Five years ago, Gregory Chaitin, a co-founder of the fascinating and mind-bending field of algorithmic information theory, offered a challenge:1
The honor of mathematics requires us to come up with a mathematical theory of evolution and either prove that Darwin was wrong or right!
In Introduction to Evolutionary Informatics2, co-authored by William A. Dembski, Winston Ewert, and myself, we answer Chaitin’s challenge in the negative: There exists no model successfully describing undirected Darwinian evolution. Period. By “model,” we mean definitive simulations or foundational mathematics required of a hard science.
We show that no meaningful information can arise from an evolutionary process unless that process is guided. Even when guided, the degree of evolution’s accomplishment is limited by the expertise of the guiding information source — a limit we call Basener’s ceiling. An evolutionary program whose goal is to master chess will never evolve further and offer investment advice.
Here I answer ten frequently posed questions about and objections to Introduction to Evolutionary Informatics.
1. Why yet another book dissing Darwinian evolution?
Solomon was right. “Of making many books there is no end, and much study wearies the body.”3 There are gobs of books written about evolution, pro and con. Many are excellent. So what’s so important about Introduction to Evolutionary Informatics? On the topic of evolution, the conclusion is in: There exists no model successfully describing undirected Darwinian evolution. Hard sciences are built on foundations of mathematics or definitive simulations. Examples include electromagnetics, Newtonian mechanics, geophysics, relativity, thermodynamics, quantum mechanics, optics, and many areas in biology. Those hoping to establish Darwinian evolution as a hard science with a model have either failed or inadvertently cheated. These models contain guidance mechanisms to land the airplane squarely on the target runway despite stochastic wind gusts. Not only can the guiding assistance be specifically identified in each proposed evolution model, its contribution to the success can be measured, in bits, as active information.
And, as covered in Introduction to Evolutionary Informatics, we suspect no model will ever exist to substantiate the claims of undirected Darwinian evolution.
2. But Darwinian evolution is so complicated, it can’t be modeled!
If this objection is true, we have reached the same conclusion by different paths: There exists no model successfully describing undirected Darwinian evolution.
3. You model evolution as a search. Evolution isn’t a search.
We echo Billy Joel: “We didn’t start the fire!” Models of Darwinian evolution, Avida and EV included, are searches with a fixed goal. For EV, the goal is finding specified nucleotide binding sites. Avida’s goal is to generate an EQU logic function. Other evolution models that we examine in Introduction to Evolutionary Informatics likewise seek a prespecified goal.
The evolution software Avida is of particular importance because Robert Pennock, one of the co-authors of the first paper describing Avida,4 gave testimony at the Darwin-affirming Kitzmiller et al. v. Dover Area School District bench trial. Pennock’s testimony contributed to Judge Jones’s ruling that teaching about intelligent design violates the establishment clause of the United States Constitution. Pennock testified, “In the [Avida computer program] system, we’re not simulating evolution. Evolution is actually happening.” If true, Avida and thus evolution are a guided search with a specified target bubbling over with active information supplied by the programmers.
The most celebrated attempt of an evolution model without a goal of which we’re aware is TIERRA. In an attempt to recreate something like the Cambrian explosion on a computer, the programmer created what was thought to be an information-rich environment where digital organisms would flourish and evolve. According to TIERRA’s ingenious creator, Thomas Ray, the project failed and was abandoned. There has to date been no success in open-ended evolution in the field of artificial life.5
Therefore, there exists no model successfully describing undirected Darwinian evolution.
4. You are not biologists. Why should anyone listen to you about evolution?
Leave aside that this question reeks of the genetic fallacy used in debate to steer conversation away from the topic at hand and down a rabbit trail of credential defense. The question is sincere, though, and deserves an answer. Besides, it lets me talk about myself.
The truth is that computer scientists and engineers know a lot about evolution and evolution models.
