Intelligent Evolution?

[cyborg]

If labels have physical consequences then I guess the ID argument has substance.

 

[quixotecoyote]

If labels have physical consequences then I guess the ID argument has substance.

The ID argument has consequences.

The cyborg argument has the consequence of irritating me in a minor fashion by relying on attempts at leading questions.

 

[cyborg]

The ID argument has consequences.

Which are?

The cyborg argument has the consequence of irritating me in a minor fashion by relying on attempts at leading questions.

Answers follow questions.

 

[quixotecoyote]

Which are?

I refuse to play your game any longer.

If you have an argument, make it and I will respond.

eta:
The consequence of ID ... take a stroll around the forum, I'm sure it's only been a topic in a few dozen threads in science and a few dozen in religion...

 

[Southwind17]

If you wish to limit our conversation to the AA I have no objection.

I'm happy to do this as the first step of a two-step approach. I strongly believe that the AA can be easily linked back to the OP analogy (or a very close, and equally valid, derivative thereof), but I think the link, which basically introduces the all-important 'design' aspect into the equation, is the gap in the reasoning that some people are struggling to bridge on the way to the AA from the OP. Let's try agreeing on the AA first, then we can attempt to cross the bridge!

If I have no intelligence, I have no ability to conceive. Intelligence is a prerequisite for thought, for anything above a stimulus-response action and possible not even that.

I agree.

From your extended questioning, I deduce you are using the term in a different way, although I'm not sure what. I confess to being slightly perplexed at your focus on intelligence in general. To my mind design is the much greater question as the ID'ers designer could be an unthinking cosmic machine and still make their theory work.

I seemingly misinterpretated your use of the word 'conception' to mean the act of physically conceiving, as a pre-cursor to the inevitable emergence of a physical entity thereafter. I realize what you mean now, and am inclined to agree. I'm not sure I agree with the notion of an 'un-thinking cosmic machine' being a 'designer', though, but I'll come onto that in a little more detail below.

And please do not descend into creationist smearing. For the first time in these debates we're approaching something resembling honest discussion instead of smears, insults, and arrogance.

I don't intend to do that, and if I've been guilty of it, or any of your other suggestions, in the past I can only put it down to frustration! I've always been seeking out honest and constructive debate throughout this thread, and I welcome the fact that one of the 'opposition' feels likewise. Thank you.

No I do not call it an intelligent methodology. The methodology itself is random, possibly excepting the selection criteria.

Do you want to discuss this further at this stage?

Before we had agreed to replace Sam with Robot Sam, Ollie also had a role in your story. In your story Ollie chose a method of constructing electronics in a preset pattern/according to the directions while Sam chose to do so randomly. In this example there were two methods of construction, just like in computer modeling you might choose an evolutionary algorithm or a more directed structure.

Unless you are going to take cyborg's route and deny the existence of intent, saying you are simply posting as a result of deterministic forces over which you assert no conscious control, then it follows inescapably that intent defines methodology in this and similar instances.

I do indeed deny the existence of intent in both the AA and in the Sam part of the Sam & Ollie story, although I'm not sure I'm sufficiently versed to offer a view as to what the actual 'drivers' are (assuming wanton replication and variation don't wash). Clarity of the term 'deterministic forces' might help, if you think it's important to the discussion.

We come again to out divergence in purpose. The ID'ers claims (if not their agendas) would be satisfied by a cosmic robot designing according to preset plans (it's builder could be totally uninvolved).

Accepted.

My focus is thus on design, as I hope you've gathered over the course of my posts. All you have done here is moved the intelligence one step backwards while the design remains in place. You have chosen the robot's method of construction to be random just as you could have chosen to have it manufacture in the style of a car factory robot. You have chosen to have it register the sale as a signal.

Depending on how you present it, you could still have it work. Limiting discussion to the AA is a vast improvement from a comparison to tech development in general and I salute you for the development of your argument.

If you could make it clear that you were limiting discussion to the process of randomness and selection itself, and not the entire system, it could work.

OK, as I've offered above, let's do that. But let's be clear what the First Step agenda is please. I believe that the AA exhibits replication, random variation and selection, although the latter resides within a separate 'process', as it does in nature. It's all part of the AA 'system' though, I guess. Are you saying that you want to keep the selection aspect out of the discussion for the time being? Are there any other aspects that you see as forming part of the 'entire system' that you want to exclude from the immediate discussion?

