Of the world and its minds

This is an extract from my second reflections books

Of the world and its minds

This is an extract from my second reflections books

Natural Hacking Thesis

An alternative vision of how natural development and evolution mechanisms work


Evolution hacks its way into what it already created.”

« If a gene has no function, then it will tend to disappear. Single genes are no more favored by evolution than single organisms are. Hence evolution does not work on the gene level, but on the function level. »

Des courants qui se maintiennent d’eux-mêmes. Et sont ardemment défendus par des individus. L’esprit des chosechoses (ch1).

The Natural Hacking Thesis is a thesis about functional development and complex organization maintenance through natural means and initial starting constraints. It is an evolution thesis that focuses on functions evolution through (and instead of) physical embodiments of functions such as genes, individuals or species – but an evolution thesis that does not restrict evolution to these embodiments only. In the Natural Hacking Thesis, the involvement of genes, individuals, etc in functional evolution is just because of initial starting constraints : it happened that functions just could start to develop slowly but surely through individuals reproduction ; and so a bit like how some bacteria “use” organisms as “host environments”, functions are “using” organisms (and organisms reproduction mechanisms) as a grounding. In this sense, genes, individual organisms, etc are just indirect constraints scaffolding functions development, constraints that progressively disappears as functional development can emancipate itself from them (and becomes more process based). Because in the end being physically embodied (in genes and such) is just a constraint to their evolution, functions will “try” as much as possible to hack these embodiments to evolve (so as to move the focus from the constrained biological evolution to a less constrained artificial evolution). But precisely because the whole system is embodied, it is tricky and risky (in regard to individual fitness and natural balance) to create hacks as it could reduce the fitness of the embodiments and hence extinct functional development that depends on it.

If we see things from a standard evolution framework, modern human beings somewhat seem to have left the main focus, reproductive success, and so seem to more and more evolve on their own, for their own reasons, not really guided by evolution. In this perspective, humanity represents the point where organisms reached such complex state that evolutionary guidance became extinct. But if we see things from the Natural Hacking Thesis framework, human beings seem to be the best support of the main focus, functional development, as they have abstract capacities and seem to completely devote their life to some kind of functional development that is beyond their own existence, transcending it. In this perspective, humanity represents the point where organisms reached such a complex state that functional evolutionary guidance could finally start to accelerate and “explode”.

All in all, the reason why it seems we human beings — with all our controlling powers and endeavors — are on top of the rules of evolution and natural selection, is because these rules never really existed. Of course evolution is a thing. But it is not exactly the « precise » thing we think it is. Hence our impression that we are no longer under the rules of natural development.

Before discussing in more details the ins out outs of a function centered evolution framework, this text will first introduce the notion of functional hack from a more classic organism/gene centered evolution framework.


Associated concepts : exaptation, natural hacking, side effetcs, emergence, fitness, inadaptationadaptation, jury-rigged design, derived proper functions, evolutionary suicide, ontogenic evolution, Baldwin effect, niche construction, memetics


Of traits and functional roles

In the classic conception of evolution, natural development is explained quite simply through natural selection : genetic errors happen aleatory and errors that prove to be useful for individuals are kept. In this conception, change is therefore purely adaptive, in regard to the individual.

In the natural hacking thesis (NHT) framework, evolution is double-layered : errors and chance does not only apply on the genetic code of organisms, they also apply on the life of organisms. Indeed, when natural selection has done its job, chance is still there. The genetic code of an organism specifies the design of the natural “tools” this organism will have. And while it also specifies the “default” functional usage this organism will make of these tools, it does not specify every functional usage this organism can (and thus will) make of these tools. After the design error (the genetic error), can happen the purpose error (the usage error). And according to NHT, this is these secondary type of errors (or rather unforeseen functional usages) that, by temporarily making natural selection lose control of the whole process and creating shifts in the selection pressure, create shifts in evolution and lead to important changes.

In other words, behaviors that make natural selection lose its control is the key to substantial evolution.


Egocentric evolution VS allocentric evolution

For a better understanding of some aspects of natural hacking thesis, one needs to be able to conceptualize evolution from two teleonomic perspectives, as two dual goal-directing forces.

First there is the individual scope perspective, from which the aim of evolution is to make organisms increase their fitness (~ procreation potential, reproductive success) by changing their traits, so that these organisms somewhat continue to exist through genetic inheritance. From this perspective, evolution is a blind and egocentric process, any solution that increases the fitness is good, even if this solution means destroying the whole world and reducing the fitness of everything else. This is the “part” of evolution that is responsible for “creativity”, bad integration and convoluted/jury-rigged design.

Then we have the global scope perspective, from which the aim of evolution is to maintain a balance by prioritizing adaptationadaptation so that all organisms continue to exist and maintain a decent fitness. From this perspective, evolution is a spatially omniscient and allocentric process, any imbalance in the selection pressure will be detected, focused on, and if possible, fixed. This is the “part” of evolution that is responsible for equilibrium, ecosystems, good integration and intelligent design. Unlike egocentric evolution which does not focus on any particular solution, allocentric evolution encourages the development of relevant solutions. It is important to note that the process only has a moment-to-moment omniscience (omniscience in space but not in time), it is not prescient (nor does it keep memory of the past history), it is just omniscient of the instant present. So both of the processes are future blind.

In natural hacking thesis, these two processes (egocentric evolution and allocentric evolution) are the two symbolic forces that allow changes and balance. They form a dualistic functional system.

I think having this distinction (instead of conceiving evolution as a unitary process) makes it easier to understand the different things evolution can produce. Note though that the Natural Hacking Thesis does not depend on these teleonomic views, so it is not a deal-breaker if you do not find this distinction convincing.

Of the interest of using teleonomic perspectives to study evolution

There is no such thing as natural development or evolution. There are just many errors of replication distributed through time that make it look like there is a development process going on. An organism does not really evolve in itself, it just produces offspring. And these offspring do not evolve either, they are just born as they are. There is no change per se at all.

In the same way, there is no such thing as natural selection. Some organisms just tend to disappear while other tend to replicate. There is no selection per se at all.

Development, evolution, selection, etc, are just metaphors that greatly help understand what the theory tries to explain.

So of course, there is no such thing as purposes, supervision, balancing, maintenance, intentionintentionality, creativity, blindness, omniscience, etc in nature. Those are just metaphors too. The thing is, we can not deny that the mere application of simple natural “laws” tend to give rise to organisms that are adapted to their environment, as if adaptationadaptation was an end. We can not deny that this application tend to create ecosystems, as if balancing was an end too. We can not deny that this application tend to produce convoluted and jury-rigged mechanisms as if sometimes evolution was desperate to add functions. And the list can go on. Of course these are « illusions ». But the human mind likes illusions (it is basically made to create illusions which we then call perceptions) and is adapted to process things like intentionintentionality & co. Yes, it can be a bad thing to apply “human” concepts to the rest of the world (~ anthropomorphism) so as to explain it, teleology is a good example of that. Realizing how wrong it can be can make us think that we should try as much as possible to avoid doing it if we are to produce rational explanations. But in my view, thinking things in such ways is not absolutely wrong and should not be completely rejected. On the contrary, I think it can be the source of great insights and explanations that would be difficult to foresee using more grounded approaches. We still create theories with our human minds, not with computers. To me, using anthropomorphic and teleonomic perspectives is a form of simplification suited for our processing capabilities, just as using colors to represent various diagrams is. Our cognition is grounded, our reasoning should then be grounded if we want to use all its power efficiently, and knowing that is is grounded if we want to keep a « scientific », meta-approach. Let’s not forget that.

Furthemore, let us note that a dual evolution is a more ungrounded, « realistic » metaphor than a single « Godlike » evolution. Because there are multiple parallel evolution processes going on at the same time. It’s not one big « mind », but an infinity of « micro-minds ». Even our meta that perceives the world has in this text has been developped by part of evolution and is supposed to bring evolutionary benefits. At one point in time, some parts of evolution can be against your organism while others are with it. It is not just binary.

