Monday, June 25, 2007

Evolution-through-natural-selection is not a reductive explanation

In the post below we discussed what evolution really means. Evolution through either natural or artificial selection is an algorithm that is based on a tautological observation: the more likely an element is to survive and reproduce the more likely it is to pass on its heritable properties to its offspring. This is simply a logical truth — or at least a logical truth in any world in which there is reproduction and heritable mutation.

Not only is it a logical truth, it is a logical truth that is not derivable from the fundamental principles of physics. Hence it is another example of a truth of a "special science" that is not reducible to physics. (See my paper "Emergence Explained: Abstractions" for why this is important. I wish I had thought of this example when writing that paper!)

This is a nice example of a non-reductive truth because the truth of evolution-through-natural-selection is a consequence of the "axioms" that define a model (i.e., the level of abstraction) in which there is reproduction and heritable mutation. Once such a world is brought into existence (or in the terminology of "Emergence Explained," once such a level of abstraction is implemented) certain consequences follow, one of the most important of which is evolution through natural selection. In other words, evolution through natural selection is a property of (a truth about) the level of abstraction itself and not of any particular implementation of the level of abstraction. In that sense evolution through natural selection cannot be expressed in a traditional peel-the-onion reductionist manner.

Evolution through natural selection is also not reductive in a more traditional sense. A reductive explanation of a property typically decomposes the entity that exhibits that property into its components and then shows how the interactions of the components produce the property. The standard examples are chemical in which one shows how a chemical compound is created by combining, according to valence theory, the components that make up the compound.

In the case of evolution through natural selection, there is no isolatable entity (other than the biosphere itself, i.e., the entire level of abstraction) that exhibits evolution. Furthermore, the explanation of the evolutionary process is necessarily open rather than closed since it depends crucially on the environment to act as a selection mechanism. Thus, again, the peel-the-onion reductionist style approach to explanation doesn't work for evolution.

Evolution through natural selection is a phenomenon that exists in any level of abstraction that supports reproduction and heritable mutations. The mechanisms of reproduction and inheritance (DNA in the case of biology on this planet) are not relevant to the process as such. So again, there is no reductionist explanation. Evolution as a property of a level of abstraction is independent of the principles of the physics that implements the level of abstraction.

What's empirical about evolution through natural selection?

Darwinian evolution through natural selection is the process whereby heritable properties establish themselves in a population as a result of the higher propensity of entities possessing those properties to survive and reproduce within a particular environment, which itself may change in time.

When stated in this form, evolution through natural selection depends on two elements.
  • the possibility of heritable variation, which may itself be subdivided into two subcomponents, variation and inheritance, although this further subdivision isn't relevant to this post, and
  • that survival and reproduction is a function of the suitability of an entity to an environment.
The second element is tautological. The entities that survive and reproduce in an environment are by definition better suited to that environment than those that don't. Darwin called this phenomenon "natural selection" to contrast it with selection performed by animal breeders. Thus if one asks what aspects of evolution is empirical, the answer would have to be whether or not heritable changes occur in nature. The rest follows logically.

Of course Darwin was quite aware of the existence of heritable mutations. We (as farmers) have been breeding plants and animals for centuries. One might call the process of plant and animal breeding "evolution through human selection." What Darwin and Wallace noticed was that selection by nature may be understood on the same terms as selection by man — that by virtue of the fact that some plants are more likely to survive and reproduce than others, the environment itself "selects" for breeding some plants and animals over others. In other words, the environment acts like the farmer. The primary difference is that farmers select variations on the basis of their value to the farmer. The environment selects variations without intention, purpose, or reason: those that survive are selected.

It's easy to see why this model led Darwin and Wallace to think that small changes built up over a long period produce major evolutionary shifts. That's the way it happens on the farm. What Darwin and Wallace apparently didn't count on was the possibility of major changes in the environment, which would lead to major changes in what survives.

