulation is a kind of cognitive plasticity that emphasizes the present, the
future, and the gestalt over the literal past – it occurs when a mind actively modifies and
fits its beliefs to a current context, altering and reinterpreting memory10 data as needed
to preserve various psychological elements in the story it tells to others and to itself. For
example, a patient with an electrode in their brain who involuntarily laughs when the
clinician unexpectedly triggers the electrode will explain this odd behavior by saying “oh I
was thinking of a funny joke”, they don’t say “Weird, I was thinking of serious things and
my mouth started laughing” (because this undermines their self-model as an autonomous
agent). The same thing happens with split brain patients, whose left hand does things

3
unexpected by the verbal mind – the language center does its best to explain what
happened, by making up stories11. This is true not only with respect to high-level thoughts,
but even very basic aspects of our perception: the reason we don’t see the big gap in our
retinas occupied by the optic nerve is that the brain fills it in – a controlled hallucination
or “inpainting” of content we didn’t see, based on the surrounding information12. This
process not only edits the current data, but also the past and future. If 2 light bulbs, yellow
and blue, are lit in rapid succession, subjects not only see the light moving but actually
report it going through an intermediate green color and position – not only does our brain
invent a new color and light we didn’t see, but the experience is felt out of sequence,
perceived as if the inserted information was back-dated (because it can’t know to make
the middle ‘dot’ green until it’s actually seen the blue one), editing the memory stream to
insert the information in an Orwellian scheme to present a coherent story [43,44]. There
are also many examples of anticipation and prediction modifying the interpretation of
incoming stimuli [45-47] and long-term memories [48]. More generally, Hoffman’s work on
the interface theory of perception illustrates how little emphasis evolution places on
veridical perception [49-51].
These dynamic features of our mind13 soften our connection to the ground truth of
what actually happened in our history [52]. The fact that we cannot really trust our
memories of the past – in a much deeper way than just the inability to remember some
details - all point to the disorienting fact that we do not perceive the past as it was but
modify and rebuild our mental model dynamically [44,53]. We have no direct access to our
past – at each present moment, to recollect we have to reconstruct it dynamically from
the engrams [54] left in our brain (and body) by the activity of a past Self. One can think
about our apparent continuous stream of cognition as a series of frames or Selflets (like
slices in the bread-loaf model of Special Relativity), each one probably a few hundred
milliseconds thick (Figure 1).

(A) (B)

Figure 1: the temporal slices of a continuous being and memories as messages pass
along between its instances. Images used with permission by Jeremy Guay of peregrine
Creative. Note that it is not claimed here that this is the correct way to think about Selves
– this is an external (3rd-person) perspective that helps to understand certain invariants
in how biology uses information. This view misses the complementary perspective of the
experiential Self (1st person experience of the persistent flow itself).

From this perspective, memories are messages between agents separated across
time – each engram is a stigmergic [55,56] note left in our body by a past version of us14.

4
A view of temporal (vertical) memories as communication between our Selflets invites us
to think of memory as parallel to the horizontal communication we do with others’
Selflets15. And like all messages, they need to be interpreted16; indeed, von Foerster [58]
emphasizes the symmetry between foresight, hindsight, and insight: the key is not the
passage of time but the need to make sense of the future and the past equally. Memories
are also an answer to “what it’s like to be past-me”, in the sense of Nagel’s “what it’s like
to be a bat” [59] – we can’t actually be our past self, but we can try to reconstruct it based
on the evidence provided by reinterpreting the clues left by them, just as we do for
heterophenomenology of other minds [60]. The task of finding optimal meanings for
memory traces is an exercise in intelligence and satisfaction of epistemic hunger –
squeezing stimuli (whether from environment or from one’s past self) for actionable
wisdom17.
All this is often considered an unavoidable but undesirable bug in cognitive
architectures: confabulation in AI, the notorious unreliability of court witness testimony,
and Humean philosophical arguments (e.g., Boltzmann brains, Descartes’ deceptive
demon, etc.) seem to undermine the solidity of our personal identity18. They are disturbing,
when people first hear about them19. Here, I argue that this is actually a powerful and
important feature, arising from our ancient origins at the dawn of life and helping to
understand many phenomena beyond brains and behavior. I propose that the necessity
for mnemonic improvisation – the active rebuilding of the content of any (proto)cognitive
system – was the source of morphogenetic robustness and eventually conventional
intelligence. The ability to improvise and make sense of your world in real-time, and the
commitment to change (not just to persistence) over an allegiance to the details of a past
history, are a fundamental biological20 strategy deployed at many scales, with massive
impact.

“No man ever steps in the same river

twice. For it’s not the same river and he’s
not the same man.”
-- Heraclitus

Remapping memories: beyond storage and simple modification

Consider metamorphosis. Caterpillars largely disassemble and remodel their brain
in becoming butterflies. The hardware is strongly refactored as the creature shifts from a
soft-bodied machine that lives in a 2D world and eats leaves to a hard-bodied one that
flies through 3D space and drinks nectar. The controller is replaced, but some memories
are known to remain [62-64], and this has also been studied in other model systems
including vertebrates [64-69]. It’s tempting to focus on the important question of “where is
the memory stored” during this process (although of course modern neuroscience has
been studying the dynamic, complex process of memory formation, encoding, and
storage as consolidation). But there is an even deeper issue here, concerning the
encoding, decoding, and interpretation of the memory engrams.
Specific memories useful to the caterpillar – what actuators to fire to make use of
the fact that leaves are to be found with a specific color cue (e.g., associative memories
cued by a bright red disk) – are entirely useless to the butterfly. The butterfly’s body is

5
controlled very differently - soft bodied robots need a different controller than ones which
have hard elements that can be pushed during actuation. Its visual system is different,
and most of all, it doesn’t eat or care about leaves – it wants nectar. The process of
memory and recall not only functionally generalizes (abstracts 21 ) the information (not
leaves, but “food”) but most importantly, remaps22 it in a way that the basic relationship
learned maintains its salience in a new body and environment. Stated another way, the
being which radically changes doesn’t bring with it specific memories into its new, higher-
dimensional life - it brings the deep lessons of its experience in a way that it can deploy
in its new embodiment.
Metamorphosing insects and planaria that maintain memories and re-imprint them
from tail fragments onto a newly-regenerating brain, and anterograde amnesia patients23
who update external notepads to keep track of their life - all seem like extreme examples
or rare outlier cases of weird memory dynamics. But to various degrees, we are all
undergoing this process of sense-making by improvisation and decoding of clues 24 .
Ricard Solé calls ant colonies “liquid brains” [74] because their subunits don’t keep a stable
connectivity pattern (like a solid) but shift around during the cognitive life of the colony
mind. Ants colonies are liquid in terms of spatial position of the ants, but we are all brains
that are liquid in time because of the shifting relationship between our Self-model and our
cognitive content across moments. The mental content of each Selflet is not welded to
that of the previous one in a linear, fixed pattern; it slides around - influenced25 by the
content of past slices but not over-committing or taking them too seriously 26 as they
reinterpret them to suit new circumstances.
Even mammalian brains change over time. The hormonally-driven brain
maturation of puberty may not be as drastic a change as happens to a caterpillar27, but
it’s significant enough to strongly re-model our preferences and priorities. At a small scale,
our brain and body are constantly turning over in terms of molecules and cellular
components (including dynamic synapses) and yet centenarians have functional
memories from their childhood. Critically, normal recall in the brain28 is not a pure “read”
operation – accessing a memory changes it [81] – a fact exploited by therapists treating
trauma, and by everyday tasks involving creativity and problem-solving [82]. This also
underscores the continuous, dynamic nature of memory. It’s much more than robustness
– it’s not about finding a way to lock a memory in place and maintain its details against
noise and perturbations29, it’s about being able to remap, adapt, and improvise to extract
the salience of memories into new contexts. Ubiquitous confabulation [41,84-86], not just
the stark examples seen in split brain patients whose language center concocts a story
about why the left hand was doing something surprising, is the central invariant across
numerous phenomena in evolutionary, developmental, and neuro-biology.
The considerable literature on (behavioral) memory transplants across animal
bodies reveals that the ability to re-interpret memories functions between Selflets which
are not somatically contiguous (not part of the same organism persisting through time).
The results of moving pieces of, or extracts of, brains of trained animals into naïve
subjects [87-95], and (as yet uncertain) claims of memories transferred by heart-lung
transplants [96,97], point to an even deeper kind of remapping capacity. First, the
movement of memories by brain extracts [89] does not require careful placement – in
David Glanzman’s experiments for example, he injects the extract in the general vicinity
of the relevant Aplysia nervous tissue. It is remarkable that it doesn’t matter where

6
precisely the molecules go, the new brain will extract their meaning and instruct the
behavior of the recipient Aplysia. In our (not yet published) work, implants of small pieces
from bioelectrically- (not genetically-) induced 2-head planaria can convert the normal,
intact host to a propensity to become 2-headed if cut; the orientation or location of the
piece doesn’t matter. The recipient tissues can interpret the aberrant information from this
weird prompt injection and remap it onto their intact body, keeping it for the future when
another round of cutting calls upon them to reveal their altered pattern memory30.
Even more remarkable is the observation that one can inject an odorant molecule
into a frog egg, and the resulting animal will seek out that odor in its search for food [99].
Consider what that must mean – mechanisms inside a single cell have to extract
information from this new input, and convert it into the processes needed to modify the
genetically-encoded hardware of the nervous system to result in a new behavior that is
functionally tied to the sensing of that molecule. Here we have: detection of novelty (how
does it know to pay attention to this molecule out of millions of other, native, molecules
inside the egg?), remapping across scales (from single cell to whole organism), and
preservation of salience across novel scenarios (from biochemical signals in cytoplasm
to the dynamics of nervous system function in guiding food-seeking behavior). The
engrams seem less like encoded memories31, and more like a kind of prompt, with a lot
of the hard work being done by the encoding/decoding process which adds interpretation,
context-sensitivity, and generalization (i.e., intelligence) to the process.

Beyond the brain: bowties everywhere

I have argued elsewhere [72,101,102] that there are profound symmetries between
the collective intelligence of neurons in navigating 3D space and that of non-neural cells
navigating anatomical morphospace. Many algorithms, competencies, failure modes, and
molecular mechanisms are shared between the self-assembly/repair of the body and the
emergence/maintenance of the mind. This symmetry allows us to think about the
relevance of the memory-remapping concept beyond brain and behavior.
The first concept that carries over naturally is the notion of triggers – low information-
content stimuli that kick off complex, spatiotemporally-appropriate responses because the
receiver is, to a degree, intelligent32. A simple voltage state imposed on somatic cells can
induce them to build a complete vertebrate eye [76] - of course that simple biophysical
state can’t contain all of the information needed to make an eye 33 . Another voltage-
modifying reagent (Monensin) triggers tails to grow in tadpoles, but legs to grow in froglets
[103,104], and never the converse. Here, the same bioelectric stimulus is functionally re-
interpreted based on the specific context, preserving salient, high-level information like
“build whatever normally goes at this location” (in the case of the appendage regeneration
trigger). Remapping is especially salient in regeneration of organs by bioelectric patterns
that were much bigger (e.g., in the case of a planarian fragment which obeys the
bioelectric pattern memory of the original body34, but only carries a small part of it when
it is cut out), or present in tissue that is very different (e.g., a salamander arm that regrows
from a body that is missing those structures and has to get its information from other parts
of the body). The importance of the Gestalt, and the functional coupling from the large-
scale target morphology information through to the molecular events inside cells are also
seen in the example of tails transplanted to the flank of a salamander, which gradually
remodel into a limb-like structure [105-107]. This transformation includes the cells at the tip

7
of the tail, which are correct in their local environment but must change to fit the global
pattern (Figure 2).

Figure 2: Remodeling of a transplanted tail into a limb-like structure in a salamander.

Image from [105].

Another aspect of this is the compression/expansion cycle of metazoan organisms.

Each organism doesn’t make a copy of itself – it undergoes a massive compression,
producing an egg, which then has to re-inflate during the process of self-assembly and
engage with a possibly different world using possibly different (mutated) parts (a kind of
de-generalization, to apply past patterns of wisdom to novel scenarios). Some somatic
information doesn’t make it through, but it’s increasingly recognized that a lot of generic
states (like “stress”) does [108-111]. The importance of a process that squeezes down
complex states into a compact, generalized representation (a so-called bowtie
architecture of information, Figure 3) is routinely used in AI (the autoencoder).