Evolutionary (abstractionabstraction) leaks and the emergence of uncontrolled and inadapted roles (side roles)

La mémoire s’est développée parce qu’elle a été utile à plein de rôles. AbstractionAbstraction cadrée ?

Un mécanisme de la théorie de l’évolution.

Leak : hitchhiking trait

More and more difficult to see what things are adapted for.

Evolutionary leaks and hacks

Novel traits (and to a larger extent novel innate abilities) of living organisms are maintained because they are “selected”, that is they are used for at least one particular functional role that is beneficial to self-procreation. But current traits (and to a larger extent current innate abilities) may not be used only for the roles they have been selected. There is no absolute bindingbinding between design and functional usage (purpose).


Of the selection of traits and their roles

If a particular trait has multiple functional roles (beneficial to self-procreation or not) since it has emerged, then everything is “fine” and adaptive. Because this trait would have past the selection process as a whole (it is like ok we let you in the system because you can do this, but we are aware and will take into consideration that you can also do this). But if a trait acquires a functional role (that is gain a new working role after it has emerged, been selected and developed) without further development, then we have something “unexpected”, that is something that is not the result of an adaptive design process. The trait did not pass the selection process as a whole (it is like we let you in the system earlier, but we were not aware and did not take into consideration that someday you could play this role). Likewise, as selected traits arise and continue to develop, the random conjunction of traits can instantly create new, “unselected” and “uncontrolled” potential abilities (it is like we individually let those guys in the system earlier, but we did not take into consideration that they could join forces and play this role together).

The conjunction of selected traits

This cover what I call “evolutionary leaks” and “hacks”, which are the basis of natural hacking. Leaks and hacks are the same thing, seen from a different perspective. From the individual perspective, when a particular trait of an existing organism is suddenly used for a new functional role, it is a hack. Whereas from the global perspective, when an existing organism suddenly starts to use a trait to do something it was not supposed to allow (that is play a functional role that is not genetically conveyed), it is a leak.

Egocentric evolution tries to create as much hacks as possible to maintain and increase individual fitness.

Allocentric evolution uses hacks only to counter the blind hacks made by egocentric evolution, so as to quickly plug the associated leaks.


Hacking potential as a genetic mutation / genetically determined. Hacks sourced by genetic mutation.

Because hacking is associated to usage, it is not confined to “conscient” beings. Examples ? Leaks arguably happen in unconscious systems.

Trait power control

Note : If one is familiar with Ruth Milikan work, I think he can understand initial selected roles of traits as what she calls proper functions, and roles that were not selected as what she calls derived functions.


Understanding hacks by distinguishing them from exaptations

Exaptation refers to a change (addition or replacement) in the selected role(s) of a trait. For example, a trait might have evolved because it served function α, but subsequently it may evolve because it serves function β (in addition or replacement to function α). In exaptation, the new function makes use of the material developed for the previous function(s). And without this available material to hijack, the new function would probably never have emerged.

At first, the concept of hack might seem strikingly similar to the concept of exaptation, to the point of being redundant. And while exaptation in the broad and loose sense can be seen as a particular type of hack, in the strict sense, exaptation is not exactly a hack.

The first distinction comes from the global perspective. Indeed, as I said earlier, the counterpart of a hack in the global perspective is a leak. But exaptations are supervised shift of development (therefore not leaks), they are caused by natural selection.

Moreover, unlike an exaptation, a hack is not really an evolution, that is a hack is not a progressive and adaptive design change. On the contrary, a hack is a transformation, that is a quite drastic and instantaneous change. Furthermore, a hack is a purely “logical” transformation, in itself it leaves no mark on the design level (that is in the genetic code). Exaptation is a particular but still classical evolution : it revolves around selected genetic changes.

When a random hack happen, there is no change in the selected roles of a trait.

All in all, these concepts are not on the same level. Exaptation is more on the design level (genetic change for a new role/usage) whereas hack is more on the usage level (no genetic change directly involved).

An exaptation can occur after a hack though, to develop, “hone” and fine-tune the hack associated usage in the genetic code (which is basically referred as a Baldwin effect). But I propose that exaptation is not necessary for a hack associated usage to be effective (and influence evolution). Hacks precede (some) exaptations.

To make the distinction clearer, let us take the example of a species in which no individual has ever been able to fly but which developed feathers for thermal regulation. If one day an individual of this species uses its feathers to fly (and succeeds in doing so) it will not be an exaptation, because nothing will change in this individual “code” (note : genetic changes arguably can not arise in an already living organism anyway, even if it is still questionable). It will “just” be a hack. If offspring of this individual develop feathers adjustments that makes flying better, then here though, it will be an exaptation.

Of course because of the time span of evolution, it is really unlikely that the individuals of a species have the potential to properly fly but do not fly. But the idea behind the example is still valid, especially for logical traits such as cognitive traits which are not (that much) constrained to physical rules.

If this hacker individual never existed but if instead another individual born with a viable genetic error that made feathers a bit different succeeds to have offspring that keep this error (without it being selected, just through neutral, basic inheritance) and if these offspring themselves produce new individuals that develop this error (still through random mutation), then if this line of descent continue to survive one descendant might eventually be able to do something that we can barely call “flying”, without it being of any usefulness. But continuing on this mutation road, one descendant of this descendant might eventually have a “flying” ability decent enough to provide an evolutionary benefit (by enhancing its hop ability for example). And from this point, an exaptation process to select and develop feathers for flying could start. And this would be a pure exaptation, that is an exaptation not preceded by a hack ; because the “proto-flying” relied on a design evolution on the one hand, and more importantly because the “proto-flying” had no role/function on the other hand.

It is questionable if pure exaptations really exist, that is if exaptations can really arise without being preceded by at least a minimal hack (because it is as hard to determine whereas a function is a hack or not as to determine whereas a change is an adaptationadaptation or not) . In our example, the most credible exaptation of feathers for flying would be initiated by an individual which (through a hack) would have used its thermal regulation adapted feathers to “fly” really inefficiently but still decently enough to add a role/function and provide a tiny evolutionary benefit. Because it flew through a hack, this individual would transmit nothing fly-related to its offspring through the design reproduction process (well actually maybe it would transmit the devious behavioral trait that made it try to use its thermal regulation adapted feathers to fly, provided it was sourced by a genetic mutation, see section “Hacking potential as an inheritable, genetically determined trait”). So flying-wise, its offspring would be just like any normal individual of this species, and in the standard evolutionary framework, there would be no evolution from this hack, the “system” would “reset” at the death of this individual.

Hack spreading and functional fitness

a hack need either some way to spread from the initial model.

“Spontaneous” Be forced to emerge by hackisition pressure. Drague, production, inventivité

So to lead to a related exaptation, a hack needs some way to spread out of the individual in which it arisen (~ the initial model), so as to simulate what genetic inheritance does. The individuals in which the hack can spread will have access to the hack associated function and will thus increase their fitness (~ evolutionary power), biasing natural selection towards them, “encouraging” exaptation of the hack associated function (in our example, flying). Hack spreading requires a mean of logical reproduction (some form of imitation, some form of basic cultural learning and inheritance). Because hack inheritance is not genetically conveyed, it is not limited to direct offspring and constrained by the rules of genetic reproduction (vertical oriented transmission VS oblique non oriented transmission). An organism can increase its fitness after it is born by acquiring hack associated usages through the mimicking of a younger individual that is not one of its offspring, for example. So in natural hacking thesis, a bit like an expanded kin selection (see inclusive fitness), the final unit of evolution is not the gene, it is the function or usage (associated to a trait). If a function/usage can survive and reproduce – whether it is through standard genetic inheritance of traits (layer 1) or through hack spreading (layer 2) – by helping some organisms (that is users) survive and reproduce, this function/usage will bias natural selection and thus evolve.

