A creation myth is a narrative mnemonic device which ties together reasonably well proven facts, earnestly held beliefs, and pure flights of fancy. If it adequately conveys the crucial facts needed to survive and prosper, and does so without serious drawbacks, then it serves its purpose. Such myths come in the form of fables, children's stories, religions, just-so stories, etc. While they serve to tie facts together and to fill in gaps, myths may of course leave some questions unanswered.
This particular creation myth is my own personal rendering, though it obviously draws on a wide range of sources. If you detect technical flaws ("scientific facts"), let me know. If you have a better (more pleasing, more plausible) narrative covering the same territory, let me know. If you just have your own narrative/myth and want to tell me I am wrong (immoral, infidel, etc.), don't bother -- publish your own page and leave me in peace.
To motivate the discussion, consider this event:
Once upon a time, a child sat watching a butterfly on a flower. The child reached out for the butterfly, but it flew away. The child turned to her parents and asked "Why did it fly away? I wasn't going to hurt it. I just wanted to see why its wings were so pretty."
Now consider its implications:
We know roughly how the universe, life, thought, and culture evolve. We don't know why (prime cause). Dozens of religions claim to know why, but frankly folks, they can't all be right -- and we must therefore assume they are all wrong. [To a first approximation, we can assume they are well-intentioned but deluded. However, any religious believer who attempts by force or intimidation to "convert" someone else must be considered dangerous. Anyone who panders to such zealots to curry political favor is deliberately wreaking havoc on the body politic. In this light, the Republican "faith-based" funding adventures are nothing short of treason.]
We do know that the resultant Reality is the instantaneous state of on-going processes. It is modeled by thinking creatures, who use those models (imperfect as they may be) to guide future action. On average, invalid models are detrimental to model users. The basic mechanism for recognizing a valid model is the Ah-Ha or Truth sensation.
Humans are obligate model builders and communicators. They cannot survive without complex models developed, improved, and conveyed from one generation to the next in the form of culture and society . Each human child relies on the Ah-Ha sensation to determine which parts of his/her native culture to accept and which to reject. The problem is that the sensation can be tricked.
Science uses repeatable processes to generate the sensation. Art uses tricks to generate the sensation, but acknowledges doing so. Religion uses tricks to generate the sensation, but doesn't realize it does so. Societal parasites use tricks to generate the sensation, do so for personal gain, do not acknowledge doing so, and pretend to be Science, Religion, or Art to camouflage their actions.
It is up to each generation to ferret out invalid models, to expose them to public scrutiny, and to purge them from the communal legacy passed to the next generation. This may include detecting and eliminating parasites.
Notice that the fundamental laws of nature probably haven't changed, nor has the instantaneous Reality reflecting those laws, but the community's perception of the Reality may change in a few years or a few generations.
Once there was nothing. It exploded and the debris flew in all directions (with some non-random variation). As it went the debris froze into more and more complex forms. Energy condensed into particles, particles condensed into atoms, and atoms begat molecules. Stars formed, burned, exploded, and reformed. Planets formed from the interstellar debris.
All this happened according to one or a few fundamental laws. The laws do not dictate specific amounts of energy, particles, stars, etc. Rather, they describe transitions, so that the outcome is the byproduct of differing rates of reactions. Thus the process is the reality.
On at least one planet near a non-descript star at the outskirts of a modest galaxy, the debris recombined in such a way that it could and did regenerate copies of itself. The same fundamental laws applied, but they now led to more and more complex systems. We call these systems life .
Why is copying (reproduction) so crucial to increasing complexity? A reproductive process assures that a useful new system feature is not lost when the particular specimen disintegrates. These features increase the probability of a specimen arising, and increase the duration of its existence. We are again looking at differences in transition rates, with specimens as the byproduct. However, the transition rates are far more favorably for complex life-systems than for complex nonlife-systems. A butterfly is absurdly improbably if it must be developed by the fundamental laws without a copying mechanism.