As we outline in Introduction to Evolutionary Informatics, proponents of Darwinian evolution became giddy about computers in the 1960s and 70s. Evolution was too slow to demonstrate in a wet lab, but thousands and more generations of evolution can be put in the bank when Darwinian evolution is simulated on a computer. Computer scientists and engineers soon realized that evolutionary search might assist in making computer-aided designs. In Introduction to Evolutionary Informatics, we describe how NASA engineers used guided evolutionary programs to design antennas resembling bent paper clips that today are floating and functioning in outer space.
Here’s my personal background. I first became interested in evolutionary computation late last century when I served as editor-in-chief of the IEEE6 Transactions on Neural Networks.7 I invited top researchers in the field, David Fogel and his father Larry Fogel, to be the guest editors of a special issue of my journal dedicated to evolutionary computing.8 The issue was published in January 1994 and led to David founding the IEEE Transactions on Evolutionary Computing9 which today is the top engineering/computer science journal dedicated to the topic.
My first conference paper using evolutionary computing was published a year later10 and my first journal publication on evolutionary computation was in 1999.11 That was then. More recently my work, funded by the Office of Naval Research, involves simulated evolution of swarm dynamics motivated by the remarkable self-organizing behavior of social insects. Some of the results were excitingly unexpected12 including individual member suicidal sacrifice to extend the overall lifetime of the swarm.13 Evolving digital swarms is intriguing and we have a whole web site devoted to the topic.14
So I have been playing in the evolutionary sandbox for a long time and have dirt under my fingernails to prove it.
But is it biology? In reviewing our book for the American Scientific Affiliation (ASA), my friend Randy Isaac, former executive director of the ASA, said of our book, “Those seeking insight into biological or chemical evolution are advised to look elsewhere.”15 We agree! But if you are looking for insights into the models and mathematics thus far proposed by supporters of Darwinian evolution that purport to describe the theory, Introduction to Evolutionary Informatics is spot on. And we show there exists no model successfully describing undirected Darwinian evolution.
5. You use probability inappropriately. Probability theory cannot be applied to events that have already happened.
In the movie Dumb and Dumber, Jim Carey’s character, Lloyd Christmas, is brushed off by beautiful Mary “Samsonite” Swanson when told his chances with her are one in a million. After a pause for introspective reflection, Lloyd’s emergent toothy grin shows off his happy chipped tooth. He enthusiastically blurts out, “So you’re telling me there’s a chance!” Similar exclamations are heard from Darwinian evolutionist advocates. “Darwinian evolution. So you’re telling me there’s a chance!” So again, we didn’t start the probability fire. Evolutionary models thrive on randomness described by probabilities.
The probability-of-the -gaps championed by supporters of Darwinian evolution is addressed in detail in Introduction to Evolutionary Informatics. We show that the probability resources of the universe and even string theory’s hypothetical multiverse are insufficient to explain the specified complexity surrounding us.
Besides, a posteriori probability is used all the time. The size of your last tweet can be measured in bits. Claude Shannon, who coined the term bits in his classic 1948 paper,16 based the definition of the bit on probability. Yet there sits your transmitted tweet with all of its a posteriori bits fully exposed. Another example is a posteriori Bayesian probability commonly used, for example, in email spam filters. What is the probability that your latest email from a Nigerian prince, already received and written on your server, is spam? Bayesian probabilities are also a posteriori probabilities.
So a hand-waving dismissal of a posteriori probabilities is ill-tutored. The application of probability in Introduction to Evolutionary Informatics is righteous and the analysis leads to the conclusion that there exists no model successfully describing undirected Darwinian evolution.
6. What about a biological anthropic principle? We’re here, so evolution must work.
Stephen Hawking has a simple explanation of the anthropic principle: “If the conditions in the universe were not suitable for life, we would not be asking why they are as they are.” Gabor Csanyi, who quotes from Hawking’s talk, says, “Hawking claims, the dimensionality of space and amount of matter in the universe is [a fortuitous] accident, which needs no further explanation.”17
“So you’re telling me there’s a chance!”
