Recently I described a possible alternative Turing test that looks for certain non-person like behavior. I’d like to try it out and see if it is robust and has any value.
Here is some data represented in an ordinary control chart.
It’s easy to see there is a recognizable pattern here. The line follows a downward trajectory till about observation 16 then it meanders around till about observation 31. At which point it begins an upward track that lasts till almost the end of the chart.
The data in question is real and public but obsolete. I won’t say what it’s source is right now to avoid any bias in your attempts see if we can infer design. I can provide the actual numbers if you like.
The question before the house is. Is the clear overall pattern we see here a record of intention or does it have a “non-mental” cause?
Using the criteria discussed in my “poker” thread I would suggest looking for the following behaviors and excluding a mental cause if they are present.
1) large random spikes in the data
2) sudden changes in the overall pattern of the data that appear to be random
3) long periods of monotony
4) unexplained disjunction in the pattern.
I have some other tests we can look at as well .
What do you say design or not? Are you willing to venture a conjecture?
peace
why, That is exactly what it was. I can’t help it that you read the word design and freak out
I don’t think so I think it’s a good test but perhaps still incomplete.
you can’t call my test a failure until you describe a test that would do a better job and we all know you are unwilling to do that don’t we
peace
on that I will agree
no, for me truth is not established but only discovered and acclaim and agreement are definitely not the ways that is done
peace
People wouldn’t have to try so hard if you operationalized your terms and provided a real testing mechanism.
That’s patently ridiculous. Your test is quite capable of failing all on its own.
That doesn’t make logical sense. If you are proposing a test that cannot work, my inability to propose a test that can does lend any credence to yours.
Sure, but unless you demonstrate that a Turing test that is not constrained by conversation is logically impossible my test is the best one we have going right now
😉
peace
did you read the paper I linked?
peace
Gary Gaulin?
Glen Davidson
Are you claiming that a Turing test that looks at behavior is impossible?
If so then you need to support your claim
If not then mine is the best option on the table right now. You can either offer suggestions to improve it or come up with your own test
I’ll be waiting
peace
Does he have a alternative Turing test that I’m unaware of ?
I know he has some kind of simulation that he goes on about.
peace
fifth:
Poor fifth. Never any real results, but always on the verge of a breakthrough.
The misunderstood genius.
A Turing test is looking at behavior. Yours is looking at a graph. Yours is not a Turing test.
Nevermind. As you were.
fifth:
To which your “solution” was to argue that God allows dogs to eat the heads of living babies, and other atrocities, because otherwise he would appear to be “winking at sin.”
I renominate you for World’s Worst Apologist.
Because Poor Misunderstood Geniuses are always right, and always on the verge of a breakthrough, but always unfairly scorned by the lazy and doctrinaire critics who can’t comprehend the Great Insights being offered by the PMGs.
The point of operational definitions is that you don’t have to think. With operational definitions, you will know whether you have a test or not. As it is, you are operatively ensuring that you don’t have a test while you claim to have it and you hope to convince the audience by bone-headed repetition. Quite an executive disorder you have there.
Silly, FMM. The way Turing test works is that you don’t know whether it’s a bot. You know that it’s either a bot or a person. You also know the behavior, and from the behavior you are supposed to judge whether it’s a bot or a person.
In contrast, you didn’t give us a behavior. You gave us a graph. Do you understand how fundamental absolute category error you are committing? Of course you don’t, otherwise you’d not be committing it. Or perhaps you’d sit ashamed in your corner. Nah, strike that last option.
The importance is this: Say you have a photograph of a dog. Examine it carefully. Are you looking at a living entity? I say you’re looking at glitzy paper. It’s a photograph.
But you are saying you’re looking at a dog because you are not making a distinction between an entity and its representation. You are not making a distinction between a behavior and a graph. The Turing test is about behavior or its simulation. It’s not about a graph.
I would not have had to say this, if you were minimally science-savvy, say on the level of a fifth-grader. Unfortunately you’re not. It would be funny if it weren’t so sad. So long.
Just to step back from the rising heat, let me offer this:
I agree that there are cases where one can conclude from inspecting a graph if the source of the data is a human being, or more likely a mechanistic process. An example of the latter could be a sine graph (sure, a human could have plotted the numbers, and students do, but apart from that there is a very good chance that you are looking at the output of some kind of generator). An example of the former could be what I already mentioned earlier, a graph of dots that, when connected, show a picture of Donald Duck.