But how do such features arise in the first place? Once the copying process begins, there is of course a chance of error. Nearly all errors are fatal. That is, the resultant specimen will disintegrate quickly. A few errors will be as good or better than the original. If there is stress which disintegrates many of the currently extant specimens, any repopulation will be with the fitter systems which survive the stress. Copying, plus errors, plus stress.
If one of the stresses is competition with similar systems, fitness may simply require being different. You eat grass on that side of the hill, and I will eat on this side. Such stresses lead to development of diversity. There are many types of butterflies, each living a slightly different life.
Note that it is the gene pool which evolves, not an individual specimen. A system feature may be more fit even if it is worse for the particular individual, so long as it helps the gene pool.
Here we mean the genetic ("nature") component of behavior, or ethology. For humans, there is a significant component, though it is overlain with learned ("nurture") components.
These genetic components include for example:
Large adult male primates (including humans) tend to be dominant -- getting the best food, the best mates, etc. This can be observed in statistical analysis of height of U.S. corporate leaders vs the labor force. It is codified in a scheme such as the Hawaiian royals vs the menehuni. Until the advent of the handgun, it was based on practical realities of expected-outcomes.
Dominance via size may have valid evolutionary basis in fixed gene pools, but goes awry under migration. Then we might see, let's say, a mediocre yet physically large Nordic male dominating an outstanding yet short-statured Asian.
Dominance via size isn't the whole story. Among chimps, a smaller male may learn to crash kerosene tins about, making enough racket to establish dominance. Among humans, these kerosene tins may take the form of boom boxes, fast cars, big bank accounts, guns, and assorted other "civilized" male accessories.
The crucial lesson here is that LEO's (Law Enforcement Officers) tend to be chosen for their physically dominating presence, and (via their "street degree") they learn to use it. If a protester refuses to accede to that primal demand for control, then the LEO experiences a natural and predictable temptation to escalate to violence, including use of lethal and non-lethal weapons.
There is nothing new or suprising about this. We must therefore assume any politician or community leader who places police in a position where their dominance will be legally challenged, does so with malice of forethought. That is one reason to look behind the scenes at any "police riot". As always, "follow the money".
Lifeforms evolved mechanisms to:
Given these tools the lifeform could flee rapid motions, seek water or food, detect and avoid poisons, etc. All animals and some plants have these tools. The actions they allow are called reflexes .
The fitness payoff for reflexes is the ability to react to the environment. In most cases it pays to react early. Given a reflex which reacts to a rapid motion, how could a reflex evolve which reacted before the rapid motion? It can be done by building an internal model of the environment, passing current stimuli through the model, and using the model's output to trigger already existing reflexes. Thus the current stimuli of a crouching lion produces the internal image of an attacking lion, and thus triggers an attempt to flee.
There can also be payoff for filtering the reflex outputs through models. Instead of instantly drinking the nearest water, pass the desire through a model which checks for signs of poisons. Don't drink if the shores of the pool are littered with dead animals.
Notice that the primitive, genetically inherited reflex is still there. It is just supplemented with models. Notice also that the models can be changed much more rapidly than the reflex. Instead of waiting generations for a new genetic mutation, the very next generation (perhaps even this one) can use a new model. This is a fitness payoff which more than offsets the costs of supporting the modeling mechanisms.
The fitness generally goes up as the models become more sophisticated (i.e., better simulations of reality). Increasing model complexity requires increasing mechanism complexity, leading to sophisticated brains. Brains are mazes of biochemical connections, apparently evolved in fits and starts to deal with specific sensors and controls. E.g., to control breathing, sexual response, vision, hearing, touch, taste, smell, motor control, etc.
Somewhere in that continuum of mental complexity, the specimen attempts to model itself. Somehow this is associated with consciousness, being aware of self.
This is expensive if each specimen must develop its own models without aid. Any system features which improve the process might pay off. Such features might be costly to individuals (e.g., teachers) but still benefit the overall gene pool.