Inclusive fitness (exemplified by kin selection) differs from individual fitness by including the ability of an allele in one individual to promote the survival and/or reproduction of other individuals that share that allele, in preference to individuals with a different allele.

Functional fitness (exemplified by hack spreading) pushes further the bounds of selection by including the ability of a function in one individual to promote the survival and/or reproduction of other individuals in which the function can reproduce (that is individuals that can acquire this function), in preference to individuals in which it can not. The promoting relation is not binary as some organisms can acquire a function more easily than others.


Evolution in standard evolution theory

Evolution in natural hacking thesis


Same thing with perfect flyer.

Hack spreading and anti-altruistic behavior selection

If kin selection can explain the natural selection of some altruistic behaviors, hack spreading can explain the natural selection of some anti-altruistic behaviors, such as information/resources hiding (i.e. keeping secret some abstract “recipe”, from a technique to harvest or hide food to encrypting and keeping closed the source code of a program) and information/resources protection (e.g. some way to watch over a territory and chase away strangers, copyright, patents). If an organism comes up with a functional solution to a problem and if it is able to keep its function to itself (or at least restrict its function use to itself), then it will increase its fitness but not the fitness of others.


An abstract and functional-based approach to reproduction

In summary, natural hacking thesis entails an abstract and functional-based approach to reproduction. In this framework, evolution is situated beyond individuals and beyond genes, it is “in the cloud” (that is to say it is not situated at all). Individuals and genes are just specific embodiments of functions, means for humans to perceive, evaluate and “touch” functions (which are purely logical). Genes are to functions what behaviors are to cognitive processes. We can learn and infer a lot about cognitive processes through the studying of behaviors. Likewise, we can learn and infer a lot about functions through the studying of genes. But functions are not limited to genes. As human history only tell about some important people, genes only tell about functions that have a long story behind them. Typically, people make history after they die, not as they are born and live. But still, people that will not make history or have not make history yet, do exist. Only people that make history somewhat live on and continue to exist after their death though. Analogically, a function is not initially a trait. A function might become a trait, or take root into a trait some day if it plays an important role in evolution. But it does not need to take root into a trait to exist. Therefore, genetic transmission is just one way (arguably the most fundamental and most reliable way) of reproducing functions (and it only encompasses functions that made Evolution History). Genetic transmission allows function reproduction through the reproduction of genetic function embodiments (i.e., genes). Cultural learning allows function reproduction through the reproduction of perceived behavioral function embodiments. Functions are wrapped in multiple types of embodiments. And it is these functions (and not only through genetic function embodiments) that in the end, evolve. Each type of embodiment allows ways of reproduction.

To make an additional analogy, we can say that innate, genetically coded functions are like hard-coded functions of a programming engine : every program that is made with this engine have access to these functions. (Note : it is interesting to note that hard-coded functions of engine are selected by developers through a process akin to natural selection). And therefore, in this analogy, non-innate acquired functions are like user-defined functions that makes the individual program. All the individual programs made with an engine would belong to a particular “species” of programs. This analogy emphasizes the genetic scaffolding in which natural hacking and hack spreading occur ; indeed, just as genes reproduction is constrained by organisms reproduction capabilities, function reproduction is also constrained by organisms learning capabilities (in the broad sense), which are genetically determined. It is undoubtedly possible to reproduce some (user-defined) function of a program into another program if both are made with the same engine. If they are not, it might be easier/harder (or even impossible) to reproduce this function depending on the function to reproduce and the engine in which the function is to be reproduced/implemented. Functions can not spread randomly through natural hacking.

A function might require multiple individuals.


Hacking potential as an inheritable, genetically determined trait

Are hacks really independent from the design level ?

No because would mean that all we can do as humans with our creativity is genetically determined. But might be genetically biased, that’s for sure… Hacks are genetically biased.

Superstition pigeon : encourage les hacks.

Are hacks really independent from the design level ? Obviously, hacks are not fully dependent of genes. Because if they were, it would for example mean that all we can do as humans with our creativity is genetically determined. Obviously it is not the case. Nonetheless, what we can do with our creativity is most probably genetically biased. Some people are naturally creative in a particular way. Therefore it is likely that hacks in a more general way, are genetically biased. Some individuals are more disposed to create a particular hack than other individuals (they have more “chance” of creating this hack). Hacks are usage (or purpose) errors that might be sourced by specific design errors. In this sense, some genetic mutations might make affected individuals “natural born hackers”, likely to use some of their traits in a “devious” way. Psycho-evolutions could thus precede and entail some morpho-evolutions. Using the previous example of the flying hack, what is important is not to have wings to fly (morpho-trait), it is to have the « idea » to use wings to fly (psycho-trait). If anything, evolution creations are the proof that it is possible to create a lot with a little, and so that is better to have a lot of “ideas” than to have a lot of material.

If every hack has a genetic source, then natural hacking thesis is somewhat implicitly included in the standard evolution theory.


Hackisition pressure, constraints, ontogenic evolution and the Baldwin effect

The Baldwin effect, as well as the concept of niche construction, are concepts I read about after developing most of the Natural Hacking Thesis ideas. The idea behind the Baldwin effect is basically the same as the idea behind what I call natural hacking evolution, and the evolution mechanism surrounding the concept of niche construction is basically the same as the idea behind what I call pressure modeling (which I will develop later in this text). I try to develop and write from scratch a lot of constructed ideas about various global things, so this (~ reinventing the wheel) does happen to me quite often. Actually I wrote about this particular phenomenon in the first chapter of my previous essay (warning : it is in french).

Why does this happen ? How come one independently “reproduce” concepts without knowing about them ? The key word is “constraints”. And the key idea is that when you apply the same constraints on similar individuals, they will naturally tend to produce similar functional responses to these constraints, even when they have no innate response to these.

It is quite common. Independently. Multiple reasons (actually I may know about them without knowing it. And constraints).

Baldwin effect and (some) niche construction = effect of hacks on evolution

So what exactly is the Baldwin effect anyway ? The Baldwin effect describes the effect of learned behavior on evolution. It basically states that an acquired behavior, if it is really useful, can progressively bias the emergence of an innate analog. Here is a quoted example from the Wikipedia page : “Suppose a species is threatened by a new predator and there is a behavior that makes it more difficult for the predator to kill individuals of the species. Individuals who learn the behavior more quickly will obviously be at an advantage. As time goes on, the ability to learn the behavior will improve (by genetic selection), and at some point it will seem to be an instinct.”. This example basically describes the combination of hack followed by exaptation I developed earlier with the flying example. I think it is quite easy to find analogous images of the Baldwin effect : to solve a particular problem, there is an ideal tool that is expansive to buy or to construct. If one face this problem once but do not have this ideal tool, he will probably just try to solve the problem without it, or improvise a makeshift tool. Buying or constructing the ideal tool will not seem worth it at this point. But if one repeatedly faces this problem, it will be a good idea to develop a proper and permanent solution instead of always relying on convoluted solutions and temporary arrangements. In software programming, it is quite common that a very useful external function becomes integrated in the main source code of a program after years of useful usages.

The Baldwin effect does not emphasize the hack itself but the capacity to create a hack. Individuals that will more easily and quickly be able to construct the ideal tool will be selected and thus bias evolution, to the point where constructing this particular ideal tool will become so easy it will seem innate. The main difference between the Baldwin effect and Natural Hacking is that in the later, a hack does not need to become innate. In Natural Hacking, what is important is the function, and therefore the evolution of functions and not the evolution of genes. So if a hack associated function works really well out of the box, there is no need to create change in the genetic code and make it innate.

The Baldwin effect also emphasizes the role of pressure on hack emergence. Here in the predator problem example, the new behavior (or in my words the hack associated function) is constrained to emerge because it is the only possible behavior that makes it difficult for predators to kill individuals of the species. We can say that the function of this particular behavior is embodied by the pressure, it is somewhat evoked by it. Just as a function can be duplicated through the perception and reproduction of behavior embodied function model, it can be duplicated through the perception and reproduction of a pressure embodied function model. The function is represented by the pressure, that is, problems are particular embodiments of solutions.