(A) (B)

Figure 3: Bow-tie architectures feature a low-dimensional compressed medium at the

center through which information must come, and active decoding and context-sensitive
interpretation on the output. Left panel – developmental processing of morphogenetic
information; Right panel – a typical machine learning architecture. Images used with
permission by Jeremy Guay of peregrine Creative.

Tadpoles produced with no endogenous eyes, but bearing an eye placed on their
tails, can see quite well – they learn effectively in visual assays [112]. These ectopic eyes
connect to the spinal cord (not brain), or sometimes the gut. Why doesn’t it require many
generations of mutation and selection to enable this new sensory-motor architecture to
work effectively? This kind of plasticity is remarkable on the one hand, as we expect the
lessons of the evolutionary history to be passed down in a deterministic fashion to the

8
next generation. On the other hand, consider development as an instance of regenerative
repair, with each developmental stage a birth defect relative to the next stage, and must
be dynamically repaired. Regeneration in adult salamanders, of a limb from the rest of
the body, is not nearly as impressive as the trick that all35 metazoa do – regenerating the
entire body form one cell.
Regeneration of limbs and eyes is not surprising when one considers that the
organism has to be able to construct itself from 1 cell, or from a mixed-up collection of
cells36. But it’s more than uncertainty of the external environment that limits the relevance
of past generations’ specific lessons: it’s also uncertainty about one’s own parts.
Engineered polyploid newts complete normal development despite their much larger cells
and extra copies of chromosomes [116,117]. Their kidney tubule diameters (Figure 4) stay
normal, being made of up fewer, but larger, cells. If their cells are made truly enormous,
then just 1 cell will bend around itself, to complete the relevant task in anatomical
morphospace. This requires using a different molecular mechanism (cytoskeletal bending
instead of cell:cell communication of tubulogenesis) – calling up different modular
subroutines to accomplish their high-level goal state as needed, while circumstances
(indeed, their inner components) change unexpectedly 37.

Figure 4: Cross-section of newt kidney tubule at different ploidy levels. Re-drawn by

Jeremy Guay of Peregrine Creative from [116].

Think of the task facing such an embryo coming into the world. It cannot take for
granted how much genetic material it will have [117], how many cells [120-122], or what size
cells [116], it will have (in addition to any external damage it might sustain, or the chemical
details of its environment – even the internal parts cannot be assumed to be the same as
those with which its past genome was forged)38. Beyond the incredible robustness of
standard form and function, we see adaptation to extreme changes of conditions, with no
need for transgenes or genomic editing, in the spontaneous emergence of Xenobots [123-
125], Anthrobots [126], and plant galls [127] (constructions of leaf cells hacked by signals
from a non-human bioengineer to produce a totally new and complex pattern, Figure 5).

9
Figure 5: Novel forms made by genetically-normal plant leaf cells, when prompted by
signals from a wasp embryo.

This is also probably why chimerism and bioengineering, at all scales, works –
coherent outcomes often result from combining not only divergent living components from
different lineages [128] but also totally un-natural and novel engineered components, from
nanomaterials to smart implants [129-134] (Figure 6). Biological components are primed to
accommodate and exploit whatever materials and processes are in their vicinity (as useful
functionality and computations).

Figure 6: A schematic of how biological systems can adaptively incorporate foreign

material (whether evolved or engineered) at every level of organization. Taken from [102];
image by Jeremy Guay of Peregrine Creative.

We don’t know how this remarkable plasticity works yet, but one thing it does, to
some extent, is free the downstream (future) agent from the restrictions of the upstream
(past) agent. If we consider genomic DNA to be the engram39 of the evolutionary scale
individual (the entire lineage as one huge time-extended agent), then its interpretation
and usage occurs via remapping and repurposing as needed, just as conventional
memories do in smaller-scale individuals.

10
 Note that this bowtie architecture – which forces a compression 40 of data to a
generative kernel that then has to get re-inflated and elaborated – is also a common
feature of biochemical, bioelectrical, and biomechanical pathways (Figure 7). That’s at
the subcellular scale; at the super-organism scale, recent data also reveal that a very
simple parameter – waves of ATP – can mediate the complex morphogenetic information
being shared and processed within groups of embryos 41 defending their anatomies
against teratogenic influence [138]. The same architecture can be used to think about
learning (when one side of the bowtie is the environment), training (when the environment
contains agents with agendas that send messages with the goal of changing the behavior
of other agents), and other kinds of communication both laterally and vertically (to one’s
future self).

Figure 7: Schematic representations of bow-tie architectures in biochemical, bioelectrical,

and biomechanical circuits. For example, causal parameters such as tension or resting
voltage potential across the membrane (Vmem), as well as specific genes, can act as
information hubs that ensure compression of signals within biological control networks.

What underlying parameter is represented by the spectrum ranging from the

hardwired, mosaic C. elegans, the intermediately plastic amphibia and (embryonic)
mammals, and the extreme plasticity of planaria? I propose it is the willingness to
confabulate in anatomical space – a pattern extraction and completion42 architecture that
does not take priors too seriously and assumes that it will have to develop models of itself,
its problem space, and its internal and external affordances on the fly (a kind of
“beginner’s mind”, emphasizing forward-looking creativity over past-constrained
structure). This suggests a strategic parameter for morphogenetic/cognitive systems akin
to “r vs. K” selection in evolutionary ecology [140]. Our in silico simulation results [141]
show how this kind of process can give rise to remarkable lineages, such as planarian
flatworms, which are extremely resistant to transgenesis, aging, cancer, and injury
despite43 their incredibly noisy genome because they have fully committed to a strategy
that over-rides genetic details with large-scale pattern completion [141].

11
Beyond biology
This idea of complex data acquiring robustness by being squeezed into a bowtie
bottleneck can be found outside the examples of development, evolution, and behavior
(see examples in Table 1). For example, communication between two humans can be
seen as the bowtie scheme, with language as the hub [143]. We cannot simply exchange
tables of neuronal state data with other humans, animals, or AI’s – those details cannot
be remapped from brain to brain directly. But squeezed down into the low-bandwidth
channel of language, the salient aspects of the message can be remapped by the
listener’s brain into whatever biophysical changes are needed to implement the
interpretation that makes the most sense to them44 [144,145].
Similarly, the scientific research process can be seen this way. Scientists
accumulate a coherent world-model after years of experimental observations and
analysis, and present it as a “published manuscript” – a low-bandwidth, impoverished
stimulus from which other scientists will extract information that may change their own
world-view and mental content (and maybe in ways other than what the author intended).
Poetry and art achieve much more drastic encoding/decoding dynamics when they pass
from artist to viewer 45 , requiring massive amounts of personal interpretation and
modification on the decoding end of the recipient, while baking recipes require less.

Table 1: diverse examples of information compression and re-interpretation

Scenario/Scale Bow-tie Hub Node Remapping Process
Developmental Morphogenetic problem-
Egg
lineage solving competencies
Increased morphogenetic
Hyper-embryo Calcium/ATP signal through
problem-solving
groups [138] medium between embryos
competencies
Hologram 46 Holographic film Laser interrogation
Planaria remapping Vmem map
Regeneration Bioelectric pattern
from whole to fragment
Single Cognitive Self Neural interpretation of
Memory medium (engrams)
across time engrams
Transplants between Extracts (RNA) or tissue Neural interpretation of
Cognitive Selves implants engrams
Neural interpretation of
Communication Language [150]
spoken/written messages
Replaying same song on a
Song Written musical score
totally different instrument
Interpretation of data by
Science, short-term Talks/manuscripts
scientists in same/other fields
Interpretation and use by the
scientific community; some
Ideas and paradigms,
Science, long-term ideas become immortalized as
explanations
engineering tech; others are
revised, or forgotten.

12
 Personal interpretation and
Art Poetry, paintings, etc.
finding personal meaning
Adapting (or not) as
Society religious frameworks technology and science
advance

There are possibly implications here beyond conventional 3rd-person science. First,
a commitment to meaning and interpretation could potentially be useful in a psychological
setting, whether personal or therapeutic, in the sense that techniques could be developed
to help reinterpret somatic and cognitive memories of past trauma47. In the short term,
two individuals may see the same event (e.g., chopping down a tree) very similarly and
agree on the micro details of what they saw. But on a longer timescale of interpretation,
each Self views this differently (perhaps, as value-creating progress vs. destruction of
ecosystems) the memory is embedded in a different mental structure in which the
meaning of the event (and thus its contributions to future thought processes) is private
and unique, while the details were the same. The answer to “what did you experience?”
is critically tied to the timescale on which the answer is meant.

What it means, and what next

The basic fact that the species with the noisiest genome is also the species with
the best regeneration, immortality, and cancer resistance has been a puzzle [142] for a
century, although is rarely talked about within a paradigm in which genes should encode,
and determine, phenotype. It seems that the underlying noise and unreliability of
biological matter is not a bug but a feature: we got morphological (and behavioral)
intelligence because of, not in spite of, the vulnerability of the substrate 48 . The
competency of our material means that selection cannot see the quality of genomes
clearly 49 . More of the evolutionary work thus has to be done on the competency
mechanisms, not the structural components, leading to a positive feedback loop – an
intelligence ratchet based on the remapping of information to new contexts50. It is an
important direction for future research to understand what factors determine where, on
the plasticity and intelligence spectrum, any given species comes to reside (with respect
to morphological, behavioral, biochemical, and other problem spaces).
While our engineering efforts build computational algorithms on top of extremely
reliable hardware, life doubled down on the “play the hand you’re dealt” strategy, not
overtraining on evolutionary priors and committing very early to sense-making at all
scales. And it’s not just about the robustness of the hardware, it’s about the semantics
of the software and its data. Current computing architectures view the data as a passive
object to be interpreted in one way, from the perspective of the user – it’s the job of a
conventional computer to facilitate that one interpretation and faithfully store and transmit
data. Biology uses a polycomputing architecture [40], committed to on-the-fly
confabulation 51 , in which a heterarchical soup of competing, cooperating, multi-scale
agents vie to develop viewpoints from which the molecular and biophysical states can be
interpreted (and hardware affordances hacked and reused) in whatever way they best
can52.
One major area for future work is the nature of the memory medium and how it
facilitates reinterpretation and invariance of saliency; if engrams are not in the synaptic

13
structures [165], where are they? My current hypothesis, driven by the above multiscale
perspective [166], is that there is no single substrate for memory53. Every component of
the system, including but not limited to those that bubble up to conscious recollections,
could be using everything in its environment as an interpretable scratchpad54. The deep
levels of biological structure and dynamics offer an incredibly high-dimensional reservoir
[171-173] (referred to as the senome in [19,174]) that can be exploited for memory
remapping55. On this view, neuronal networks are not so much used for holding memory
as they are for learning to interpret the engrams embodied by subcellular components.
An immediate opportunity for future development concerns the mechanisms and
algorithms of informational remapping. By modeling networks of self-organizing
observers (and tracking the emergence of autopoietic perspectives [177] and
polycomputing competencies) we may arrive at testable hypotheses about how memories
can be remapped to new contexts. Reservoir computing may also offer interesting tools
for understanding how rich biological structures can be interpreted on the fly. Model
systems for research on moving memories range from in silico (transfer of learned
memories across gene regulatory network models with different geometries) to in vitro/in
vivo, such as movement of pattern memories from 2-head planaria to wild-type hosts via
tissue implants (we have already done this), and movement of behavioral memories (such
as nicotine addiction) from human donors through Anthrobots (will they self-medicate?)
to the rat hosts in which we can implant them. The lessons we learn will likely be
actionable for making new kinds of truly bio-inspired AI architectures [178], and also help
design strategies for regenerative biomedicine that target how cells and tissues interpret
the interventions we provide in the form of biochemical and bioelectrical stimuli
[101,179,180].
Computational models of this process can investigate whether for example, spatial
information can be remapped into temporal information. Some great work is being done
in this area in the field of Alife, starting with the original Bongard experiments of a robot
which did not know its shape and had to discover it, and then re-use that information when
its body shape was abruptly changed by damage [181], and other work on transfer of skills
in dynamical recurrent neural networks [182] or deep learning in embodied robotics [183].
Another possible model system is that of sleep, and the difficulties of the waking self to
make sense of the memory traces left by the sleeping self. Perhaps some of the tools of
dream analysis could shed light on the more generic memory reinterpretation process56.
A specific hypothesis that can be investigated is the extent to which memories,
especially non-local ones, actually reify the agent to whom they belong, in a sort of
feedback loop in which the Self elaborates and maintains memories, which in turn
reinforce its existence beyond that of its parts. Can they serve as one of the cognitive
glue mechanisms that makes the whole more than the sum of its parts? For example, a
rat trained to associate pressing a lever to get the reward. No individual cell has both
experiences – the skin cells of the paws experience the lever, the cells in the gut get a
metabolic reward. This associative memory belongs to the “rat” – the collective, and to
none of the cells individually. Could having such memories actually help create the
organism in a real sense? The body houses and maintains them, but perhaps they in turn
make the virtual governor [188,189] that we call the cognitive Self more real? We are
currently investigating this in the context of our training of associative memories into gene-
regulatory networks and other substrates, with measures of integrated information to

