eric: The fact that "there is a causal link" is a truism. It is always true.

Now you got it! But note that many Intelligent Design advocates disavow this view.

eric: Therefore, that bare fact contributes nothing to the fact that we do infer intelligence.

Exactly wrong. Knowing there is a causal link, we can make specific and distinguishing empirical predictions.

eric: There would be a causal link whether "Bradford" was human or not. There would be a causal link whether "Bradford" used the internet or not. There would be a causal link whether "Bradford" was an intelligent agent or not.

Now you got it! And we can use this knowledge to help determine "Bradford" identity by forming and testing hypotheses. In a legal investigation, we call this "forensic science".

eric: Nevertheless, without making any forensic investigation whatsoever and without needing to elaborate the nature of the causal link, people will readily infer that the messages come from some intelligent source. The truism that there was a causal link contributes nothing to this inference.

You just admitted to abandoning science.

Forensics is a branch of science (having to do with legal investigations). You are saying that without making any scientific investigation, people may very well 'intuit' that "Bradford" is intelligent. We have a great deal of experience with messages written in English, and from this great store of knowledge, we would probably make a tentative claim that "Bradford" was a human or a human proxy (such as a computer). That's why forensic scientists keep vast libraries of crime cases. But forensics is a systematic and scientific approach to what people often do intuitively.

eric: The truth of a fact that is true in any and every case cannot be used differentiate between the cases.

That's just ridiculous. Think about it again. You're saying that knowing there is a causal link between the sender, the medium and the message is irrelevant. But that's the whole point of forensics"”to trace the causation between the act and the perpetrator.

eric: The point about "unless science turns up something" is to acknowledge that science continues "” not stops "” and that our inferences to the best explanation are tentative in nature.

No. It's just how you wave your hands. If you were doing science, you would point to specific and distinguishing empirical predictions.

eric: 1. Consistent positive observation
2. Consistent absence of negative observation
3. Theoretical considerations favoring
4. Theoretical considerations against the contrary hypothesis
5. Search for testable counter-examples

These do not qualify as testable hypotheses. Do you know why?

About the "Bradford" Turing Test, eric wrote: In short, we don't know as much about the "causal link between the message, the medium and the sender" as you assume.

Zachriel: We know there is a causal link"”a fact that is used by forensic scientists everyday.

The fact that "there is a causal link" is a truism. It is always true. Therefore, that bare fact contributes nothing to the fact that we do infer intelligence.

There would be a causal link whether "Bradford" was human or not.
There would be a causal link whether "Bradford" used the internet or not.
There would be a causal link whether "Bradford" was an intelligent agent or not.

Nevertheless, without making any forensic investigation whatsoever and without needing to elaborate the nature of the causal link, people will readily infer that the messages come from some intelligent source. The truism that there was a causal link contributes nothing to this inference.

The truth of a fact that is true in any and every case cannot be used differentiate between the cases.

Zachriel: You are STILL putting forth this odd view of science that we can reach conclusions and stop. Your job is done unless "science turns up something" that you are willing to consider. But that is not how science reaches its conclusions.

Nope. I'm observing that an inference to the best explanation is reasonably based on uniform observational evidence combined with consistent indications from considerations in principle. The point about "unless science turns up something" is to acknowledge that science continues — not stops — and that our inferences to the best explanation are tentative in nature.

Face it — you are the only one who keeps claiming that there is some kind of stop.

eric wrote, concerning the basis of making an inference and the opportunities to look for counterexamples:

1. Consistent positive observation: The fact that intelligent agents can produce and use language has been observed to the point of being beyond doubt.

2. Consistent absence of negative observation: We can also observe that there are no established contrary cases to the universal observation that language depends on intelligent agency.

3. Theoretical considerations favoring: We can consider in principle what enables intelligent agents to use symbols and symbolic agents, which point us to abilities of the mind such as imagination.

4. Theoretical considerations against the contrary hypothesis: We can consider in principle that non-intelligent, undirected processes do not have the attributes of imagination and mind and so do not have access to the methods of creating and using language that intelligent agents employ.

5. Search for testable counter-examples: Finally, we can continue the investigations into the possibility that there could be some other path by which an undirected, non-intelligent process could create and use symbolic language. So far, such attempts have been so barren that we cannot yet even devise a testable scenario whereby such processes could be reasonably expected to mindlessly build the necessary machinery for symbolic translation.

At present the evidence, both observational and theoretical, are entirely in support of the Turing Test and the link between symbolic language and intelligence.

