I am very sorry that you think that I am practicing “intellectual dishonesty”, but you are free to think as you like. It is perfectly true that I had not understood well the context of your point in my post #27 (maybe because I often write in a hurry). I understood it better after your post #28 (as I have explicitly said in my post #29), and in that same post I tried to “make some order”. What is there of intellectually dishonest in that?

You may believe that I pass everything under a ““it must be right, so how can I interpret it that way” filter”, but luckily you are not my epistemology teacher, nor my analyst. I can only say that I believe very much in a sincere search for truth, that I take very seriously the criticism of others, and that I would never consciously support something which I believe is not true, exatly at the level where I am debating it.

What I tried to express in my post #29 is simply that I had not understood what you meant because I had not understood that you had shifted to a purely logic discussion, while I was focused on an empirical onr. For me, that is a very important distinction. Again, I am perfectly aware that you acnnot prove an hypothesis logically false by a negative test of the kind we were discussing, but that was never my purpose. I quote from my post #29:

“I think there are still some misunderstandings, so I will clarify better my position.

First of all, I am discussing empirical science, and not a mathematical theorem. You should understand the basic difference. In empirical science, and especially in biology, we are looking for “best explanations”, not for absolute demonstrations.”

Does that make sense? Is that intellectually dishonest? I mean simply that, while to prove something false in mathemathics you would need an infinite set of tests (and that’s why mathematicains try to demonstrate theorems logically, and not empirically), a series of repeated failure in trying to support a theory empirically in empirical sciences has great importance for the credibility of the theory. In other words, empirical theories need empirical support. If they cannot find it, that’s not a good sign, even if one can always hope that sometime, somewhere, that support will be found.

Why do you object to that concept? I find it simple enough. Again, I am in no way denying your point that “negative tests are not, in fact, logically valid tests of a hypothesis”. I fully understand this, and I admit it. As you correctly point out, I am so convinced of that that I “pointed out so clearly exactly what the logical flaws are”. And that was not because I “forgot that we were analyzing a statement made by you rather than one made by me”. It was just because I believe that true. But I had not realized what you implied about those conclusions, and I have tried to clarify why we were arguing different things from a similar conclusion. Again, I don’t think there is anything intellectually dishonest in that. The key is in your word “logically”. If you have read something of what I have written here at UD in the last months, you should know that I always make clear distinctions of the level of the discussion, strictly separating logical arguments from empirical ones.

Moreover, I was frank just from the beginning, affirming that some of my tests were positive and other were negative. For some reason, you have stuck to the “logical” (and IMO non relevant) problem with the negative ones, and never discussed the positive ones. I could suspect you of intellectual dishonesty for that, but I don’t, and I accept your explanation that you had not time.

Finally, I have tried many times to point out that predicitons are not the “basis” of a hypothesis. Explanation of known facts and internal consistency are the true basis of a scientific hypothesis. Good predictions are a “support” for a hypothesis. I can’t understand why you don’t want to recognize that a negative prediction, if repeatedly verified, is an empirical support for the predicting hypothesis, a potential falsification of the same, and an empirical lack of support for the alternative hypothesis, which would require the negative test to become, sometime somewhere, positive.

My best wishes to you.

Now, I freely admit that there are many points that you brought up throughout this debate that were simply ignored by me, not the least of which is the fact that not all of your tests are “negative”, which I will go ahead and acknowledge now. I ended up ignoring so many points because the time involved with addressing all of them would be immense. Entire books have been written addressing those points, both for and against. I could not hope to address them all in this forum, so I chose instead to focus on one point that I thought could be made particularly clear in the relatively small amount of time that I have for this kind of exchange.

Having said all of that, I don’t think that I will continue our debate on any of those points, nor will I debate you in the future. “Why?” you ask. Because in the process of this exchange it was clearly demonstrated why negative tests are not, in fact, logically valid tests of a hypothesis. I am quite certain that you fully understand this, whether you admit it or not, because it was you who pointed out so clearly exactly what the logical flaws are. In part I think it was because for a brief moment you forgot that we were analyzing a statement made by you rather than one made by me. The result was that you did not pass it through your “it must be right, so how can I interpret it that way” filter, but actually looked at it critically. The filter was obviously reinstated in your very next post, when you suddenly reject your own stated conclusion as an “extreme and unwarranted logical possibility,” reverting instead to your previous stance as though the point had never been made. This indicates a level of intellectual dishonesty that I simply cannot accept and will not implicitly condone by continuing to pretend that you are taking an honest look at ID and CSI.

Please feel free to interpret this as “I’m afraid to debate with you on those other points,” if it helps.

I think I understand what you are saying better now. But still I cannot agree.

Let’s try to make some order.

In #16, answering your question about my affirmation that

““We have no reason to assume that the designer of biological information must similarly exhibit CSI.”

I gave a somewhat more detailed model of why I affirmed that.

In #17 you, without debating my model in any way, asked for testable predictions. Wel, I gave you a few of them in #18, and we have been debating a little about those.

I think there are still some misunderstandings, so I will clarify better my position.