14
determine if the forming of memories increases the self-coherence of composite systems.
Many such experiments suggest themselves.
Another area for development concerns the blurring of the boundary between
passive data and active cognitive architectures which hold them. For example, it has been
argued that what persists are algorithms [190], which is a powerful way to think about
active information. But what if we go further on the continuum, beyond passive and even
active data, to basal agency. Perhaps there is no principled, sharp distinction between
data and algorithm, between memories and minds – just a continuum of different degrees
of agency of understander and understandant [191]? What if, in James’ words, “thoughts
are thinkers” in the sense that they actively help (perhaps, cooperate and compete for
opportunities57 or use each other as affordances in a heterarchy) to be remapped and
utilized by cognitive systems58? What if memories, which are not static details but active
deep patterns, can resonate59 with a cognizer or even a group of cognizers (in the case
of federated inference and belief sharing [196]) in a kind of circular causality [197-199] in
which they exert some minimal agency as they shape the mind of the thinker and thus
help construct the niche60 within which they will be utilized in subsequent time steps61?
But isn’t it crazy to think that an agent, even a minimal kind, can be just a
metastable pattern in an excitable medium – a temporarily persistent pattern62? But that
is what we are too – temporarily persistent patterns which self-reify our boundaries from
the outside world via active inference and interpreting our environment to tell coherent
stories (models) that hold us together and make us more than the sum of our parts. And
on an evolutionary scale, the thoughts of a lineage mind, of which each individual creature
is a hypothesis about the outside world, are definitely active agents (they are the
conventional, medium-scale agents we recognize every day as behaving life forms).
Thus, there is the possibility to develop a rigorous model for how thoughts can
scale up to become thinkers of their own63, spawning off new thoughts as a virtual stack
of resonances (perhaps by closing some “strange loop” [36] of self-reference and self
reification). But on a more practical level, these ideas could be tractable now in a
biomedical context, to develop tools to model, recognize, and control persistent
physiological patterns that represent stresses, priors, setpoints, etc. as basal thought
forms in the mind of the collective intelligence of the body operating in physiological,
metabolic, and transcriptional state spaces. It is not hard to imagine how managing such
self-perpetuating states could be useful for biomedicine [80]. We are using emerging tools
of 4D physiology, in vivo imaging, and optogenetics to test these hypotheses.
Although a full account will be given elsewhere, it is tempting to mention the possible
links to consciousness. Could consciousness simply be what it feels like to be in charge
of constant self-construction, driven to re-interpret all available data in the service of
choosing what to do next? In this sense, cognition is essentially freedom from the past;
cognitive Selves could be systems that are not committed to their own past and their own
memories. Paradoxically, biological Selves do not take themselves too seriously in the
sense that they are not committed to a fixed set of meanings established by their prior
Selflets – their freedom consists not only in actions, but in forward-looking sense-making
of their own mental content. Letting go of the past Self and living life forward is the
commitment to make the best of internal, not only external, information. This is in broad
agreement with Solms’ idea that consciousness is palpated uncertainty about the outside

15
world [46,202]; I propose to expand this idea, with the hypothesis that consciousness is
palpated uncertainty about your own memories and internal states.

Conclusion
This work is a continued exploration of classical ideas in which the construction of
the mind and that of the body are tightly linked [38,39,203-208]. Here, I’ve tried to provide
the strongest versions of claims along these lines, to help crystallize how far these ideas
could possibly go and what they may imply. It is entirely possible that the most useful lens
on the topics discussed above will turn out to be a more moderate version – an
intermediate hybrid between the conventional picture and the full-bore picture proposed
here.
Organisms that cannot handle novelty – of their own parts and memories, not just
of their environment – will not be evolvable and are likely not to persist. The symmetry of
precariousness and uncertainty in the environment and in the internal milieu underscores
the tenuous line between organism and outside world – both are mysteries that must
continuously be grappled with, by life’s essential adaptive sense-making competencies. I
hypothesize that remapping memories onto a new body (and indeed, maintaining
memories across a lifetime) works because development is also regulative and able to
remap genetic information into new scenarios, as in the salamander kidney tubule and
similar examples. Morphogenetic and cognitive intelligence (problem-solving, adaptive
competencies and plasticity) arise because of the constant need to battle dispersive,
degrading tendencies of the material. This cannot be done in an invulnerable, static
substrate but requires active reinterpretation in real-time. Regeneration, regulative
development, hackability, and the interoperability of life in weird new configurations are
all consequences of repair and remapping processes. Thus, life, at all scales, from the
microscopic to the trans-personal, is all about repairing and defending the Self, because
the Self is a construction 64 , in the most powerful and useful sense - an adaptive,
actionable, embodied story which holds components together and enables intelligent
navigation of a problem space.
Selves are simultaneously a construct in the mind of observer(s), including itself,
and real, causally-important agents that live, suffer, die, strive, and matter. I think this also
helps us understand what an Observer is [209]. An Observer is real and significant to the
extent that the content of what it observes makes a difference to them and their future
behavior – they will act differently 65 based on their interpretation of the signals they
receive (unlike for example, a telescope, or even photographic film, which forms a
memory record but does not analyze in a way that links up to any cybernetic perception-
action cycles). Observers interpret what they sense from their own perspective; their
allegiance is to extracting meaning, not preserving accurate details. It is the mark of
significant Observers that they exert their agency not in seeing an event or set of states
as they are, but in weaving a coarse-grained compression that adaptively captures what
it sensed in a process of autocrine story-telling that will be easy to exploit in future
behaviors. The criterion for being an Observer is that it is fundamentally committed to re-
interpretation and meaning, not micro-scale realism [49-51,210]. They bring their own
history, perspective, biases, hedonic valence, and predictive coding strategies about
what’s important in sensory and interoceptive experiences [29,44,46,52,202,211,212].
Currently, only biological beings are clearly recognized to be able to do this significantly,

16
but these capacities do not require protoplasmic substrates or an origin through random
mutations, and likely can be engineered in an endless range of novel forms of embodied
mind [74,213-215].
Tracking back the causality in this space of ideas suggests the following.
Behavioral intelligence in 3D space evolved from morphogenetic problem-solving
competencies in anatomical space [73,102]. Those competencies in turn were required by
the inevitability of evolutionary and physiological damage (mutations and injury). That
capacity is implemented by the polycomputing property – ability to see the same physical
process as computing and providing different functions depending on perspective66. And
that, in turn, comes from our multiscale architecture: at every level, we are a collection of
minimal agents which are sense-making and hacking everything around them (their parts,
their neighbors, etc.) as best as they can to fulfill their minimal but vital agendas [216].
Much of biology and cognitive science can thus be seen from the perspective of this
fundamental paradox: "do I still exist if I change"? Creatures, whether biological or
technological (or both), that resolve this Zen-like riddle, don’t just persist – they thrive.
The inorganic world, and much of today’s engineering, are stuck in an object-centered,
matter-first view [217]. Biology, from the start, embraced a process ontology in which
perspectives and agency are primary; thus, change is the driver of intelligence, and
perspectival storytelling is a primary mechanism by diverse minds transform and grow. I
think the lesson to take from this is to embrace the dizzying freedom of breaking with the
goals and structures handed down to us from our evolutionary and personal past, and
take on the responsibility of writing our own, improved somatic and mental patterns and
values for the future. What engrams do you want to leave to your own future Self, and to
humanities collective future? Knowing that they too will not interpret them in the way you
may envision now, it is still wonderous to imagine every act as a benevolent
communication event to a future being.

Acknowledgements: I thank the following people for comments on the manuscript

and/or helpful conversations: Reed Bender, Josh Bongard, Douglas Brash, Anna
Ciaunica, Pranab Das, Daniel C. Dennett, Thomas Doctor, Bill Duane, Chris Fields, Karl
Friston, Sam Gershman, Adam Goldstein, Gregg Henriques, Eva Jablonka, Timothy
Jackson, Bernardo Kastrup, Stuart Kauffman, Katherine Peil Kauffman, Benjamin Levin,
Perry Marshall, Nicolas Rouleau, Anil Seth, Richard Silberstein, Ricard Solé, Mark Solms
(especially for the Freud quote), Elizaveta Solomonova, John Vervaeke, Richard Watson,
Olaf Witkowski. I also thank Patrick Erickson, Ben Hartl, Hananel Hazan, Santosh
Manicka, Jordan Strasser, and Yanbo Zhang, as well as other members of the Levin lab
for many helpful discussions.

17
References

1. Baluska, F.; Reber, A.S.; Miller, W.B., Jr. Cellular sentience as the primary source of
biological order and evolution. Biosystems 2022, 218, 104694,
doi:10.1016/j.biosystems.2022.104694.
2. Baluska, F.; Miller, W.B.; Reber, A.S. Cellular and evolutionary perspectives on
organismal cognition: from unicellular to multicellular organisms. Biological Journal of
the Linnean Society 2022, doi:10.1093/biolinnean/blac005.
3. Reber, A.S.; Baluska, F. Cognition in some surprising places. Biochem Biophys Res
Commun 2021, 564, 150-157, doi:10.1016/j.bbrc.2020.08.115.
4. Baluška, F.; Levin, M. On Having No Head: Cognition throughout Biological Systems.
Front Psychol 2016, 7, 902, doi:10.3389/fpsyg.2016.00902.
5. Lyon, P.; Keijzer, F.; Arendt, D.; Levin, M. Reframing cognition: getting down to biological
basics. Philos Trans R Soc Lond B Biol Sci 2021, 376, 20190750,
doi:10.1098/rstb.2019.0750.
6. Levin, M.; Keijzer, F.; Lyon, P.; Arendt, D. Uncovering cognitive similarities and
differences, conservation and innovation. Philos Trans R Soc Lond B Biol Sci 2021, 376,
20200458, doi:10.1098/rstb.2020.0458.
7. Lyon, P. Of what is “minimal cognition” the half-baked version? Adapt Behav 2020, 28,
407-424, doi:10.1177/1059712319871360.
8. Lyon, P. The cognitive cell: bacterial behavior reconsidered. Front Microbiol 2015, 6,
264, doi:10.3389/fmicb.2015.00264.
9. Lyon, P. The biogenic approach to cognition. Cogn Process 2006, 7, 11-29,
doi:10.1007/s10339-005-0016-8.
10. Goodwin, B.C. A cognitive view of biological process. J. Social Biol. Struct. 1978, 1, 117-
125.
11. Solch, D. Wheeler and Whitehead: Process Biology and Process Philosophy in the Early
Twentieth Century. J Hist Ideas 2016, 77, 489-507, doi:10.1353/jhi.2016.0021.
12. Whitehead, A. Process and reality : an essay in cosmology; Free Press: New York, 1978.
13. Kapstein, M. Self, Non-self, and Personal Identity. Oxford Bibliography 2021,
doi:10.1093/obo/9780195393521-0147.
14. James, W. The principles of psychology; H. Holt and company: New York,, 1890.
15. Vervaeke, J.; Lillicrap, T.P.; Richards, B.A. Relevance Realization and the Emerging
Framework in Cognitive Science. Journal of Logic and Computation 2009, 22, 79-99,
doi:10.1093/logcom/exp067.
16. Fields, C.; Levin, M. How Do Living Systems Create Meaning? Philosophies 2020, 5,
doi:10.3390/philosophies5040036.
17. von Foerster, H. Logical Principles of Information Storage and Retrieval. In Electronic
Handling of Information, testing, and Evaluation, Kent, T., Beltzer, and Goldstein, Ed.;
Academic Press: London, 1967; Volume BCL Publication 150, pp. 123-147.
18. Beer, R.D.; Di Paolo, E.A. The theoretical foundations of enaction: Precariousness.
Biosystems 2023, 223, 104823, doi:10.1016/j.biosystems.2022.104823.

