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
Richardthughes,
Good job,
So what does that tell you from a statistical standpoint?
Also what is your impression, Does the pattern exhibit the behavior I listed?
Finally do you think that the cause of the data is a mental or algorithmic “non-personlike” process?
peace
Test, to see if I can link to a graph.
EDT: I guess that’s a ‘no’ then. Can someone help? I uploaded the graph to the file library here.
After you’ve uploaded it to the library, click on the thumbnail within the library. You’ll see an ‘Attachment Details’ page that includes a URL for the uploaded file.
faded_Glory,
You managed to upload the two images OK. Here are the links
Slide 1
Slide 2
Thank you, but how can I then make them show up in a posted reply here?
Sorry for being a noob 😉
faded_Glory, In a comment, you should be able to use the “upload file” facility. It will only display one image at low resolution, however. You could always author an OP. I’ll amend your member status – it seems posting multiple high res images into comments is only available in admin status. Otherwise, I can edit the images into your comment. You can PM me the details.
ETA I see you already have “new author” status.
fifthmonarchyman:
Below you see two graphs, A and B representing two different data sets. I won’t tell you much about them except that I did not make up the data. Either or both data sets could be natural, or man-made, or random.
What is your impression, do the patterns exhibit the behavior you listed?
Do you think that the cause of the data is a mental or algorithmic “non-personlike” process?
After you have given us your views, and others have perhaps chipped in, I will tell you the source of the data sets.
faded_Glory,
Your graphs are less than half the length of the one I posted and the most common complaint I received was that we needed more information to make a determination.
So, do you have anything that is longer? The old rule of thumb is that you need at least 30 data points to even see if the Data follows a normal distribution.
Surely we could agree that we need more than that to try and make a determination if the pattern exhibits certain behaviors
peace
fifthmonarchyman,
Why don’t you give it a try, or, alternatively, explain the minimum data requirement for your approach to work?
Do you see patterns in these graphs or is it random data?
We are trying to establish a minimum data requirement. I have no idea what it is
That is why I asked you all to comment on whether the 5 min cold conversation in the standard Turing test was sufficient. Apparently you could not be bothered to do so
That is one reason Ive concluded that you are not interested in helping with this enterprise.
I can give you my impressions of your charts if you like the first of which is that you have less than half the data represented that mine does.
Common sense suggests that when it comes to your charts my hunch should half or less as accurate as yours when looking at my data.
What was your determination again? I’ll need it so we can test that assumption
peace
Here is a longer extract from the same data set ‘A’. It now has more points than yours.
Go right ahead, we’re waiting.
Graph B shows all the points in that particular data set.
Edited: Actually, here is a data set related to B, this one has 33 points.
I don’t see any discernible pattern in that chart. So my first impression is that I can’t say that it was the result of a mental cause.
To be sure that there is no pattern I would need the raw data to see if I could distinguish your chart from randomized groupings of the same data.
There is an apparent pattern in graph B but with so few data points I can’t be sure whether it exhibits any of the behaviors we are looking for.
now what about your impressions of my chart?
If you don’t feel you have enough information then please comment on the sufficiency of a 5 min “conversation” in the standard Turing test
peace
OK
It looks as though perhaps this new set is exhibiting the “long monotony” behavior typical of nonmental causes though it’s still a very small amount of data.
Therefore from this sample size I could not tentatively affirm that the cause was mental
peace
I never cease to be amazed at the apparent fear some folks have of simply making a guess.
What do you think will happen to you if you are wrong?
peace
Interestingly, I do see a pattern it this chart. Check out
So here we have a bit of a problem for your approach: subjectivity.
To get round that you will have to come up with some objective tools to establish if a data set has a ‘pattern’ or not. Do you know of any such tools?
You really need to pay attention if you want to participate. I offered a tool in the very post you quoted me from
quote:
To be sure that there is no pattern I would need the raw data to see if I could distinguish your chart from randomized groupings of the same data.
end quote
peace
Lol, so far you have not made any guesses yourself – saying ‘I can’t be sure it is X’ is not making a guess, it is expressing ignorance.
If I put a gun to your head and you had to make a choice, what would you guess for my data set A – ‘mental’ or not?
And my data set B?
And I am not the one claiming it is possible to infer ‘mental’ input from graphs like these, so why would I make a guess? It would be based on nothing and therefore be worthless.
Subjectivity will never be eliminated when it comes to this.
After all subjective experience is what it is that makes us persons in the first place.
If it were wholly objective we could create an algorithm to do it
All we can do is try and minimize our own bias where possible. That is what this enterprise is attempting to do
peace
So far we haven’t really got anywhere, haven’t we?
I showed you two graphs of different data, and all you can say is ‘I don’t know’.
If you had shown me the same graphs without me knowing their background, I would also have said ‘I don’t know’.
So what exactly has your ‘methodology’ contributed to our collective knowledge?
quote:
If you choose not to decide you still have made a choice
end quote:
Getty Lee
If you put a gun to my head I would say I can’t say for sure it’s mental. 😉
I’m not claiming anything. I’m exploring the possibility of developing an alternative Turing test.
