That Homo sapiens evolved as the obligate symbiosis of organism and extended phenotype


Hominin evolution has not been fully explained.  Material evidence of the evolutionary paths travelled by the hominin organism increases by the week, but accounts of how the content of hominin cognition developed have not yet produced any visualised model (except in the case of linguistics).  This absence is surprising.

Human beings do not doubt that their abnormally big brains are the basis of cognitive powers unique in the animal kingdom.  While this still seems self-evident, our increasing appreciation of various modes of animal cognition make it correspondingly harder to define exactly what it is that makes human cognition unique.  The distinction seems to be increasingly quantitative rather than qualitative.

Cognition can be described as the correlation of systems of information physically present, down to the quantal level, in the architecture of the brain, and the nervous system more generally.  The more numerous the systems of information, the more complex the cognition.  The smallest bit of information, the Shannon S bit, can only be adduced by such a correlation, by its irreducible difference with at least one other bit.  At this level of irreducibility, human cognition is the correlation of vastly more numerous systems of information than exist in the brain of any other animal.

The substantive irreducible bit of information in animal cognition is the difference between two things as perceived through the nervous system, as between, for instance, a stone and a nut.  The unique ability of human beings is to make innumerable distinctions of this sort, and to be able to operate their durable registrations in the brain without the presence of external stimuli (that is, think of a stone without being able to see a stone) in hierarchies of abnormal complexity (as expressed in cooking an egg, or building the Taj Mahal or writing or reading Cervantes’ Quixote).  This irreducible locus of culuture I call the lep.

This ability is expressed, entirely by way of muscle contraction, as the material human extended phenotype; the totality of human culture. Where an evolved species, as for example a web-weaving spider, produces through its own organism an extended phenotype upon which the organism is dependent for survival, the relation between the two is one of obligate symbiosis.  Clearly, if one part of this symbiosis evolve by the Darwinian process of evolutio; replication, variation, selection; then the other part must evolve by the same process.  As long as it is accepted that Homo sapiens is an evolved species (and we must acknowledge that this is only accepted by a minority of human beings) then a first assumption must be that the human extended phenotype, today a mass of some thirty trillion tons, emerged by a process of Darwinian evolution.

This paper will outline a physical model for the evolution of the content of hominin cognition, including Homo sapiens cognition.  The model is speculative, since the evidence for what happened over the long duration of hominin development is sparse.  If we had a fly-on-the-wall video of a day in the life of a Homo erectus family, it would help immeasurably.  We don’t.



Many significant theories first emerge as physical models, which are models at least partially constructed in the visual cortex; this was how Faraday, Darwin and Einstein worked.  The physicality of biological evolution is beyond doubt, and the evidential data for it must initially be perceived by the senses.  The same should be true of the content of all cultural evolution, which is not mere behaviour mediated by conspecifics, but also individual interaction with insentient structures.  In some non-human taxa these insentient structures are in a relationship of obligate symbiosis with their constructors; the webs of spiders, the nests of birds, the hives of bees.  Neither organism nor extended phenotype could persist without the other.

The extended phenotype of Homo sapiens is the present thirty trillion tons of the technosphere (Zalasiewicz, 2016)[i].  Cultural evolutionists seldom take account of the fact that the survival of Homo sapiens depends entirely upon the extended phenotype of the species.  The reciprocal is exactly true of the extended phenotype.  At the instantaneous cessation of all humanity, the extended phenotype would cease to proliferate and begin a process of increasing entropy.  On this evidence alone the most parsimonious account of hominin evolution, and thus of hominin cognition, would be that it occurred in the context of obligate symbiosis between organism and extended phenotype.  If the organism evolved by the consensual Darwinian process, then the extended phenotype, its obligate symbiont, must have done the same.  Thus the 3my old collective of hominin organisms was the immediate environment in which the technosphere evolved.

The first step is to “specify the exact question under investigation” (Smaldino, 2016)[ii].  The current question is consensually perceived as something like, “Account for unique status of hominin cognition”.  However, cognition is a metaphysical trope.  Our perception of what cognition is depends entirely on the content, the physical extension of whatever is being known.

Human thought, another metaphysical trope, can only be transmitted between brains via muscle contraction.  I’ll repeat that in another way, because it’s a rather unusual axiom.  Muscle contraction is the immediately perceivable and only output of human thought.   “A physical system manifests itself only by interaction with another” 216ff. (Rovelli, 2016)[iii].  Much of this muscle contraction is expressed in behaviour; smiling, throwing, speaking.  In the hominin clade we have no material evidence of such behaviour from the time before the emergence of graphic art, and very little until the emergence of movie photography.

The other fraction of  the history of hominin muscle contraction, that which is not unmediated inter-individual behaviour, is expressed as material extended phenotype; a flaked stone used for cutting, the Taj Mahal.  The aggregate of this fraction of the realisation of hominin cognition has been called the technosphere (Zalasiewicz, 2016); the insentient novel content of the recently coined Anthropocene.