One of the best ways I grasped evolution was a little program where little dots moved in a series of individual and random directions. After a given time if the dots were in given position they had a better chance of going into the next round than those that didn't. If the dot went into the next round it had a chance of duplicating itself (if there was a spot open from a dot that didn't make it) with the duplicate having the original's movement script and the possibility of a changed part of the move.

You can probably guess how that went. But if someone had told me that the program was proof of evolution, I'd have been skeptical. "You made the program," I'd have said, "so how does that prove anything?"

I would have tended to focus my scepticism on the notion of the dots having to be in 'given position' to go on to 'replicate' which, to me, does sound analogous to artificial selection, and distinct from the AA. I'd be pretty comfortable with the idea of the program itself, as it's simply a model seeking to mimic a real-life process.

Once we've adopted just the AA, I can see it working if, IF, you make the distinction between the behaviors/actions within the system and the system itself.

If you meet me that far, I can agree with it being a semi-useful example. Semi-useful instead of fully useful only because I doubt that once you've crafted the elaborate scenario of the electronics building robot going to market, you'll get people who don't understand evolution to separate the element. I'm a reasonably intelligent person who shares the knowledge you are trying to express and look at the battles I've had on this thread. Imagine the issues with an ID'er.

OK, I hope you've already concluded that I'm prepared to meet you part way, but you'll have to clarify to me exactly what you mean by and see as the distinctions between behaviours/actions within the system and the system itself first.

As an aside, I have no idea how fast your robot is working, but if you expect it to duplicate sat. nav. systems, radar installations, mass spectrometers, MRI scanners, etc. you are going to end up confirming the ID'ers arguments about evolution being fantastically unlikely in the time frames involved. You might get a working circut or two, but expecting highly specialized machines from basic building blocks would be like watching a bacteria and expecting it to morph into a duck. The timeframe just isn't there. Minor point.

It is a minor point, because I have in no way suggested or even inferred previously that the AA would evolve complex electronics devices any faster than natural evolution works. Indeed, I have on numerous occasions argued that one of the key drivers for humans applying intelligence to the 'design' process (assuming that intelligence alone isn't the defining feature of 'design', which I think we might need to discuss and agree on as a precursor to continuing the wider discussion) is to 'short-circuit' (and by that I mean expedite, not wantonly cut out some key part of the process) the otherwise convoluted, uncompetitive and wholly uncommercial scenario that purely random variations would derive. That said, I would expect the AA certainly to work faster than natural evolution in terms of replication frequency, for obvious reasons, but given that we'd be comparing electronic evolution with organic evolution I'm not sure that we could measure progress or complexity on the same scale.

My reference to standards of success was not simply whether it sold or not. It was deciding that sales would be the deciding factor rather than aesthetic appeal, or utility, or whether grandma liked it enough to put on her mantle.

Well to my mind sales would be a defining measure, as all of your other alternatives above, plus any others, are essentially sub-sets of sales given that the devices have to sell before the benefits to the owner can be realised. Sales, to me, is an appropriate representation of a device's success in surviving its environment, although to be precise we would probably have to say that overall sales, in the long term, would be the true measure, as some people could buy an electronic device and immediately thereafter conclude that it doesn't suit their purpose as well as an alternative product might. Either way, given that overall sales in the long term is effectively determined by combining all of the individual sales over the same time period, I think that each individual sale, or more to the point the proceeds thereof, can be considered an appropriate trigger for replication. If you want to 'scale up' the process and hence frequency of replication we can consider volume sales instead, but I think that's academic to the validity or otherwise of the AA.

I'll stand by it, but as I said above, if we do the legwork in preparing a specific scenario rather than a broad category comparison, I think it can be made to work.

OK

The first system was similar enough to the current version of the AA that further explaination is unnecessary. My reference to a system that mimics evolution was to something like this:

http://www.truthtree.com/evolve.shtml

Which is similar to the program I described earlier.

OK

Again, I fail to appreciate your focus on intelligence. I am more focused on the "D" in ID so I'm not grasping your point here at all.

As I wrote above, I think we need to clarify what we mean by 'design'. Clearly, most people don't ordinarily consider design synonymous with 'manufacture' or 'assembly' alone (and I'm one of them), which, admittedly, could well be seen as a flaw in the AA, were it not for the rather subtle variation part of the 'assembly' process. I guess my focus on intelligence derives from my belief that 'design' inherently requires intelligence, and by that I mean intent and forethought. I suppose it could also include researching, prototyping, trialling, testing and the like, but to my mind they're all essentially manifestations or extensions of intent and forethought. If you remove intent and forethought, then you essentially remove intelligence, and 'design' then evaporates.