18
19. Miller, W.B., Jr.; Baluska, F.; Reber, A.S.; Slijepcevic, P. Why death and aging ? All
memories are imperfect. Prog Biophys Mol Biol 2024, 187, 21-35,
doi:10.1016/j.pbiomolbio.2024.02.001.
20. Man, K.; Damasio, A. Homeostasis and soft robotics in the design of feeling machines.
Nature Machine Intelligence 2019, 1, 446-452, doi:10.1038/s42256-019-0103-7.
21. Jablonka, E.; Ginsburg, S. Learning and the Evolution of Conscious Agents. Biosemiotics-
Neth 2022, 15, 401-437, doi:10.1007/s12304-022-09501-y.
22. Froese, T. Life is Precious Because it is Precarious: Individuality, Mortality and the
Problem of Meaning. In Representation and Reality in Humans, Other Living Organisms
and Intelligent Machines, Dodig-Crnkovic, G., Giovagnoli, R., Eds.; Springer International
Publishing: Cham, 2017; pp. 33-50.
23. Rosen, R. Anticipatory Systems in Retrospect and Prospect. Gen Syst 1979, 24, 11-23.
24. Rosen, R. Anticipatory systems : philosophical, mathematical, and methodological
foundations, 1st ed.; Pergamon Press: Oxford, England ; New York, 1985; pp. x, 436.
25. Pietsch, P. Shufflebrain; Houghton Mifflin: Boston, 1981; p. 273 p.
26. Torday, J.S.; Miller, W.B., Jr. The resolution of ambiguity as the basis for life: A cellular
bridge between Western reductionism and Eastern holism. Prog Biophys Mol Biol 2017,
131, 288-297, doi:10.1016/j.pbiomolbio.2017.07.013.
27. Friston, K. Is the free-energy principle neurocentric? Nature reviews. Neuroscience 2010,
11, 605, doi:10.1038/nrn2787-c2.
28. Constant, A.; Ramstead, M.J.D.; Veissiere, S.P.L.; Campbell, J.O.; Friston, K.J. A
variational approach to niche construction. J R Soc Interface 2018, 15,
doi:10.1098/rsif.2017.0685.
29. Allen, M.; Friston, K.J. From cognitivism to autopoiesis: towards a computational
framework for the embodied mind. Synthese 2018, 195, 2459-2482,
doi:10.1007/s11229-016-1288-5.
30. Niizato, T.; Nishiyama, Y.; Sakamoto, K.; Kazama, T.; Okabayashi, T.; Yamaguchi, T. The
ambiguous feeling between "mine" and "not-mine" measured by integrated information
theory during rubber hand illusion. Sci Rep 2022, 12, 18002, doi:10.1038/s41598-022-
22927-1.
31. Ramakonar, H.; Franz, E.A.; Lind, C.R. The rubber hand illusion and its application to
clinical neuroscience. J Clin Neurosci 2011, 18, 1596-1601,
doi:10.1016/j.jocn.2011.05.008.
32. Kieliba, P.; Clode, D.; Maimon-Mor, R.O.; Makin, T.R. Robotic hand augmentation drives
changes in neural body representation. Sci Robot 2021, 6,
doi:10.1126/scirobotics.abd7935.
33. Shull, P.B.; Damian, D.D. Haptic wearables as sensory replacement, sensory
augmentation and trainer - a review. J Neuroeng Rehabil 2015, 12, 59,
doi:10.1186/s12984-015-0055-z.
34. Hassan, A.; Josephs, K.A. Alien Hand Syndrome. Curr Neurol Neurosci Rep 2016, 16, 73,
doi:10.1007/s11910-016-0676-z.
35. Miller, W.B., Jr.; Torday, J.S.; Baluska, F. Biological evolution as defense of 'self'. Prog
Biophys Mol Biol 2018, 142, 54-74, doi:10.1016/j.pbiomolbio.2018.10.002.

19
36. Hofstadter, D.R. I am a strange loop; Basic Books: New York, 2007; pp. xix, 412 p., 414 p.
of plates.
37. McGilchrist, I. The master and his emissary : the divided brain and the making of the
Western world, New expanded edition. ed.; Yale University Press: New Haven, 2019; pp.
xxvi, 588 pages, 588 unnumbered pages of plates.
38. Varela, F.J.; Thompson, E.; Rosch, E. The embodied mind : cognitive science and human
experience; MIT Press: Cambridge, Mass., 1991; pp. xx, 308 p.
39. Maturana, H.R.; Varela, F.J. Autopoiesis and Cognition the Realization of the Living.
Boston Studies in the Philosophy of Science, 1980, 1 online resource (180 pages).
40. Bongard, J.; Levin, M. There's Plenty of Room Right Here: Biological Systems as Evolved,
Overloaded, Multi-Scale Machines. Biomimetics (Basel) 2023, 8,
doi:10.3390/biomimetics8010110.
41. Keeling, S. Confabulation and rational obligations for self-knowledge. Philosophical
Psychology 2018, 31, 1215-1238, doi:10.1080/09515089.2018.1484086.
42. Sacks, O.W. The man who mistook his wife for a hat and other clinical tales; Simon &
Schuster: New York, 1998.
43. Dennett, D.C. Consciousness explained; Little, Brown and Co.: Boston, 1991.
44. Seth, A.K. Being you : a new science of consciousness; Dutton, an imprint of Penguin
Random House LLC: New York, New York, 2021; p. 340 pages.
45. Jahn, C.I.; Markov, N.T.; Morea, B.; Daw, N.D.; Ebitz, R.B.; Buschman, T.J. Learning
attentional templates for value-based decision-making. Cell 2024,
doi:10.1016/j.cell.2024.01.041.
46. Solms, M. The statistical mechanics of felt uncertainty under active inference. Behav.
Brain Sci. 2023, 46, e108, doi:10.1017/S0140525X22002783.
47. Friston, K. The free-energy principle: a unified brain theory? Nat Rev Neurosci 2010, 11,
127-138, doi:10.1038/nrn2787.
48. Osorio-Gomez, D.; Miranda, M.I.; Guzman-Ramos, K.; Bermudez-Rattoni, F.
Transforming experiences: Neurobiology of memory updating/editing. Front Syst
Neurosci 2023, 17, 1103770, doi:10.3389/fnsys.2023.1103770.
49. Hoffman, D.D. The case against reality : why evolution hid the truth from our eyes; 2019.
50. Hoffman, D.D. The Interface Theory of Perception. In The Interface Theory of Perception,
Stevens, Ed.; 2017.
51. Prakash, C.; Fields, C.; Hoffman, D.D.; Prentner, R.; Singh, M. Fact, Fiction, and Fitness.
Entropy (Basel) 2020, 22, doi:10.3390/e22050514.
52. Friston, K.J.; Stephan, K.E.; Montague, R.; Dolan, R.J. Computational psychiatry: the
brain as a phantastic organ. Lancet Psychiatry 2014, 1, 148-158, doi:10.1016/S2215-
0366(14)70275-5.
53. Seth, A.K. Interoceptive inference, emotion, and the embodied self. Trends Cogn Sci
2013, 17, 565-573, doi:10.1016/j.tics.2013.09.007.
54. Semon, R.W.; Simon, L. The mneme; The Macmillan company: London, New York,, 1921;
p. 304 p.
55. Heylighen, F. Stigmergy as a universal coordination mechanism II: Varieties and
evolution. Cognitive Systems Research 2016, 38, 50-59,
doi:10.1016/j.cogsys.2015.12.007.

20
56. Heylighen, F. Stigmergy as a universal coordination mechanism I: Definition and
components. Cognitive Systems Research 2016, 38, 4-13,
doi:10.1016/j.cogsys.2015.12.002.
57. Chater, N. The mind is flat : the illusion of mental depth and the improvised mind; Allen
Lane: London, 2018; p. 251 pages.
58. von Foerster, H. What is memory that it may have hindsight and foresight as well? In
Proceedings of the The Future of the Brain Sciences, New York, 1969; pp. 19-65.
59. Nagel, T. What is it like to be a bat? Philosophical Review 1974, 83, 435-450.
60. Dennett, D.C. Who's On First? Heterophenomenology Explained. Journal of
Consciousness Studies 2003, 10, 19-30.
61. Hofstadter, D.R.; Mitchell, M. The Copycat project: A model of mental fluidity and
analogy-making. In Analogical connections.; Advances in connectionist and neural
computation theory, Vol. 2.; Ablex Publishing: Westport, CT, US, 1994; pp. 31-112.
62. Blackiston, D.J.; Silva Casey, E.; Weiss, M.R. Retention of memory through
metamorphosis: can a moth remember what it learned as a caterpillar? PLoS One 2008,
3, e1736, doi:10.1371/journal.pone.0001736.
63. Blackiston, D.; Shomrat, T.; Levin, M. The Stability of Memories During Brain
Remodeling: a Perspective. Communicative & Integrative Biology 2015, 8, e1073424,
doi:10.1080/19420889.2015.1073424.
64. Alloway, T.M. Retention of Learning through Metamorphosis in Grain Beetle (Tenebrio-
Molitor). American Zoologist 1972, 12, 471-472.
65. Sheiman, I.M.; Tiras, K.L. Memory and morphogenesis in planaria and beetle. In Russian
contributions to invertebrate behavior, Abramson, C.I., Shuranova, Z.P., Burmistrov,
Y.M., Eds.; Praeger: Westport, CT, 1996; pp. 43-76.
66. Ray, S. Survival of olfactory memory through metamorphosis in the fly Musca
domestica. Neuroscience Letters 1999, 259, 37-40.
67. Punzo, F.; Malatesta, R.J. Brain-Rna Synthesis and the Retention of Learning through
Metamorphosis in Tenebrio-Obscurus (Insecta, Coleoptera). Comparative Biochemistry
and Physiology a-Physiology 1988, 91, 675-678.
68. Goldsmith, C.M.; Hepburn, H.R.; Mitchell, D. Retention of an Associative Learning-Task
after Metamorphosis in Locusta-Migratoria-Migratorioides. Journal of Insect Physiology
1978, 24, 737-741.
69. Miller, R.R.; Berk, A.M. Retention over metamorphosis in the African claw-toed frog. J.
Exp. Psychol. Anim. Behav. Process. 1977, 3, 343-356.
70. von Foerster, H.; Inselberg, A.; Weston, P. Memory and Inductive Inference. In
Proceedings of the Bionics Symposium 1966: cybernetic problems in bionics, New York,
1968; pp. 31-68.
71. von Foerster, H. Memory without Record. In Learning, Remembering, and Forgetting:
volume I, Kimble, D.P., Ed.; Science and Behavior Books: Palo Alto, CA, 1965; Volume BCL
Publication 120, pp. 388-433.
72. Fields, C.; Levin, M. Competency in Navigating Arbitrary Spaces as an Invariant for
Analyzing Cognition in Diverse Embodiments. Entropy (Basel) 2022, 24,
doi:10.3390/e24060819.