The thing is, hack spreading (reproduction of hack from an initial working model by another individual) is not the only way for a particular hack to emerge at a population level. As pressure forces particular traits to develop, it also forces particular hacks to emerge. I call this the hackisition pressure, which is the pressure to learn and acquire particular functions (to better cope with the environment) that are not genetically coded (hard-coded), but embodied in pressure. If an individual is confronted to a problem (new predator, etc) and if its genetically coded functions are useless to face this problem, it will be forced, if it can, to figure out and come up with a hack (that is a new function) using its existing traits. Or at least it will have better chances of solving the problem if it is able to create new functions. And if there is few functional solutions to the problem, these solutions will arise spontaneously in multiple individuals facing that same problem, as if they had the same model. That way a hack can emerge through independent instances, without reproduction of another individual (see also convergent evolution, recurrent evolution).

The combination of hack spreading and hackisition pressure

Search for truth = one solution.

Once again, I think this process is analogical to other. Instead of modifying the code of a program, a lot of for a new usage. Have not the skills to change the code of a program, they use the existing materials in a convoluted way to achieve

“Spontaneous” Be forced to emerge by pressure. Drague, production, inventivité

Re-emergence fonctionnelle, republier abstractionabstraction et mettre lien


When hacking becomes integrated – a brain programmed to hack the world

Alternate titles:

programmed to hack the world

When hacking enters supervision

the hacks zones

Animals can hardly do that.

These zones require conscious

prefrontal cortex




Capacité de raisonner et comprendre les relations causales

tool use and obtaining goal

Our ability ro reason (which is different from our decision-making ability)

abstract reasoning, mechanical problem solving, technical reasoning

anterior supramarginal gyrus

Our abstractionabstraction ability / decomposed

regarder les TODONEXT

TODO gyrus supra-marginal antérieur. Raisonnement d’usage, outils. Not in chimp.

TODONEXT ontogenic adaptationadaptations

phylogenic adaptationadaptation


The main role of some parts of our brain seems to be… the creation of roles. This is the case for example, of the anterior supramarginal gyrus (aSMG). The aSMG grants human beings the ability to understand mechanical logic and work out in advance how physical tools could be used or modified to solve a multitude of new problems (that are not genetically encoded), to obtain a goal. In other words, the role of the aSMG as a trait is to allow mechanical problem solving. Non human-primates all lack this particular brain “area” and can not do that. They can use a tool as some kind of extension of their arms, but can not figure out a new use for a tool. If we acknowledge that traits evolve to solve a specific yet abstract problem, that traits are a particular solution to a problem, then the problem that aSMG solves is mechanical problem solving.

A trait that evolved for some abstract processing (such as mechanical/technical reasoning) is a trait that will create functions that are not genetically encoded when used. In other words, it is a trait that will naturally create hacks (it is not really the same type of hacks though, because they are somewhat programmed and therefore “approved” by evolution ; there is no usage error with these traits). It is of no surprise that such traits have emerged. Indeed, hacks can allow individuals to gain immediate advantages. Therefore, natural selection should favors traits that have a strong inclination to be used to create hacks (~ Baldwinian evolution), traits that do not have a strictly fixed and restrictive role and better, traits that can create roles. Moreover, hackisition pressure favors the development of organisms than can “decipher” and imagine functional solutions to a pressure shaped problems, and hence favors the evolution of hacking capabilities. Instead of relying on fortuitous or forced hacks to gain immediate benefits, it is better to have the power to create flexible hacks. For egocentric evolution, developing a flexible ontogenic adaptationadaptation is the holy grail. Being able to generate new functional roles without additional genetic changes is the “nuclear weapon” that allows it to win the war against allocentric evolution.

Memory as a trait is a good example of supervised natural hacking. Memory has been developed because it was used for many roles. So memory did not develop for any particular role.

La mémoire s’est développée parce qu’elle a été utile à plein de rôles. AbstractionAbstraction cadrée ?

Some hacks don’t increase reproductive success, on the other hand. Natural Born Hackers diverge from their reproductive objectives.

hackers sont engagés par le gouvernement. But still as hackers can be to work on a particular job, they can parralelly work on their own “hacker” projects and screw up. It’s basically an error to trust hackers.


To me, such abstractionabstraction oriented traits are what differentiate humans from the rest of animals. Unlike other (still existing) species, human species is a natural born hacker species that is hacking the world it lives in by creating mental tools. The common organism is basically confined to genetically coded usages of its traits. That is it has a relatively fixed set of usages. Humans do not have to rely only on the programmed usages of their traits to survive and reproduce, they can define their own. Not only we have the aSMG, but most of our prefrontal cortex is devoted to abstract reasoning and functions creation. Actually, in my view the core role of language is abstract reasoning, and not communication.

To me language (the ability), in its core and in all its complexity, is more part of an abstractionabstraction system than part of a communication system. That is, what language is to me is mainly an internal system allowing thought orchestration and articulation (for hacking).

Besides, human beings do not really use language to communicate complex thoughts. The main usage of language for communication is basically simple social interaction. On the other hand, some individualistic activities such as reflection, creation, problem solving and such relies heavily on the capabilities of language. The aSMG area allows us to understand the causal relations of mechanical tool usage from a high-level. Likewise, language allows us to understand the relations between various mental representations.

Pre-humans arguably did communicate using some of their traits before language was a trait. When language emerged, it was then integrated in that same communication role because it is a powerful trait than can improve almost any role. Language made this communication role reach a new level, without really transforming it into something qualitatively different. We use language to talk about “little things”.

It is important to note that the communication role of language involves a lot of additional mechanisms than its (main) abstractionabstraction role. The communication role of language involves externalizing thought, interpreting externalized thought, establishing shared signifiers across individuals, etc. The mechanisms of production (externalizing inner language) and comprehension (internalizing external language) are not part of the core language ability in itself. Those forms additional layers.

To me, the parallel between our language ability and programming languages is strong. We created and use programming languages to construct logical structures (~ ideas) that solve some problem. We did not create them nor use them to communicate. Likewise, I think language initially emerged for constructing mental structures (no matter how they are actually constructed) that solve some problem. This trait allowed individuals to figure out adaptationadaptations and better cope with the environment.

Evolution in standard evolution theory

Of course now, because of functional fitness, language is also selected for its communication roles. Because language communication allows large-scale development and mass functional spreading (through shared signifiers). The conjunction of the core language role (abstractionabstraction) and its communication roles made possible development of things through space and time (~ the Tower of Babel). I will come to the implications of such usages later in this text.


Sexual selection as the main “hacking force”

Initially, evolution strives towards survival and reproduction. To do that, it creates systems. Then, after there is at least one system created, evolution can also strives towards these systems. For example, we have evolved to appreciate bodies with particular shapes because these bodies were positively linked to survival and reproduction (and likewise, evolved to dislike bodies which were negatively linked to survival and reproduction). Then I think that we also evolved to appreciate/dislike bodies because they are appreciated/disliked by others. At some point the links became so rooted that it became really advantageous to just have a particular body shape (especially regarding reproduction, but regarding survival too). And this is why some people “artificially” shape their body through plastic surgery, bodybuilding, etc, without any functional objective but an aesthetic one. By focusing on the aesthetic function, they do not acquire the survival/reproduction traits that were implied by such bodies in a natural and ecological environment ; but their bodies elicit the appreciation system of other individuals nonetheless, as if. The initial link between shapes and survival/reproduction is exploited. I think this “snake biting its tail” effect, or circling evolution, can actually lead to loss of fitness, especially in the long run. It also explains why we are over-focused (in regard to actual survival and reproduction) about things that elicit our systems (beauty, knowledge, suspense, drama, …).