One vision that comes to my mind by way of developing this thought is a person deciding to make, say, an item of furniture using left-over pieces of timber, screws, hinges, etc, essentially 'making it up' as they go along, compared to somebody turning to the drawing board first and then obtaining the materials to match the drawing. To my mind, the latter clearly involves design, manifest in the drawing. The former might not involve design, because forethought is absent even though intent is clearly present. Assuming both items of furniture turn out to be somewhat functionally impressive (putting aside craftsmanship), which, admittedly, is more likely in the case of the latter, I should not be surprised to hear somebody asking each person if they 'designed' it themselves. What would the first person say? I'm not sure, but as I write this I'm beginning to think that intent and forethought, and hence design, are present in all endeavours of human construction other than any that could be shown to be random, even if occurring instantaneously before acting upon it.

I'll stand by this one as well. If you don't start by explaining the differences at all, but instead have a scenario that doesn't invoke conventional tech design at all and instead explains evolutionary principles themselves in a more understandable format (AA) then you have avoided my objections.

That's good, but I still reckon I can get you 'over the bridge' later!

I would disagree with what Sam knew, but as he's been assimilated, I agree the point is mute.

OK - you don't know my sons though!

In your original example Sam's entire motivation was the sale of electronics.

In the AA, it is much easier to agree with so long as we can avoid discussing why the robot is doing his mindless doings.

To be accurate, it wasn't. Sam didn't even have a motive, which is why we're best morphing him into an automaton with no human traits!

Quite. Although my implication was quite different from your inference (cough cough).

My point was that Sam was not acting in a truly random manner. If he was, he would be just as likely to destroy the electronic kit entirely as to ever produce something with it and the analogy would sink completely.

I disagree. He was acting completely randomly within the context of the analogy, which pre-supposes Sam is pre-disposed to acting constructively and not destructively, analogous to biological reproduction. Sam, essentially, is simply acting as the equivalent of the biological mechanism within two mating organisms that initially causes them to copulate then controls the replication, with mutation.

This is however, FYI, as Sam has been turned into a mindless cyborg and no longer has the ability to behave as randomly.

Not quite true. Although the automaton will generally replicate the electronic devices exactly, the variations (mutations) that it makes to each generation will be essentially random. In reality, this would be part of the programming logic that causes the automaton to select a new component randomly from a list of possible components, or connect a wire randomly between any two components, for example.

In your original example those conceptualizations were Sam's motivation. Now in AA we don't have to deal with that. Again, with the proviso that we can avoid taking the analogy to a higher level.

Again, for the record, Sam had no motivation (he's a simple lad ).

Meh. A little close to denying the existence of intent, which I feel is an ontological blunder, but we'll move on.

We might need to discuss this further right away, as I do, as clarified above, deny the existence of intent from the AA. Why do you feel that that's an ontological blunder?

So how did we go?

I think you and I, at least, can develop a meaningful debate and make some progress. I'll make one condition, though. If the likes of ID, mijo, jimbob and President Bush (although Mr President's contributions add so little, if anything, to the debate they can generally be dismissed out of hand!) choose to chip in with irrelevancies (even if they don't see them as irrelevancies!), distractions and obfuscations, I'm declaring here and now that I will, in all likelihood, completely ignore their comments and reject any inference that that constitutes tacit acceptance. This is between you and me, plus anybody else, including the foregoing, who is prepared to stick to the script.

One can define Intent (and I am) as a result or part of the mental phenomenon produced by the brain. It is intangible, yet completely material, and completely real. If you are conscious, which I assume you are, then that consciousness is the overarching intangible yet material mental phenomenon under which concepts such as intent, thought, feeling, emotion, desire, pain, etc. are placed.

For the purpose of this discussion I'm inclined to agree with this.

 

[cyborg]

If you have an argument, make it and I will respond.

I've made my arguments. It is simple: the decision to label some physical process as being X where X is a composite of sub processes is arbitrary. This applies to all X whether that be "weather" or "intent" or "sun" or "eating" etc... This is highlighted most strongly when it comes to the edge cases of deciding if something "is X" or "is not X" - as exemplified earlier by reference to the difficulty there with categorising some basic elements of "life" as to whether they are actually "alive" or not: e.g. viruses.