21
73. Fields, C.; Bischof, J.; Levin, M. Morphological Coordination: A Common Ancestral
Function Unifying Neural and Non-Neural Signaling. Physiology 2020, 35, 16-30,
doi:10.1152/physiol.00027.2019.
74. Sole, R.; Moses, M.; Forrest, S. Liquid brains, solid brains. Philos Trans R Soc Lond B Biol
Sci 2019, 374, 20190040, doi:10.1098/rstb.2019.0040.
75. Levin, M. Bioelectric signaling: Reprogrammable circuits underlying embryogenesis,
regeneration, and cancer. Cell 2021, 184, 1971-1989, doi:10.1016/j.cell.2021.02.034.
76. Pai, V.P.; Aw, S.; Shomrat, T.; Lemire, J.M.; Levin, M. Transmembrane voltage potential
controls embryonic eye patterning in Xenopus laevis. Development 2012, 139, 313-323,
doi:10.1242/dev.073759.
77. Bedard-Gilligan, M.; Zoellner, L.A.; Feeny, N.C. Is Trauma Memory Special? Trauma
Narrative Fragmentation in PTSD: Effects of Treatment and Response. Clin Psychol Sci
2017, 5, 212-225, doi:10.1177/2167702616676581.
78. van der Hart, O.; Bolt, H.; van der Kolk, B.A. Memory fragmentation in dissociative
identity disorder. J Trauma Dissociation 2005, 6, 55-70, doi:10.1300/J229v06n01_04.
79. van der Kolk, B.A.; Fisler, R. Dissociation and the fragmentary nature of traumatic
memories: overview and exploratory study. J. Trauma. Stress 1995, 8, 505-525,
doi:10.1007/BF02102887.
80. van der Kolk, B.A. The body keeps the score: memory and the evolving psychobiology of
posttraumatic stress. Harv. Rev. Psychiatry 1994, 1, 253-265,
doi:10.3109/10673229409017088.
81. Bridge, D.J.; Paller, K.A. Neural correlates of reactivation and retrieval-induced
distortion. J Neurosci 2012, 32, 12144-12151, doi:10.1523/JNEUROSCI.1378-12.2012.
82. Thakral, P.P.; Barberio, N.M.; Devitt, A.L.; Schacter, D.L. Constructive episodic retrieval
processes underlying memory distortion contribute to creative thinking and everyday
problem solving. Mem Cognit 2023, 51, 1125-1144, doi:10.3758/s13421-022-01377-0.
83. ͡ A.R. The mind of a mnemonist : a little book about a vast memory; Harvard
Luriıa,
University Press: Cambridge, Mass., 1987; pp. xxv, 160 p.
84. Trivers, R. The folly of fools : the logic of deceit and self-deception in human life; Basic
Books: New York, NY, 2011; pp. xvi, 397 p.
85. Nisbett, R.E.; Wilson, T.D. Telling more than we can know: Verbal reports on mental
processes. Psychological Review 1977, 84, 231-259.
86. Stammers, S.; Bortolotti, L. Introduction: Philosophical Perspectives on Confabulation.
Topoi 2020, 39, 115-119, doi:10.1007/s11245-019-09668-z.
87. Byrne, W.L.; Samuel, D.; Bennett, E.L.; Rosenzweig, M.R.; Wasserman, E. Memory
transfer. Science 1966, 153, 658-659, doi:10.1126/science.153.3736.658.
88. McGaugh, J.L. Analysis of Memory Transfer and Enhancement. Proceedings of the
American Philosophical Society 1967, 111, 347-351.
89. Bedecarrats, A.; Chen, S.; Pearce, K.; Cai, D.; Glanzman, D.L. RNA from Trained Aplysia
Can Induce an Epigenetic Engram for Long-Term Sensitization in Untrained Aplysia.
eNeuro 2018, 5, doi:10.1523/ENEURO.0038-18.2018.
90. Reinis, S. Analysis of Memory-Transfer by Puromycin. Activitas Nervosa Superior 1970,
12, 289-&.

22
91. Frank, B.; Stein, D.G.; Rosen, J. Interanimal "memory" transfer: results from brain and
liver homogenates. Science 1970, 169, 399-402, doi:10.1126/science.169.3943.399.
92. Maldonado, H.; Tablante, A. Behavioral transfer in praying mantis by injection of brain
homogenate. Physiol Behav 1976, 16, 617-621, doi:10.1016/0031-9384(76)90223-7.
93. Miller, B.E.; Holt, G.L. Memory transfer in rats by injection of brain and liver RNA.
Journal of Biological Psychology 1977, 19, 4-9.
94. Holt, G.L.; Bentz, G. Interanimal memory transfer of a barpress response through brain
and liver RNA injections. B Psychonomic Soc 1983, 21, 51-53, doi:10.3758/BF03329952.
95. Ungar, G.; Irwin, L.N. Transfer of Acquired Information by Brain Extracts. Nature 1967,
214, 453-.
96. Liester, M.B. Personality changes following heart transplantation: The role of cellular
memory. Med Hypotheses 2020, 135, 109468, doi:10.1016/j.mehy.2019.109468.
97. Pearsall, P.; Schwartz, G.E.; Russek, L.G. Changes in heart transplant recipients that
parallel the personalities of their donors. Integr Med 2000, 2, 65-72.
98. Hepper, P.G.; Waldman, B. Embryonic olfactory learning in frogs. Q. J. Exp. Psychol. B
1992, 44, 179-197.
99. Hepper, P.G.; Waldman, B. Embryonic Olfactory Learning in Frogs. Quarterly Journal of
Experimental Psychology Section B-Comparative and Physiological Psychology 1992,
44B, 179-197.
100. von Foerster, H. What Is Memory that It May Have Hindsight and Foresight as well? In
Proceedings of the The Future of the Brain Sciences, Proceedings of a Conference held
at the New York Academy of Medicine, New York, 1969; pp. 19–64.
101. Pezzulo, G.; Levin, M. Re-membering the body: applications of computational
neuroscience to the top-down control of regeneration of limbs and other complex
organs. Integr Biol (Camb) 2015, 7, 1487-1517, doi:10.1039/c5ib00221d.
102. Levin, M. Technological Approach to Mind Everywhere: An Experimentally-Grounded
Framework for Understanding Diverse Bodies and Minds. Front Syst Neurosci 2022, 16,
768201, doi:10.3389/fnsys.2022.768201.
103. Tseng, A.S.; Beane, W.S.; Lemire, J.M.; Masi, A.; Levin, M. Induction of vertebrate
regeneration by a transient sodium current. J Neurosci 2010, 30, 13192-13200,
doi:30/39/13192 [pii] 10.1523/JNEUROSCI.3315-10.2010.
104. Tseng, A.; Levin, M. Cracking the bioelectric code: Probing endogenous ionic controls of
pattern formation. Communicative & Integrative Biology 2013, 6, 1-8,
doi:10.4161/cib.22595.
105. Farinella-Ferruzza, N. The transformation of a tail into a limb after xenoplastic
transformation. Experientia 1956, 15, 304-305.
106. Farinella-Ferruzza, N. Risultati di trapianti di bottone codale di urodeli su anuri e vice
versa. Riv. Biol. 1953, 45, 523-527.
107. Holtfreter, J. Transformation of a Tail into a Limb or Gill-Like Structures. Journal of
Experimental Zoology 1955, 129, 623-648.
108. Jablonka, E. The evolutionary implications of epigenetic inheritance. Interface Focus
2017, 7, 20160135, doi:10.1098/rsfs.2016.0135.
109. Bourrat, P.; Lu, Q.; Jablonka, E. Why the missing heritability might not be in the DNA.
BioEssays 2017, 39, doi:10.1002/bies.201700067.

23
110. Jablonka, E.; Lamb, M.J. The inheritance of acquired epigenetic variations. Int. J.
Epidemiol. 2015, 44, 1094-1103, doi:10.1093/ije/dyv020.
111. Jablonka, E.; Raz, G. Transgenerational epigenetic inheritance: prevalence, mechanisms,
and implications for the study of heredity and evolution. Q Rev Biol 2009, 84, 131-176.
112. Blackiston, D.J.; Levin, M. Ectopic eyes outside the head in Xenopus tadpoles provide
sensory data for light-mediated learning. The Journal of experimental biology 2013, 216,
1031-1040, doi:10.1242/jeb.074963.
113. Lovas, J.R.; Yuste, R. Dissociation and reaggregation of Hydra vulgaris for studies of self-
organization. STAR Protoc 2022, 3, 101504, doi:10.1016/j.xpro.2022.101504.
114. Martín-Durán, J.M. General Principles of Planarian Embryogenesis and Its Analysis by In
Situ Hybridization and Immunohistochemistry Methods. In Planarian Regeneration:
Methods and Protocols, Rink, J.C., Ed.; Springer New York: New York, NY, 2018; pp. 405-
421.
115. Blackiston, D.; Lederer, E.K.; Kriegman, S.; Garnier, S.; Bongard, J.; Levin, M. A cellular
platform for the development of synthetic living machines. Sci Robot 2021, 6, eabf1571,
doi:10.1126/scirobotics.abf1571.
116. Fankhauser, G. Maintenance of normal structure in heteroploid salamander larvae,
through compensation of changes in cell size by adjustment of cell number and cell
shape. Journal of Experimental Zoology 1945, 100, 445-455,
doi:10.1002/jez.1401000310.
117. Fankhauser, G. The Effects of Changes in Chromosome Number on Amphibian
Development. The Quarterly Review of Biology 1945, 20, 20-78, doi:10.2307/2809003.
118. Thornton, C. Predictive processing simplified: The infotropic machine. Brain Cogn 2017,
112, 13-24, doi:10.1016/j.bandc.2016.03.004.
119. Falandays, J.B.; Nguyen, B.; Spivey, M.J. Is prediction nothing more than multi-scale
pattern completion of the future? Brain Res 2021, 1768, 147578,
doi:10.1016/j.brainres.2021.147578.
120. Cooke, J.; Webber, J.A. Dynamics of the Control of Body Pattern in the Development of
Xenopus-Laevis .1. Timing and Pattern in the Development of Dorsoanterior and
Posterior Blastomere Pairs, Isolated at the 4-Cell Stage. Journal of Embryology and
Experimental Morphology 1985, 88, 85-112.
121. Cooke, J. Scale of body pattern adjusts to available cell number in amphibian embryos.
Nature 1981, 290, 775-778.
122. Cooke, J. Cell number in relation to primary pattern formation in the embryo of Xenopus
laevis. I: The cell cycle during new pattern formation in response to implanted
organisers. Journal of Embryology and Experimental Morphology 1979, 51, 165-182.
123. Blackiston, D.; Kriegman, S.; Bongard, J.; Levin, M. Biological Robots: Perspectives on an
Emerging Interdisciplinary Field. Soft Robot 2023, doi:10.1089/soro.2022.0142.
124. Kriegman, S.; Blackiston, D.; Levin, M.; Bongard, J. Kinematic self-replication in
reconfigurable organisms. Proc Natl Acad Sci U S A 2021, 118,
doi:10.1073/pnas.2112672118.
125. Kriegman, S.; Blackiston, D.; Levin, M.; Bongard, J. A scalable pipeline for designing
reconfigurable organisms. Proc Natl Acad Sci U S A 2020, 117, 1853-1859,
doi:10.1073/pnas.1910837117.