You still putting forth this odd view of science that we can reach conclusions and stop. That may work fine for someone who isn't all that inquisitive. But in science, we make tentative conjectures, infer specific empirical implications, then test these implications. From these results, we modify our hypothesis accordingly, then repeated the process in order to refine our knowledge. If you just reach a conclusion and stop, then you aren't doing science. But you might be doing ID.

Or a charicature of ID. We might, for example, explore, as have others have, the question of what would constitute a minimally functional genome? One could have an ID perspective on this or a standard approach. Either way experimental results providing the identity of an organism and the minimal number of genes by which it is able to reproduce, could be viewed as favorable to one of the two distinct perspectives and tentative.

eric: Consequently, the inference is warranted, unless and until science turns up something that overturns the uniformly solid evidence for this relationship.

Hence, on that basis, intelligence is also reasonable as an inference for the symbolic language needed by biological life.

You are STILL putting forth this odd view of science that we can reach conclusions and stop. Your job is done unless "science turns up something" that you are willing to consider. But that is not how science reaches its conclusions.

Bradford: This is a good example of an inference based on matching the properties of a biological entity to a causal factor linked with the generation of systems having this property.

So you have a hypothesis. Now what specific and distinguishing empirical predictions can you make and test?

Examples: Halley predicted that a comet would reappear, naming the year and quadrant. Darwin predicted that a few million years ago there existed not-quite human apes, and that they probably orginated in Africa. Mendeleev predicted the existence of heretofore unknown chemcial elements and their properties. Einstein quantitatively predicted the curvature of starlight as it passes near the Sun.

eric: The question remains "” is it realistic?

Avida isn't meant to be a realistic simulation of biology. Rather evolutionary algorithms are a mathematical class and Avida is a member of that class.

eric: In one evolutionary model, acquiring surplus instructions counted as "food" regardless of whether the instructions did anything or not. Though useless, they were treated as beneficial. Why?

You really need to read the primary sources. Longer genomes take proportionally longer to replicate, so the SIZE_MERIT_METHOD, if proportional to genome length, simply makes all genomes take the same time to replicate. By adjusting this value, we can also experiment with the tendency of the genome to collect junk.

As most people would also acknowledge, it is the our universally consistent experience (with reasons in principle to support it) that symbolic language requires and therefore implies intelligence. Consequently, the inference is warranted, unless and until science turns up something that overturns the uniformly solid evidence for this relationship.

Hence, on that basis, intelligence is also reasonable as an inference for the symbolic language needed by biological life.

I agree Eric. This is a good example of an inference based on matching the properties of a biological entity to a causal factor linked with the generation of systems having this property.

Raevmo Says: Mathematical models have shown that mutation and natural selection can produce solutions to complex problems. I wouldn't call that feeble at all. I would call it amazing and beautiful. People without some minimal training in mathematics tend to underestimate the complexity that can be generated by simple rules.

I commented a bit about this earlier in this post. Behe has something to say as well.

I don't know if you've gotten your The Edge of Evolution yet, but when you do, don't miss the last section of the last appendix, i.e. section Computer Assumptions within Appendix D: The Cardsharp.

In it, Behe discusses how mathematical models of biology and evolution can easily be unrealistic.

In one evolutionary model, acquiring surplus instructions counted as "food" regardless of whether the instructions did anything or not. Though useless, they were treated as beneficial. Why?

As explained on the Avida website, the counterbiological feature was needed, "Otherwise there is a strong selective pressure for shorter genomes." In other words, otherwise the program wouldn't give the desired results. The computer programmers remark, "This isn't the most elegant fix, but it works."
- The Edge of Evolution, p. 276

This is an example of why I am not overly impressed by the ability to make a model that behaves in a manner that fulfills expectations. The question remains — is it realistic?

eric: The Turing Test says "No, it does not. We can infer intelligence from the message content, regardless of whether the sender is human."

This is exactly contrary to Turing. He sidesteps the problem of defining intelligence by proposing a modification of a game of purposeful deception.

eric: Even in the face of the possibility of a non-human A.I., the Turing Test would agree with most people's normal assessment, i.e. that we can reasonably infer intelligent agency from the content of the messages themselves, regardless of whether "Bradford" is human or not.

Again, this is false. The Turing Test does not test for some vague notion of "intelligence", but specifically for human-like intelligence.

eric: What matters is the capacity to meaningfully use symbolic language. This is the central point of the Turing Test.

I don't think you have read, or if you have read, understood Turing's article.

eric: In short, we don't know as much about the "causal link between the message, the medium and the sender" as you assume.