The question ignored by anthropic principle enthusiasts is whether or not an environment for even guided evolution could occur by chance. If a successful search requires equaling or exceeding some degree of active information, what is the chance of finding any search with as good or better performance? We call this a search-for-the-search. In Introduction to Evolutionary Informatics, we show that the search-for-the-search is exponentially more difficult that the search itself! So if you kick the can down the road, the can gets bigger.
Professor Sydney R. Coleman said after the Hawking’s MIT talk, “Anything else is better [than the ‘Anthropic Principle’ to explain something].”18 We agree. For example, check out our search-for-the-search analysis in Introduction to Evolutionary Informatics.
7. What about the claim that “All information is physical”?
This is a question we have heard from physicists.
In physics, Landauer’s principle pertains to the lower theoretical limit of energy consumption of computation and leads to his statement “all information is physical.”
Saying “All computers are mass and energy” offers a similar nearly useless description of computers. Like Landauer’s principle, it suffers from the same overgeneralized vagueness and is at best incomplete.
Claude Shannon counters Landauer’s claim:
It seems to me that we all define “information” as we choose; and, depending upon what field we are working in, we will choose different definitions. My own model of information theory…was framed precisely to work with the problem of communication.19
Landauer is probably correct within the narrow confines of his physics foxhole. Outside the foxhole is Shannon information which is built on unknown a priori probability of events which have not yet happened and are therefore not yet physical.
We spend an entire chapter in Introduction to Evolutionary Informatics defining information so there is no confusion when the concept is applied. And we conclude there exists no model successfully describing undirected Darwinian evolution.
8. Information theory cannot measure meaning.
A hammer, like information theory, is a tool. A hammer can be used to do more than pound nails. And information theory can do more than assign a generic bit count to an object.
The most visible information theory models are Shannon information theory and KCS information.20 The consequence of Shannon’s theory on communication theory is resident in your cell phone where codes predicted by Shannon today allow maximally efficient use of available bandwidth. KCS stands for Kolmogorov-Chaitin-Solomonoff information theory named after the three men who independently founded the field. KCS information theory deals with the information content of structures. (Gregory Chaitin, by the way, gives a nice nod-of-the-head to Introduction to Evolutionary Informatics.21)
The manner in which information theory can be used to measure meaning is addressed in Introduction to Evolutionary Informatics. We explain, for example, why a picture of Mount Rushmore containing images of four United States presidents has more meaning to you than a picture of Mount Fuji even though both pictures might require the same number of bits when stored on your hard drive. The degree of meaning can be measured using a metric called algorithmic specified complexity.
Rather than summarize algorithmic specified complexity derived and applied in Introduction to Evolutionary Informatics, we refer instead to a quote from a paper from one of the world’s leading experts in algorithmic information theory, Paul Vitányi. The quote is from a paper he wrote over 15 years ago, titled “Meaningful Information.”22
One can divide…[KCS] information into two parts: the information accounting for the useful regularity [meaningful information] present in the object and the information accounting for the remaining accidental [meaningless] information.23
In Introduction to Evolutionary Informatics, we use information theory to measure meaningful information and show there exists no model successfully describing undirected Darwinian evolution.
9. To achieve specified complexity in nature, the fitness landscape in evolution keeps changing. So, contrary to your claim, Basener’s ceiling doesn’t apply in Darwinian evolution.
In search, complexity can’t be achieved beyond the expertise of the guiding oracle. As noted, we refer to this limit as Basener’s ceiling.24 However, if the fitness continues to change, it is argued, the evolved entity can achieve greater and greater specified complexity and ultimately perform arbitrarily great acts like writing insightful scholarly books disproving Darwinian evolution.
