Paranormal parrot in New York!

[Dids]

Got this from the BBC:

http://news.bbc.co.uk/1/hi/sci/tech/3430481.stm

Along with actual "real" speech, apparently N'kisi even does telepathy! Can a parrot take the JREF Million Dollar Challenge? Could it also pass a polygraph test?

[mummymonkey]

A quick Google reveals the hand of Rupert.

http://www.sheldrake.org/nkisi/

[Goshawk]

...(We discarded several trials with the target "camera", as N'kisi often made direct comments about the cameras we were using). As an animal, N'kisi could not be expected to fully understand the experiment parameters, and there was no guarantee of his participation. Our experiment design left him free to say whatever he wished during the sessions. Non-scorable comments consisted of N'kisi's attempts to contact Aimee, or unrelated chatter about events of the day. "Hits" were key words corresponding to the image Aimee was looking at during a particular trial. As there was no way for N'kisi to understand the need to restrain his comments to the strictly timed two-minute test period for each image, many of the "misses" scored were his continued repetitions of "hit" comments from previous images in the session. So, they hold up pictures for Aimee to look at and beam thought waves about them to the parrot several rooms away. "Flower". Bird says "flower". Bird also says other words, which are "unrelated chatter", "non-scorable", and "discarded".

And...Aimee claims that this proves her parrot has ESP?

I see.

[AlienX]

Quite an amazing bird by the sound of it and the stuff about telepathy is not required.

The testing procedure and interpretation is highly questionable and coming to such a conclusion as the parrot is telepathic is unjustified and to be blunt incompetance.

Being realistic it would be very difficult to produce meaningful data - a bird chattering using a 950 word vocabulary?

It would be intersting to have a computer randomly select phrases and comments from the same 950 word vocabulary and see how well it does ;-). Maybe we would find out that computers are also telepathic - I can just see Intel and AMD producing chips with various "PSI" levels as well as mhz etc ;-)

IMO its a shame such an amazing creature is being used for what is seemingly pointless science. Why use and exception in the animal kingdom for such experiments when we have billions of people who are capable of doing the same test and produce "better" results.. what a waste.

AX

AX

[MRC_Hans]

The stastics of the experiment are seriously flawed. I don't know exactly how they came to their target probability, 7.4 (I cant get exactly the same number in any way), but anyway, it is wrong.

The parrot has a vocabulary of some 950 words. First crucial figure is, how many words does it say during each two-minute period? This info does not seem to be available, and I suspect it is not used, which is a one serious fault.

The parrot "scored" in 32% of the instances. This would require him to say 304 words during each 2 min. phase, if he spoke words completely random. That is more than two words per second, which is unlikely even if he talks a blue streak. However, it is very unlikely that he uses his vocabulaty even partly randomly. The 19 test cards were selected based on his vocabulary, and if they were selected among the words he uses mostly, the stastistics look very different.

Let us assume that 80% of the parrot's chatter is covered by 200 different words, and the 19 keywords are among those, then it only needs to say about 35 words per minute to account for the hit rate.

Another fault, of course is the obvious reporting bias; they deem hits and misses in a subjective way. For one thing, negative hits should also count; if it says wrong keywords, it would counterindicate telepathic contact.

Hans

[Undodog]

N'kisi got the picture of 'a gottle of geer' spot on.

(Polygraph.. I like what you did there, Dids)

[Goshawk]

Originally posted by Undodog
N'kisi got the picture of 'a gottle of geer' spot on. Where was that? I don't see it mentioned in either link.

[Cecil]

This is one remarkable parrot. He seems to understand tenses, grammatical rules (generating "flied" out of "fly", like most children do) and over 3 times as many words as any other animal (I believe the chimps that know sign language are limited to about 300 "words"). His reaction to meeting Jane Goodall seems to indicate that he can recognize faces and has a memory. It's disappointing that nearly everyone (including most of the sites I found on Google) are focusing on the belief that he's telepathic. Isn't it enough that this is one of the most prolific users of human language that we've ever seen?

However, it doesn't appear that he can understand spoken English. Here (http://www.sheldrake.org/nkisi/nkisi1_text.html)'s a transcript and audio of a session with N'kisi. It looks he's only reacting to his environment, not to anything Aimee (his trainer) says.

[Suezoled]

There are other stories of people who have parrots who seem to have a grasp of human languge. They don't involve "psychic" powers, others. Am trying to see if I can find those links...

[Suezoled]

http://www.alexfoundation.org/index.html

[Interesting Ian]

Originally posted by Goshawk

From Sheldrake link

(We discarded several trials with the target "camera", as N'kisi often made direct comments about the cameras we were using). As an animal, N'kisi could not be expected to fully understand the experiment parameters, and there was no guarantee of his participation. Our experiment design left him free to say whatever he wished during the sessions. Non-scorable comments consisted of N'kisi's attempts to contact Aimee, or unrelated chatter about events of the day. "Hits" were key words corresponding to the image Aimee was looking at during a particular trial. As there was no way for N'kisi to understand the need to restrain his comments to the strictly timed two-minute test period for each image, many of the "misses" scored were his continued repetitions of "hit" comments from previous images in the session.