All the things that elicit our system, disjointed

Distal to proximal. in the first place. Hot traits. Because there is not actual supervision system, evolution can differentiate

A progressive switch from adaptationadaptation to function.

So there is adaptationadaptation for survival and there is adaptationadaptation for reproduction. And traits are parallely developed for one or both of these adaptationadaptations. AdaptationAdaptation for survival tends to create “cold” traits, that is useful traits without anything superfluous, traits that are well-fitted, well “thought-out” for the environment, using simple and direct functional approaches. There is some kind of pure rational efficiency with cold traits ; we could say that adaptationadaptation for survival is like an engineer mind. On the other hand, adaptationadaptation for reproduction tends to create “hot” traits, that is aesthetic, extravagant, “showy & catchy” traits that, with the exception of reproduction, are not really useful (might be the opposite actually) nor well thought-out for the environment. We could say that adaptationadaptation for reproduction is like an artist mind.


Unlike cold traits physical impact vs logical impact

Instead of relying on true survival and reproduction

It is as if the system understands that survival and reproduction is not the decisive factor, cunning.


Note : not all reproductive traits are “hot” traits though. A lot of them are “cold” traits, particularly the low-level reproductive traits (better fertility rate, …).

The thing is, I think it is not directly possible to create a “hot” trait. “Hot” traits use indirect and convoluted approaches, they are the result of some “cold” traits hacking. Take for example the famous peacock feathers. I think peacock feathers (a “hot” reproduction trait) relies on an exploit (see later section for differences between hacks and exploits) of the visual perception system functioning (a “cold” survival trait).


Peacock feathers display

This example is very interesting because it shows a lot of characteristics of natural hacking. First it shows that natural hacking is not limited to the individual scope. Hot traits are good examples of traits conjunction. The hot trait in itself (the feathers appearance in our example) is not functional. It is the conjunction of the feathers appearance and the functioning of the visual perception system of others that creates a hack. So it is an abstract function that spans on multiple traits located in different individuals. These conjunctions possibilities greatly increase the ways for natural hacking to happen.

Next, this example shows that some traits can passively hack the traits of other individuals by creating an exploit. A trait can gain a new role without being used for this new role, that is without requiring any new use at all. In our example, the “passive” exploit of the (female peacocks) visual system by the (mutated male peacock) feathers appearance creates a passive hack (because feathers gain a new role). And this passive hack will bootstrap the actual hack : using feathers for basic courtship display. And this hack will be followed by an exaptation to enhance this beneficial use and create a sophisticated courtship display.

Indeed, here is a story possibility for the peacock feathers. At first there was just a random mutation that changed feathers appearance of one peacock. And this appearance change made this peacock stands out, it made it more attractive (a bit like having a special eyes color). At this point it was just an exploit. But at later point in time, this peacock or one of its descendant made active use of the passive attractive power of its feathers to seduce females peacocks. This was a hack, bootstrapped by an exploit. As generations passed, feathers progressively exadapted for this new role (courtship display).

Because it would be an exaptation.

Survival it is the individual VS the world. You don’t really need anyone. With reproduction it is a different story. This is not your traits that gains usages, but the traits of others.


Peacock feathers : a force on the appreciation system.


Evolution from the global perspective : uncontrolled usage of traits as leaks

Quick remainder : hacks and leaks are the same thing, seen and studied from a different perspective. When a particular trait of an organism is suddenly used for a new role, there is a leak in the global allocentric evolutionary process.

Considering the increasing complexity of organisms, evolutionary leaks are bound to happen. Most of the time though, they are “handled” through “just-necessity” principle and create a new base for further evolution. In other words, leaks are “integrated” in the evolutionary process and balance is restored.

The allocentric evolution supervision process and the just-necessity principle

[TODO article, concept related to selection pressure]

Agents (or organisms) do not evolve aleatory ; they evolve mostly in response to selection pressure. If a new pressure arises, agents start to evolve and develop traits so that they can resist this pressure. And if a pressure is no more a pressure (because of some evolution in the agent or simply because the pressure source disappeared), agents start to evolve so that they lose (or decrease) traits linked to this pressure. In other words, evolution respect the just-necessity principle, that is to maintain and develop (and continue to maintain and develop) only what is needed to stay in the business. Because of this, agents co-evoluate and auto-organize themselves in a harmonious balanced way. More importantly, this principle avoid super and under powered entities.

That is why we can say, when evolution is purely adaptive and progressive, that there is some kind of supervision process to evolution, some kind of awareness.

But because of leaks (among other causes, such as sudden environment shifts), evolution is not purely adaptive and progressive. Typically, leaks suddenly decrease or remove pressures for agents in which they happen (and are), and suddenly increase or create pressures for the other agents (in which they don’t happen). Moreover, because it can allow agents to change their environment, leaks can also open agents in which they occurred to new pressures.

And then, after a relatively long period of feedback reactions (to the changes made by the leak) in every affected agents, the leak is fully integrated (“plugged”, or “fixed” we might say) and therefore is no longer a leak.

Supervised allocentric evolution process, “ordinary” adaptationadaptation → leak happens and breaks supervision → extraordinary adaptationadaptation → leak is integrated in allocentric evolution process → Supervision is back

(Note : clic droit sur l’image, ouvrir dans un nouvel onglet pour la voir en gros. Comment ça j’écris en français dans un article anglais ?)

How leaks are processed

Leaks tend to be more and more abstract, and thus more and more difficult to integrate in the existing system. But I’ll come to that later.

Typical feedback reaction can not happen with such leaks.


Of design and usage problems – the adaptationadaptation paradox

Thus, hacks/leaks are not an uncommon phenomenon. In fact, the principle of natural hacking might be the main basis of function, or role emergence (see also : exaptation). But the main problem with this principle is that new traits (~ new abilities) are “implemented” by other traits not “designed” to serve such role. Hence why some see evolution design as imperfect, jury-rigged (F. Jacob used the term “tinkering” in his 1977 article). Adaptive traits end up serving “inadapted” roles (or uses). This is what I call the adaptationadaptation paradox : the whole system (~ organism) is adapted for what it has to do (~ survive to procreate) but the traits that makes this adaption are not adapted for what they do. That is because adaptationadaptation is only made in regard to the increasing of self-procreation.

Derivative Indirect inadaptationadaptation.


Raisonnement par marqueurs somatiques.

D’une manière générale, un changement évolutif (càd une variation sélectionnée) est presque toujours bon quand on considère l’individu (ce qui est attendu puisque c’est “dans” l’indivudu – unité de base du réseau – que le changement apparaît). Mais quand on considère les conséquences d’un changement évolutif au niveau de la population, c’est autre chosechose. Cf suicide évolutif.


Quand le potentiel d’un changement évolutif est utilisé de manière distribuée par la population.


Individual value VS population value

A trait, no matter why and how it became a trait, is almost always “good” when we consider an individual organism. Which is expected because it is “in” the individual that evolutionary changes arise. But when we consider the consequences of an evolutionary change at a wider level (e.g. population level, world level, offspring level) things are different. Traits that are positive to individuals can be negative to individuals’ population (and thus, indirectly, to individuals themselves) or to individuals’ offspring. This lead to what is called evolutionary suicide.

Evolution is blind to everything that is not “directly” related to individual procreation. For example, evolution designs organisms to consume particular natural resources that are available without taking into consideration that these natural resources might or might not be renewable. Evolution designs as if what is available now will always be.

Because it spans over years, a typical progressive evolution has less impact power on ecosystems than a hack, because the environment will coevoluate to this evolution as it happens. When a natural hack happens, there is no coevolution possible as the hack instantly creates a developed ability. Environment will start to adapt but it might already be too late. That is why I think evolutionary leaks are strongly linked to evolutionary suicide, and why agents that develop hacking capabilities in evolutionary-like systems develop them very progressively (so as to not cause their own extinction).


