A Replicator

That

Generates, Selects And Copies

Traits 

Is Intelligent

By Kevin Aylward B.Sc.


Index

Definitions

Overview

It is argued here, that a Replicator has evolved, as a replicator that generates, selects and copies traits, such that such traits maximize that Replicators numbers, is an Intelligent Replicator.

That is, there exists Darwinian Machine machines.

Of course, there may be more to intelligence than that stated above, so it is not intended to exclude any other aspects of intelligence. The point here is to address an aspect of  intelligence that may be considered a major factor in explaining many aspects of human behavior. Aspects such as conscious awareness are not addressed.

Trait Generators, Trait Selectors and Trait Replicators

The environment of Replicators can be very arduous. Environmental conditions can change so quickly that Replicators with fixed, physical hardware based traits can not adapt and survive in such changed environments. That is, the same action, in different circumstances, could result in either negative or positive replication consequences. In addition, a Replicator might not have the required traits to enable it to survive in a given environment. It would obviously be advantages to have a Replicator that could change its traits as required in order to match its environment.  Combined trait, generator-selector-replicators, as described in Special Replicators, are the obvious solution to this problem.

It would appear that as a natural phenomena, the basic axioms of General Replicator Theory, that is, selection, variation and replication, is all we have. That's your lot. So, we are pretty much forced to the conclusion that however the brain works in detail, it is fundamentally limited by the laws of physics to physical entities that can only be varied/generated, selected and replicated. It is a system such that its internal properties satisfy the same axioms as its external properties.

Intelligence 

Consider a Replicator that has evolved a combined Trait Generator, Trait Selector and Trait Replicator, as described in Special Replicators. This Replicator would have the ability to generate traits,  recognize traits that could aid its replications, and the ability to copy such traits, and hence use those traits for its own replication advantage. 

It is therefore held, that a fundamental function of intelligence is as a Replicator that has "trait generator" "trait selector", and "trait replicator" features that act to increase its replication numbers. That is:

An Intelligent Replicator is a Replicator that generates, selects and copies its traits.

That is, the Intelligent Replicator, is a Darwinian Machine.

By assumption, it is highlighted that such an Intelligent Replicator can generate new traits simply by varying exiting meme traits.

It is a matter of simple observation that intelligent individuals do indeed generate, select and copy traits. It is now held that the fundamental reason for this action, is an attempt to maximize the numbers of an individual's Replicators by utilizing such generate, select and copy traits. However, whether it actually does so in practice, is another matter.

It is also noted that Darwinian Machine "genetic algorithms" simulations have been performed that show how powerful such variation, replication, and select process can be in solving intrinsically hard problems e.g. the traveling salesman problem.

It should be noted that an alternative view that the brain functions as a Turing Machine is fundamentally flawed and is rejected.

It is therefore held that the human Brain is, essentially, a Darwinian Machine.

Comment - typical meme traits that aid replication would be, making tools, planting crops, cooking food, selecting stronger mates etc...

Meme Machine

In order to obtain the optimum traits for replication, the Intelligent Replicator will select for both memes and genes. To do this successfully requires an "intelligent" Intelligent Replicator. That is, one that generate, selects and copy traits that have a high rate of replication. Clearly, the ability to select another Intelligent Replicator to mate with, that is also "intelligent", will result in offspring similarly endowed, since such offspring would be better able to replicate due to such intelligence. Selection of this "intelligence" trait will instigate an evolutionary drive to increase the "intelligence" of  Intelligent Replicators to that limited by the physical constraints of the environment. That is, although the genes initiated intelligence, once initiated, the advantage of being a Intelligent Replicator was such that it led to the conscious selection of better genes to further increases the abilities of such intelligence, essentially forming a positive feedback loop. In this way, it may be said that the "intelligent" memes, are selecting for their own genes.

A point here, is that once meme trait generate-replicators-selectors appear i.e. a meme machine, they will do anything that is required to increase their numbers, in accordance with the General Theory of Replicators. That is an appearance of a Intelligent Replicator, necessarily forms an evolutionary drive for better Intelligent Replicators. These Intelligent Replicators instigate selection of genes that will better replicate their "copy me" meme, not the genes themselves. However, in the long term, any Intelligent Replicator that had memes that eliminated its own physical embodiment, would not be a happy camper. 

The essential feature of the Meme-Gene Machine is the ability to adapt to all environments and is addressed in environmental adaptability.

Note: There is no assumption here that all of the generate/select/replicate features are required for all aspects of intelligent behavior. There may be some Replicators with only some of these features.

Replicator Physical Storage

By assumption, there is no magic. A Replicator that has select and copy traits must therefore have a physical embodiment in which such traits can be stored. For simplicity, it is assumed that such storage is on a similar basis as to that in computers, that is random access memory, or RAM. However, it is not suggested that this is the exact physical construct that such Replicators utilize, and this is not a necessary assumption. All that matters is that copied traits must be stored and accessed in some manner in/from a real physical medium. It is a trivially obvious assumption that the volume of such storage must increase with the number of traits able to be stored. Therefore, a Replicator must increases in volume as its intelligence increases.