The first possible evidence of deliberate shaping of a stone with another stone does not, at the moment, go back further than about 3.2my, when hominin brains were not much bigger than apes’.  Human brains at 200kya were apparently big enough to optimise the persistence of production and use of the extended phenotype (EP) of the hominin clade.  It seems that further brain enlargement, with its increasing energy consumption, did not further enhance that utility.

It is customary to omit this hominin EP, the thirty trillion tons of the technosphere, from current accounts of cultural evolution.  It might be productive to do the reverse, and bring it into primary focus.  The study of cultural evolution becomes the study of the evolution of the hominin EP.

A thought experiment: if every atom of human organism were to be annihilated on the instant, the still existent hominin EP would contain much information available to an extra-terrestrial intelligence.  If every atom of the hominin EP were to be annihilated on the instant, relict humanity would be reduced to apehood in days.  Our persistence as a species is entirely dependent on the hominin EP, which includes all domesticated species of animal and plants.

So, two axioms: 1)The hominin extended phenotype has information for any information-accepting and processing system, human or non-human.  2) Current Homo sapiens persistence is entirely dependent on the hominin EP.

This dependency saves energy for the organism.  A digging machine replaces the muscle power with which human beings wielded pick and shovel.  It also has an energy cost to the organism.  The resting metabolic rate of a brain of sufficient capacity to produce and utilise the present hominin EP consumes at least twice the energy of the brains of our nearest related primates, the chimpanzees.  This energy cost could not have evolved if there was no reciprocal energy gain.  There is no possible evolutionary model whereby the big brain evolved over six million years, and only then “invented” the marvels of the hominin EP.

The gain that compensated for increased metabolic rate is the utility value of the hominin EP.  If the two things could not have emerged sequentially, small brain to big brain, and only then scant extended phenotype to vast extended phenotype, then they must have emerged diachronically.  At any stage, the extended phenotype could only increase to the extent that the brain could accommodate it, and the brain could only increase to the extent that the extended phenotype could satisfy the increasing energy demand.

A mutualistic system is one where two systems correlate in a way that benefits both systems.  Obligate symbiosis is where each system is dependent for persistence on the persistence of the other.  Trees persist in obligate symbiosis with fungi.  The fungi provide the trees with vital minerals that tree roots have too little surface area to accumulate, the trees provide the subterranean fungi with energy from photosynthesised carbon compounds.

In the absence of any counter-argument I shall proceed as if hominin organisms and their extended phenotype persist in a relationship of obligate symbiosis, as the evidence suggests.  Thus what we call a human being is not an organism exhibiting behaviour that is entirely independent of its material environment (the Cartesian view), nor is it the insentient technosphere, but a mutualistic co-persistence of the two; not the rough Lomekwi 3 stone tool or the Taj Mahal, not a hairless ape, but a superposition of two correlating systems describable at every scale that the rest of the universe is describable.

Extended phenotype is a metaphysical trope, like trait.  It can only be perceived as its content.  There is no causal link between mass of extended phenotype and brain size.  Ants have much smaller brains than dolphins but, in terms of persistent mass, much bigger extended phenotypes.  Dolphins have slight physical means (their beaks and flippers, in fact) of manipulating the material world, so modify little persistent mass, whereas ants produce ant-hills.  Dolphins do have large energy-hungry brains, and a material extended phenotype in that part of an E=mc2 universe which is  the transmission and reception of acoustic energy, possibly as language; but every atom of every dolphin were to be annihilated on the instant, nothing with significant mass would persist. Many protists, stromatolites, ants, bees, earthworms, birds, octopuses and beavers, annihilated (or merely dead), do leave significant altered mass behind them after their passing; as do human beings.

Materials and Methods

(This is a meta-study, in the sense that the writer is not discovering anything new, merely synthesising some empirical data fully available to anybody interested.  The only excuse for its presentation is that it paints with a coarse brush an unusual picture of hominin evolution that might be worth some refinement by more accomplished practitioners.  The materials and methods are too distributed and, being part of the extended hominin phenotype, too complex to be catalogued in a few paragraphs.)



As an extended phenotype which is far from human, take the web of the web-spinning spider clade.  Clearly this was not a result of saltation,  one day no spiders spinning any kind of web and then the next day a spider emerging via a huge and complex genetic mutation fully equipped to weave a complex web of the sort we visualise when we visualise a spider’s web; and then spinning it.  The web-spinning of a modern spider must have evolved in minute particulars, and those which increased the spider’s available energy were selected while those that decreased it were extinguished, all by selective factors distributed between the spider organism and the environment beyond spider and web.  At the same time as the web evolved, the organism of the spider also evolved in a way that could produce, and utilise, the web; in step, minute step by minute step.