One system feature which enhances learning is mimicking. The young observe the adults and try to perform the same tasks. For this to work, the young generally must bond to some mature exemplar. Apparently this bonding is largely genetically determined though it can be tricked.
The learning process is enhanced if there is also a teaching process. The teacher arranges a sequence of experiences which facilitate model-building in the young. E.g., a lioness demonstrates stalking increasingly difficult prey. There is additional evolutionary cost for this behavior. There is no payback for performing it for young outside the gene pool, so there is some payoff for bonding to specific young. Again, this is apparently genetically defined behavior with the potential for being tricked.
Some models are useful only when shared. There is a diffuse payoff for teaching lion cubs to hunt individually. But there is a dramatic payoff for teaching them to hunt as a pack. For this to work, all players need to absorb either the same model (e.g., peer-to-peer cooperation) or compatible models (e.g., landlord/tenant). We can call these shared, learned models culture .
Multi-organism coordination based on shared, learned models should be distinguished from that based on genetic inheritance. The results may be similar, but the mechanisms are different. Because the mechanism is non-genetic, learned models can evolve much more rapidly. Such models are subject to the standard evolutionary drivers (copy errors, testing under stress, proliferation of successful variants) but do so in years rather than generations.
A gene pool which relies on teaching shared models one by one pays a high cost. If the only means of transmitting a model is to arrange experiences, it takes a long time to convey complex models. Or, more to the point, the complex models never evolve. A teaching shortcut could have a big payoff.
One approach is inheritable codes. For humans, a smile is an inherited code meaning roughly "All is well". However, inherited codes are inadequate for ad hoc communication of arbitrarily complex models.
Another approach is to build a shared model of abstract symbols which can be manipulated to convey any idea as needed. Language and its derivatives are such shared models.
Apparently only humans, other great apes, and possibly dolphins and whales have the genetic tools necessary for learning language. Apparently, only humans learn abstract language routinely. The young must learn language itself the old way (mimicking, arranged experiences).
There is nothing magical about language. The student must still build models from his/her experience. The experience just happens to include sounds (or texts) generated by a teacher. Language only works if the student and teacher have a baseline of direct hands-on eye-witness experiences to use as submodels. Field trips, summer camps, grand tours, apprenticeships, etc. formalize the hands-on process. Student and teacher can then refer to these snippets of experience in discussing new topics.
Notice that paradigms laid down in the field-trip phase structure subsequent abstract discussions. This is the basis for kings insisting that the sons of their nobility live at the king's castle. It is the basis for Rhodes scholarships and more generally for prestigious scholarships for those who might otherwise grow up to oppose the established order.
Vocal teaching/learning does not take as long as arranging experiences, but it is still costly. There would be a payoff if the teaching could be captured once, for all students to use. Writing is such a mechanism.
We have seen model building to be the crux of culture, and that models can be conveyed by writing. The obvious next step is to write down a model of models. However:
Unanswered Questions :
We will sidestep these questions by claiming only to provide an adequately useful model of models.
Model building seems to be a process of abstraction. We are bombarded with sensations and events. Our first line of defense weighs these and renders abstracted signals. The next line weighs a collection of such signals and abstracts once again.
Individual neurons work this way. Vision apparently works this way. Considering that evolution typically works by ramifying existing mechanisms instead of creating anew, it is a fair guess that the rest of the model building process works in this manner.
Intuitive experience tends to support this guess. When we enter a strange room, there is a vague sense of confusion. We manage to stand on the floor and usually avoid tripping over the larger objects, but our visual and aural fields are jumbled. Within seconds, we begin perceiving things. That is a lamp, That is a whirring fan. There is a painting of a sailboat. Those seconds where spent building models.
Given a mechanism for model building, why do we do it? Why not walk into a strange room and stand there dumbfounded? Apparently we have a genetically transmitted seek-and-model behavior. We can call it curiosity.