As such labelling a process as "intented" is arbitrary and may be pushed as far out or as far in as you wish. If you push it in then the physical is inherently intentional and you move towards constructing an atomic entity for it. Some people call it "God". If you push it out then the physical is inherently unintentional and you move away from constructing this entity.

I am arguing away from unnecessary entities.

The consequence of ID ... take a stroll around the forum, I'm sure it's only been a topic in a few dozen threads in science and a few dozen in religion...

I've been over this before. If you have been reading then you know my stance.

 

[Belz...]

I am not going to respond to content free one-liners

Yeah, lots of people do that. I suspect that's because there's a lot less to misinterpret or nit-pick in a single sentence.

 

[Belz...]

Yes - the point you seem to have missed is that the whole point of solidifying the analogy is to show the idempotence of "intent" and "forethought" - that is to say that you are not being considered as a "human": your output is "random" by virtue of our ignorance as to the variables upon which "intent" and "forethought" are dependant.

"Intent" and "forethought" are being torn down from the pedestal that you and the IDers put them on. That is why none of us on the other side can comprehend why you think we are supporting the ID argument. We are showing that "intent" and "forethought" are totally unnecessary as existential entities to explain our universe.

They don't really exist in your mind: "intent" and "forethought" are a series of physical triggers that represent these abstract concepts. There is no more "intent" in "you" than there is an intention for an atom to wander about its business. There is no more "forethought" to your actions than there is to where that atom winds up on its merry journey.

If you think there is more "substance" then you become an dualist and open to "God" as the "originator" of the "forethought" and "intent".

Yep. This summarizes the thread pretty well, methinks.

Before I go any further, are you willing to stipulate that some things can be more random than others.

Actually, they are either random or not. Barring quantum fluctuations having an effect on the macroscopic world, nothing is random.

 

[Belz...]

How did the preceding four sentences come to be written?

Certainly not by being invented out of thin air. You'd swear they were assembled from previous experience !

The reason the analogy falls is because the particular properties of "intelligence" that make it differ from biological evolution are important to the point made in the analogy.

Actually, the analogy is made from a standpoint of information being passed on because it has an adaptative advantage. Intelligence is irrelevant from that point of view.

So you deny that physical systems can be more than the sum total of their individual part?

I know _I_ do.

The water molecule, for example, is a direct result of the properties of its constituents. Nothing more.

 

[Belz...]

I'm merely pointing out that there is a difference between the intensive and extensive properties of a system and that the extensive properties of a large system may not be fully explicable from the intensive properties of its most fundamental part.

I strongly disagree. If you're correct, then whence do those properties come from ?

 

[Walter Wayne]

As an aside, I have no idea how fast your robot is working, but if you expect it to duplicate sat. nav. systems, radar installations, mass spectrometers, MRI scanners, etc. you are going to end up confirming the ID'ers arguments about evolution being fantastically unlikely in the time frames involved. You might get a working circut or two, but expecting highly specialized machines from basic building blocks would be like watching a bacteria and expecting it to morph into a duck. The timeframe just isn't there. Minor point.

I would go further, and say that his robot probably wouldn't create it ever, not unless its "mutation" rules were unrecognizable when compared to biological mutation.

The reason is that in technological development we can have irreducibly complex products. Why, because intermediate products don't have to be "successful". If I build a non-functioning MRI scanner, I can debug it and say either "what a stupid idea" or "the concept was sound, but I need such-and-such to fix it". So my dud-parts successfully reproduce.

Since biological evolution works only on success of the concrete product, it requires viable intermediate steps. There is a theoretical limit on biological evolution that it won't produce irreducibly complex products, but no such limit on design.

However, if you did run such automaton I think you would gain some interesting results. The products resulting would resemble biology so much more than they resemble what we think of as technology, that you would have a concrete example of how life arose by a means other than design.

Walt

 

[cyborg]

Barring quantum fluctuations having an effect on the macroscopic world, nothing is random.

But given the world that we exist in that randomness leads to order - natural order.

It is natural for some things to have more "intent" than others by virtue of the fact that these "things" can be seen to operate in a contiguous way from one period of time to the next. I think the examples of cellular automata like Conway's Game of Life illustrate this. In it are periodic patterns - stable, moving, generating etc... These all seem more "intentful" by virtue of the fact that they, naturally, have "functions" in the simple universe in which they inhabit. In fact it is able to compute any function as it is Turing Complete.