As we observed above, the mechanism of evolution by natural selection has two components: variation and selection. Variation, for the most part, involves relatively small changes. It is these small changes that Darwin and Wallace focus on. But selection, i.e., what actually fits the environment, depends on the environment. If the environment changes significantly — the temperature rises or falls significantly, the oceans cover or uncover land masses, the bio-environment itself changes as a result of activities of successful species, etc. — what will survive under the changed conditions is likely to be significantly different from what survived under the previous conditions. Major environmental changes are likely to effect many species at once, leading to a cascade of changes. It is this process that produces the punctuated equilibrium model we now understand as characterizing evolution.

Here's how Darwin put it in the conclusion to "On the origin of species."
Under domestication we see much variability …

Variability is not actually caused by man … But man can and does select the variations given to him by nature, and thus accumulates them in any desired manner. He thus adapts animals and plants for his own benefit or pleasure. He may do this methodically, or he may do it unconsciously by preserving the individuals most useful or pleasing to him without any intention of altering the breed. It is certain that he can largely influence the character of a breed by selecting, in each successive generation, individual differences so slight as to be inappreciable except by an educated eye. This unconscious process of selection has been the great agency in the formation of the most distinct and useful domestic breeds. That many breeds produced by man have to a large extent the character of natural species, is shown by the inextricable doubts whether many of them are varieties or aboriginally distinct species.

There is no reason why the principles which have acted so efficiently under domestication should not have acted under nature. …

If, then, animals and plants do vary, let it be ever so slightly or slowly, why should not variations or individual differences, which are in any way beneficial, be preserved and accumulated through natural selection, or the survival of the fittest? If man can by patience select variations useful to him, why, under changing and complex conditions of life, should not variations useful to nature's living products often arise, and be preserved or selected? … I can see no limit to this power, in slowly and beautifully adapting each form to the most complex relations of life. …

On the view that species are only strongly marked and permanent varieties, and that each species first existed as a variety, we can see why it is that no line of demarcation can be drawn between species, commonly supposed to have been produced by special acts of creation, and varieties which are acknowledged to have been produced by secondary laws. …

As natural selection acts solely by accumulating slight, successive, favourable variations, it can produce no great or sudden modifications; it can act only by short and slow steps. Hence, the canon of "Natura non facit saltum," [nature makes no leap] which every fresh addition to our knowledge tends to confirm, is on this theory intelligible.

Wednesday, June 13, 2007

Plants 'recognize' their siblings

From McMaster University
Researchers at McMaster University have found that plants get fiercely competitive when forced to share their pot with strangers of the same species, but they’re accommodating when potted with their siblings.

The study appears today in the Royal Society journal Biology Letters.

“The ability to recognize and favour kin is common in animals, but this is the first time it has been shown in plants” Susan Dudley, associate professor of biology at McMaster University in Hamilton, Canada, said. “When plants share their pots, they get competitive and start growing more roots, which allows them to grab water and mineral nutrients before their neighbours get them. It appears, though, that they only do this when sharing a pot with unrelated plants; when they share a pot with family they don’t increase their root growth. Because differences between groups of strangers and groups of siblings only occurred when they shared a pot, the root interactions may provide a cue for kin recognition.”

Though they lack cognition and memory, the study shows plants are capable of complex social behaviours such as altruism towards relatives, says Dudley. Like humans, the most interesting behaviours occur beneath the surface.

Monday, June 11, 2007

The spammers are developing the next AI systems

From the New York Times
Captchas are the puzzles on many Web sites that present a string of distorted letters and numbers. These are supposed to be easy for people to read and retype, but hard for computer software to figure out.

Most major Internet companies use captchas to keep the automated programs of spammers from infiltrating their sites.

There is only one problem. As online mischief makers design better ways to circumvent or defeat captchas, Web companies are responding by making the puzzles more challenging to solve — even for people. …

“You can make a captcha absolutely undefeatable by computers, but at some point, you are turning this from a human reading test into an intelligence test and an acuity test,” said Michael Barrett, the chief information security officer at PayPal, a division of eBay. “We are clearly at the point where captchas have hit diminishing returns. …

The emergence of the technology started a wave of research into ways to make computers smart enough to crack the puzzles.