24
126. Gumuskaya, G.; Srivastava, P.; Cooper, B.G.; Lesser, H.; Semegran, B.; Garnier, S.; Levin,
M. Motile Living Biobots Self-Construct from Adult Human Somatic Progenitor Seed
Cells. Adv Sci (Weinh) 2023, e2303575, doi:10.1002/advs.202303575.
127. Mani, M.S. Ecology of plant galls. Monographiae Biologicae, 1964, 1 online resource (XII,
434 p. 369 illus, doi:10.1007/978-94-017-6230-4.
128. Nanos, V.; Levin, M. Multi-scale Chimerism: An experimental window on the algorithms
of anatomical control. Cells Dev 2021, 169, 203764, doi:10.1016/j.cdev.2021.203764.
129. Roth, J.G.; Huang, M.S.; Li, T.L.; Feig, V.R.; Jiang, Y.; Cui, B.; Greely, H.T.; Bao, Z.; Pasca,
S.P.; Heilshorn, S.C. Advancing models of neural development with biomaterials. Nat Rev
Neurosci 2021, 22, 593-615, doi:10.1038/s41583-021-00496-y.
130. Liu, Z.; Wan, X.; Wang, Z.L.; Li, L. Electroactive Biomaterials and Systems for Cell Fate
Determination and Tissue Regeneration: Design and Applications. Adv Mater 2021, 33,
e2007429, doi:10.1002/adma.202007429.
131. Bernheim-Groswasser, A.; Gov, N.S.; Safran, S.A.; Tzlil, S. Living Matter: Mesoscopic
Active Materials. Adv Mater 2018, 30, e1707028, doi:10.1002/adma.201707028.
132. Gil, E.S.; Panilaitis, B.; Bellas, E.; Kaplan, D.L. Functionalized silk biomaterials for wound
healing. Adv Healthc Mater 2013, 2, 206-217, doi:10.1002/adhm.201200192.
133. Muskovich, M.; Bettinger, C.J. Biomaterials-based electronics: polymers and interfaces
for biology and medicine. Adv Healthc Mater 2012, 1, 248-266,
doi:10.1002/adhm.201200071.
134. Cohen-Karni, T.; Langer, R.; Kohane, D.S. The smartest materials: the future of
nanoelectronics in medicine. ACS Nano 2012, 6, 6541-6545, doi:10.1021/nn302915s.
135. Jovasevic, V.; Wood, E.M.; Cicvaric, A.; Zhang, H.; Petrovic, Z.; Carboncino, A.; Parker,
K.K.; Bassett, T.E.; Moltesen, M.; Yamawaki, N.; et al. Formation of memory assemblies
through the DNA-sensing TLR9 pathway. Nature 2024, doi:10.1038/s41586-024-07220-
7.
136. Krakauer, D. Cryptographic Nature. 2015, arXiv:1505.01744,
doi:10.48550/arXiv.1505.01744.
137. Levin, M. The evolution of understanding: A genetic algorithm model of the evolution of
communication. Biosystems 1995, 36, 167-178, doi:10.1016/0303-2647(95)01557-2.
138. Tung, A.; Sperry, M.; Clawson, W.; Pavuluri, A.; Bulatao, S.; Yue, M.; Levin, M. Embryos
Assist Each Other’s Morphogenesis: calcium and ATP signaling mechanisms in collective
resistance to teratogens. in review 2023.
139. Levin, M.; Yuste, R. Modular cognition. Aeon Essays 2022.
140. Pianka, E.R. On r- and K-Selection. The American naturalist 1970, 104, 592 - 597.
141. Shreesha, L.; Levin, M. Cellular Competency during Development Alters Evolutionary
Dynamics in an Artificial Embryogeny Model. Entropy-Switz 2023, 25, 131.
142. Levin, M.; Pietak, A.M.; Bischof, J. Planarian regeneration as a model of anatomical
homeostasis: Recent progress in biophysical and computational approaches. Semin Cell
Dev Biol 2018, 87, 125-144, doi:10.1016/j.semcdb.2018.04.003.
143. Dor, D. The Instruction of Imagination : language as a social communication technology;
Oxford University Press: Oxford ; New York, 2015; pp. xii, 259 pages.

25
144. Veissière, S.P.L.; Constant, A.; Ramstead, M.J.D.; Friston, K.J.; Kirmayer, L.J. Thinking
through other minds: A variational approach to cognition and culture. Behavioral and
Brain Sciences 2020, 43, e90, doi:10.1017/S0140525X19001213.
145. Vasil, J.; Badcock, P.B.; Constant, A.; Friston, K.; Ramstead, M.J.D. A World Unto Itself:
Human Communication as Active Inference. Front Psychol 2020, 11, 417,
doi:10.3389/fpsyg.2020.00417.
146. Wess, O.; Roder, U. A holographic model for associative memory chains. Biol Cybern
1977, 27, 89-98, doi:10.1007/BF00337260.
147. Poggio, T. On holographic models of memory. Kybernetik 1973, 12, 237-238.
148. Gabor, D. Improved holographic model of temporal recall. Nature 1968, 217, 1288-1289.
149. Chopping, P.T. Holographic model of temporal recall. Nature 1968, 217, 781-782.
150. Yngve, V.H. A Model and an Hypothesis for Language Structure. Proceedings of the
American Philosophical Society 1960, 104, 444-466.
151. Feduccia, A.A.; Mithoefer, M.C. MDMA-assisted psychotherapy for PTSD: Are memory
reconsolidation and fear extinction underlying mechanisms? Prog
Neuropsychopharmacol Biol Psychiatry 2018, 84, 221-228,
doi:10.1016/j.pnpbp.2018.03.003.
152. Inserra, A. Hypothesis: The Psychedelic Ayahuasca Heals Traumatic Memories via a
Sigma 1 Receptor-Mediated Epigenetic-Mnemonic Process. Front Pharmacol 2018, 9,
330, doi:10.3389/fphar.2018.00330.
153. Raut, S.B.; Marathe, P.A.; van Eijk, L.; Eri, R.; Ravindran, M.; Benedek, D.M.; Ursano, R.J.;
Canales, J.J.; Johnson, L.R. Diverse therapeutic developments for post-traumatic stress
disorder (PTSD) indicate common mechanisms of memory modulation. Pharmacol Ther
2022, 239, 108195, doi:10.1016/j.pharmthera.2022.108195.
154. Sarmanlu, M.; Kuypers, K.P.C.; Vizeli, P.; Kvamme, T.L. MDMA-assisted psychotherapy
for PTSD: Growing evidence for memory effects mediating treatment efficacy. Prog
Neuropsychopharmacol Biol Psychiatry 2024, 128, 110843,
doi:10.1016/j.pnpbp.2023.110843.
155. Muller, G.B. Why an extended evolutionary synthesis is necessary. Interface Focus 2017,
7, 20170015, doi:10.1098/rsfs.2017.0015.
156. Futuyma, D.J. Evolutionary biology today and the call for an extended synthesis.
Interface Focus 2017, 7, 20160145, doi:10.1098/rsfs.2016.0145.
157. Calvey, T. The extended evolutionary synthesis and addiction: the price we pay for
adaptability. Prog Brain Res 2017, 235, 1-18, doi:10.1016/bs.pbr.2017.07.004.
158. Lindholm, M. DNA Dispose, but Subjects Decide. Learning and the Extended Synthesis.
Biosemiotics-Neth 2015, 8, 443-461, doi:10.1007/s12304-015-9242-3.
159. Laland, K.N.; Uller, T.; Feldman, M.W.; Sterelny, K.; Muller, G.B.; Moczek, A.; Jablonka,
E.; Odling-Smee, J. The extended evolutionary synthesis: its structure, assumptions and
predictions. Proc Biol Sci 2015, 282, 20151019, doi:10.1098/rspb.2015.1019.
160. Cisek, P. Beyond the computer metaphor: Behavior as interaction. Journal of
Consciousness Studies 1999, 6, 125-142.
161. Cisek, P.; Pastor-Bernier, A. On the challenges and mechanisms of embodied decisions.
Philos Trans R Soc Lond B Biol Sci 2014, 369, doi:10.1098/rstb.2013.0479.

26
162. Powers, W.T. Behavior: the control of perception; Aldine Pub. Co.: Chicago,, 1973; pp. xi,
296 p.
163. Buss, L.W. The evolution of individuality; Princeton University Press: Princeton, N.J.,
1987; pp. xv, 201 p.
164. Brash, D.E. Cellular proofreading. Nat Med 1996, 2, 525-526, doi:10.1038/nm0596-525.
165. Gershman, S.J. The molecular memory code and synaptic plasticity: A synthesis.
Biosystems 2023, 224, 104825, doi:10.1016/j.biosystems.2022.104825.
166. Fields, C.; Levin, M. Scale-Free Biology: Integrating Evolutionary and Developmental
Thinking. BioEssays 2020, 42, e1900228, doi:10.1002/bies.201900228.
167. Shapiro, J.A. Evolution : a view from the 21st century; FT Press Science: Upper Saddle
River, N.J., 2011; pp. xi, 253 p.
168. Froese, T.; Izquierdo, E.J. A dynamical approach to the phenomenology of body
memory: Past interactions can shape present capacities without neuroplasticity. Journal
of Consciousness Studies 2018, 25, 20-46.
169. Biswas, S.; Clawson, W.; Levin, M. Learning in Transcriptional Network Models:
Computational Discovery of Pathway-Level Memory and Effective Interventions. Int J
Mol Sci 2022, 24, doi:10.3390/ijms24010285.
170. Biswas, S.; Manicka, S.; Hoel, E.; Levin, M. Gene Regulatory Networks Exhibit Several
Kinds of Memory: Quantification of Memory in Biological and Random Transcriptional
Networks. iScience 2021, 24, 102131, doi:https://doi.org/10.1016/j.isci.2021.102131.
171. Nakajima, K. Physical reservoir computing—an introductory perspective. Japanese
Journal of Applied Physics 2020, 59.6, 060501.
172. Tanaka, G.; Yamane, T.; Heroux, J.B.; Nakane, R.; Kanazawa, N.; Takeda, S.; Numata, H.;
Nakano, D.; Hirose, A. Recent advances in physical reservoir computing: A review.
Neural Netw 2019, 115, 100-123, doi:10.1016/j.neunet.2019.03.005.
173. Seoane, L.F. Evolutionary aspects of reservoir computing. Philos Trans R Soc Lond B Biol
Sci 2019, 374, 20180377, doi:10.1098/rstb.2018.0377.
174. Miller, W.B., Jr.; Baluska, F.; Reber, A.S. A revised central dogma for the 21st century:all
biology is cognitive information processing. Prog Biophys Mol Biol 2023, 182, 34-48,
doi:10.1016/j.pbiomolbio.2023.05.005.
175. Bird, J.; Layzell, P. The evolved radio and its implications for modelling the evolution of
novel sensors. In Proceedings of the Proceedings of the 2002 Congress on Evolutionary
Computation. CEC'02 (Cat. No.02TH8600), 12-17 May 2002, 2002; pp. 1836-1841
vol.1832.
176. Lehman, J.; Clune, J.; Misevic, D.; Adami, C.; Altenberg, L.; Beaulieu, J.; Bentley, P.J.;
Bernard, S.; Beslon, G.; Bryson, D.M.; et al. The Surprising Creativity of Digital Evolution:
A Collection of Anecdotes from the Evolutionary Computation and Artificial Life
Research Communities. Artif. Life 2020, 26, 274-306, doi:10.1162/artl_a_00319.
177. Friston, K. Life as we know it. Journal of the Royal Society, Interface / the Royal Society
2013, 10, 20130475, doi:10.1098/rsif.2013.0475.
178. Goertzel, B. The General Theory of General Intelligence: A Pragmatic Patternist
Perspective. 2021, arXiv:2103.15100, doi:10.48550/arXiv.2103.15100.
179. Lagasse, E.; Levin, M. Future medicine: from molecular pathways to the collective
intelligence of the body. Trends Mol Med 2023, doi:10.1016/j.molmed.2023.06.007.