Goshawk
So, they hold up pictures for Aimee to look at and beam thought waves about them to the parrot several rooms away. "Flower". Bird says "flower". Bird also says other words, which are "unrelated chatter", "non-scorable", and "discarded".

And...Aimee claims that this proves her parrot has ESP?

I see.

You think one should include chatter about the events of the day and the camera used in the scoring?? I would suggest that this would not be sensible.

And yes this could well be the real thing. The article states:


Aimee found that her state of mind was critical, and if she intentionally tried to "send" the information, it wouldn't work. N'kisi responded best when Aimee's full attention was genuinely immersed in exploring the images, without any thought of the experiments.


This certainly corresponds to my experiences.

[Interesting Ian]

Originally posted by AlienX
[B]Quite an amazing bird by the sound of it and the stuff about telepathy is not required.



It's not required for what? Interesting that you should state that compelling evidence destroying your view of the world, and how it works, is "not required".



The testing procedure and interpretation is highly questionable and coming to such a conclusion as the parrot is telepathic is unjustified and to be blunt incompetance.



Yes it would be. Any skeptic would state the same no matter how tight the experimental protocols were. And liikewise, a skeptic would view any experiment showing negative results as demonstrating an exemplary experimental protocol.




IMO its a shame such an amazing creature is being used for what is seemingly pointless science.



Pointless science?? How is smashing the materialist metaphysic and all the beliefs the pseudo-skeptics irrationally cling on to pointless science?? :rolleyes:



Why use and exception in the animal kingdom for such experiments when we have billions of people who are capable of doing the same test and produce "better" results.. what a waste.



What are you talking about?? You don't believe in anomalous cognition, but if it does exist, human beings should be better at it than parrots and other non-human animals?? :rolleyes:
:rolleyes: :rolleyes:

Since you do not believe in telepathy in the first place, then how the f*ck do you come to this bizarre conclusion?? {SIGHS VERY HEAVILY}

Are there any skeptics on this board who are not complete airheads?? :eek: :confused:

Complete and total unbelievable a*seholes the lot of you.

[Suezoled]

Are there any skeptics on this board who are not complete airheads??

Complete and total unbelievable a*seholes the lot of you.

Ian must be drunk. Again.

[Interesting Ian]

Originally posted by Suezoled


Ian must be drunk. Again.

So? Anything wrong with being drunk??

[Suezoled]

Originally posted by Interesting Ian


So? Anything wrong with being drunk??

You do get a good deal more abusive when you drink.

[Yahweh]

Ian, please forgive the Evil Skeptics for being so closeminded toward Parrot Telepathy.

[Goshawk]

You think one should include chatter about the events of the day and the camera used in the scoring?? I would suggest that this would not be sensible.I suggest that it would have been honest. They should have provided a transcript of exactly what the bird said, so we could all see for ourselves. However, they didn't do this--they merely announced that the bird had said the correct words.

And, by their own account, they admitted that they had purposely ignored some of the words the bird had said. They had decided for themselves that some of the words the bird was saying (i.e. "some of their test results") were irrelevant. People who are serious about carrying out an experiment don't throw out some of their test results because they decide, for themselves, that those test results are irrelevant.

Raw data, Ian. That's what it's about. Serious researchers have no hesitation about posting their raw data for all the world to see.

But people with an axe to grind--people who have already made up their minds that their experiment worked and who wish to announce this to the world--never publish their raw data. We're expected to take their word for it. Me, I don't. Call that "skepticism"? You betcha.

I'd be delighted if they could prove to me that this parrot has ESP. But the only way they're going to prove it to me is if they let me see for myself what the parrot said. I'm not even requiring any special test protocols, because actually I thought their protocol sounded good. Aimee was way down the hall, several rooms away, obviously out of earshot. There was no way the bird could have been picking up either sound or visual cues from her, so if it did get any words right, it must have been from ESP. However, they admitted that they ignored many of the words the parrot said. Which is cheating, to me.

We have only their own personal calculations as to how many "hits" the parrot got. That's not "science", no matter what the test protocol was.

[asthmatic camel]

Reminds me of dog owners who claim that their pet "understands every word I say".

[Interesting Ian]

Originally posted by Goshawk
I suggest that it would have been honest. They should have provided a transcript of exactly what the bird said, so we could all see for ourselves. However, they didn't do this--they merely announced that the bird had said the correct words.