The obligate symbiosis between a web-spinning spider and its web is not just dependent on information flow from spider to web, but also from web to spider.  For the small-brained spider the algorithms of web-construction depend on the last angle between filaments that the spider encountered.  If just the information processing system correlated to the web-weaving system of every web-weaving spider were to be terminally disrupted on the instant, there would be a lot of dead spiders within however long it takes an unfed spider to die.

We who accept Darwinian evolution as the process by which life emerged and diversified have no difficulty in accepting this obligate symbiosis of organism and extended phenotype in the case of spiders.  We do have immense difficulty accepting the same obligate symbiosis of hominins, including ourselves, and their extended phenotype.  If we overcame this difficulty, then the emergence of Homo sapiens would be easier to account for.

As with the web-weaving spider, it is likely that hominin brains and the hominin EP evolved by way of obligate symbiosis.  And as with the web weaving spider, it is likely that the information flow is not one way. “A physical system manifests itself only by interaction with another.”  Like the spider, the individual human organism gains information from the extended phenotype.  A child doesn’t learn what a cup or a house is via some sort of abstract explication by one or more con-specifics.  A child comes to know what a cup or a house is by the correlation of a physical system, delimited as a cup or a house, with the physical system contained by its skin.

The hominin EP contains orders of magnitude more information than any individual human brain.  Even the World Wide Web, a fraction of the hominin EP, contains more information than any individual human brain.  The aggregate of human knowledge stored in every now existent human brain is clearly also much larger than that in any single brain, but it can only be correlated at any instant by human networks much smaller than the total population.  There is no meaningful aggregate of all the knowledge in human brains, even on the instant, and knowledge changes continuously with information flux within and between brains.

As has been said, information in a human brain can only get out into the world via muscle contraction.  There is clearly unexpressed information in each brain, but it is effectively non-existent to all other human brains.  A passing bullet might merely make this non-existence permanent.

It can be well argued that there are certain somatic processes, weeping, secreting, excreting, that are part of the phenotype and not the hominin EP; though each of these is clearly must also be part of the hominin EP, because I have just expressed their representations by way of muscle contraction, and you have just perceived them via language, which is certainly part of the hominin EP.  Every bit, and the word bit is pregnant, of human knowledge has its hominin EP equivalent.

This necessity for the correlation of at least two systems, the necessity of muscle contraction to get information from one human brain to another, leads to the logical conclusion that the hominin EP contains as much information as the aggregate of all human brains.

So, to put together the following reasonably secure suppositions:  The web of a web-weaving spider is it’s extended phenotype.  It is external to the spider organism.  It evolved very gradually from a blob of goo to its present typical structure by small, gradual steps, each step being selected, or not, by Darwinian “external factors”.  This evolutionary path was diachronically interdependent with the evolution of the spider organism, muscular, neural and excretory; that is, behavioural.  Therefore an extended phenotype, as a bird’s nest, a beaver’s dam, a termite’s mound, is an evolved structure; structure in the sense of something with extension in the material universe.  This structure persists in mutual dependence with the spider organism and its behaviour.

It is a small speculative step to test the same proposition for hominins.  Hominin material culture is the hominin extended phenotype.  It evolved very gradually from, probably, sticks and stones to the present technosphere by small, gradual steps, each step being selected, or not, by Darwinian “external” factors.  This evolutionary path was diachronically interdependent with the evolution of the hominin organism, skeletal, muscular and neural; that is, behavioural. Therefore the hominin EP, just as a bird’s nest, a beaver’s dam, a termite’s mound, is an evolved structure; structure in the sense of something with extension in the material universe.  This structure persists in mutual dependence with the hominin organism and its behaviour.

It will be argued that the hominin EP is unique, both in its mass and its complexity, and therefore is qualitatively different from the phenotypes of all other animals.  It is indeed different, and it is indeed many orders of magnitude more complex than the extended phenotype of any other animal.  Take a coral reef.  It is the product of three correlated systems of information, the coral polyp, the substance and form which constitutes the reef, and the wider ocean, substance and ecology.  The substance and form of the reef can be an imposing edifice, but it does not correlate with any other entity in the polyp’s extended phenotype because there is no other entity.  The hominin EP on the other hand constitutes a very large number of correlations of systems of information which persist in a continuum.  This continuum persists as a distribution across a space of the same dimensions as cyberspace.  The nodes that define the space are every material occurrence of the hominin EP, and their persisting registrations in the architecture of every living human brain.  The almost empty cup of coffee by my hand is connected back in time to the first ancestral cups, across the universe by its own gravitational field, through the hominin EP to the fact that coffee drinkers throughout the world know what a cup of coffee is; and by agricultural and trading practice, meteorological change… a universal continuum.  How can this magnitude be accounted for by mere evolution?

