Comp. Sci. / Eng. Informatics

Pioneer of non-Darwinian, Intelligent Evolutionary Design passes away

Spread the love

[Comp Sci. /Eng, Informatics]

I felt compelled to offer a tribute…..

A great pioneer of Intelligent Evolutionary Design and non-Darwinian evolutionary computing passed away September 20, 2007. See this Washington Post Article: Ryszard Michalski; Shaped How Machines Learn.

Michalzki created the notion of Intelligent Evolutionary Design and advanced the hypothesis of non-Darwinian evolutionary computing.

From Foundations of Intelligent Systems: 12th International Symposium

In contrast to Darwinian evolution, an intellectual evolution is guided by an “intelligent mind,”….

Every Easter at George Mason, Campus Crusade would flood the campus bulletin boards with a list of professors offering to share their wisdom and knowledge with students. Ryszard S. Michalski’s name was always on the list.

17 Replies to “Pioneer of non-Darwinian, Intelligent Evolutionary Design passes away

  1. 1
    rockyr says:

    Interesting person, interesting design and engineering ideas wrapped in evolutionary terminology. Especially the word – “Entelligence” (evolutionary intelligence) is full of paradoxes and contradictions.
    So is his paradoxically sounding “Learnable Evolution Model (LEM)”, (“… Unlike conventional Darwinian-type methods that execute an unguided evolutionary process, the proposed method, called LEMd, guides the evolutionary design process using a combination of two methods, one involving computational intelligence and the other involving encoded expert knowledge.”) Just like in real-life processes.

    Phrases like “Intelligent Evolutionary Design” really make one think about the meaning of those words, don’t they?

  2. 2
    scordova says:


    Thank you for your thoughts.

    I have some disjointed thoughts and disconnected observation to offer….

    People like Michalski inspired me to study computer science at his school, especially the topic of Artificial Intelligence and Machine Learning….

    If there is such a thing as “Artificial Intelligence”, then what is the essence of “Real intelligence”?

    Hofstadter’s book, Godel, Escher, Bach explored the whole issue of what is intelligence. (The subtitle of Hofstadter’s book is “A metaphorical fugue on minds and machines”). After 800 pages, Hofstadter offered no definition of intelligence!!!

    What we have been able to do in computer science is get computers to act like “real intelligence” in small ways, even though to this day, I don’t think there is a stable definition of intelligence (nor do I think there has to be).

    Michalski created machines capable of “learning”, the very activity which we attribute to intelligence……

    One issue I have with the computer science community is idea of a Darwinian evolutionary algorithm. Is there really a Darwinian algorithm in the strict sense in computer science? After all, such algorithms are goal-directed and full of teleology….

    Un-adulterated Dawkins/Dennett type Darwinism is brainless, mindless, stupid, aimless, blind and undesigned. In contrast, computerized “Darwinism” has to be designed and purposeful, and in accordance with the purposes of the intelligent designers that crafted it. So in a strict sense, even “Darwinian algorithms” in computer science are not really Darwinian. I would suggest the proper term would be “Blythian” not “Darwinian”.

    Michalski’s work raises scientific questions such as “what is the implication of No Free Lunch on Machne Learning?”

    No Free Lunch theorems still place a bound on what systems with the ability “learn” (such as Michalski’s mahcines) might actually be able to learn. I do not have answers to the implication of No Free Lunch theorems on Michalski’s work but that would be a good topic of exploration for an Evolutionary Informatics research program, exactly the kind of research program which scientists like Robert Marks pioneered.


  3. 3
    lars says:

    Thanks for coming back out of the woodwork for this one.