Bogus good ideas – “Oops, I didn’t mean to create this !”

Moreover, the idea of a link between evolutionary leaks and evolutionary suicide, is also that evolution intelligence (through the egocentric process) has bogus good ideas : if there is a far-fetched way to do something interesting (that will increase individual fitness) using traits that are already available, it will do it. From a human perspective, this “tinkering” used by nature is both genius and stupid. Genius because nature somehow manages to reuse unrelated specific traits as a base to create a totally new trait, which is really economic and greatly speeds up development. And stupid because such an implementation is bound to create problems (let alone be unreliable and deficient), at least in some conditions. (Sure, once a hack has emerged, it might start an exaptation process and get improved through natural selection. But its functional approach will remain the same, with all the associated flaws. And its functional presence will probably interfere with the development of a proper solution.)

And actually, the main problem is not that a trait is deficient or harmful, but that a trait can be deficient or harmful in some conditions and that there is no way to prevent using it in these conditions.

“Oups” effect : didn’t mean that.

Deprecated traits and out of date adaptationadaptation – Exploits

Earlier in this article, I mentioned that in addition to leaks, sudden environment shifts can break evolution supervision. That is because evolution is a process that works on the basis that environment is globally stable (~ that it is more or less the same anywhere the organism can be) and changes very slowly and progressively. Organisms traits and traits’ parameters are precisely fitted to what their environment used to be ; and because the typical environment evolves very slowly, these parameters are probably pretty well fitted to the current environment nonetheless.

When for some reason an element that was not taken into consideration (in the definition of parameters) suddenly appears/enters in the global situation/environment of organisms, the presence of this element will probably mess up with their regulation processes. The typical examples are drugs and supernormal stimuli. Drugs can either be like super stimuli, that is elements that are processed through natural ways but which have a concentration level that (far) exceeds the normal level of the exogene elements the organism has evolved to process, and thus create a supra-reaction when processed. This type of drug usages belong to what I call “value” exploits. But drug usage can be like “path” exploits, that is processing elements which contain a property (~ molecule type) internally used by the organism through a pathway that was not naturally developed. A particular yet frequent case being breaching a pathway that did not naturally exists ; for example a path exploit happens when an individual ingests elements which contain a property (~ molecule type) internally used by its organism and if in the design process of this organism that molecule was not supposed to be present/available in the environment.

This cover what I call exploits. Exploits are a particular type of hacks/leaks. Exploits do not happen when an existing trait is hacked, that is when an individual suddenly use a trait for a new role. Exploits happen when an existing role or usage is hacked, that is when an individual gets to process things its ancestors have not evolved to (from the global perspective, things it was not supposed to process, things that were not supposed to exist with that concentration or/and in that quantity for this organism).

For example, if over the years evolution forges some organisms to process some natural resources “A” based on the concentration level of a molecule “X” present in these resources, these organisms will develop a response to this molecule “X” based on its normal level in “A”. If some years later in the history of evolution these organisms discover some other natural resources “B” that contains 10 times more X than the typical “A” resource they are used to process, processing “B” will severely mess up with their behaviorial traits. And because the evolutionary response associated to molecule “X” is probably some kind of attraction and addiction, the exploit will spread quickly in the population.

Because evolution is a very slow process, a lot of previously positive traits become unadapted when the global situation/environment evolves. Their adaptationadaptation fitness becomes out of date and their usage becomes deprecated in regard to evolution.

Exploits and hack spreading

Hack spreading and inheritance (chocolate, macdonald).

Elements that create exploits, and exploit associated usages, have great chance of spreading through space and time, invading whole populations. Things like chocolate and fast-food consumption are typical example of “value” exploits. It is also no wonder that drugs (“path” exploits) are so widely used.

Only path exploits strictly involve a new usage of traits. But with value exploits, a “naturally conceived” usage can be so much exploited that it can hardly be considered to be the same natural usage it was before being exploited (that is for example, over-eating is not eating).

So by distinguishing general usage (i.e., ingesting food) from particular usage (i.e., ingesting chocolate), we can arguably say that it is not whole functions that spread and are reproduced… but particular execution of functions. Eating chocolate will spread easily, unlike eating something with a bad taste.


Thus, the main problem is not that a trait can be deficient, it is that there is no way to prevent it from being deficient and creating problems (that is, basically, used for something other than what it was selected). On the contrary, other traits (like needs) tend to force “wrong” usage of a particular trait (that is force the emergence of unadaptive hacks). Good traits can become bad traits as they acquire usages that decrease fitness (e.g., path exploits).

And then real allocentric (global) problems can arise because the presence of these “unreliable” traits allow the emergence of new (complex) traits. And therefore a leak can become worse and worse, more and more out of feedback control. That is why it is risky to create hacks.


The case of human cognitive traits (abilities)

Grounded cognition, embodied cognition, extended cognition : it is easy to see how all these modern conceptions of the mind revolve around design constraints, natural hacking, exaptation and such.

But are human cognitive abilities better described as supervised hacking abilities (that is traits that evolved for their abstract usage through natural selection) or as leaked abilities ? That is a difficult question to answer.

I think that because of natural hacking,

TODONEXT Our abstract oriented traits… are not that abstract.

Indeed, as it is conceivable to explain the emergence of such hack-associated traits through natural selection (see previous section “When hacking becomes integrated”), it is also possible to defend that most of our advanced (abstract) cognitive abilities are leaked abilities. That is to say that these abilities do not mainly exist (nor did evolve) as such, that they are mainly a twisted usage of other traits, of other systems. Of course because of transcendence, usage makes function, but as they are now I think it is difficult to say if advanced cognitive abilities are still better described as particular usage of sub-systems or as transcendence of these sub-systems.

traits that implement these abilities were not developed .

The thing is, even if our abstract cognitive abilities might not be leaked abilities, they are important leak sources nonetheless. And the usages we make of them right now can more certainly be qualified as hacky (that is not naturally selected) as the opposite. In the same way as over-eating can not be seen as the same usage as eating, the over-usage of our cognitive traits can be seen as a functional exploit.

(corsair that revert to pirate)

Because hacks are logical transformations, mental (cognitive) traits are a lot more subject to hacks than physical traits, for they are not constrained by physical rules.

We cannot overcome our arms inadaptationadaptation to fly, but we can overcome our mind inadaptationadaptation to handle big problems by abstracting and distributing resolution (~ development of solutions). Indeed, our capacity to find solutions is overdriven by abstractionabstraction (that is problems are recursively decomposed into smaller problems) and distribution (that is decomposition and resolution of problems are handled by several individuals over multiple periods of time). Still, even if we can tackle big problems (creating complex systems, etc), our abilities are not adapted to handle them.

We are makeshift machines in regard to what we are doing now, only just capable of doing it because of hacking strategies. We are doing everything we can to overcome our limits and make the most of the little we have at our disposal, so as to develop bigger and bigger things.

Work partitioning – The flying elephant problem

It is as if we could store the energy of each arms fluttering movement we do (each one being insufficient to fly) and release this energy at once. Flying with arms would be possible that way. But because of physical constraints, it is not possible to do this, we cannot pause such physical process. Flying with arms can not be distributed in time, nor can it be distributed in space (storing the energy of other individuals arms fluttering for our usage). It must be done in a single continuous period of time by the same individual. Therefore if there is no preadaptationadaptation for flying (whether it is the fruit of selection or a fortuitous, incidental preadaptationadaptation), there is no way an individual succeed to fly, it requires way too much effort for its organism (and training won’t help reducing the amount of effort significantly). Because cognitive traits allow efforts to be diluted, they do not require specific adaptationadaptations.

Partitioning allows us to live in the “somewhere and sometime” paradigm with “here and now” traits. Being able to do this opens individuals to new functional dimensions. But it is not only for the best it is also for the worst.


When too much adaptationadaptation becomes unadaptive – Over-adaptationadaptation and under-qualified traits

We plan things.