The question is answered by the answer to another, what is the ability of the human brain that all other animals lack?  It is possible that the answer is relatively straight-forward.  It is not the ability to imitate, nor is it sociality or social complexity, nor is it the ability to recognise the difference between two perceptual objects, as it may be a lion and a lion-sized rock, or an edible and a poisonous fruit.  Nor is it language.  Language is certainly part of the hominin EP and did, I suggest, evolve in obligate symbiosis with it, but did not initiate it.  The evolutionary roots of the hominin EP go back a long way, and are certainly there in macaque monkeys that diverged from the hominid line about 25mya.  The early evolving hominin EP both provided the ecology in which language could evolve, and had reached a stage of complexity where it could not evolve further without language.

The ability manifested by hominins at the time of the emergence of the specifically hominin extended phenotype was the holotype of human competence.  Macaques had hammered stones with stones, and licked the resultant powder, ignoring the adventitious flakes.  At some protracted time after 3.5mya, hominin individuals recognised the chance products of their hammering as entities good for cutting, of the same type as the found stones good for cutting they were already using.  Macaques have never expressed this act of recognition.  The evolving hominins had the capacity, not of invention, but of recognition.  And they had the capacity to shut their eyes, look away, and still have that entity stored in the brain.  They had a durable registration of all that is a >flake that is also a cutting tool<, and everything else in the world that is not a >flake that is also a cutting tool<.

This is the hominin genius.  Chimpanzees can enact it in the context of the already evolved extended human phenotype, but it seems they can’t perform autogenic acts of type recognition, the sequence distraction (attention moving away from the object) and “re-engagement where you left off”, which necessitates not only a persisting passive registration of the object in the brain, that which triggers simple recognition, but a replicable registration of the object in the brain which triggers anticipation of the object being present even when the context of its probable appearance is not immediately available to the senses.  If the worked stones of Lomekwi 3 really are the product of Australopithecus afarensis or a similar taxon, then they could perform this act of durably registered recognition and anticipation over three million years ago.

The significant word in all this is recognise.  The competence is not one of invention,  it is one of persisting registration of a type in the brain such as will trigger recognition of that type even when an immediate stimulus for acquiring it is absent.  It is collecting behaviour, but not the same as a squirrel storing nuts.

Collecting without an evolved stimulatory pathway, as in a hominin picking up a stone good for cutting and carrying it home (a manuport), is significant, but it did not build the Taj Mahal.  It constituted only a part of the evolving hominin competence.  The other part was to recognise, initially probably only as a brain-neuro-muscular registration, the spatiotemporal relationship (the semantic space where the verb would emerge) between hand, hammer stone, core, anvil stone and flakes durably registered as cutting stones, such that after a period of distraction, a few seconds or a whole day, the operation of striking flakes off a core resting on an anvil stone — the same operation as a chimpanzee cracking a nut — could be repeated.  That is to say, they knew you had to hit the core with the hammer and that would produce flakes with a cutting edge.

This unique competence, of not just recognising a potential specific dynamic relationship between two objects when the objects are present, but of retaining a durable and addressable representation of this relationship when the physical referents are absent, is clearly described in Iain Davis’s game-changing Carta lecture, University of California, UCSDTV [iv].  It constitutes anticipation, or foresight.

Clearly anticipation, foresight, are not uniquely human, or even uniquely warm-blooded.  An African hunting dog can superimpose durable brain registrations of a particular dynamic environment upon an in-the-world dynamic environment being instantaneously registered by its perceptions; it can anticipate that when a colleague is moving in from the left, the prey will move to the right, where possible.  But the human foresight is thus.  It can superimpose durable brain registrations of a certain static environment, not upon a simultaneously perceived in-the-world environment, but upon the persistent registration in its brain of another dynamic environment that has no simultaneously perceived correlate in the immediately external world.  The correlation is between two information systems both within the organism; organism advisedly rather than just brain, because the whole organism is needed to perform acts such as using a stone to cut with, or knapping a flake.  This sounds abstruse, but in naturalistic terms it just means seeing a stone of a certain type and thinking, that will come in useful, with a context, largely represented in the visual cortex, of how this may be so.

This is the competence which has the potential for building the Taj Mahal.

This evolutionary departure point, Australopithecus afarensis or Homo habilis, required the (very very slowly) increasing ability to divide the world up into more and more discrete things, >?<, that might much later become the >cord<, the >awl<, the >chisel<, the >bowl<, the >basket<, the >spear<, the >cloak<, all initially adventitious morphological loci, each with a function that contributed to more efficient nutrition or other means of balancing the energy equation, and each a durable registration in the brain; and to locate each of those loci in a matrix of dynamic spatiotemporal relationships, or actions.

The remaining and more intractable problem is this.  If the hominin EP evolved in a Darwinian process, by which I mean one of observed phylogenetic heritability, continual replication with fidelity, an envelope of fractional variation, selection by external factors; how can this evolution be described?