Yet some of that activity can be ethically murky. Aleksey Kolupaev, 25, works for an Internet company in Kiev, Ukraine, and in his spare time, with his friend Juriy Ogijenko, he develops and sells software that can thwart captchas by analyzing the images and separating the letters and numbers from the background noise. They charge $100 to $5,000 a project, depending on the complexity of the puzzle.

Mr. Kolupaev said he had worked both for legitimate companies that want to test their own security and for spammers who seek to infiltrate Web sites.

“Nothing is unbreakable, and each system has its own weakness,” he said. “If you create a program that only recognizes one picture from a hundred, it’s not a problem. You just hit the site 100 times, and you break through.”

On his Web site, ocr-research.org.ua, Mr. Kolupaev boasts of cracking the captchas of companies like MySpace and PayPal; the site also ranks the effectiveness of each captcha. He says he believes that his work makes the Internet more secure because companies tend to improve the captchas that he critiques.

Friday, June 08, 2007

Bob Park reports that
Stephen E. Straus, 60, Died of brain cancer. The first director of the National Center of Complementary and Alternative Medicine at NIH, his task was to turn the quack-dominated Office of Alternative Medicine, created by Congress, into a scientific center. He did it with grace, the only way possible, subjecting one quack cure after another to randomized double-blind tests, while enduring attacks from scientists who thought he moved too slowly.
The Center seems to take an open-minded but clear-eyed view of alternative medicine. It notes, for example, that tests show the
efficacy of acupuncture … in adult postoperative and chemotherapy nausea and vomiting and in postoperative dental pain.

Tuesday, June 05, 2007

Are we safer now?

Has Bush's anti-terrorist policies made us feel safer? American business doesn't think so. The following is from an article by Daniel Gross in Slate Magazine.
Judging by the latest report from insurance giant Marsh, Americans—at least American businesses—are feeling less and less safe. According to Marsh's 2006 report, 59 percent of companies bought terrorism insurance last year, up from 58 percent in 2005, even as prices for such policies rose by about 9 percent. In 2003, only 27 percent of Marsh clients bought such policies. 'I think it's fair to see this data as a reflection of greater concern,' said Jill Dalton, managing director at Marsh Inc. 'Our clients only have a certain amount of money to spend on insurance. And more of them are choosing to spend more on terrorism insurance.' …

Insurance is generally local. The damage wrought by Hurricane Katrina wouldn't necessarily cause flood insurance rates for a home in Minnesota to rise. But in terrorism, she notes, "pricing responds to global events rather than purely local events." In other words, the higher rates and higher volume of sales in the United States probably doesn't reflect concerns by American companies that the nation is less safe from terror than it used to be. Rather, it may reflect concern that the world is less safe from terror than it used to be. This concern is backed up by hard data. The State Department reported that terrorism attacks worldwide rose 25 percent in 2006.

Word of the day: socialize

To promote an idea by talking to people about it.
Now that the committee has agreed, we have to socialize the plan to the rest of the company.

Monday, June 04, 2007

Global Peace Index

According to Vision for Humanity Norway is the most peaceful country
in the world. Iraq is the least peaceful. The top 5 are Norway, New Zealand, Denmark, Ireland, and Japan. The bottom 5 (out of 121) are Nigeria, Russia, Israel, Sudan, and Iraq.

The index is based on 24 indicators divided into three categories: ongoing domestic and international conflict, safety and security, and militarization.

Sunday, June 03, 2007

The web is becoming a video medium

I guess the YouTube phenomenon was a give-away. But I hadn't thought about it much. I have noticed that I've been watching a lot more talks.

Google has lots of talks such as Google Tech Talks and more Google TechTalks.

The IBM Almaden Institute has made its talks available: Cognitive computing and Comlexity.

TED has made its talks available.

Here's Linus Torvolds on GIT, his new SCM system. Even though this talk was given at Google it is posted on YouTube and not on Google TechTalks.

Here are Authors@Google.

I just came across a startup site that has advice in simple video segments about how to live your life. Here's one on giving your cat medicine.

Most of us find it much easier to watch a video than to read text.