We know there is a causal link"”a fact that is used by forensic scientists everyday.

Zachriel: The Turing Test assumes that the computer is designed to "imitate an adult human mind". So the test is actually to distinguish between a human mind and an imitation human mind.

The test sees whether a distinction can be made. If the artificial participant has sufficient mastery of language such that the distinction cannot be made reliably, then it is considered to be intelligent. It is a test for artificial intelligence.

Zachriel: You still putting forth this odd view of science that we can reach conclusions and stop.

Nope. Straw man, and already dealt with here. Are you running out of responses that you need to repeat false ones?

The scientific issue is whether symbolic language facility requires intelligence.

The fact that this relationship is recognized formed the basis for the Turing Test, and for our own inferences from language to intelligence. If this relationship were ever overturned by new evidence, that could undermine the basis for the Turing Test.

As I point out in the same post I linked to, science can continue to test this relationship. Although science is never done, at present the relationship between language and intelligence is strongly supported and without any known counter examples. For example, no one expects that static caused by pigeons or any other source would yield meaningful symbolic messages.

Zachriel: We immediately infer "Bradford" is a human (or its proxy) or a clever imitation of a human. We further infer that he accessed the Internet through the normal means.

If you mean to say we can infer "Bradford" accessed the TelicThoughts servers via the internet, that is an invalid inference. You have no knowledge that tells you how "Bradford" accessed those servers. If "Bradford" is an A.I., it wouldn't need to use the internet at all. (The fact that you use the internet to fetch the stored messages from the servers is, of course, irrelevant.)

In short, we don't know as much about the "causal link between the message, the medium and the sender" as you assume. You don't know the physical nature of the sender or the manner of construction of the message or the nature of the transmission.

Neverthess, it is still obvious to everyone without further investigation that there is an intelligent source. The messages provide sufficient evidence.

Zachriel: When we read the message, we are making valid inferences about the sender far beyond a simplistic assertion of "intelligence".

Since you reckon that the valid inferences extend "far beyond a simplistic assertion of "intelligence"", I will accept that as an acknowledgment that it is valid to at least make the inference from language to intelligence. Thank you, that is exactly what I was looking for.

As most people would also acknowledge, it is the our universally consistent experience (with reasons in principle to support it) that symbolic language requires and therefore implies intelligence. Consequently, the inference is warranted, unless and until science turns up something that overturns the uniformly solid evidence for this relationship.

Hence, on that basis, intelligence is also reasonable as an inference for the symbolic language needed by biological life. Tentative, yes. Science still working, yes. But already a reasonable inference today. The established relationship stands until it is overturned.

Zachriel: How do we know it's a message? Is it random gibberish? Shakespeare? You're positing a situation without any detail whatsoever. "¦

eric: I posited the very detailed situation of the "Bradford" messages as an example of the Turing Test.

Would it be fair and accurate for me to say that your "Yes" means you think it is reasonable to be skeptical that the "Bradford" messages are from an intelligent source (whatever else may be true), and that you would want to carry out an investigation (e.g. doing traces on the messages) before we could reasonably infer that the source was intelligent?

When we read the message, we are making valid inferences about the sender far beyond a simplistic assertion of "intelligence". We immediately infer "Bradford" is a human (or its proxy) or a clever imitation of a human. We further infer that he accessed the Internet through the normal means. And that we can investigate the message in more detail if we choose. That's because there is a causal link between the message, the medium and the sender.

It's not that complicated and is a day-to-day occurrence in the forensic sciences. But because in normal life, much of this process is intuitive, you forget you've done it.

Zachriel: "¦ If we receive a message in English from a pulsar, we would normally check our equipment. This claim can be investigated.

eric: ..most reasonable people would still conclude that whether the message came from the pulsar or from Earth or from somewhere else in between, it still would have come from an intelligent source.

They would infer this even without the benefit of further investigation.

They would infer that it probably came from Earth, and then would investigate further to confirm this fact. Maybe it's the pigeons.

You still putting forth this odd view of science that we can reach conclusions and stop. That may work fine for someone who isn't all that inquisitive. But in science, we make tentative conjectures, infer specific empirical implications, then test these implications. From these results, we modify our hypothesis accordingly, then repeated the process in order to refine our knowledge. If you just reach a conclusion and stop, then you aren't doing science. But you might be doing ID.

eric: This is the central point of the Turing Test.

The Turing Test assumes that the computer is designed to "imitate an adult human mind". So the test is actually to distinguish between a human mind and an imitation human mind.