Here information theory may be useful.  The smallest unit of information is the Shannon bit S, the minimum number of alternatives between two possibilities.  N = the number of actual possibilities in a system. The Shannon bit is conventionally the base 2 logarithm of N:S = log2N, so that when N = 2, S = 1.  The S bit is the fundamental , irreducible unit of all information transmission.

There have been half-hearted attempts to identify such an irreducible entity within the metaphysical trope culture.  These have been made in a context where human material culture is conceived of as a discrete, downstream, informationally inert system that did not evolve, but was in some undescribed way the product of another trope, human cognition.  Thus the task is made impossible.  Dawkins (Dawkins, The Selfish Gene, 1976)[v], rather off the cuff, suggested the meme as the replicating entity, but the meme too is a metaphysical object.  It cannot replicate.

The possibility of such a replicating entity is therefore denied, without further discussion.  Boyd and Richerson in their seminal Not by our Genes Alone (Boyd, 2005)[vi] discount it in a few unargued assertions. However, the possibility of replication with fidelity is inherent in the uniquely hominin competence outlined above.  That competence is based on recognition of type, and that in turn is derived from the recognition of the difference between two entities; between a stone good for cutting and a stone good for bashing, where N=2 and S=1. The whole environment of this early hominin, not just stones, could be reduced to such distinctions.  A stone good for cutting is not just the irreducible difference between itself and a stone good for bashing, or throwing, or scraping.  It is also the irreducible difference between a stone good for cutting and everything else in the universe.  That is the information in the system that is the type of a stone good for cutting when it correlates with its durable registration in a human brain.  That correlation is a distributed persistence, not just within the stone or the brain, but across the space-time continuum that connects the two.  Dragging in the space-time continuum may appear to be the gratuitous appeal to the authority of a higher power, a rhetorical trick.  And we know now that space-time is not a continuum, but quantal.  Nonetheless it is a pedestrian fact that the correlation of two systems requires a field between them.  An insentient object like a stone good for cutting may be perceived by a hominin as light.  The atomic structure of the surface of the stone receives light in one set of frequencies and radiates it in another, which is received in the visual cortex as a system, an image, which correlates with a persistent registration of the type stone good for cutting.  As long as this process continues across the space-time continuum, so long will the image of that particular, unique stone correlate with the system in the brain: type, stone good for cutting.  Correlation necessitates modulation.  Thus the briefly perceived particular stone  modulates its persisting registration as a type in the brain.

The type stone-good-for-cutting will be populated by many variations, both found and recognised before the emergence of controlled striking, or knapping; and after its emergence, when struck flakes will vary within one envelope delimited by the nature of the material, and another envelope delimited by the skill which constitutes the technique of the knapper.  The selecting environment in which such cutting tools function, as it might be cutting meat, type A, or cutting plant fibre, type B, will lead to differences being recognised within the type: stone-good-for-cutting.  In this way new loci of irreducible difference will emerge.  Within the categorical type, there are sub-types.

This ability for categorisation is again not peculiar to hominins.  It can be diagrammatically represented by a branching tree.  Any sequence of differentiation in an animal brain: prey animal: worth chasing/not worth chasing; is, within the brain, categorisation and sub-categorisation.  But in the process of knapping stone, the human competence is to retain in the architecture of the brain a persistent registration of the differences between types, and a persistent registration of the dynamic relationships between types, as it might be between a hammer stone and a core. That durable registration can initiate and modulate, through so far unknown complexities of feed-back, a sequence of neural, muscular and skeletal actions that eventuate in, on average, flakes of either type A, or type B.

The locus of irreducible difference upon which this model is based is difference between types, not between objects.  Within a type there may be hundreds, billions, of in-the-world examples, each at some scale different from all the others; ball bearing for instance.  Type is a hard word to pin down.  Here it is used it to signify, for example, the set of all knives in the world; their in-the-world signifiers (words, pictures); and their persisting and distributed registrations in all human brains.  I don’t think we yet know enough about the dynamic state of the brain to describe the precise nature of this distributed persistence, knife.

Here is a worked example.  There is a type knife.  Every instance of that type has a unique morphology at the microscopic level.  There are more significant differences of morphology that divide sub-types; knife with a  20cm blade, sharp point and serrated back for stabbing / knife with a rounded tip for pushing food onto a fork.  There are manifold differences of substance in the various parts, steel, bronze, ceramic, wood, ivory, leather.  We can assert that something is a knife, and something else is a dagger.  If any human being is asked to visualise a knife, each human being will visualise something unique within a uniquely connected neural architecture.  And yet we all know what a knife is.  And we know how it differs from a dagger morphologically.  A knife is a backed blade with one cutting edge, a dagger has two.  Morphologically there is likely to be a continuum between knife and dagger, and therefore at the margin arguments will happen, but for most people most of the time there is no confusion.