  4. 4
    scordova says:

    Here is further memorial to Michalski at a weblog run by various people inlcuding Ftk, myself, and others:

    Tribute to Michalski and Machine Learning: “My Way”

  5. 5
    rockyr says:

    Salvador, There must be something about Polish AI and robotics engineers and mathematicians that inspires people. Like you, long time ago I was drawn to robotics, AI and engineering by another Polish guy whose name I have, sadly, forgotten, (hopefully it will come back to me), but he had a great book about practical robotics, cybernetics and AI. Even Hofstadter was quite fond of S. Lem and his sci-fi. I bought & read Hofstadter’s book long time ago, a lot of interesting stuff and food for thought, (although I haven’t revisited it in the last 10 years or so), but I know I disagreed with his philosophy of intelligence and how things acquire meaning. You are right, defining intelligence is not easy, and the bottom line is that the term “AI” is really an oxymoron. Not sure about “Blythian”, but I agree with you about the meaning (or unmeaning) of the “Darwinian” computing algorithms — correctly speaking, there is no such thing, and neither is a “free lunch”. It is just common sense, no need to devote a lifetime to the study of the free lunch computing “science”, unless one gets payed for it. So what has Michalski concluded about the “free lunch”?

  6. 6
    larrycranston says:

    Mr. Cordova,

    Did you mean to cut off the quote in the original post? I think the full text is

    “In contrast to Darwinian evolution, an intellectual evolution is guided by an ‘intelligent mind,’ that is, by humans who analyze advantages and disadvantages of previous generation of solutions and use the developed understanding in creating next generation of solutions.”

    It sounds like he is talking here about something other than the Designer in the broader context of ID. Can you clarify?

  7. 7
    scordova says:

    Did you mean to cut off the quote in the original post?

    I certainly did because it highlighted the difference between mindless Darwinism and intelligent design. Human designs are also instances of intelligent design, or to use his phrase, intelligent evolutionary design…..

    Michalski may or may not believe Darwinism in biology can create biological complexity. His writings express the assumed mainstream beliefs, but does he really believe the world is a product of mindless processes?

    But even granting he might believe Darwinism to be true, the irony would be that he did he not try to get machines to emulate brainless Darwinian processes, but rather he tried to get machines to emulate an intelligent mind.

    His actions speak louder than words….

  8. 8
    lotf says:

    You mention GED has no definition of intelligence after 800 pages which is correct in a formal sense but the aim of the book was to introduce the concepts in the study of intelligence.

    What is your definition of intelligence? Maybe I can help you understand what Hofstadter was trying to tell you!

    It’s one of my favourite books and what pointed me to a life in computer programming.

  9. 9
    scordova says:


    I’m so glad to meet others who like Hofstadter’s book! It is a stealth ID classic, imho.

    Hofstadter pointed out that certain highly primitive concepts are either undefined in a system or implicitly defined by all the theorems of the system.

    I think intelligence, like points in Euclidean geometry, or force in physics, is best left as an undefined term.


    At this point critics of intelligent design often protest that design theorists have yet to provide a careful definition of intelligence. While I agree that terms need to be defined as carefully as possible, the call for definition can itself become a subterfuge. Thus the call for definition can become a way of avoiding the challenge posed by an idea by endlessly requiring further clarification of key terms. The later Wittgenstein certainly thought the call for definition was overrated. Indeed, the finiteness of language itself implies that the call for definition must at some point either end or issue in circularity. Within intelligent design, intelligence is a primitive notion much as force or energy are primitive notions within physics. We can say intelligible things about these notions and show how they can be usefully employed in certain contexts. But in defining them, we gain no substantive insight.

    The very word intelligence derives from the Latin words “inter” (a preposition meaning “between”) and “lego” (a verb meaning to “choose” or “select”). Thus strictly speaking intelligence refers to the capacity to choose or select. Yet unlike natural selection, which operates without goals or purposes, ordinarily when we think of an intelligence as choosing or selecting, it is with a goal or purpose in mind. We could therefore define intelligence as the capacity for rational or purposive or deliberate or premeditated choice. Have we therefore defined intelligence to the satisfaction of the critics of intelligent design? Hardly. When Howard Van Till, for instance, issues his call for definition, his worry is not what intelligence or design means as such, but what these terms mean in contexts where no embodied intelligence was acting and thus where his view of nature as a complete system of natural causes (cf. his fully gifted creation and robust formational economy) comes under pressure. Invariably I’ve found that the call to define intelligence by critics of intelligent design is not a call for clarification but a defensive move to relieve pressure from some aspect of the critic’s own worldview that intelligent design calls into question.