27
180. Mathews, J.; Chang, A.J.; Devlin, L.; Levin, M. Cellular signaling pathways as plastic,
proto-cognitive systems: Implications for biomedicine. Patterns (N Y) 2023, 4, 100737,
doi:10.1016/j.patter.2023.100737.
181. Bongard, J.; Zykov, V.; Lipson, H. Resilient machines through continuous self-modeling.
Science 2006, 314, 1118-1121, doi:10.1126/science.1133687.
182. Izquierdo-Torres, E.; Bührmann, T. Analysis of a Dynamical Recurrent Neural Network
Evolved for Two Qualitatively Different Tasks: Walking and Chemotaxis. In Proceedings
of the IEEE Symposium on Artificial Life, 2008.
183. Cully, A.; Clune, J.; Tarapore, D.; Mouret, J.-B. Robots that can adapt like animals. Nature
2015, 521, 503-507, doi:10.1038/nature14422.
184. Saniova, B.; Drobny, M.; Sulaj, M. Delirium and postoperative cognitive dysfunction after
general anesthesia. Med Sci Monit 2009, 15, CS81-87.
185. Veyckemans, F. Excitation and delirium during sevoflurane anesthesia in pediatric
patients. Minerva Anestesiol. 2002, 68, 402-405.
186. Wentlandt, K.; Samoilova, M.; Carlen, P.L.; El Beheiry, H. General anesthetics inhibit gap
junction communication in cultured organotypic hippocampal slices. Anesth Analg 2006,
102, 1692-1698, doi:10.1213/01.ane.0000202472.41103.78.
187. Peracchia, C. Effects of the anesthetics heptanol, halothane and isoflurane on gap
junction conductance in crayfish septate axons: a calcium- and hydrogen-independent
phenomenon potentiated by caffeine and theophylline, and inhibited by 4-
aminopyridine. J Membr Biol 1991, 121, 67-78.
188. Dewan, E.M. Consciousness as an emergent causal agent in the context of control
system theory. Globus, G., Maxwell, G., Savodnik, I., Eds.; Plenum Press: New York,
1976; pp. 179-198.
189. Pezzulo, G.; Levin, M. Top-down models in biology: explanation and control of complex
living systems above the molecular level. J R Soc Interface 2016, 13,
doi:10.1098/rsif.2016.0555.
190. Goldstein, A. Algorithms Are the Matter.
https://www.adamjuliangoldstein.com/blog/algorithms-are-the-matter/ 2021, 2023.
191. Fields, C.; Friston, K.; Glazebrook, J.F.; Levin, M. A free energy principle for generic
quantum systems. Prog Biophys Mol Biol 2022, 173, 36-59,
doi:10.1016/j.pbiomolbio.2022.05.006.
192. Pai, V.P.; Cervera, J.; Mafe, S.; Willocq, V.; Lederer, E.K.; Levin, M. HCN2 Channel-
Induced Rescue of Brain Teratogenesis via Local and Long-Range Bioelectric Repair.
Front Cell Neurosci 2020, 14, 136, doi:10.3389/fncel.2020.00136.
193. Blackiston, D.J.; Vien, K.; Levin, M. Serotonergic stimulation induces nerve growth and
promotes visual learning via posterior eye grafts in a vertebrate model of induced
sensory plasticity. npj Regenerative Medicine 2017, 2, 8, doi:10.1038/s41536-017-0012-
5.
194. Dawkins, R. The selfish gene, New ed.; New York: Oxford, 1989; pp. xi, 352 p. : ill. ; 322
cm.
195. Manicka, S.; Levin, M. The Cognitive Lens: a primer on conceptual tools for analysing
information processing in developmental and regenerative morphogenesis. Philos Trans
R Soc Lond B Biol Sci 2019, 374, 20180369, doi:10.1098/rstb.2018.0369.

28
196. Friston, K.J.; Parr, T.; Heins, C.; Constant, A.; Friedman, D.; Isomura, T.; Fields, C.;
Verbelen, T.; Ramstead, M.; Clippinger, J.; et al. Federated inference and belief sharing.
Neurosci Biobehav Rev 2024, 156, 105500, doi:10.1016/j.neubiorev.2023.105500.
197. Vernon, D.; Lowe, R.; Thill, S.; Ziemke, T. Embodied cognition and circular causality: on
the role of constitutive autonomy in the reciprocal coupling of perception and action.
Front Psychol 2015, 6, 1660, doi:10.3389/fpsyg.2015.01660.
198. Tasaki, K.M. Circular causality in integrative multi-scale systems biology and its
interaction with traditional medicine. Prog Biophys Mol Biol 2013, 111, 144-146,
doi:10.1016/j.pbiomolbio.2012.09.005.
199. Thomas, R. Circular causality. Systems biology 2006, 153, 140-153.
200. Jones, P.D.; Holding, D.H. Extremely long-term persistence of the McCollough effect. J.
Exp. Psychol. Hum. Percept. Perform. 1975, 1, 323-327, doi:10.1037//0096-1523.1.4.323.
201. Doolittle, W.F.; Booth, A. It’s the song, not the singer: an exploration of holobiosis and
evolutionary theory. Biol Philos 2017, 32, 5-24.
202. Solms, M. The Hard Problem of Consciousness and the Free Energy Principle. Front
Psychol 2018, 9, 2714, doi:10.3389/fpsyg.2018.02714.
203. Maturana, H.R.; Varela, F.J. The tree of knowledge : the biological roots of human
understanding, Rev. ed.; Distributed in the U.S. by Random House: Boston, 1992; p. 269
p.
204. Maturana, H.R.; Varela, F.J. Autopoiesis and cognition : the realization of the living; D.
Reidel Pub. Co.: Dordrecht, Holland ; Boston, 1980; pp. xxx, 141 p.
205. Varela, F.G.; Maturana, H.R.; Uribe, R. Autopoiesis: the organization of living systems, its
characterization and a model. Curr Mod Biol 1974, 5, 187-196.
206. Bateson, G. Mind and nature : a necessary unity, 1st ed.; Dutton: New York, 1979; pp.
xii, 238 p.
207. Piaget, J. Behaviour and Evolution; 1976.
208. Walsh, D. Piaget’s Paradox: Adaptation, Evolution, and Agency. Hum Dev 2023, 67, 273-
287, doi:10.1159/000534306.
209. Wigner, E.P. Remarks on the mind-body question. In The Scientist Speculates, Good, I.J.,
Ed.; Heineman: 1961.
210. Dietrich, E.; Fields, C.; Hoffman, D.D.; Prentner, R. Editorial: Epistemic Feelings:
Phenomenology, Implementation, and Role in Cognition. Front Psychol 2020, 11,
606046, doi:10.3389/fpsyg.2020.606046.
211. Paulson, S.; Hustvedt, S.; Solms, M.; Shamdasani, S. The deeper self: an expanded view
of consciousness. Ann N Y Acad Sci 2017, 1406, 46-63, doi:10.1111/nyas.13403.
212. Friston, K.J.; Wiese, W.; Hobson, J.A. Sentience and the Origins of Consciousness: From
Cartesian Duality to Markovian Monism. Entropy (Basel) 2020, 22,
doi:10.3390/e22050516.
213. Clawson, W.P.; Levin, M. Endless forms most beautiful 2.0: teleonomy and the
bioengineering of chimaeric and synthetic organisms. Biological Journal of the Linnean
Society 2022, doi:10.1093/biolinnean/blac073.
214. Pio-Lopez, L. The rise of the biocyborg: synthetic biology, artificial chimerism and human
enhancement. New Genetics and Society 2021, 40, 599-619,
doi:https://doi.org/10.1080/14636778.2021.2007064.

29
215. Orive, G.; Taebnia, N.; Dolatshahi-Pirouz, A. A New Era for Cyborg Science Is Emerging:
The Promise of Cyborganic Beings. Adv Healthc Mater 2020, 9, e1901023,
doi:10.1002/adhm.201901023.
216. Ororbia, A.; Friston, K. Mortal Computation: A Foundation for Biomimetic Intelligence.
2023, arXiv:2311.09589, doi:10.48550/arXiv.2311.09589.
217. McGilchrist, I. The Matter with Things; 2021.

Notes
1
The below analysis of this problem is performed in the context of the emerging field of
Diverse Intelligence, which seeks symmetries among agents regardless of their
composition, spatiotemporal scale, or origin (evolved vs. engineered). This field also
seeks to understand the relationship between life and cognition, to understand the deep
nature of embodied minds beyond contingent facts of evolutionary lineage on Earth [1-
9].
2
Definitions, for the purpose of this paper: Self = a process by which an emergent
perspective on the world persists and embodies, but feels (from the inside, in 1st person
perspective) as a persistent structure. Intelligence = some degree of competency at
reaching the same goal by different means as necessary [14].
3
Iain McGilchrist points out that snapshot selves are particularly a “left hemisphere”
invention – a framing used because it does not understand flow.
4
A parallel paradox with respect to free will can be analyzed similarly: if you are bound
by the constraints of your past experiences, preferences, and beliefs, as implemented
in the mechanisms of your cognitive apparatus, are you truly free? But if not bound by
them, how can you implement learning, moral convictions, etc.? The solution is, again,
to embrace a definition of “freedom” that focuses not on snapshots of action and their
antecedent causes, but on the long-range process by which a Self weaves its way
through the world constructing an extended narrative that is best understood and
interpreted as choices driven by coherent large-scale patterns and commitments vs. by
local forces. Our main kind of freedom is the freedom to continuously reinvent ourselves
– freedom from our past Selves, whether genetic or cognitive, and their stale thoughts.
5
This is related to the concept of “precarious” in the enactive cognition field [18] and to
ideas on the relationship between the different kinds of disorder and death in living
systems and the need for freedom from fixed interpretations of the past [19]. Note that it
is a matter of time-scale, with respect to how quickly various components degrade
relatively to the job they do within the agent. See [20-22] for more on the relationship
between vulnerability, agency, and plasticity.
6
This idea is also discussed with respect to intelligence and uncertainty in [25], with
specific ideas around phase codes and tensors as the mechanisms. There are other
candidates for such over-arching concepts, including resolution of ambiguity [26] and
free energy minimization [27-29].
7
This maintenance requires constant adjustment of the estimated boundary between Self
and world, and of the sensors and effectors an agent thinks it has. The plasticity of this
process, and lack of allegiance to evolutionary priors, is revealed when humans, with
millions of years of history as a tetrapod, adopt a new limb as their own within minutes

30
 of experience in the “rubber hand illusion” experiments and other cases of sensory-
motor augmentation [30-33], or in syndromes where patients deny that their conventional
bodyparts belong to them [34]. The ongoing maintenance of the cognitive Self is highly
analogous to the ongoing maintenance required of the somatic self, as the Ship of
Theseus of the body resists aging and degenerative processes while remodeling as
needed to fit environmental demands.
8
Polycomputing [40] is a kind of observer-first way to see living systems, in which each
subsystem can and must interpret the actions and outputs of their neighbors and their
own parts. In this framework, computations do not have a single, objective, correct
meaning: the same set of physical events can be seen as performing different
computations by different observers. By focusing on different ways to interpret the same
processes, biological subsystems can adaptively make use of each other; within a body,
this allows evolution great freedom to innovate – add new functionality by adding new
interpretations of mechanisms, rather than having to change those mechanisms and
risk ruining all of the other systems that depend on them (thus avoiding catastrophic
forgetting – a problem that plagues machine learning, and preserving hard-won gains).
This freedom from how others in the past intended something to be used or understood
is an essential component of the notion of hacking, which is ubiquitous in biology.
9
This can be further unpacked into a multi-scale vision of endless stacks of virtualization,
where we – as patterns – have thoughts, and thus thoughts can have thoughts of their
own.
10
Here, I focus on the functional definition of memory as history-dependent behavior, not
the perceived qualia of mental representation.
11
This is also true for intact individuals, who often attribute fabricated reasons for
decision-making [41,42].
12
Both laterally, and in time (because of saccades).
13
Which begin at the very front of the perception cycle, during active processing by the
retina.
14
The corresponding memory medium for another collective intelligence – ant colonies –
is the ground on which they leave chemical messages that mediate the colony-level
decision-making.
15
Memory is communication, and the mechanisms used to assemble those
communications into an emergent Self-model that seems to persist over time is
analogous to the mechanisms that bind individual agents (like cells) into an emergent
Body-model, because in both cases a continuous stream of signals (message-passing)
underlies a temporally- and spatially-extended whole (as perceived by itself and by
external observers).
16
“The Mind is Flat” [57] is an excellent discussion of how this happens in real-time.
17
The model of intelligence implemented as the analogy-making Copycat system [61] is
a great example of this: learning that “abc” maps to “123” will later have to be used to
guess what “def” (or even “deg” or “456”) maps to, and it’s not the original stimulus, but
the extracted mapping that has to be stored and applied as best as one can to a future
situation.
18
Because our memories are the reference point for everything else we think we know
about ourselves – if our memories were changed, in a consistent way, we would never
know.