And here a remarkable fact about human cognition, and about language, becomes significant.  A knife is an in-the-world object which a hundred thousand word monograph could analyse and anatomise in exhaustive detail, and yet—

“He laid the knife on the table and smiled, radiantly, as if the whole evening had been a joke.  Then the smile, it was hard to tell the exact instant, stopped spreading across his face, and he picked it…”

—and yet when you just read that, you did not have to refer to such a compendium of knowledge as the monograph to work out what a knife is.  You knew on the fly, stored it almost instantaneously in short term memory so you could pay attention to the phrase “on the table” [table ditto], and when you got to the word “it” you could immediately correlate it with “knife” rather than with “table”.  There must be a unique, dynamic but persisting structure that is a functioning element of this episode of recognition.  It cannot be a widely distributed synthesising system like long-term memory which can take days to come up with a name you have forgotten.  The processing speed has to be high enough to get the whole thing done in a fraction of a second.  That unique, persisting, dynamic structure is the locus of irreducible difference signified by the word “knife”.

The locus of irreducible difference, a dynamic structure small enough to be almost instantaneously addressable, is a discrete functional element of what happens when you read knife in continuous text.  It is also a necessary nexus in decoding language.  It is probable that “knife” actuates a complex almost instantaneous pulse around the brain which provides context and thus a more particular knife morphology (abattoir, café, a grape to be peeled) which certainly involves the visual cortex, but that pulse is cued by the material locus of the registration of irreducible difference that is knife and nothing else.  That discrete locus may itself be distributed, but will be small.  This recognition cannot address the whole of the brain potential that could write the monograph on knife, which one way and another will correlate with the entirety of human knowledge, because that would require virtually infinite processing power to read a sentence containing only four nouns, knife, table, evening, joke.  Information theory, thermodynamics, take precedence over infinity.

An information system that is one (S)bit must be located in at least one human brain, but it is derived from objects in the world (say, the set of all knives) and from the distributed system of “knowing what a knife is” across a continuum of all human brains.  It is the smallest possible set of attributes which differentiates a knife from everything that is not a knife, inherent in the correlation of these two distributed systems, that is the type.

I return to another set of correlating systems of information; the web-weaving spider, its web, and the wider environment.  The replication of the web weaving spider organism is concordant with the necessary conditions for evolution: observed phylogenetic heritability, continual replication with fidelity, an envelope of fractional variation, selection by external factors.  The replication of its web across many generations of both web and spider is also concordant with those conditions.  The external factors which acted selectively upon web morphology were the spider organism and the wider environment.  So the web also evolved; and it is a general condition of extended phenotypes that they evolve alongside their symbionts; octopuses and their little cities, birds and their nests, termites and their mounds.  So it is parsimonious to suggest that the hominin extended phenotype also evolved.  Also, apart from Just So stories, there is not other current explanation.

From this evidence, knives evolved according to the Darwinian model.  It is I think irrefutable that, in order to be replicated, a knife has to go through a human brain; or did before the age of AI and robotic machines.  It is widely agreed that insentient objects do not spontaneously replicate.  Leave a knife, or even two knives, in complete isolation in a drawer and the quantity will never increase.  The same is not true of rabbits.  But replication of knives does take place.  There is a variation of form, size and substance both across time, and at any instant across the world, but within that envelope knives have been replicated for tens of millennia.  Go anywhere in the world, use the correct words to ask to borrow a knife and people will know what you want to borrow.  Knife is a stable type with a phylogeny that goes way back beyond the emergence of Homo sapiens.  Subtypes are replicated with fidelity.  The knives I use to eat with at home have been replicated in their hundreds of thousands, identical at the natural visual scale.

The problem that remains is to identify the irreducible locus which carries the information necessary for replication with fidelity of the hominin EP.  Once that is done, variation and selection by external factors raise no problem.  At the onset of the production of stone tools, excessive variation, both in the structure of the raw material and the competence of the organism, was something that evolution might suppress, leading to uniformity and economy in the replicating process. Likewise the environmental factors which selected certain types of stone tool over others, and certain morphologies over others, are already well described and argued.

An irreducible locus is not infinitesimal.  Both an actual knife and the dynamic neural structure which is its locus of irreducible difference in the brain, have extension.  They persist in three dimensions and time.  The difficulty is that we are looking at three correlated systems of information, the knife, the organism, and the continuum through which they correlate.  We know that if every human organism was annihilated on the instance, knives would no longer replicate, so the organism is part of the environment of replication.  We also know that if the entire hominin EP were annihilated on the instant knives, being thus non-existent, would fail to replicate.  Thus the information upon which the replication with fidelity of knives is conditional is distributed between the organism and its extended phenotype.