First of all, I am discussing empirical science, and not a mathematical theorem. You should understand the basic difference. In empirical science, and especially in biology, we are looking for “best explanations”, not for absolute demonstrations.

So, I repeat again: let’s start from the facts:

Fact 1: Human beings easily generate CSI

Fact 2: Human beings are conscious

Fact 3: Machines and computers known to us cannot generate CSI and, notwithstanding the incredible increase of power and complexity in our computers and software, we are in no way near to accomplish that.

Fact 4: Machines and computers known to us are not conscious.

Fact 5: Humans include CSI in their structure.

Fact 6: Machines and computers known to us include CSI in their structures.

Now, you can interpret those facts according to different explanatory models. Your model, if I am right, is that both consciousness and the capacity to generate CSI are in some way dependent on the existing CSI, on its quality and quantity. If you find the right algorithm, then consciousness will arise, and new CSI will come out of the existing one.

What can I say about that? I believe that model is both inconsistent and unsupported.

It is inconsistent because there is no explanation of how some special algorithm, still to be found, should be capable of generating CSI, while all the others known cannot do that. After all, algorithms are base on the same logical principles. If, as I believe, the impossibility to generate new CSI is soundly based on mathematical and statistical reasons (see Dembski), then no mathematical tool, no algorithm, can do that.

But the main point is that your theory is empirically unsupported: there is no evidence that any existing algorithm can generate consciousness or CSI, so why should we believe that?

So, no convincing rationale, plus no supporting evidence = no credible empirical theory.

And is your theory falsifiable? No, because, as you said, you can always appeal to the remote possibility that some unknown algorithm could one day succeed. That’s exactly what makes of your theory not a scientific theory in a Popperian sense.

But if you accept that your theory could be falsified in a empirical sense, in other words if you accept that your theory must be supported in a reasonable time by results in the field of AI which must show some relationship between existing CSI, consciousness and the ability to generate new CSI, and if you rnounce the appeal to extreme logical unproved possibilities, then your theory becomes a true scientific theory, although not a credible one (for the reasons expounded before).

Now, let’s go to my model: my model assumes that two of the observed facts (consciousness and the capacity of generating CSI) are necessarily connected, in the sense that consciousness is necessary to generate new CSI.

Is that model consistent? You bet. It is a very simple causal connection between two events which are always observed together: each time we observe a design process (generation of new CSI) there is a conscious being behind it. You could object that biological information exhibits CSI, but we are not sure if a conscious being is behind that or not. I agree, but that’s exactly the core of the ID – darwinism question, so I think we have to leave that alone for now, and suppose we just don’t know. But any time we can observe the real process of CSI generation, a conscious being is behind it.

Is my model supported by facts? You bet. If you believe that humans are conscious intelligent beings, you can observe that they can very easily output CSI in almost any condition. And no contrary evidence, of CSI coming into existence by itself, or just by previous non conscious CSI, has ever been observed.

So, my model is consistent with known facts, logically consistent, explains very well what we daily observe, is a good potential explanation for biological information, and is supporte by known facts.

Is it falsifiable? You bet. We can certainly agree that the deminstration of new CSI being generated by a non conscious machine would falsify it. So, I am perfectly at ease with Popper.

And the predictions? I insist that all my negative predictions are very significant empirically, in the real scientific world, where we are looking for best explanations according to the facts we have or we can reasonably gather. In the real world, any new failed attempt to generate consciousness and/or CSI by a machine is a further argument in favor of my model and against yours. But if you want to stick to extreme and unwarranted logical possibilities, that’s your choice.

In that case, I would just leave you alone, or remind you one last time that anyway some of my predictions are not negative at all. I paste here my argument about that from my post #23 just to show that I am always sticking to the same arguments, and on those arguments I would like to get a response from you.

From my post #23:

1) Not all the “tests” I proposed are “negative”. Point c), for instance, is positive and empirical, and can certainly be understood more deeply as our knowledge of those kind of phenomena increses. Point d)
is positive too, and open to empirical investigation, although our scientific approach to subjective experience is still rudimental. Point f) is empirical too, although probably more difficult to model. Finally, the last point (which should have been g), I realize, but was labeled e) again for a mistake) is again empirical and positive, and your objection that it would not demonstrate the “intelligence without CSI” hypothesis is not wholly correct, because the purpose of the test would be that intelligence, even if operating through existing CSI, can generate “new” CSI, which is exactly what machines cannot do.

So, while we wait for machines which are conscious, or which can produce genuine new meaning and language and intelligence, the failure of the materialist model is the only empirical clue we have. Any new model is welcome. In the meantime, independent existence of a simple consciousness and design remain the “best explanation”, also in the light of the above positive points, which must certainly be actively pursued in further investigations.

“isn’t “the set of all possible algorithms usable by non conscious machines” an infinite set?” “So, affirming that “the test is necessarily incomplete until all possible algorithms have been tested for CSI generation” is the same as saying that the test will always be incomplete (unless we find a positive solution)”

You make my point for me. Basically I’ve only restated your test case in different words, and the problem that you highlighted for my statement applies equally to yours. After all, if you say that “Non conscious machines (like computers) will never be able to generate new CSI,” you must necessarily consider all possible non conscious machines to be able to say “never”. If you limit your considerations to anything less than all possible non conscious machines, then you cannot say “never”, but only “not for those machines being considered”. In that case, I think you would need a very good logical reason for excluding any given non conscious machine from consideration.