We analyze exactly this case in Introduction to Evolutionary Informatics and dub the overall search structure stair step active information. Not only is guidance required on each stair, but the next step must be carefully chosen to guide the process to the higher fitness landscape and therefore ever increasing complexity. Most of the next possible choices are deleterious and lead to search deterioration and even extinction. This also applies in the limit when the stairs become teeny and the stair case is better described as a ramp. As Aristotle said, “It is possible to fail in many ways…while to succeed is possible only in one way.”
Here’s an anecdotal illustration of the careful design needed in the stair step model. If a meteor hits the Yucatan Peninsula and wipes out all the dinosaurs and allows mammals to start domination of the earth, then the meteor’s explosion must be a Goldilocks event. If too strong all life on earth would be zapped. If too weak, velociraptors would still be munching on stegosaurus eggs.
Such fine tuning is the case of any fortuitous shift in fitness landscapes and increases, not decreases, the difficulty of evolution of ever-increasing specified complexity. It supports the case there exists no model successfully describing undirected Darwinian evolution.
10. Your research is guided by your ideology and can’t be trusted.
There’s that old derailing genetic fallacy again.
But yes! Of course, our research is impacted by our ideology! We are proud to be counted among Christians such as the Reverend Thomas Bayes, Isaac Newton, George Washington Carver, Michael Faraday, and the greatest of all mathematicians, Leonard Euler.25 The truth of their contributions stand apart from their ideology. But so does the work of atheist Pierre-Simon Laplace. Truth trumps ideology. And allowing the possibility of intelligent design, embraced by enlightened theists and agnostics alike, broadens one’s investigative horizons.
Alan Turing, the brilliant father of computer science and breaker of the Nazi’s enigma code, offers a great example of the ultimate failure of ideology trumping truth. As a young man, Turing lost a close friend to bovine tuberculosis. Devastated by the death, Turing turned from God and became an atheist. He was partially motivated in his development of computer science to prove man was a machine and consequently that there was no need for a god. But Turing’s landmark work has allowed researchers, most notably Roger Penrose,26 to make the case that certain of man’s attributes including creativity and understanding are beyond the capability of the computer. Turing’s ideological motivation was thus ultimately trashed by truth.
The relationship between human and computer capabilities is discussed in more depth in Introduction to Evolutionary Informatics.
In Introduction to Evolutionary Informatics, Chaitin’s challenge has been met in the negative and there exists no model successfully describing undirected Darwinian evolution. According to our current understanding, there never will be. But science should never say never. As Stephen Hawking notes, nothing in science is ever actually proved. We simply accumulate evidence.27
So if anyone generates a model demonstrating Darwinian evolution without guidance that ends in an object with significant specified complexity, let us know. No guiding, hand waving, extrapolation of adaptations, appealing to speculative physics, or anecdotal proofs allowed.
Until then, I guess you can call us free-thinking skeptics.
Thanks for listening.
Robert J. Marks II PhD is Distinguished Professor of Electrical and Computer Engineering at Baylor University.
(1) Chaitin, Gregory. Proving Darwin: Making Biology Mathematical. Vintage, 2012.
(2) Marks II, Robert J., William A. Dembski, and Winston Ewert. Introduction to Evolutionary Informatics. World Scientific, 2017.
(3) Ecclesiastes 12:12b.
(4) Lenski, R.E., Ofria, C., Pennock, R.T. and Adami, C., 2003. “The evolutionary origin of complex features.” Nature, 423(6936), pp. 139-144.
(5) ID the Future podcast with Winston Ewert. “Why Digital Cambrian Explosions Fizzle…Or Fake It,” June 7, 2017.
(6) IEEE, the Institute of Electrical and Electrical Engineers, is the largest professional society in the world, with over 400,000 members.
(7) R.J. Marks II, “The Joumal Citation Report: Testifying for Neural Networks,” IEEE Transactions on Neural Networks, vol. 7, no. 4, July 1996, p. 801.
