Editor’s note: ENV is pleased to welcome Stephen A. Batzer, Ph.D., as a contributor. He is a forensic engineer with licensure in Michigan and Arkansas. His expertise includes the fields of materials selection, design, and failure analysis. Dr. Batzer frequently gives invited lectures and short courses on a variety of topics, to include evolution, forensic engineering, and expert witnessing. He has testified nationally regarding disputes large and small. He has over 60 peer-reviewed technical papers. He is currently an adjunct assistant professor of mechanical engineering at the University of Arkansas. Dr. Batzer is a retired US Army Reserve Lieutenant Colonel of Ordnance.
When I was in the Army in the 1980s, my battalion was equipped with I-HAWK surface-to-air missiles. In addition, we had rugged launchers, cramped control stations, and powerful guidance radars. The guidance radar illuminated the bogey with radiation so that the on-missile receiver could clearly recognize it. This radar was particularly capable. After acquiring the target, the assembly would rotate and elevate as necessary, staying locked on to the moving target like a spectator at a tennis match. If lock was broken for whatever reason, the device would instantly begin searching for the target, frantically moving up and down, back and forth, as if it had a mind of its own, in an eerily human way.
Of course, this was a machine simply following a search algorithm that told it what success looked like. The radar looked intelligent. It was not intelligent.
Daniel Dennett has a similar story to tell, but draws different conclusions. His talk is publicly available. He speaks of another faux-intelligent system. He plays a video for his audience that documents the work of digital media artist Karl Sims showing evolution in action. Actor Alan Alda pleasantly narrates, and playful Disneyesque music fills the background.
Sims’s work is useful for showing how computer algorithms behave, automating a task that could be done manually. In this simulation, the program starts with two eager to please, but mostly immobile digital blocks. These two blocks are jointed, and wobble to the best of their ability, but are stymied by their lack of complexity. These ur-blocks are then mutated to produce daughter organisms of increasingly complex design. Blocks of different sizes and orientation are added, and differing motion types are input. These changes produce (what else?) differences in mobility! Better organisms have evolved, all “without human intervention.” Presumably, the unfortunate less-fit block creatures have been humanely euthanized. Dennett is positively radiant with this proof of Darwinism in action.
Of course, a little skepticism is in order. We’re just staring at a computer screen here, and then talking about biology. Everything that is actually alive is physical, not virtual. Calling a thing biological doesn’t make it so. Since we live in a physical world, it would have been helpful for Mr. Sims to build physical, not virtual, organisms. Obviously, introducing this level of added realism would bring the whole enterprise to a swift halt. It is hard to imagine shoeboxes that are bungee-corded together reproducing the motion prescribed by Sims.
The Sims-ulated organisms do not model life in any meaningful way. This program is modeling a very simplistic random search algorithm to produce an output, like a radar searching for an aircraft, or a robo-call computer punching out all the numbers inside of one area code, looking for a mark. The information, process, and therefore success have all been pre-loaded.
That we’re only talking about blocks with magic joints cannot be overstressed. Think about a steam-powered locomotive, which was one of the first non-biological devices to show autonomous motion. A steam engine has a frame, tank, water, fuel, boiler, exhaust, cylinders, pistons, valves, regulators, wheels, pitman arms, axles…the list goes on. None of these highly interdependent and complex components arose via Darwinian processes applied to an ox cart.
Darwinism is a “bottoms up” view of life, which can work for components, but not to originate complex systems. A system is a coherent and sophisticated grouping of interrelated parts that all work to achieve a goal. Like an orchestra. The actions of the violinist are dependent upon those of the flautist and vice versa. Suppose each of the members of an orchestra were to compose and record a demo piece independently, and then the collection of 25 audiotapes were to be run concurrently. It would be gibberish, complete incoherence, as there was no top-down goal specified.
Years ago, there was a hilarious segment of the British television show Jeeves and Wooster with Hugh Laurie. The Drones, Wooster’s man club, decided to make some money by playing banjo for parties. The Drones rehearsed, but the piece sounded awful and incoherent. As they played, the Drones looked at one another in discomfort. When the song concluded and everyone was feeling dismayed, one of the Drone banjo players jumped up and declared triumphantly, “I finished miles ahead of you fellows!” Clearly, he didn’t get the memo.
As a last example of this, there is an informative video on YouTube that shows the sobering complexity of mobility. The Defense Advanced Research Projects Agency (DARPA) intelligently designed an ersatz mule to help soldiers carry equipment in combat. This was not designed in any Darwinian fashion.
To his credit, Dennett realizes that Sims’s production is unreasonably simplistic, and he pokes fun at it. Why he uses it is clear. It is convincing, but only to the already convinced. The “bottoms up” method has already been tried experimentally in biology. You can take a mutagen such as a low-level poison or x-rays and artificially mutate a genome. What is found isn’t astonishingly rapid progress coupled with moving music, but rather damaged or altogether unviable organisms.