How do we know a model is complete? How do we know it is time to dissemble one model and build another?
There is no way for an evolved lifeform to know some absolute, metaphysical TRUTH. The best we can do is build a model and use it until it leads to surprises. For example, my model may be:
Only couching lions attack
If a sitting lion suddenly attacks (and I escape) I will certainly revise that model.
Note that surprises may be at various levels of abstraction. I have a model of a telephone. If I find one while digging in the garden I will be surprised. I will recognize it as a telephone, but I will be surprised because it does not fit my model of gardens. At a different level, if I find a telephone with Roman rather than Arabic numerals, I will be surprised no matter where I find it.
Apparently, models are generated willy-nilly ("brainstorming"). Then they are subjected to internal testing. That is, we pass the current store of experiences through the model to see if it correctly predicts the known outcomes. If there is a surprise, we frown, wrinkle our brows, and try again.
Sometimes a model is so efficient and effective that all available experiences whiz past. This exhilarating experience is known as the Ah-Ha experience. This is a visceral sensation which is generally indicative of an exceptionally good model. To a lesser extent, it is the sensation felt whenever a model passes its tests. The sensation is most profound when:
It is tempting to think that this sensation is an isolated phenomenon, reserved for physicists shouting "Eureka!". No so. It is the fundamental process by which we recognize true and false.
For example, I may go through an intricate mathematical proof to decide the truth of a proposition. How secure is the result? At each step I knew I was right or wrong by the sensation of truth. At the end I cannot remember all the steps but I do have a sensation of truth when I state "I have done all the steps needed, and done them all correctly."
Or I may be choosing among wines on the menu. I mull over the meal, the wine list, the palates of my guests, and eventually make a selection. My trust is placed in that gut feel that I made the right choice -- the truth sensation.
It is of course possible to make mistakes. A mistake is a submodel which was right at the time it was made (it gave the truth sensation), but upon reexamination (a sitting lion did attack) it no longer provides the truth sensation. Sometimes this is recognized by re-running current experiences past the model and noting a surprise that was missed the first time. More often it is done because the world provides a surprise gratis.
Science is a collection of carefully tested models. Some of the models describe the world, others describe effective testing procedures. The act of scientific research is the process of using some of the models to test aspects of others. It differs only in degree from normal seek-and-model behavior.
Perhaps the key is the testing method, otherwise known as the scientific method. This model says:
The testing model is a fairly recent cultural invention. It has been so successful that nearly all we know with any certainly has been discovered or at least verified in the last couple of centuries.
Before looking that the testing model, notice what it does not do:
The testing model's requirements are:
A scientific model must make testable predictions. Other types of models may or may not be accurate, but science has nothing to say about them. Following Occam's Razor, the standard approach is to simply ignore such models as a waste of time.
A test consists of comparing the model's predictions to reality. First provide an input to the model and note its prediction. Then provide the same input to the real world and note the result. The model passes if we can say with a sense of truth that:
The prediction and model agree too closely to be just by chance
This awkward phrasing (the null hypothesis) is required because what we want are not just models which might be right, but models which can't be wrong.
To properly note the results of the reality check, all but the test input must be held constant.
Technically this is difficult to do. Much of scientific methodology is an attempt to control (or account for) non-test inputs. Still, we may miss some, known as random factors. We run a test several times in hopes that such factors will average out.
A test must be repeatable by other investigators. This is a crucial requirement. It effectively rules out the inner world of awareness. That is not to say accurate models cannot be developed in that world. It just means the scientific method does not apply.
A model is true if it is the simplest available model which passes all the tests competent researchers can devise.
Several models may pass the tests. We choose the simplest as a matter of economy, because it is typically more art-ful, and because history has generally shown that simple models are more robust. That is, they last longer before surprises destroy their credibility.