“Should we therefore believe by faith for an infinite time that non conscious machines can generate CSI , while we wait for a test which cannot be completed to be completed?”

The hypothesis being considered is not that non conscious machines can generate CSI but that they cannot do so. We need not assume that they can just because we cannot prove that they cannot.

I don’t understand: isn’t “the set of all possible algorithms usable by non conscious machines” an infinite set?

So, affirming that “the test is necessarily incomplete until all possible algorithms have been tested for CSI generation” is the same as saying that the test will always be incomplete (unless we find a positive solution). Should we therefore believe by faith for an infinite time that non conscious machines can generate CSI , while we wait for a test which cannot be completed to be completed?

Shouldn’t you check your logic? Or just your epistemology?

By the way, for those interested, some discussion is going on between me and Mark at his blog (link at #22).

“You comment makes some sense, but it is just another way of “sticking to a vague and unsupported hope that, sooner or later, you could find some support for your model”.”

Allow me to restate test (a) more precisely so as to highlight exactly why I think that this is the only valid stance: “Given the set of all possible algorithms usable by non conscious machines, no particular algorithm will be found among them which is capable of generating new CSI.” Stated in this way it should be obvious that the test is necessarily incomplete until all possible algorithms have been tested for CSI generation, or until one has been found which does, in fact, generate new CSI. What this means in practice is that each attempt to generate new CSI using a non conscious machine is not an independent test itself, but is in fact only a single step in the larger test. In the meantime, we are asked to assume that the hypothesis is valid by default while the test is being run.

In the interest of time and space I will address other points in future posts.

thank you for your note. It is late here now, but I will certainly read your piece, and hopefully give some feedback, tomorrow.

You comment makes some sense, but it is just another way of “sticking to a vague and unsupported hope that, sooner or later, you could find some support for your model”.

The materialistic model is that consciousness and CSI can be generated by algorithmic methods. If no known algorithmic method can do that, then the model, for the moment, is unsupported.

I agree with you that, in that case, “a paradigm shift may be necessary”. Well, acknowledging the independent existence of consciousness, and the reality of design, is indeed a paradigm shift. You may not like it, but it is a kind of paradigm which is very ancient and has been shared by most thinkers in human history. And which explains very well what we observe.

Just a couple further notes:

1) Not all the “tests” I proposed are “negative”. Point c), for instance, is positive and empirical, and can certainly be understood more deeply as our knowledge of those kind of phenomena increses. Point d)
is positive too, and open to empirical investigation, although our scientific approach to subjective experience is still rudimental. Point f) is empirical too, although probably more difficult to model. Finally, the last point (which should have been g), I realize, but was labeled e) again for a mistake) is again empirical and positive, and your objection that it would not demonstrate the “intelligence without CSI” hypothesis is not wholly correct, because the purpose of the test would be that intelligence, even if operating through existing CSI, can generate “new” CSI, which is exactly what machines cannot do.

So, while we wait for machines which are conscious, or which can produce genuine new meaning and language and intelligence, the failure of the materialist model is the only empirical clue we have. Any new model is welcome. In the meantime, independent existence of a simple consciousness and design remain the “best explanation”, also in the light of the above positive points, which must certainly be actively pursued in further investigations.

2) I have complied with your requests for predictions and tests, because that is certainly a valid aspect of the discussion. But I would like to remind you that predictions and tests are in no way the only criterion of science. The main properties of a scientific theory remain:

a) to be able to explain as well as possible known data

b) to be internally consistent (logically and mathemathically)

There are a lot of reasons why the theory of an independent transcendental and simple I explains the phenomena of consciousness and of subjective experience much better than any materialistic theory: we have discussed them extensively in other threads, and they are pertinent to the adequacy and consistency of the models.

Another property, if we accept the Popperian view, is falsifiability. Well, I must emphasize again that all the “negative” tests I have suggested are potential ways to falsify my view: in other words, if and when you will produce a machine which freely generates consciousness and CSI as the human brain is supposed to do, you will have falsified my assumption that machines cannot do that, and that a transcendental I is necessary to have that result.

On the contrary, your assumption that machines can accomplish that becomes really unfalsifiable, as long as you stick to your position that:

“It is possible that our current method of computation is fundamentally flawed if the goal is a conscious machine that can create CSI. If that is the case, a paradigm shift may be necessary before we even begin to approach such a thing, and for all we know the necessary shift may not occur for generations.”

That invocation of a vague possibility of a future knowledge, about which we have no idea or clue, and which could realize what our present knowledge cannot do, is exactly the kind of position which becomes unfalsifiable, and therefore, according to Popper, non scientific.

I wrote this small entry in my own blog which you might be interested in:

http://mark_frank.blogspot.com.....alism.html

Mark