And, by their own account, they admitted that they had purposely ignored some of the words the bird had said. They had decided for themselves that some of the words the bird was saying (i.e. "some of their test results") were irrelevant. People who are serious about carrying out an experiment don't throw out some of their test results because they decide, for themselves, that those test results are irrelevant.

Raw data, Ian. That's what it's about. Serious researchers have no hesitation about posting their raw data for all the world to see.

But people with an axe to grind--people who have already made up their minds that their experiment worked and who wish to announce this to the world--never publish their raw data. We're expected to take their word for it. Me, I don't. Call that "skepticism"? You betcha.

I'd be delighted if they could prove to me that this parrot has ESP. But the only way they're going to prove it to me is if they let me see for myself what the parrot said. I'm not even requiring any special test protocols, because actually I thought their protocol sounded good. Aimee was way down the hall, several rooms away, obviously out of earshot. There was no way the bird could have been picking up either sound or visual cues from her, so if it did get any words right, it must have been from ESP. However, they admitted that they ignored many of the words the parrot said. Which is cheating, to me.

We have only their own personal calculations as to how many "hits" the parrot got. That's not "science", no matter what the test protocol was.

They should have provided a transcript on the net? :confused: I'm not sure that people generally tend to put all their raw data on the net do they? Obviously, for whatever reason, not much details are given over the net. But perhaps they are available some other way? Are you claiming it isn't? How do you know this? Have you contacted Sheldrake and he has told you this perhaps? Has he explicitly said that the transcripts are private and nobody is allowed to see them apart from a few select individuals? Could you be good enough to reproduce this correspondence?

Thanks.

[Interesting Ian]

Originally posted by asthmatic camel
Reminds me of dog owners who claim that their pet "understands every word I say".

Which just goes to show how stupid you are. Clearly they do not mean this literally. They mean their dogs can sesne their emotions.

[Tanja]

Originally posted by asthmatic camel
Reminds me of dog owners who claim that their pet "understands every word I say".

I remember a woman in my neighbourhood who always talked to her dog. OK, most people when walking a dog would say things like "common Fido, let's go" but she would always say things like "ok, Fido, now I have to quickly go to the supermarket as we ran out of croissants, and then we'll go back home" .

[showme2]

As a believer in many aspects of the paranormal, I just have to tell you guys that "You can't be serious!" even discussing this subject.

A paranormal parrot ? ...... Get outa here !!!!

[Suezoled]

Originally posted by Interesting Ian


Which just goes to show how stupid you are. Clearly they do not mean this literally. They mean their dogs can sesne their emotions.

Ian, you are a social mistake.

[LeFevre]

Ian is like Carlos, he is sometimes a nice guy. But get a little drink in him (or deny that the space alien hat was really a space alien hat in Carlos' case) and he drops his facade.


Go back to the Matrix Neo.

[Goshawk]

They should have provided a transcript on the net? I'm not sure that people generally tend to put all their raw data on the net do they? Obviously, for whatever reason, not much details are given over the net. But perhaps they are available some other way? Are you claiming it isn't? How do you know this?
Researchers who have nothing to hide--have nothing to hide. Meaning, they are quite open on the subject of "where our original data may be found".

You can go to websites such as the NIH PubMed, and find out "where the original data may be found" for thousands of studies, dating back to the 1950s.

However, Sheldrake and Aimee & Co., by the simple fact of their having NO information on the website about "where our original data may be found", automatically disqualify themselves in my mind from being "honest researchers". They don't want anyone to see their original data--because anyone with any sense and a seventh-grader's grasp of logic would understand that what was actually happening was that the parrot was saying a whole stream of words--that it was merely reciting a chunk of its extensive vocabulary--and that the researchers were simply choosing which of the words they would consider the "hits".

[Interesting Ian]

Originally posted by Goshawk

Researchers who have nothing to hide--have nothing to hide. Meaning, they are quite open on the subject of "where our original data may be found".

You can go to websites such as the NIH PubMed, and find out "where the original data may be found" for thousands of studies, dating back to the 1950s.

However, Sheldrake and Aimee & Co., by the simple fact of their having NO information on the website about "where our original data may be found", automatically disqualify themselves in my mind from being "honest researchers". They don't want anyone to see their original data--because anyone with any sense and a seventh-grader's grasp of logic would understand that what was actually happening was that the parrot was saying a whole stream of words--that it was merely reciting a chunk of its extensive vocabulary--and that the researchers were simply choosing which of the words they would consider the "hits".

Well the parrot telepathy study is published in the Journal of Scientific Exploration (vol 17 issue #4, 2003).
Would they not have all the original data there?

[asthmatic camel]

Originally posted by Interesting Ian


Which just goes to show how stupid you are. Clearly they do not mean this literally. They mean their dogs can sesne their emotions.

And I'm stupid ? I "sesne" a little duplicity here.

[asthmatic camel]

Sorry, double post. Perhaps I am stupid atfterall.