Note that a model must pass all the tests. Many (infinitely many?) models are almost right. But one verified failure brings them tumbling down. Of course, if you think you have disproven conservation of energy in a shop project, it makes sense to reexamine the test before rejecting the conservation model.
Note also that the tests must be by competent researchers. It takes training and skill to devise and run significant tests.
A scientific model can be wrong. When and how do we dismantle it?
A scientific model is a cultural invention which is generated, tested, used, and conveyed by fallible humans. Unlike many such inventions, it happens to have been screened rather carefully before passing into the public domain. But it could still be wrong. It could make predictions which are not borne out by reality.
There are two ways this can happen. First, the universe could change. There is no guarantee it will not or does not. Possibly it is fickle. A model that works today may not work tomorrow. Fortunately this does not seem to be the case.
NOTE: Religious groups sometimes claim to know of instances when the universe changed its rules. These moments are know as miracles. By definition they are not repeatable, and by examination, they seem to never be verifiable (e.g., forensic analysis of the scene). It is more economical to accept these stories as well-intentioned parables than to believe the universe is fickle. The stories may well be true, but without repeatability we cannot use them and thus ought not convey them in our cultural legacy.
The second and more probable case is that we did not devise a crucial test. Test generation is a creative act -- there is no way to be sure we have covered all possibilities. So if our tools become more powerful or a new collection of experiences becomes available, we may be using a model which no longer applies to the new context. When this happens, we need to generate and test new models until we find a good one, and then rebuild the house of cards.
Of course, rebuilding is hard work. Experienced researchers have invested enormous effort in absorbing the old models. Young rebels with crazy ideas do not get grants. No-one else cares. It is a credit to the intellectual integrity of modern scientists that fundamental changes are ever made. In the past problems or surprises would have been hidden or disguised. (E.g., Pythagoreans and irrational numbers.)
It appears that a Kuhnian revolution is required to make major changes in scientific truth: The troubling data points keep bubbling up, oddball hypotheses are offered, gradually a school of thought develops which better handles the full set of data, and this school eats the turf of the old school.
Math is not a science. It is a language designed to reduce ambiguities. Once ambiguities have been reduced, all competent investigators can agree or disagree on precisely the same points. Once they can do that, the scientific method's requirement for repeatability can be met.
The design of the language itself is also sharply improved by this reduction in ambiguity. With few errors in transmission, quite complex models can be built out of carefully tested submodels.
The derivations are models. They can be laid out with great care, but in the final analysis it is the truth sensation which gives the go/no-go decision. It might seem that math is unique in that its "experiences" are all abstract. Not so. Just like any other model building exercise, it is subject to mistakes (it was right until shown to be wrong).
It is possible, though, that math deserves some credit as a psychological model. It is fundamentally a statement of how to proceed in manipulating models. It may therefore be an implicit model of how our modeling processes operate.
The evolutionary fitness payoff of the truth sensation is that we thereby recognize good models. We enjoy and seek that sensation.
But given that such a sensation exists and can be triggered by certain types of models, we are not above cheating. Why bother making a useful, predictive model which happens to give a strong truth sensation? That is hard to do. Why not instead concentrate on models which trigger the sensation directly. Such models are art.
The conditions for truth sensation are met:
The artist's task is not simply to generate and transmit a good model. The first step is to generate a good model. The next step is to find a way for the viewer to seek-and-model it. The artist directs the viewer through a series of experiences which result in the building and rapid testing of the model. Establishing the base of experiences can be via:
Start with lifeforms which build models. Given them a truth sensation which operates when a model is not contradicted by experience. [This is not the same as saying it operates when the model is confirmed by experience.] Then make the lifeforms so genetically helpless that they must rely on one another and on learned, shared, transmitted models to survive.
The result will be culture. A collection of testable and untestable models. Some of these models describe the world; others prescribe rules of behavior. Some have survival value; some do not. As a whole, the collection of models has survival value. It is taught to all children in a process known as acculturation. It is the basis for coordinated behavior.