Information that gets itself selected becomes the information of tomorrow. That which persists is selected. Chaos is order.

 

[cyborg]

The reason is that in technological development we can have irreducibly complex products. Why, because intermediate products don't have to be "successful". If I build a non-functioning MRI scanner, I can debug it and say either "what a stupid idea" or "the concept was sound, but I need such-and-such to fix it". So my dud-parts successfully reproduce.

Oh I don't know... maybe we could invent a Marketing bot that would take successful, but simpler, inventions like say, a clock and a radio, and mash them together to see what happens?

Then if only his Worker bot invented a succession of increasingly better watches...

...and maybe when he got round to it he could get to producing Worker bots...

 

[Belz...]

But given the world that we exist in that randomness leads to order - natural order.

Are you refering to real randomness or apparent randomness ?

No, wait. Scratch that. They both lead to order.

 

[jimbob]

Fair enough. What kind of evidence are you prepared to accept?

Evidence that novel inventions are the results of an intelligent process that requires no iterative development or accumulated a priori knowledge that arises from the a posteriori.

No learning. No experiment. Just pure "whole cloth" (as jimbob put it) design please.

Firstly, I have no recollection of using that term anywhere in this (or any other) thread, and a search confirms this too.

I have frequently said that both biological evolution and technological design *are* iterative processes, and that there is optimisation over time in both cases.

The analogy falls down after that. Even Intelligent Design proponents accept "change over time", it is just that they ascribe this to a deity.

Technological "development" typically is an iterative process that involves learning from experiment.

How does this help the the analogy?

If I use calculations to determine design parameters, then I performe an intelligent action. The resulting information comes about in response to the experiment, and in no way resembles random changes to the information that would occur if the design parameters simply "mutate".

Saying that we can use evolutionary algorithms does not prohibit other techniques.

If technological development can produce artifacts that would not evolve in the lifetime of the universe; than I am happy to say that the processes are different, and the products of the processes can be different too.

 

[cyborg]

How does this help the the analogy?

*Ugh.*

Technological "development" typically is an iterative process that involves learning from experiment.

Does this ability "learning from experiment," just exist? Is it computational? Is it evolutionary? Is it random? Or is it oracle like?

 

[Walter Wayne]

Oh I don't know... maybe we could invent a Marketing bot that would take successful, but simpler, inventions like say, a clock and a radio, and mash them together to see what happens?

Then if only his Worker bot invented a succession of increasingly better watches...

...and maybe when he got round to it he could get to producing Worker bots...

Yes, since I claim some inventions don't follow from simple alteration/combination of previous generation products I am claiming that all products came about that way. Yup, you got me.

 

[cyborg]

Yes, since I claim some inventions don't follow from simple alteration/combination of previous generation products

Name some and then tell me how they came about without any simple alteration or combination of previous generational products.

 

[jimbob]

Southwind, I am merely pointing out that the evolutionary algorithm that you have described still has implicit selection criteria. If you tried to impliment such a system without any selection criteria, then the system would not work.

The instructions are to build what has sold, and make small changes. I woul dsay that you have an evolutionary algorithm devoted to "making stuff that sells".

How exactly do you determine a failure?

Suppose, as is likely, the first random variant did not sell.

Without a process or an algorithm to determine that this is unsucessfull, the system would remain waiting.

If you add an algorithm to determine the fialure criteria, then you have added the required intelligence.

Of course there could be an interesting case, when an imperfect copy actually has randomly acquired the ability to self-replicate. This part woould then, without any further input make a further (imperfect) copy of itself, and its descendents would, from this point actually evolve.

Rather than talking about hypothetical cases, I think it would be better to talk about real case studies where evolutionary algorithms have been used. They have been used to produce circuits that human designers couldn't, and which are currently very hard (impossible with present modelling techniques) to model, and in only about 100 generations.

Your story would also work for this, but it is probably better to have real examples.

This shows the power of evolutionary algorithms when harnessed to an intelligently defined goal. One can then point out that organisms show "imperfect copying", and that those which manage teo reproduce are obvioulsy "fit enough" to reproduce. This is just an evolutionary algorithm where the selection criterion is the ability to reproduce. Unsurprisingly, this results in good reproducers.

 

[cyborg]

If you add an algorithm to determine the fialure criteria, then you have added the required intelligence.

And if you evolve one in a system where there are many factories, with many worker bots producing many things that are all being consumed by the environment and firing off triggers...?