As with any system of information down to the node of a grain of space, there can be no singularity.  Such a node can only be described as its relationship to other nodes.  The same goes for a gene.  A gene is conventionally described as a sequence of nucleotides with the information required for a ribosome to build one protein.  A nucleotide is therefore the irreducible entity within the system of life.  But a single nucleotide contains no information useable by another system.  Its useful information only emerges in its correlations with other nucleotides.  Likewise, within the same system, a single gene carries information within the environment of its chromatin infrastructure, and of its correlated messenger RNA; and that in the context of available ribosomes, and so on.  The information necessary for the production of a protein, let alone a whole organism, is distributed throughout the phenotype, which necessitates two-way correlation.  This leads to a very complex picture of biological evolution.  But biologists do not say as a result, biological evolution is an impossibility.  They say, this means there is a lot of work to do.  Then they get on with the work.

Distribution is not itself a problem.  An (S)bit is distributed.  All information is distributed.  Any description of a system, such as the locus of irreducible difference as a structure in the brain, knife/dagger, or knife/rest of the universe, “is also therefore always the description of the information which a system has about another system, the correlation between the two systems” (Rovelli, 2016).  In the case of the irreducible locus of difference, it is this correlation which is the basis of replication with fidelity.  The correlation is a dynamic persistence across a continuum.  It is very difficult to grasp.  Physics can handle such concepts without boggling, natural human understanding has more difficulty.

I might help to give this correlation a name.  Let us call it the lep, which is a monosyllable, is not yet a lexeme in English[1], and might have as an acronym Locus Extended Phenotype.  A lep is the (S)bit of hominin culture.  1 lep = the irreducible difference between two systems of information.

The extension of 1 lep in the dynamic structure of the brain is very small; the systems with which it correlates can be of any size.  You can read the word knife on the fly and decoding will occupy a fraction of a second because correlation of small systems is much faster than that of big systems.  Think of the respective reproductive processes of E. coli and elephants.  Nonetheless, if I give you some matches, charcoal, a copper ingot, some clay, some wood and leather and tell you to make a knife, your registration of the lep knife is necessary to the outcome, that you make a knife and not a  spearhead.  The lep knife will correlate with at least three distributed systems: 1) the information latent in the stuff I’ve given you, 2) all information about knives in your brain, and 3) the rest of the world and all the information it contains useful for making a knife, like the presence and significance of oxygen.

This correlation is a dynamic process, not a statistical property. The information about knives in your brain will be many orders of magnitude larger than the irreducible lep knife; fires, bellows, moulds, and the hoped-for morphology of the knife you are about to make.

It is likely that all these correlations in the brain do not persist as bigger and bigger conglomerates of discrete dynamic neural structures each of which is absolutely identical to all the others.  They persist as dynamic alliances between differentiated neural structures, right down to the finest possible reduction, the lep.  Some leps will be identical, they may even form groups of two or more, but in “Knife”, the monograph, the lep knife will correlate with a number of leps upon which at the moment we cannot set a limit.

Each large complex correlation (a concept in natural language) is also different from every other, and can be addressed via the nexus of a lep.  We know the difference between mathematics and chemistry (discrete leps), but you cannot reduce mathematics to an essence, you can only reduce it to its particles; ultimately, to leps.

If we regard these correlations as within a continuum, then information travels through the continuum as various material structures. You’ve been challenged to make a knife.  You look at the gear on the ground.  The copper ingot receives information from the sun, the excitation of its surface atoms emits this information both as heat and light at transposed energies.  The eye receives this light through the lens, and at the retina it is transposed into electro-chemical information, goes through a process of selective destruction for increased informational efficiency and travels to the brain, where it correlates with the kinds of conglomerates of leps outlined above, which correlate with many other systems.  One of these systems will be the lep ingot, and another the lep copper, which as a pair tell you that this is a copper ingot (not a copper pipe) and this is a copper ingot (not a silver ingot).  Further on in your exploration ingot will correlate with the information that you have the components of a furnace but you do not have a hammer.  Melt and cast, not bash.

In this process information moves from the sun to the lep through many transpositions.  Information will be gained and information will be lost.  Overall the whole system will lose information.  Entropy increases.  One definition of life is that it is an enclosed system that can transfer entropy from inside itself to outside itself.  It gains information (such as 1 lep) but the universe loses more.  The hominin EP does the same, with the concomitant gross loss of information from the biosphere which already haunts its hominin component.

To conceive of the watch on my wrist or the International Space Station as composed by leps seems like the outer reaches of fantasy.  In biology things are different.  Darwin did not know what were the basic “factors” that composed life, what it was that  selection acted on.  Mendel produced genes as a verbal and visual model.  This was gradually co-related with continuing  biological discoveries until in the same short period, and respectively, Rosalind Franklin produced X-ray crystallographic images, and Watson and Crick a 3D model, of the DNA sequence structure.  Electron microscopes are rematerialising clearer and more complex images of the genome by the day.  Through this evolutionary process we can now see DNA, and that makes belief much easier.  There is no way at the moment that we can see the structure of the distributed system of a lep as it is encoded in the dynamic structure of the brain.  In that sense it is still an “unknown factor”.  But it occupies a space which can be filled by further information.  There is no reason why in the future we should not be able to see a “lep”, just as we can see the sequence of nucleotides that make up a “gene”.