    I suppose the best we can do then is enumarate necessary but not sufficient characteristics of intelligence, we can’t formally define it.

  10. 10
    lotf says:


    Euclid himself defined points as “That which has no part” so I am not sure why you say we can’t define points.

    But anyway – are you saying ID can basically be rewritten as Undefined Designed? I expect I am misunderstanding you otherwise we’d be struggling to detect such design wouldn’t we?

    Apologies if this is a stupid question, I am still trying to catch up with all the literature.

  11. 11
    scordova says:

    Hi Lotf,

    Unfortunately I have to run and won’t be back till Tuesday as I have a plane to catch!

    We don’t have sufficient definitions of Intelligence or Life, but it does not make the entities less real. You and I are alive, we accept it as a fact even though we don’t have a definition for what life really is….

    The inability to describe something with only a few or finite axioms is exactly the problem of incompleteness. Reality may be greater than our ability to adequately describe it (see page 19-20 GED, “truth is greater than provability”).

    I think perhaps intelligence cannot be described with any finite number of axioms. That’s what Penrose suggested in Emperor’s New Mind (another Stealth ID Classic, imho).

    Perhaps, “Undefined” is too strong at term. We might say intelligence is “incompletely defined”, much like all axiomatic systems capable of arithmetic.

    Sorry I must run. Feel free to disagree or comment more. I look forward to reading your ideas when I return.


  12. 12
    allanius says:

    Douglas Hofstadter is a materialist. The purpose of his famous book is to delimit intelligence as a “strange loop.” There is no thought of design here; the loop is purely a quirk of nature. His appeal to strangeness enables him to cloak the book in an appearance of mysticism, but his theory amounts to little more than the modest assertion that intellect is self-reference. But then what about the “I”? When you wake up in the morning, do you know who you are? If intelligence is nothing more than a strange loop, then shouldn’t the “I” appear as a stranger each time it reappears? And yet the “I” seems so familiar that we never think to question it. It is not double upon its return, like an Escher drawing, but perfectly unified. The absence of strangeness in the “I” is truly strange; “strange loops” seem to amount to little more than a parlor game.

  13. 13
    lotf says:

    Hi scordova,

    Well I read GED as one big essay detailing how intelligence, or at least the appearance of, can arise from a collection of what could be described as un or semi intelligent processes. How can this be helpful for ID? It suggests the opposite.

    ID should not consider GED a book that supports it.

  14. 14
    sane person says:

    Censorship is a bad thing. How do you people sleep at night with your constant lying and your censorship of the truth? You have no moral values. You are worse than muslim terrorists. I hope the designer you fuckwits believe in send you to hell where you belong.

    I didn’t censor this one, Bob. Happy now? Tell Bob what you think of his conduct at -UD moderator

  15. 15
    professorsmith says:

    With a moniker like “sane person” one would think that a sane comment was to follow. Alas, it was not the case. That has to be the only explanation for why this person could complain about censorship of the truth and our moral values while simultaneously calling us “f-wits” and damning us to hell. I usually find that personal attacks of this nature are usually employed because the attacker can not defend his/her position.

  16. 16
    dl says:

    The tone of the comment by “sane person” certainly undermined any arguments (legitimate or not) that he may be trying to make.

    I hope this is not out of line, but making his email address public seems a little excessive. Should that be taken down?

  17. 17
    scordova says:


    Well I read GED as one big essay detailing how intelligence, or at least the appearance of, can arise from a collection of what could be described as un or semi intelligent processes. How can this be helpful for ID? It suggests the opposite.

    State the page where he offers the formal derivation of his claim. He only hand waves, and that in itself makes the point.

    Euclid himself defined points as “That which has no part” so I am not sure why you say we can’t define points.

    Define part. Then define the words that define part. That demonstrates the regress toword something undefined.

Leave a Reply