Having said this, there is a huge grey middle area where simply looking at a graph will not tell you this, with any degree of confidence. Mechanistic processes can come in any form, regular, not so regular, irregular, with short patterns, long patterns, subtle patterns, obvious patterns. Likewise, human activity can take many forms, again, regular, irregular, with clear patterns or without. In this middle zone, any guess is likely as good as any other in the absence of context.
Your graph falls squarely in this grey area, as do mine.
What doesn’t help is leaving your key terms undefined, so that there are no objective criteria by which to classify either the graphs themselves, or their sources. Because of this the exercise boils down to the trite and tiresome ‘I think it is designed because it looks like it’ argument.
Let me make these suggestions on how this could be developed into a more useful exercise:
As possible sources, the choice would have to be between graphs displaying processes that involve human beings (somewhere, anywhere), and processes that occur in nature without any human involvement. Both could be further subdivided and classified. Drop the term ‘mind’ – it creates far more confusion that clarity.
Machine outputs are problematical, because one could always claim that a machine is a human product and therefore shows traits of ‘mind’.
Instead of one graph, you would want to test a great many – dozens, hundreds, from different sources in both categories.
Instead of looking at the test results of individual people, you would use many testers (dozens, hundreds, from various backgrounds) and aggregate their results, so that one can do statistical analysis to see how often they get it right, and on what types of graphs.
If you set up your exercise like that, it might be interesting. As presented, it is useless for reasons many here have explained.
By the way, my data set ‘A’ is the monthly average rainfall in Oxford, UK, in the years 1853 – 1859. Data set ‘B’ (the longer version) is voter turnout in UK Parliamentary Elections between 1945 and 2015.
‘A’ would fall squarely in the class of natural sources unaffected by humans (leaving aside human induced climate change for now!), whereas one can argue about ‘B’ – sure, the individual choice to go and vote is a mindful action, but when does the mindful action of an individual turn into the mindless (and somewhat predictable) behaviour of a crowd (which is actually an interesting question in its own right). Classify ‘B’ as a ‘human process’ instead and the issue disappears.
Graphs are produced by humans, so both graphs are designed, products of “person-like behavior” or whatever FMM wants to call it. Without rigorous delimitations in the procedure, such as where precisely the “source of data” phase ends and “processing of data” begins, the entire exercise is blatant nonsense.
As you yourself say,
For the exact same reason, any output whatsoever is “problematical”. A dog is a living entity. For theists, dogs (like all nature) is God’s creation (“designed”). A photograph is a piece of paper, a lifeless thing, but manufactured by humans. Also designed. What is not designed?
In another thread, I was told to affirm something like “we know that cars are designed and from there you can go on detect other things that are designed”. This suggestion would have a meaning if, similarly as we “know” that cars are designed, we would be offered some example of which we “know” that it’s not designed. Is there such an example? What is that thing which is not designed? Can we ever get an answer what “(not) designed” means in the framework of this allegedly scientific theory? Context and definitions, sir!
exactly !!!!.
Turing tests are all about thinking. If you had operational definitions an algorithm could do it and that would defeat the purpose of the test.
Yet Turing tests work just fine
peace
Graphs map behavior that is their purpose
In a standard Turing test when you are reading the pixels on a screen are you observing the behavior of the bot or just pixels.
Your problem is that every objection you raise against my test works against a standard Turing test and yet Standard Turing Tests work just fine.
peace
That is an excellent comment thanks. It provides the sort of feed back I’m looking for and wished would have come at the beginning
I completely agree. My test like any Turing test will never be infallible and will only work for obvious cases not borderline ones.
The perfect should not the enemy of the good
Peace
You may be a bit harsh here. What is under discussion is the data stream that is the input into the graph. The graph is just a visual display to make it easier to see trends and patterns. Assume a machine that has a facility to output diagnostics about whatever it is doing. I don’t think it matters for this discussion if there was someone who read the diagnostic values off a dial and plotted them manually, or if the machine had a printer attached to it that produces the graph automatically.
Perhaps theists have it harder here than non-theists? I would consider natural data such as average monthly rainfall a non-designed variable. Do theists have a different view?