31
19
and can actually lead to existential crises – “a thought that breaks the thinker” – for
some.
20
There may be a useful definition of “life” that could be developed with respect to this.
We certainly don’t call unchanging informational media (e.g., a magnetic disk) living,
and change is not enough to signal the presence of life, but change in a way that
preserves saliency for some observer across transformative steps may be a signature
of life in unfamiliar embodiments.
21
Von Foerster [17,70,71] talks about the role of inductive inference, and suggests that
living organisms’ and human organizations efforts to use past experience to predict the
future, and memory, are fundamentally a form of generalization and not inherently
temporal. This emphasis on sense-making, over time as a uniquely privileged aspect,
is consistent with previous suggestions that evolution readily pivoted the competencies
of bioelectric networks’ control of spatial pattern during morphogenesis into nervous
systems control of temporal pattern during behavior in 3D space [72,73].
22
In other words, not error correction (de-noising), aiming to reconstruct the low-level
details of the message, but re-interpreting it as needed to extract useful meaning, with
allegiance not to the original sender and their intent, but to the current recipient and their
perspective. This is related to the notion of universal hacking of biological parts
exploiting neighbors and subcomponents [40], in the sense of a lack of commitment to
others’ interpretations, and even your own past interpretation (an essential ingredient of
a growth in wisdom).
23
Such patients can’t remember a doctor they just met but will refuse to shake hands with
him if he poked them with a concealed pin the first time they shook hands, rapidly
concocting stories like “I don’t like to shake hands” [42] – an alternate form of memory
that also requires confabulation to fit it into their sense of Self.
24
A great game to illustrate the process is Bananagrams, where you compose a sensible
Scrabble-like pattern with your tiles, but the winners are the ones who do not fear the
plasticity needed to destroy your pattern and re-create a better one each time a new tile
appears.
25
From this perspective, memories can be seen as prompts – just extremely rich versions
of trigger stimuli – that push the cognitive system into specific movements in linguistic
or behavioral spaces. We have studied similar phenomena in anatomical morphospace,
where simple bioelectric states likewise trigger memories of organogenesis [75,76].
26
There may be parallels here to “over-training” as seen in ANNs, as well as changes to
memory in trauma and PTSD [77-80].
27
However, the same issue comes up for human practices leading to self-transcendence.
28
Perceiving pattern memories in morphogenesis (the anatomical collective intelligence)
is also not a pure read operation, as many biochemical signals are modified/destroyed
by receptors that bind to them during cell-cell instructions in the navigation of anatomical
space [72].
29
One thing that this model does not yet deal with however is exceptional cases of eidetic
memory, and failures of generalization [83].
30
The injection of odorant molecules into an egg that change the behavior of an adult
animal [98] is another example of a small information structure taking over the
functionality of a much larger whole, with a past history and its own information
structures in place. Why does it work? One possibility is that it’s triggering a built-in

32
 drive meant to overcome the dark room problem of active inference: if nothing new
happens in a long time, it’s possible that the lack of surprise is due to a maladaptive
failure of infotaxis; thus, there may be selection for a module that counteracts boredom
with excessive attention paid to any new information that disagrees with the too-
successful world model. Maybe this is why ~20% of the planarian host decide to go with
the aberrant opinion of an implant from a 2-headed worm and become 2-headed despite
the normal information present in the larger host. The ability of tissues to change on a
dime with small trigger inputs makes systems hackable (and susceptible to exploitation)
but also enables cognition and evolution (mutations, like ideas, can make huge
differences). A formalism that might be useful is that of tiling the plane (because of the
relevance of alignment as cognitive glue), where one abnormal shape nucleates
changes and geometric frustration that can spread through the whole system.
31
Von Foerster emphasizes the difference between recognition/recall (a fundamentally
computational, dynamic, inference process) and mere storage/retrieval [71,100].
32
In this scheme, we can think of the Left and Right sides of the bowtie architecture
(Figure 2B) as two agents, having to communicate via a narrow interface in-between,
like language that helps complex beings cooperate through a low-bandwidth interface.
Being agential/intelligent means, you do not track microstates – the communication
channel doesn’t preserve the lowest level – bits – it preserves meaning or saliency. In
biology, this is essential because the parts are never guaranteed to preserve low-level
details; engineers make cables where the slightest change of pinout can destroy the
functionality; biology assumes there will be “pins” missing, reversed, hacked, etc. and
that it will have to spend effort to decode and extract meaning not dependent on the
microstates of the message. This is why one can inject brain homogenate blindly into a
recipient in a memory transfer experiment, and perhaps why you can randomize the
pixel elements of a photo of a dog and the ANN still recognizes the dog-ness of it.
33
As Douglas Brash points out (personal communication), re-writable memory could be
viewed as the art of building a new triggerable module.
34
The process of planarian regeneration, in which tiny pieces are able to restore the entire
animal, is especially reminiscent of holographic properties [25], but all reproduction can
be seen this way – restoring the entire body from 1 egg cell.
35
People often ask why mammals aren’t as regenerative as planaria etc. At the root, we
are.
36
Such as reconstitution of dissociated Hydra [113], planarian chaotic mode of
development [114], Xenobots assembling from dissociated frog embryo cells [115], etc.
37
In effect, a kind of top-down causation or pattern completion (low-level mechanisms
affected by the higher-level Gestalt) akin to predictive processing in cognitive systems
[29,118,119].
38
As will be argued below, uncertainty about your own physical parts has scaled up to
uncertainty about your own memories. Here, as in behavior, biology solves these
unsolvable Humean skeptical arguments by refusing to halt and resolve them, but
accepting change and moving forward with living.
39
Indeed a recent paper shows DNA damage and repair in the context of memory
formation [135].
40
The enormous compression is achieved because of the intelligence of both sides of the
bow-tie, and their incentive for mutual understanding. In general there is a fascinating

33
 dynamic here between forces that incentivize clear communication (encodings that are
understandable by future instances of oneself) and cryptographic resistance to hacking
by future instances of others [136,137].
41
It’s possible that large groups work better than small ones for the CEMA effect because
the bigger groups provide a more compelling, more fault-tolerant, context for
interpretation of the simple signals. With increased distance between the individuals,
the latency makes it hard for the collective to maintain the bowtie architecture.
42
Pattern completion – filling in gaps (physical and informational), and out-painting – is a
key aspect of machine learning and evolution of cognition [139].
43
It is actually not despite them, but because of them. Asexual planaria accumulate
mutations due to their somatic inheritance reproductive mode, and largely ignore them
– evolution put all of the effort into an algorithm that can reliably make and remake a
planarian body despite the extremely variable (mixoploid!) and unpredictable hardware.
We have hypothesized that this is due to a competency ratchet in which the remapping
process hides information from selection, putting all the pressure onto the competency
mechanisms, not the structural genome [141,142]. Because of their chaotic hardware,
planaria commit to a large-scale body-plan, not the details of the structural genome –
an extreme type of morphogenetic intelligence in the sense of “same goal by different
means”. They epitomize top-down control and sense-making in the form of extracting
high-level wisdom from the low-level details provided by their genomic data.
44
Chris Fields points out: “this kind of boundary/Markov blanket thinking says that a
shared semantics can never be inferred from observations. It turns Wittgenstein's
private language argument on its head - languages are useless if they are only private,
but mechanistically they are private. "Shared semantics" is a constantly evolving
negotiation between language users.” (personal communication)
45
This connects to on-going debates about the status of an author’s intended meaning
vs. that assigned by consumers of their work. We can visualize how semantic autonomy
and deconstructive interpretation extends beyond written works to their antecedents –
the dynamic patterns inside minds which progressively broaden out and change the
more they interact with other agents, and even their author’s future Self, in a kind of
(non-quantum, but perhaps related) decoherence.
46
Note the model of the holographic mind [25,146-149].
47
Perhaps research on psychedelics/plastogens, trauma, and memory are relevant [151-
154].
48
In the same way that reservoir computing needs materials with high degrees of freedom.
49
Much like the egg-shell and the egg’s maternal resources protect embryo from the
environment and harsh selection until maturity, competency protects fragile genome
from selection too (and assimilation may “mature”) it.
50
Some of this dynamic is captured by the emphasis of the extended evolutionary
synthesis approach [155-159].
51
This may be partly why effective transfer of learning is difficult in machine learning today.
52
Prediction-as-projection, and perceptual control theory speak to this extensively [160-
162]. This also has links to cell lineage competition in evolutionary developmental
biology [163,164].

34
53
Paul Pietsch suggests that memory phase codes can be encoded in the resonance
between atoms within living tissues [25]. James Shapiro [167] points out that there is no
line of demarcation between informational and functional molecules in biology.
54
Note that unconventional memory media can include dynamical states, rather than
physical substrates, as discussed here: [168-170].
55
Computational simulations in the field of Artificial Life show the willingness of
evolutionary processes to exploit all of the affordances of the environment [175,176].
56
A related model system would be general anesthesia; not only cases of post-operative
delirium [184,185], but the broader question of how it is even possible for a mind to
recover its state at all, after the bioelectric network is interrupted and rebooted [186,187].
57
Perhaps small implants are sometimes able to over-ride the patterns of a much larger
body [192] because they work harder – being at risk for extinction, maybe they exert
more effort to hack their host (their niche) to persist. There are data showing that cells
which are in their correct positions with their anatomical setpoints satisfied complacently
ignore global signals which are actively processed by cells that are in a precarious or
uncertain state [103,104,193].
58
And the more robust, remappable, deep thoughts are the most agentially-potent ones
(it’s not the physical agent that is robust, it’s the behavioral, physiological, and
anatomical memories that are). This view is the opposite of the superficially similar
Dawkensian memes [194] because, I think, like in other aspects of biology, limiting
oneself to a view in which propagating patterns are pawns of a purely mechanical
process leads to many missed opportunities afforded by an agential lens [189,195].
59
The shelf which helps Huygens clocks get into synchrony is a very basal example of a
low-bandwidth hub node that helps active agents transfer information. This leads to a
further question of whether wave patterns need a medium that is waving, or whether,
like with electromagnetic waves and their back-and-forth between the electric and
magnetic components, it can be a self-reinforcing pattern that needs no “cogniferous
aether” in the form of a brain in which to persist. While resonance is one way to
understand the interaction of the elements of thought, Von Foerster suggests another,
more geometric formalism for the chemistry of thought: tiling, in which “cognitive tiles”
tesselate into larger patterns [100].
60
Niche construction is basically like stigmergy (memory), when we look at the
evolutionary timescale. The actions and perceptions of an agent leave information in its
environment, which then alters how it acts in the future – the environment is like a
memory medium, external from the perspective of individual organisms but internal from
the perspective of the lineage mind.
61
This is also a possible framework for understanding addiction to processes that
remodel the mind to facilitate their own continuation. Another potential risk of systems
that enable the data to manipulate the system is that it makes systems susceptible to
permanent changes [200].
62
It has already been proposed [201], in the case of evolutionary thinker, that the
fundamental units of evolution are metabolic and developmental interaction patterns. A
relevant science fiction story concerns some extremely dense beings emerging from
the center of the Earth, to whom we surface-dwellers are part of the gas phase – they
can’t see us. One of them claims that he has been studying ripples in this gas that
almost look like they have a degree of agency, despite the fact that they only persist for

35
 ~100 years and are easily disrupted by his swirling, but the others don’t believe him as
patterns in a gas cannot be a “thing”, much less an agential being, and certainly not
ones that are formed and disappear in what to them is a blink of an eye. Thus, whether
one sees patterns, or the physical medium in which they exist, as the agent, is
dependent on the cognitive properties of that observer. Extending this idea, one might
imagine alien beings wanting to communicate with life on Earth – would they try talking
to people, or to our cells, or to our financial system, or our ecosystems? Every cognitive
being will have a preferred vantagepoint from which to recognize other minds, and it’s
essential to note that such vantagepoints are just evolutionary user-interfaces [49-51],
not privileged/correct ways to pick the right level of organization and cognition.
63
Perhaps the dichotomy between goals and goal-havers could be made into a
continuum in the same way.
64
Perhaps all the “stable objects” (including organisms) that we see are just the low-
dimensional hub nodes – the medium bearing the pattern, and it’s the persistent, deep
pattern that is the true agent. This links to ideas we are developing on the active Platonic
Space concept, which focuses attention on the patterns and their interactions, not the
media which they temporarily animate. On this view, we are just slightly coarser variants
of our memories, all essentially denizens of Platonic space, projected temporarily into
low-dimensional media.
65
How differently and for how long? That is a matter of degree, determined by (and
determining) the system’s cognitive light cone – the spatio-temporal distance of events,
from the here-and-now, that are needed to have a powerful understanding of a system.
66
Polycomputing emphasizes both, the read side of the interaction (where agents
perceive and interpret others’ activity in whatever way they want) and the write side of
the interaction (where agents use this information to optimally direct other systems’
behavior in ways that bring adaptive benefit to them).

36