The argument presented here, is direct support for the view suggested by Dr. Susan Blackmore1 that the size of the brain is increased due to meme-gene evolution, and why it is suggested as the basis as to why human brain size is as large as it is.

Darwinian Purpose

Consider a Darwinian Machine processing data, that is, generating trait outputs, from trait inputs, by its, initially, random trait selector algorithm. These outputs will be selected for by the environment such that, after sufficient Replication generations of the Darwinian machine, the mostly observed overall traits of surviving Darwinian Machines, will be such that whatever selection algorithm, controlled by such traits, maximizes it numbers. That is, the Darwinian Machines selection algorithms themselves will be globally selected by the environment such that they will be optimized to maximize the numbers of its Replicator. In this way, a blindly evolved Darwinian Machine may be said to have a goal or purpose, even though no conscious awareness is implied by recognition of such a purpose. In addition, clearly, if an intermediate goal can be identified, that ultimately acts to maximize numbers, than this intermediate goal may well also evolve. In this sense a goal or purpose may be defined as:

Definition of Darwinian Purpose - generating a specified trait output from arbitrary inputs.  

It is now held that a significant level of evolution is directed and implemented by such already evolved sub systems. That is, irrespective of any notion of conscious awareness, much of evolution is "designed" by "purpose" directed sub systems, as defined above. That is larger system blocks are developed from blocks that already have the essential basic properties that allow the Replicator to be maximized. 

It should be noted that such directed evolution is equivalent to the field of Memetic Algorithms, of which research has indicated a speed up of the order of 1000 over simple Genetic Algorithms in problem solving.

Darwinian Irrationality

It is held that the Darwinian Machine evolves to solve the maximization problem of its own interests. However, this machine can not, in general, have all the information that it requires to solve such a maximization problem. It can not therefore always take the appropriate action, in a given environment to maximize its own interests. It has no choice therefore than to do nothing, or take its best "guess". Its best guess, being simply to generate a somewhat selected semi-random trait or action by it Darwinian random trait generator. Such a random action is manifested as the notion of irrationality.

Definition of Darwinian Irrationality - generation of a random trait that impedes its own interests. 

Electronic Engineer as a Darwinian Machine

An Electronic Engineer copies existing circuits into his brain. He than varies those circuits to obtain new ones. He than analyses them to select the good ones. It is held that no other fundamental process are required. An Electronic Engineer is thus a Darwinian Machine, as are all of us.

Comment - there was a time when this author had simply never examined what procedure he used to design new electronic circuits. The original notion held, was that there was something unexplainable going on. However, a deep examination of the facts, resulted in the inescapable conclusion that the above method formed the essentials of the process. For example, where did the knowledge of what a transistor itself come from? Where did the majority of the transistor configurations come from? The reality is that the bulk of any design is obtained from existing designs, that is simple copying (replicating) existing knowledge. This knowledge is then varied and compared with the desired results.

The Daft Designer and the Blind Watchmaker

Intelligent Creator advocates argue that, just as a watch is "surely" designed by a creator, so must humans be so "designed". The fundamental fallacy with this argument, is that they fail to understand that a watch design has indeed evolved from the same basic Darwinian process of variation, selection and replication as that proposed for human evolution.

Although we can not trace back the evolution of a human, we can do so for a watch. A finished watch is essentially a design copied from an existing design, with possibly some minor variations. The first mechanical clock was a very rudimentary state of affairs, such as a simple swinging pendulum. This first mechanical clock, clearly, can be explained without requiring to be intelligently designed by a creator. It was simply an observation that an elementary process such as a bob on string repeated its motion when swung. Simple variations on this concept were made to improve this basic clock. That is, trial and error. Variations that were tested and performed badly were rejected, better variations were kept. Thus, a continuous process of generation, select and replication was used such that the final article appeared to be intelligently designed. It wasn't. Its that simple.

In may also be noted for reference that it is well known that Edison tried 10,000's of variations before the "design" of a practical light bulb was achieved. 

Illusion of Creativity

Since, the brain is an Intelligent Replicator instructed by its genes and memes, all artistic, musicianship, inventiveness, etc must be the result of a random process. That is any new idea (trait), must be due to a randomly generated trait.

Support for the claim:

There are two options

1 A new idea is derivable from existing ideas - i.e. not intrinsically new

2 A new  idea is not derivable from existing ideas - i.e. intrinsically new

If "1" is correct for an idea, then by construction it is not a new idea, it is simply a logical consequence of existing traits.

If "2", the idea is new, that is not derivable from existing traits, then that idea must therefore be non predictable. However, a non predictable trait is, essentially, a random trait, by definition.

Thus, those of us that believe that having a "bright" idea is something to be intrinsically valued, actually value a coin toss as a "superior" way to go. Indeed, it is being able to derive unsuspected results from existing knowledge that would appear to have the greater worth. 

So, creativity is nothing more that the brains random trait generator feature of its Intelligent Replicator kicking in.

References

1Dr. Susan Blackmore, "The Meme Machine",  http://www.susanblackmore.co.uk/.


These papers may be freely copied only for non commercial use,

provided full credit is given to the author.

© Kevin Aylward 2004 - all rights reserved