It would be nice If I had a little help in that regard 😉
peace
For one thing we know we need more data points than you provided. Progress is progress no matter how slight 😉
you provided 35 data points the Poker players had 1,500 the minimum data requirement should be somewhere in the middle
peace
How many did you provide in the original ?
ETA: what were those data points, outcome of a hand?
I am not sure of the point of this excercise. Are you suggesting that plotting data points can tell you whether or not the cause of the data points was intentional (ie, designed)?
I agree that this may be possible, but only if we know what the data represents. For example, simple air or water temperatures can provide graphs with very clear trends. Ignoring the concept of AGW, nobody would suggest that there was any intention involved. However, of the temperatures were taken within a heated house, any trends would be do to design (even daily fluctuations from a poorly insulated house or an undersized furnace would still be the result of design.
I would say there are only two ways to reliably determine design. The first (often hard) is to investigate the design process in enough detail to determine that the observed design was produced by that process. The second (part of human nature) is to PRESUME design by simply rejecting any disagreement as “unreliable”.
This second technique is illustrated daily by the new US President – simply DECLARE that any uncongenial story is “fake news” – and that “real news” is what is excreted by sources that fail all attempts at fact checking. He is using the “religious method” of knowledge.
Just to say thanks to the Admin who posted my graphs into the thread. I don’t understand why I can’t do that myself, it would save you the bother.
faded_Glory,
It’s just how the permission levels work in WordPress. There may be a free plugin other than the basic image option we already have. I’ll have a look and if anyone happens to know of one, a PM would be appreciated.
To let the cat out of the bag the graph is the estimated percentage of the German U boat fleet destroyed by month from September 1939 till the end of 1944.
The Enigma cipher intercepts along with High-frequency direction-finding allowed the Allies to find the location of German subs with much greater ease starting at about observation 23 (summer 1941).
The German Naval command assumed that the enigma was unbreakable and this bias probably prevented them from seeing the evidence of “mental” influence in the pattern till it was to late
peace
The problem is that if we know what the data represents we open ourselves up to our own bias.
If I know the data is from a bot I will assume that there is no intention. The same thing goes for you all when it comes to “natural” things
The hope of the exercise is to try and minimize this sort of personal bias.
Suppose we had 80% of respondents infer design in a pattern then we could at least say that we need some reason to rule out design.
On the other hand if only 15% of respondents inferred design in this sort of blind test it would mean that we would need some reason to claim the pattern exhibits mental behavior.
It’s the “blind” aspect that makes Turing tests so appealing
peace
You need to read more carefully. No one is worried about being wrong, it’s simply that you haven’t defined your terms or explained how this graph is supposed to help you achieve your goal. There’s simply no point in hazarding a “guess”.
why did you need to know my “goal” before you made a guess?
peace
The inadequacy of human perception is one of the reasons that the scientific process was developed. Give any pre-school kid a magnet and I am sure that 80% would think that it was magic.
Your reasoning is what ID is relies on. My intuition is that it is designed, so we must accept that as the best explanation unless you can provide a step by step process to explain it and demonstrate through testing that this is what happened. We don’t require that burden of proof for plate tectonics, why is it required for evolution?
Not your goal, how your graph is supposed to help achieve it. Without operational definitions and a clear plan to test whatever it is you want to test, all you have is vague handwaving. There’s simply no point in participating.
Humans perception is uniquely suited to the task of recognizing other minds. It’s what we do. Nothing else has been demonstrated to come close in that regard.
That is why all Turing tests rely on human perception.
Actually I would be interested to know how plate tectonics would fare in my graph test. I’m not giving it a free pass 😉
peace
Again Mind by it’s very nature can not be defined with mathematical precision. If it could be defined in that way we could build an algorithm to produce it
I suppose you could claim that consciousness is nothing but vague handwaving. But that is simply a cop out. No offense to Dennett intended
We all know consciousness exists because we experience it ourselves and recognize it in others even if we can’t define it mathematically
Turing tests work just fine with out the sort of operational definitions you demand.
peace
That’s cute.
You dismiss Patrick’s claim of hand-waving, and you sandwich it in-between two waves of your hand.
If you are going to make that claim you need to provide an operational definition of hand-waving 😉
peace
I’m not so sure that he is demanding them.
He is saying that definitions are necessary to make sense of your exercise.
He is not saying that definitions are necessary unconditionally.
I provided dictionary definitions of all the terms that Eric requested. If he he has any other requests all he has to do is ask. But he needs to be specific,
peace
I’m still confused. You are suggesting that we can perceive intention in a graph because we are good at detecting patterns. Yet, there is an entire field of research around statistical approaches to detecting patterns, presumably because the accuracy of our perception is not as good as you claim it is.
Of greater danger is our ability to perceive patterns when none exist, because of inherent biases. That is why we insist on double blind randomized processes for clinical trials.
Regardless, I don’t see how you can make the link between a pattern on a graph to intention without knowing something about the intention you are looking for. It’s possible mechanisms. The nature of the being doing the designing. Things that ID absolutely refuses to do. Without these, you have nothing more than a leap of faith.