Descriptive models explain the how and why of experience. Some of these models are carefully tested scientific models. Others are guesses or flights of fancy.
Given time and better tools, many of the guesses can come under the scope of scientific testing. Cosmology, psychology, and archaeology all began in the realm of religion. Aspects of each field remain there. For example, it is a matter of faith rather than proven fact that the rules of the universe have not changed since the Big Bang.
On the other hand, given time and the best tools we can conceive, some descriptive models would remain untestable. They appear to be flights of fancy. Religious gods are prime examples. What test could we possibly make to prove that the Christian trinity, the Hindu pantheon, the Hellenic gods of Olympus, or the animist spirits of North America are correct or incorrect? Any effect they have on the testable world can be traced to the act of believing, not to the model itself.
The act of believing is no small matter. Since the models per se have no effect, it does not directly cost to believe them (ignoring Occam's Razor). It is testably true that such beliefs can calm people, and that calm people may handle crises more successfully than agitated and confused people. Of course, religions can calm people into accepting tyrannies. As the Polyneasians say about the Christian missionaries, "They asked us to kneel and pray. As we knelt, they preyed."
There are also artistic reasons to believe the untestable. Such beliefs can tie together bits and pieces of the cultural collection into a unified whole, making it easier to remember and convey. And such models can "explain" the unexplainable, resolving the sense of unease we have whenever we cannot build Ah-Ha worthy models for our experiences:
Such beliefs may not be rigorously provable now or in the future, but they can still provide the truth sensation. We can pass all known experiences through these models and not once find a failed prediction.
The only problem is that we can find people who believe just as fervently that something else (diametrically opposed) is true. In other words, the crucial test of any religion model is that it produce a plethora of such models, with intense beliefs immune to disproof.
Prescriptive models (taboos, commandments, superstitions, etc.) are essentially predictions that the society will collapse if the members fail to follow certain rules. These models are not only difficult to test, but the very act of testing them may be more costly than erroneously following unnecessary rules.
Actually prescriptive models are often partially tested. Human societies have collapsed at various times for various reasons. Local unpleasantries (murder, theft, etc.) are common. Rules such as "Do not kill" or "Do not steal" have been more or less tested in the school of hard knocks. Even seemingly absurd rules ("Thou shalt wear designer jeans") may serve to identify members of a given culture.
One potential problem with cultures is that they may be out of synch with physical reality, including genetic inheritance. For example, primates (including humans) recognize dominance hierarchies. These are usually based on physical size and ability to bluff. But as Jane Goodall noted, low-level chimps can use tools such as kerosene tins to make enough racket that they can move to the top of the hierarchy. Humans take this to interesting extremes. Our kerosene tins include weapons, loud music, large bank accounts, large cars, and intellectual prowess. The problems are that:
A troubling phenomena is arising among those considered nerds and geeks, who traditionally were low in the dominance hierarchy. In this internet age, they are finding they are in some ways more powerful than physically larger men, and are finding this at a young enough age to impact personal image. Yet without societal acceptance of this dominance (as in, the girls still want the jocks), they remain in limbo. The tension escalates beyond bluffing to the kill (literally and figuratively). This may help explain both Columbine-like attacks on jocks and Microsoft's exterminate-all-competitors corporate mentality.
We now have a vocabulary for discussing parasitism. In both the biological and cultural context, the host depends on certain system features. E.g., a steady supply of oxygen and nutrient rich blood. The parasite uses that feature, harming the host but (usually) not killing it outright. E.g., living off the blood supply.
The host may develop detection and containment responses, but cannot completely eliminate the parasite without destroying itself. There is an arms race. White blood cells attack the parasite, parasite exudes toxins, host halts before parasite can be killed.
This same arms race occurs in societies. Examples:
Having concluded it is possible and even likely there will be social parasites, how do we detect them and eliminate them? That is the world of politics
Creator: Harry George