Thanks to the conjunction of our minimalist rational abilities and our memory systems (which includes external memory systems such as written paper), we human beings managed (and manage everyday) to leave the “here and now” paradigm that all natural agents were previously strongly bound to. Which means some of the things we do (that is things we develop over space and time) can hardly be controlled by the allocentric evolution supervision process (~ maintenance, co-evolution equilibrium) ; our adaptationadaptation capacity is abstract, and thus unpredictable. In some way, we can say that we are over-adapted. That is our capacities allow us to adapt to a lot of situations as they are, without any additional evolution (through genetic fine-tuning). We do not need to (genetically) adapt and evolve as we use our abstract capacities to do things that change the world for it to be adapted to us. Through niche construction, we adapt and model the pressure so that we do not have to adapt. And because we model and transform pressure (and because pressure is what defines adaptive evolution direction) we somewhat influence and “control” evolution direction.

Under leak-proof development (that is purely progressive development were every particular function is genetically embodied), evolution would have never created agents capable of doing what we are doing so early. It would have created truly rational agents, which we are not. It would have naturally “waited” (a very long time) for agents to develop a proper rational (and abstract) decision system, which we have not.


Rational decisioning (decision thinking and making) as an hack associated usage

Option-outcome scenario

During the second half of the 1980’s, Antonio Damasio and his colleagues unveiled an interesting cognitive system that they called the somatic markers. Each time we make a choice (that is select a particular option among multiple options) with an ambiguous outcome (that is not knowing what option is the best/worst, what option is advantageous/dangerous), an association between the “representation” of this particular choice and the emotion it led to (satisfaction, regret, …) is stored. This [choice of option X → emotion Y] type of association is what a somatic marker is.

Thereafter, each time we will be confronted to this choice again, when we will evaluate the option that we chosechose in the past, the associated emotion will be automatically reactivated, causing a bias in the decision making process. If the leading emotion was positive, then it will bias us to select this option again, whereas if the leading emotion was negative it will bias us to not. Somatic markers are additive, and that is why they are so powerful. If an option is sometimes extremely profitable (++++) and sometimes extremely unprofitable (- – – -), it will not be associated to the last emotion elicited (+ + + + or – – – -), but to the average (that is a somewhat neutral value). If an option is sometimes profitable (++) and sometimes moderately unprofitable (-), its activationactivation will elicit a rather positive emotion (+). If an option is sometimes highly profitable (+++) and sometimes extremely unprofitable (- – – -), its activationactivation will elicit a rather negative emotion (-). It is the final outcome that counts.

We do not realize it, but in every ambiguous situation, our brain probes the reactivated state of our body (throughout the vagus nerve) in order to influence, if not define the decisions that we will consider and/or make (by the way, I think this is also how familiarity feelings are implemented).

The accumulation of somatic markers ends up forming a formidable adaptationadaptation to complex life (and especially to social life) : a powerful shortlisting system capable of approximating the future consequences of this or that decision, and this, in an instant (nothing is explicit, everything is compressed as a “little” emotional influence). It is this system, for the most part, that allow us to detect “trap” choices, that is the selection of options that can lead to immediate attractive gains but also lead to future and greater losses. Our consciousness lives in the present, but not our intelligence.

Therefore, the failure of this system is much, much more incapacitating than what we would think. Indeed, cases of subjects no longer having access to this cognitive function shows seemingly good rational abilities but have a “myopia” for the future, they cannot foresee the future based on the past and so keep considering and selecting options based on their short-term appearance. These subjects forced researchers to admit that pure reason was clearly insufficient to face many situations of everyday life. Without the support of emotion, pure reason is completely overwhelmed and lead to many failures, even to “tragedies” (cf., subject E.V.R., Eslinger & Damasio, 1985 + start of the book “Descartes’ Error”).

It is actually very rare to decide a choice by a conscious assessing of the future outcomes of various options and alternatives through some type of cost-benefit analyses. And even when we do it, emotion still weigh in and is probably what makes the final decision. Explicitly assessing every possibility takes time and attentional resources. It is an anti-evolutionistic solution. Due to survival constraints, evolution favors fast, reliable and efficient systems (fight-or-flight response). And due to design constraints, evolution favors efficient, economical and easy implementations. Somatic markers provide implicit and instantaneous global viewing. It was the perfect hack grounding for a complex decision making ability. And without it, I think humanity would never have succeeded to reach such an advanced state of development.

But the somatic marker system (SMS) is not a generic rational decision making system, nor has evolved to be such a system. Why did the SMS was initially developed, what problem (from an evolutionary perspective) does it solve ? In my view, it solves the problem of memorizing (and knowing) the long-term value of things in regard to our own fitness (~ survival) in a complex environment. When environment interaction was still simple and direct (or of short duration), things just had a fixed valence, so it was not necessary to develop such a system. But as species evolved and environment interaction became complex, symbolic and indirect (or of long duration), valence alone became insufficient and not relevant to evaluate the actual value of things.

SMS is a system that solves this complex evaluation issue. And it solves it the “evolution way”, that is efficiently (as in fast and reliable) and economically (as in involving little information processing and implemented through a grounded design). The costly, anti-evolutionistic solution would require to store every experience and associated “stories” in details and act on this huge memory as some kind of expert system. It would require a lot effort, memory and attention. The functional approach of somatic markers allows to automatically and effortlessly guide the decision process, to take a weight off attention and to form meticulous adaptationadaptations to our habitual environment (activities, work, relation management, etc) once sufficiently developed.

As it is, the SMS is adapted to select decisions within our scope and feelings though. It is of no use to distinguish options of choices from which we cannot get any emotional feedback. That is, if an individual has two working options “A” and “B” to solve a problem, but if option “A” has a lot more future negative consequences on things that does not directly and causally impact this individual than option “B”, the somatic marker for selecting option “A” (the bad one) will nonetheless be the same as the one for selecting option “B” (the good one). The SMS can not help differentiate (and bias) these options. SMS provides an implicit global viewing that is egocentric. Not allocentric.

Modern man lives in decentralized societies, organized in “abstract” structures. In such an environment, the egocentric body grounding of decision making is probably obsolete and problematic. For example, purchasing almost anything in our societies has an indirect global impact, disconnected from the body.

I do believe that many problems faced by human societies (astronomic global debts, bad management of natural resources and world-level priorities, profit culture) are indicators that high-level decision making still crucially relies on somatic markers, and thus are indicators of the inadaptationadaptation of this system to the current environment. Normal functioning humans are like brain-damaged subjects when it comes to the global future : they have a myopia for the global future. The somatic markers system was developed for a world in which organisms act on and change their life but do not act on and change the world. And yet this is what we do now.

SMS explains why we struggle to actually make some decisions that we know are (rationally) good decisions. For example, it explains why people that smoke struggle to quit smoking. People that smoke know than smoking is causally linked to health problems. Figures demonstrates it. But until it is their feelings that demonstrate it, until they actually feel health problems (than can not be explained by something other than smoking), they probably will not quit smoking. Because on the other hand, smoking actually makes them feel good. The somatic marker for smoking is a positive one that will slowly turns into a negative one (or rather a mixed one) as associated health problems become perceptible. It is really hard for us to act on things that are directly quite good and insidiously very bad, even though we know they are insidiously very bad.

We do not really feel affected by (health) problems that do exist until they actually happen to us, until they actually damage our integrity. Thus, we do not weight them as we should.

Studying mathematics might make a child feel bad even though he knows that it will make his future better. The SMS is really bad for decisions that have consequences we will not feel until a long period of time.