Whether or not there are obvious and observable patterns and trends in a graph means absolutely nothing with regard to interpretation unless you know in advance that it is a plot of data that is the consequence of some intention (eg, a classic control chart) or of some natural factors (eg, climate measurements). The degree and nature of “intention” is something that must be known before you can effectively interpret the data. But you can’t travel in the other direction.
Unless more submarines sank because of increased bad weather effects or similar, all possible reasons for the increase in losses can be attributed to ‘mental influence’ – be it Enigma, or better sonar, or better depth charges, or an increase in Allied destroyers, or better bomb sights in Allied aeroplanes, or…..etc. etc. etc. None of those are ‘natural’ causes, all of them include ‘mental influence’.
Do you seriously think the Germans attributed the increased losses to natural (‘algorithmic’) causes? Of course they realised that somehow the Allied had stepped up their game (‘mental influence’). To claim that they didn’t is ridiculous. The idea that a ‘less biased’ interpretation of the data would have made a difference in their response is equally absurd.
I can’t point to a reliable source, but what I saw on TV (probably History Channel) was that the Germans thought–at least at first–that the allies might be using some sort of magic, dowsing or some such thing. They tried magic in return, with the expected results.
Some of those guys were pretty nutty.
Glen Davidson
Nope that is not what I’m suggesting at all.
I’m suggesting that we can recognize non-personlike behavior. That is what the poker players did.
We can in fact detect patterns that are apparently invisible to statistical analysis. Check it out
https://arxiv.org/abs/1002.4592
right, That is also why Turing tests have multiple judges and are based on short cold conversations with very little context.
We are trying to duplicate this with a test that is not constrained by verbal communication
You need to elaborate on this.
For instance when looking at the poker bot what is the intention? What is the mechanism? Why do we need to know this?
Every time we ascribe mind to an entity other than ourselves we are making a leap of faith. We get along just fine
if that was true then Turing tests would be impossible. We know that they are not
peace
I don’t have any inside information but it’s obvious that the Germans did not have a counter to the increased losses they were taking,
If they had realized what was happening they could have changed their strategy accordingly. We also know that they did not ever realize that enigma had been broken.
besides
If you like we can grant that the Germans did realize that something “mental” was going on and that would only serve to bolster the idea that you can infer intention by behavior in processes
peace
I can think of all sorts of non-mental causes for increased submarine losses such as
1) faulty raw materials
2) different battle locations
3) poorer quality control in newer subs
4) simple random fluctuation, remember there is very little to verify statistically there is any pattern at all in the data
5) etc
peace
You aren’t inferring anything you know from historical records that it intentional. The question for your test is if a present day computer existed given the same data whether the pattern of the sinkings would change.
That they did not have a counter has nothing to do with your idea that graphs can differentiate ‘algorithmic’ from ‘mental’ causes.
Yes, if they knew that Enigma had been broken they could have changed their coding system. Hello??? What on earth has this to do with your idea that graphs can differentiate ‘algorithmic’ from ‘mental’ causes?
Your graph analysis doesn’t tell anybody anything that they wouldn’t already know anyway – that at some point in time their losses took a pretty serious turn for the worse. The reasons why would have to come from further analysis of the available data.
We can grant that the Germans realised that something ‘mental’ was going on
Those clever Germans, figuring out that the increase in losses wasn’t caused by acts of nature. I’m sure that without your brilliant graph analysis that realisation would never have entered their minds.
Now you will have to explain very carefully why these things are not ‘mental’:
2) different battle locations
3) poorer quality control in newer subs
I think that an operational definition of ‘mental’ would come in quite handy here.
With regards to:
4) simple random fluctuation
I would have liked to see Hitler’s response had you offered that as the explanation for the big increase in submarine losses.
5) etc
Right.
In both the Turing test and your poker bot example, we are asked to distinguish a well known human activity from a computer programmed to mimic the same specific human activity (a conversation and playing poker).
I don’t see how this helps you identify ID in the biological realm. All you have arguably shown is that humans are reasonably adept at distinguishing human conversation and poker from a computer programmed to do both. And, I might add, that probably has more to do with computer power and programming skills than it does with any innate human ability to make these distinctions. Not to mention the fact that the classic Turing test has been beaten by a computer.
Yes, people requested more information – not more data. You know the difference, don’t you? People asked for content and context. You consistently failed to grasp this.
And that showed you have nothing. Nobody does science with dictionary definitions. They do it with operational definitions – relevant to the hypothesis and the method at hand.
Unless you are specific, you have no exercise, test, experiment, practical or otherwise. You weren’t.
Anyway, I see the data is about German losses. So it’s designed. I was right. I knew it was designed when I saw the zeroes around the dots in the raw data. Is there a medal?
What??
I have no idea what this means?? Could you explain?
peace
Graphs can’t differentiate anything. That is certainly not “my idea”
My “idea” is that mental behavior is different than non-mental behavior.
That is what the poker players saw while playing against the bot. We are trying to see how we can universalize that
Exactly !!!! They should have been able to identify something mental was afoot by looking at the data. That is what we are trying to duplicate.
peace