Most human output would be designed (with exceptions like doodling during a meeting, or talking in one’s sleep etc.). Machines are harder to classify, I agree, which is why I don’t think they should be used to test for a distinction between ‘designed’ and ‘undesigned’, or ‘algorithmic ‘ and ‘mental’ if you like, before the principle of the test has been validated on unambiguous data.
Now, if such experiments done on natural and human processes show that it is possible to distinguish between them from simple inspection of output data patterns, it would be interesting to apply the test to machines and see if they are closer to ‘natural’ or ‘human’ entities. Personally, I fully suspect that you would see a mixture of both.
But first the test needs to be validated, and that hasn’t happened yet.
I’m a Calvinist. I don’t believe in randomness. But I recognize that there is such a thing as apparent randomness.
In the same way I believe that everything is designed but I recognize that sometimes it’s less apparent and other times it’s obvious.
peace
faded_Glory,
I like this new Faded Glory 😉
peace
Instead of random examples, give a definition that makes the relevant distinction and nobody regardless of affiliation will have a problem. Just a matter of common sense. But we don’t have common sense here. It boils down to “I’m a Calvinist. I believe this and I don’t believe that…”
It is the same one. I entered the conversation by requesting that you define your terms, provide operational definitions. You claim that this is unnecessary. I maintain that this makes your entire exercise a waste of time. I’m not the only one.
Stretched out over a number of posts, put up in response to various criticisms, you have ended up at the position that just from inspecting your graph, without other info, it is somehow possible to conclude that the increased losses of the submarines are more likely to be caused by breaking the Enigma code than by poor QC during ship building. I am telling you that this breathtaking nonsense, and really, you ought to accept that.
Perhaps the main thing holding your back is your defensiveness. Critics are not always wrong. You might end up in a better position if you could get yourself to agree with them when they are reasonable, and modify your claim. The alternative is to keep digging more and more and deeper and deeper holes, until you end up in a ridiculous position. It puts people off.
We already established that you have no idea what a Turing test is.
But the official source recognizes apparent design, “The theory of intelligent design is simply an effort to empirically detect whether the “apparent design” in nature acknowledged by virtually all biologists is genuine design (the product of an intelligent cause) or is simply the product of an undirected process such as natural selection acting on random variations.”
So there’s apparent design and genuine. The difference between the two is…?
Yet another question nobody ever answered. The distinction is thrown out there, but nobody defines it. That’s the only method there is to the entire ID movement.
I know you think I’m repeating myself but the possibility of philosophical zombies means that there is no line you can draw and say that everything on the left is designed and everything on the right is not.
Despite that and despite the huge gray area that FG mentioned there are still some things which everyone can agree exhibit the behavior we associate with mind
peace
Again I think you are too fixated with ID. Your disdain for certain individuals is hindering your ability to step back and look at this with an open mind
It would be a lot better if you acted like this had nothing to do with ID
peace
I have no idea nor do I care. It’s irrelevant to the task at hand
I do think that there are cases where it’s more apparent that an object was designed.
peace
Does a dog that buries a bone exhibit the behavior we associate with mind?
Does a plant that turns its flowers towards the sun exhibit the behavior we associate with mind?
Does a crow that makes a tool to reach food exhibit the behavior we associate with mind?
Does a psychiatric patient who bangs his head against the wall for hours at the time exhibit the behavior we associate with mind?
Do you see where I’m going with this?
Do you think that Turing provided operational definitions in his original paper?
If so please show me where.
If not do you think that this makes Turing tests a waste of time?
peace
Section 2, heading Definitions
https://www.cs.virginia.edu/~robins/Turing_Paper_1936.pdf
You can stop pretending you know what you’re talking about. For your own good.
ETA Or if we take the 1950 article, read the second sentence. What does it say? Warns you of dictionary definitions, doesn’t it? https://www.csee.umbc.edu/courses/471/papers/turing.pdf
The only dubious consolation here is that you’ve probably wasted your own time more than ours.
The Turing test is meant to see if we can tell humans and machines apart when all we have is their output.
He doesn’t give an operational definition of ‘humans’, and I hope we can agree to forgive him for that.
He goes at considerable length to clarify what he means by ‘machines’ in the context of his game. See sections 3, 4 and 5 of the paper.