It is good system when the impact of something is distributed like this : +, +, +, – – -, +, +, – – –

It is bad when it is distributed like this : +, +, +, +, +, +, +, +, +, +, …, +, – – – – – – – – – – –


And yet, I think the SMS system is responsible for humanity biggest developments, I think that humanity would never have succeeded to reach such an advanced state of development without it. Because working on a project makes individuals build-up implicit decision knowledge and thus progress in the development of solutions. Some problems are virtually impossible to solve using only our innate abilities, their resolutions require too much “thinking” (organization, explicit memory, etc) at the same time. But as we work on a problem, test solutions, etc, we feel things and accumulate somatic markers that will bias our future work and simplify the resolution of the problem. We learn from our mistakes and failures. That way, if an individual works on a virtually unsolvable problem (for the common human) for a long time, the accumulation of somatic markers might create an adaptationadaptation to this problem that will sufficiently reduce its complexity to make it solvable. SMS might be the key component that allows individuals to manage large-scale projects during their life.

Our language ability, our working memory, our gigantic and abstract memory, our ability to conceive tools and usages, etc : those are certainly very good working tools, but I do not think they allow medium term development (let alone long-term development). I do not think that these tools are alone responsible for the realization of humanity biggest developments. To me it is really this instant and implicit (egocentric) global viewing and guidance system that allowed humanity to get that far through major individualistic developments. Too far for what we can virtually manage, in fact.

Those are consciousness driven abilities.

Et c’est bien là tout le problème ; la bonne idée a mal tourné. En effet, le fait que notre bon fonctionnement décisionnel repose sur l’ancrage corporel est selon moi l’une des principales raisons des « choix » totalement aberrants de l’humanitéhumanité (dettes globales, gestion des ressources naturelles, cf chapitre 23, 4 et 1). Le procédé n’est plus du tout adapté au monde moderne, dans lequel la moindre action a une portée qui dépasse le cadre corporel ; et pourtant c’est le seul dont on dispose !

Néanmoins – et comme le montre l’étude précédente – ce système n’est pas infaillible, il peut se faire piéger

Dettes globales, mauvaise gestion des ressources naturelles et des priorités mondiales, culture du profit => inadaptationadaptation des marqueurs somatiques à l’environnement actuel?

Then, what is wrong if this system linking somatic states with reasoning is that powerful ?

Like brain-damaged patients, we know it is bad, but can not help but do it because it is tempting (invasif word ? ).


I reckon that it is a bold and stretched assumption to say that the SMS is the key component that allowed artificial development. What makes me believe in this idea is that in some abstract situations our normal human behavior is similar to the typical behavior of brain damaged people, which are people that can not develop anything for themselves without failing at some point and are unable to correct their behavior (which show that our whole toolset minus SMS is just not enough and too limited).


Because our decision making ability is rooted and derives from an individualistic system, as individuals, agents make wrong decisions for their population. As a population, individuals engage into developments that are negative to population survival in the long run. That is because our abilities are still designed to fulfil the here and now (which a particular somewhere and sometime will never be) and thus are inadapted to handle global things (e.g. resources management, properly constructing complex things, …). One would think that our ability to choose rationally is somewhat independent from the here and now paradigm. But it is not.

In other words, we have no standalone and fully-fledged ability to choose rationally in a “somewhere and sometime” paradigm (that is to rationally evaluate abstract possibilities). Because thanks to hack associated usages of some traits, we managed to functionally do and build humanity without such a system.

Indeed, our rational thinking ability is in part derived from the usage of somatic markers. Somatic markers are part of an unconscious and probabilistic evaluation system that depends on body state. Which means it is not conscious. What emerged as a genius “idea” to boost development is revealing to be a bogus good “idea” as its usage expanded way too much. This is the main problem with abstract traits. They are so powerful range wise that they can be used for many roles, even though they arguably did not evolve for such roles and are not really efficient and flawless solutions to the associated problems.

Using it to evaluate and develop complex solutions.

We hijacked tools to do inadapted things with them.

We use tools that are inadapted to do things we do with them.

Evolutionary dead-end

Bottom-up adaptationadaptation VS top-down adaptationadaptation

In standard evolution theory, evolution is basically the result of progressive design adaptationadaptation to pressure. In natural hacking theory, design adaptationadaptation to selection pressure is just the starting force of evolution, which initiates the slow progressive transformation of species. But it is not what(solely) causes the biggest changes. Once the slow adaptative design evolution process is started, a new layer of changes emerges on top of it. And this is how big changes emerge. The biggest changes are not the result of progressive design adaptationadaptation to pressure, they are the result of pressure adaptationadaptation by hacks, they are the result of pressure transformation by life (~ usage).

Le gros de l’évolution ce n’est pas quand la sélection naturelle adapte progressivement les individus (pour qu’ils procréent toujours mieux !), c’est quand le fonctionnement des individus (à la manière de catastrophes naturelles) adapte brutalement la sélection naturelle.


Every hack is a transformation that creates a pressure modification which in turn creates a shift in evolution and lead to important changes. If a hack is an isolated leak that adds a concrete and predictable function to some organisms, pressure will naturally adapt back and integrate this new “parameter”and eventually return to a “normal” state. But if it is not isolated, as is the case of every usage of our abstract cognitive abilities, pressure will notbe able to adapt back and integrate anything,nor have the time to do so. Our advanced cognitive abilities are abstract abilities: they are not adapted to anything in particular. They can not be (fully) integrated because their usage can not be (fully) predicted. And we use them everyday in unpredictable manners. So our whole lives are filled with hacks, our lives shape evolution.This is a great responsibility for a species to have such a power on pressure. For I believe transforming pressure can be unadaptative and lead to evolutionary suicide (especially when individuals are not emancipated from old values, see following section). Because pressure transformation prevents some changes to naturally happen, it stops the evolution of some traits. Paradoxically, it is the environment that evolves and develops, becoming more and more complex, as if functional evolution moved from biological evolution to artificial evolution. The slow biological tools (~ traits) evolution has been replaced by a faster external artificial tools evolution. And this artificial evolution is self-handled by the human species. Innate biological traits therefore become more and more under-qualified as the environment continues to develop.

In such situations, functional fitness is arguably more relevant that genetic fitness. Complex societies evolution makes it clear that it is functions that evolve and not genes or individuals (see previous section “An abstract and functional-based approach to reproduction”). Because individuals do not biologically evolve (that much) whereas the environment evolves quickly. Humanity does not require any additional hard-coded functions to create functions, functions does not have to rely on genetic embodiments any longer to emerge and reproduce through human beings. Evolution is clearly beyond genes and individuals in such environments. Sure individuals than can acquire and understand advanced skills and be part of artificial evolution have more chance of reproducing because they are vectors of functions ; but high-level functional creation is clearly not inheritable (at least not reliably inheritable). Great inventors are not necessarily the sons of other great inventors. It is the created functions (inventions, methods, theories, …) that survive and spread. (It is interesting to note that individuals devote their life to their baby product. Standard evolution theory would say it’s because fame can lead to a good position).

Self-guidance or drives exploit-guidance ?

We could think that humanity took control of its “destiny” and created its own ecosystem, out of evolutionary guidance. But this is not really the case. We are still driven andenslaved” by historical evolutionary values/drives. Only our means, because they allow so much abstractionabstraction, are not. Having such power on pressure without being totally abstracted from evolutionary “drives” is dangerous (~ evolutionary suicide). To me, humanity development is largely emergent and uncontrolled and reflects the exploitation and exacerbation of drives that typically happens when individuals that initially evolved to have an auto-organized life get abstract controls over the the satisfaction of these drives. In a way, humans are more guided by their needs than other animals because they have unlimited ways of satisfying them.

Steering evolution by taking control of Natural Hacking

Still, we can’t deny that we humans have some kind of control. We can understand evolution. We can understand the pitfalls and opportunities of Natural Hacking. Thus we can make us responsible for avoiding these pitfalls, and responsible for finding ways to « hack » nature so as to shape evolution in a good way.

In the end, we can be either victim or master of Natural Hacking.

Did you find this article interesting ? If so then you can share it with your acquaintances by clicking the buttons below :

Twitter Facebook Google Plus Linkedin email

Commentaires fermés.

Laisser un commentaire