My watch is not composed only of metal, jewels, glass and leather.  It contains information.  An exhaustive explication of the properties of a lep will have to wait for a book, but each component will have a lep, and each component will also be composed of leps (a cog wheel has teeth, as does a saw) and every lep has a phylogeny, some short, some going way back before Homo sapiens.  So yes, a watch is complicated, but what did we expect?  Biological forms are complicated, as our own organism.  We have to deal with it.

Another worked example; the question is, “What was the process  of replication, variation and selection by which the burin evolved?”  The following answer is purely diagrammatic, it does not pretend to be a literal history.  A burin precursor can be simplified as a rectangular lithic flake with a short edge retouched so it slopes.  This  tool is one of the upper Palaeolithic tool types and is replicated with fidelity over millennia.  Occasionally in its manufacture a shoulder might be  struck off by mistake at the acute angle.  This mistake was a variation upon the type burin precursor, a rectangular scraper or blade.  It would be uneconomical to reject this broken ‘second’ as mere waste, since it will still have some residual use as a kind of chisel.  It is at some time discovered that the miniature transverse edge that results from this mistake can be used for engraving and scouring square-sided grooves in hard materials such as antler or bone.  The variant is recognised as a sub-type, >precursor with missing shoulder and resultant spall<.  This recognition of the spall becomes a lep spall.  This is selection by the wider environment, proximately the organic material which the spall would cut, without which the spall would carry no useful information to the knapper.

The inadvertency of striking off the shoulder to produce a spall develops into a deliberated strike.  Skilled flint knappers recognised the dynamic relationship between hammer stone and flake blade that produced the spall, and thus the burin.  They start deliberately striking burins, rather than recognising them on the odd occasion when they fortuitously occur (nothing is ever “invented”). The resultant now preconceived tool becomes a type in its own right, a lep, burin.

The phylogeny of a burin is clear.  It is derived only from its precursor.  It is not derived from a flint projectile point, or a quern, or a stone wall.  And it fulfils the conditions of observed phylogenetic heritability, continual replication with fidelity, an envelope of fractional variation, selection by external factors.

Where there had been one lep, there were now three, distributed between replicating objects in the world, blades, spalls, burins, and persisting as discrete dynamic encoding structures in hominin brains.  And the whole process conforms at every level with the necessary conditions of Darwinian evolution, within a context of obligate symbiosis, from which the burin fortifies its persistence by having a place in this paper, and human beings, through the bone needle, got warmer and better-fitting clothes.

The same goes for my watch.  The evolution of complex machines requires further explanation.  It is dependent on reticulation and intromission.  I conclude with a short observation on the bicycle, which is descended from the hobby horse, diagrammatically a beam supporting two in-line wheels at either end, but with no drive chain.  This is a four wheel cart, morphologically  squashed laterally as far as it will go without vanishing into two dimensions.  The drive chain is an intromission.  One component is the chain ring, a large cog turned by a crank.  The mechanics of a crank depend on rotation about a fulcrum, and far back in the phylogeny of a crank we can see the first trimmed branches stuck under a large stone, or into a fissure in a promising slab of rock, and acted upon by one or more hominin organisms, their efforts synchronised with rhythmic grunts ancestral to the word, in English, and its attendant lep (because verbs too are leps derived from a dynamic relationship of their respective nouns)  “Heave!!!”

The model I propose is speculative and full of gaps.  It also seems to me to fill a gaping explanatory hole without any competition.  It might serve as a default postulate until the competition shows up.






[1] In Ireland it is an alternative to leap.

[i] Scale and diversity of the physical technosphere: geological perspective

Jan Zalasiewicz,(et al.) The Anthropocene Review 1–14 © The Author(s) 2016 Reprints and permissions: OI: 0.1177/2053019616677743


[ii] August 12, 2016   Draft of chapter to appear in: Computational Models in Social Psychology, edited by R. R. Vallacher, A. Nowak, & S. J. Read. Forthcoming in 2017 from Psychology Press.


[iii] Reality Is Not What It Seems, Carlo Rovelli Penguin Books London 2017


[iv] (accessed 30/11/17) Iain Davidson (Univ of New England, Australia) CARTA – Behaviorally Modern Humans: The Origin of Us Premiere Date: 8/2/2013; 21 minutesCARTA: Behaviorally Modern Humans: The Origin of Us Iain Davidson: Stone Tools and Cognition: Lessons from Australia.


[v] The Selfish Gene, Richard Dawkins, Oxford University Press London 1976 ISBN


[vi] Not By Genes Alone, How Culture Transformed Human Evolution, Peter J. Richerson and Robert Boyd, University of Chicago Press June 2006




Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.