COMPUTING MACHINERY AND INTELLIGENCE
So yes, he does very much define and clarify his terms.
You, on the other hand, are talking about ‘intention’ and ‘non-mental cause’ without further clarification. As such these are poorly defined concepts, controversial, and will lead to massive subjectivity and endless debate before you even begin your test. Failure to define these terms up front means your test will never deliver.
Yet, they have been beaten. At some level of computing power and programming sophistication, the Turing test becomes useless.
The same holds for your graphical Turing test. I doubt very much that you can provide a graph of some mindful process that will not closely resemble a graph of some natural process. And if this is the case, you don’t have a valid test.
Nothing Acartia wrote could be remotely interpreted to say that. Read what is written, not what you wish were written.
Your “test” as it stands is so vague as to be useless, regardless of whether or not any other option is available.
You completely missed Erik’s point. You’re assuming your conclusion that thinking is non-algorithmic and using that as an excuse to not operationalize your terms. That leaves you spouting literal nonsense.
There is no support for the idea the philosophical zombies are possible.
There are some things that most people will probably recognize as normal human behaviors. Leaping from there to claims about some abstract “mind” requires more support.
Data streams are produced by humans, too — or by devices designed by humans.
Sure, but that doesn’t mean we can’t isolate the human factor from the natural data we want to analyse.
So instead of design detection, we are really doing non-design extraction. Not quite sure if FMM wanted his exercise to end up this way.
no I think that is the right approach.
Remember the poker players identified non-person like behavior not behavior not behavior associated with persons.
IMO It’s easier to say what a person would not generally do than what he would do
peace
Ok then you can draw and say that everything on the left is designed and everything on the right is not. So much the better for the test.
Let me know where that line is and your presence here will be of some value. I won’t hold my breath
cool that is what the test is looking for.
It a good thing I’m not leaping from there then
peace
Thinking is non-algroythymic. When I say something “thinks” I mean that it arrives at conclusions in ways that are nonrandom and non non-algroythymic.
On the other hand when I say zombie I mean a being that’s actions are determined by an algorithm.
That is the reason that I can’t give a mathematically precise definition for things like mind and person.
Turing did not need these kinds of definitions when he originally conceived of his test and therefore alternative versions of the test don’t need them either,
Your continued insistence on these sorts of definitions for my test but not all Turing Tests is evidence that you are not acting in good faith in this conversation.
peace
I think that there is no level of computer sophistication that will be indistinguishable from a person. As computers advance will become more and more difficult but given enough time you will always be able to recognize the non-personlike behavior.
No matter if I correct or not as long as the test has not been demonstrated to be useless it is still of use
peace
What claim? I am making no claim
I am trying to develop an alternative Turing test. That is all
As part of that process I am looking for constructive criticism and am grateful when you stop trying to defeat some imagined claim and instead engage in a little conversation
peace
fifth:
The title of fifth’s OP:
Overpromise and underdeliver. It’s the PMG way.
When the wrong guesses equal the correct guesses, the test will be useless. Exactly where you currently are with your graphical Turing test. As I said earlier, there is no graph of an intelligently caused process that is not similar to graphs of perfectly natural processes. Patterns and trends don’t help because there are plenty of natural phenomena that display observable patterns and trends.
Without providing a fairly detailed context of what the graph is plotting, which pretty well tells you if an intelligence is involved, you have no hope.
This is supposed to explain something to us? Did Patrick misspell “non-algroythymic”, the term which you really prefer instead of non-algorithmic? Can we ask for a definition? Of course not, because when we do, your answer is (has been) “Ive see the lengths that normally rational people will go to avoid even thinking about things they don’t like”.
With all due respect, you are the only who is failing to think about your own exercise. Everybody else is doing the thinking for you which was really your job in the first place.
Erik:
Even Frankie eventually learned what a red squiggly line means in the comment box. Perhaps God still hasn’t revealed the secret meaning to fifth.
Perhaps there is a deeper game here than we realise? Are we performing an actual Turing test in this thread? Is Fmm posting the bot’s responses to our criticisms?
Inquiring minds would like to know…
faded_Glory,
Judging from the overall behavior, he’d be rightly called Calvinist Biblebot, but I’d say that the recent surge in his spelling errors indicates a little vulnerable human soul trapped in that shell.