(8) Fogel, David B., and Lawrence J. Fogel. “Guest editorial on evolutionary computation,” IEEE Transactions on Neural Networks 5, no. 1 (1994): 1-14.
(9) R.J. Marks II, “Old Neural Network Editors Don’t Die, They Just Prune Their Hidden Nodes,” IEEE Transactions on Neural Networks, vol. 8, no. 6 (November, 1997), p. 1221.
(10) Russell D. Reed and Robert J. Marks II, “An Evolutionary Algorithm for Function Inversion and Boundary Marking,” Proceedings of the IEEE International Conference on Evolutionary Computation, pp. 794-797, November 26-30, 1995.
(11) C.A. Jensen, M.A. El-Sharkawi and R.J. Marks II, “Power Security Boundary Enhancement Using Evolutionary-Based Query Learning,” Engineering Intelligent Systems, vol. 7, no. 9, pp. 215-218 (December 1999).
(12) Jon Roach, Winston Ewert, Robert J. Marks II and Benjamin B. Thompson, “Unexpected Emergent Behaviors from Elementary Swarms,” Proceedings of the 2013 IEEE 45th Southeastern Symposium on Systems Theory (SSST), Baylor University, March 11, 2013, pp. 41-50.
(13) Winston Ewert, Robert J. Marks II, Benjamin B. Thompson, Albert Yu, “Evolutionary Inversion of Swarm Emergence Using Disjunctive Combs Control,” IEEE Transactions on Systems, Man and Cybernetics: Systems, v. 43, #5, September 2013, pp. 1063-1076.
Albert R. Yu, Benjamin B. Thompson, and Robert J. Marks II, “Swarm Behavioral Inversion for Undirected Underwater Search,” International Journal of Swarm Intelligence and Evolutionary Computation, vol. 2 (2013). Albert R. Yu, Benjamin B. Thompson, and Robert J. Marks II, “Competitive Evolution of Tactical Multiswarm Dynamics,” IEEE Transactions on Systems, Man and Cybernetics: Systems, vol. 43, no. 3, pp. 563- 569 (May 2013).
Winston Ewert, Robert J. Marks II, Benjamin B. Thompson, Albert Yu, “Evolutionary Inversion of Swarm Emergence Using Disjunctive Combs Control,” IEEE Transactions on Systems, Man and Cybernetics: Systems, vol. 43, no. 5, September 2013, pp. 1063-1076.
(15) Review of Introduction to Evolutionary Informatics, Perspectives on Science and Christian Faith, vol. 69 no. 2, June 2017, pp. 104-108.
(16) Claude E. Shannon, “A mathematical theory of communication,” Bell System Technical Journal 27: 379-423 and 623–656.
(17) Gabor Csanyi “Stephen Hawking Lectures on Controversial Theory,” The Tech, vol. 119, issue 48, Friday, October 8, 1999.
(18) The bracketed insertion in the quote is Csanyi’s, not ours.
(19) Quoted in P. Mirowski, Machine Dreams: Economics Becomes a Cyborg Science (New York: Cambridge University Press, 2002), 170.
(20) Cover, Thomas M., and Joy A. Thomas. Elements of Information Theory. John Wiley & Sons, 2012.
(22) Paul Vitányi, “Meaningful Information,” in International Symposium on Algorithms and Computation: 13th International Symposium, ISAAC 2002, Vancouver, BC, Canada, November 21-23, 2002.
(23) Unlike our approach, Vitányi’s use of the so-called Kolmogorov sufficient statistic here does not take context into account.
(24) Basener, W.F., 2013. “Limits of Chaos and Progress in Evolutionary Dynamics.” Biological Information — New Perspectives. World Scientific, Singapore, pp. 87-104.
(25) Christian Calculus.
(26) See, e.g., Penrose, Roger. Shadows of the Mind. Oxford University Press, 1994.
(27) Hawking, Stephen. A Brief History of Time (1988). AppLife, 2014.
Photo credit: Postman85, via Pixabay.