[Interesting Ian]

Originally posted by asthmatic camel


And I'm stupid ? I "sesne" a little duplicity here.

I have no idea what you are talking about. I am denying your assertion that people who talk to their pet dogs believe their dogs understand every single word.

Now what is your problem?

[asthmatic camel]

Well Ian, you're the expert on semantics, but when I hear someone say "He understands every word I say", I take it that they believe that their pet does, indeed, have that ability.

Now, call me stupid or a prick again if you wish, but I've never heard anyone say about their pet that "He "sneses" my emotions".

I don't think I have a problem, has it ever occured to you that you may ?

[69dodge]

Originally posted by MRC_Hans
Let us assume that 80% of the parrot's chatter is covered by 200 different words, and the 19 keywords are among those, then it only needs to say about 35 words per minute to account for the hit rate.Sheldrake's web page doesn't give nearly enough detail about the statistical analysis, I'll agree, but he does claim that a Randomized Permutation Analysis was done, which would seem to rule out this particular explanation.

If your explanation were correct, the hit rate should be about as high even if the targets were shuffled.

[Interesting Ian]

Originally posted by asthmatic camel
Well Ian, you're the expert on semantics, but when I hear someone say "He understands every word I say", I take it that they believe that their pet does, indeed, have that ability.



Oh come on. You cannot be serious!

[Goshawk]

Originally posted by Interesting Ian
Well the parrot telepathy study is published in the Journal of Scientific Exploration (vol 17 issue #4, 2003). Would they not have all the original data there? Volume 17, Issue 4, has not had its abstracts posted online yet.
http://www.scientificexploration.org/jse/abstracts.htmlVolume 17, Number 4 (Winter 2003)

COMING SOON And when it does, I don't expect to see the original data. According to http://www.sheldrake.org/nkisi/ :
The results of our telepathy study with nkisi have been published in the Journal of Scientific Exploration (vol 17 issue #4, 2003), a peer reviewed journal. [bolding mine]
I'm reading this as only the results will be published, not the raw data.

Sheldrake's other abstract is "A Dog That Seems to Know When His Owner Is Coming Home: Videotaped Experiments and Observations" (http://www.scientificexploration.org/jse/abstracts/v14n2a4.html), which involves over 100 hours of videotapes. All the abstract features is a short summary of what happened. The testing protocol is not described.

However, if you go to a place like PubMed and look up an abstract, it has much more information about the actual experiment than Sheldrake's abstract does.

For example, this abstract of a study (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=14748947&dopt=Abstract) of iron deficiency in Crohn's patients gives you much more information about their experiment, describing the test protocol as well as a summary of their findings. It's in their own best interests to have their peer group know exactly what they did, because in order for their experiment's results to be accepted by their peer group, those results have to be reproducible by other researchers.

And the only way they'll be reproducible by other researchers is if they publish exactly what they did, so folks trying to reproduce your experiments will know, for example, what kind of iron intake your Crohn's patients were ingesting.

And it's just professional courtesy to let your fellow researchers know, up front, what you already did in your experiment so as to save them from wasting time and money on similar experiments--researchers want, and need, to explore new scientific territory. They don't want to spend a couple of years doing tests that wind up proving what the guys down the block already proved last year. Therefore, even your abstract is going to give much more information about exactly what you did.

So it's in the best interest of serious researchers to have everything out in the open. However, it's not in the best interest of people like Sheldrake to have everything out in the open--because his experiments are not going to be reproducible. The times when someone does try to reproduce a non-mainstream experiment, they somehow never get the same results as the original researchers. The best example I can think of off-hand is dowsing (http://skepdic.com/dowsing.html).

So non-mainstream researchers have learned the hard way to be cagey and coy about exactly how they got their results, because the chances are good that if they publish exactly how they did it, some skeptic is going to come in and try to reproduce it, and he won't be able to.

[!Xx+-Rational-+xX!]

In their delusions they can deny it but in reality kooks can’t hide from the truth since materialism is everywhere and everything!

[showme2]

..... and I thought this was a forum for SERIOUS discussion. (sigh)

Paranormal parrots my arse.
Why waste your time even commenting on this ?

[TLN]

Originally posted by showme2
..... and I thought this was a forum for SERIOUS discussion.

And where the hell did you get that idea?

[Goshawk]

This is a forum for serious discussion. And we're seriously discussing whether this parrot has telepathy.

[TheERK]

Originally posted by showme2
..... and I thought this was a forum for SERIOUS discussion. (sigh)

Paranormal parrots my arse.
Why waste your time even commenting on this ?

So, you're prepared to believe that humans have telepathy, but not that an another animal might?

[Mike D.]

The complete Journal of Scientific Exploration paper can now be found here:

http://www.sheldrake.org/papers/Animals/parrot_telepathy.html

[Frederick Troteville]

Originally posted by TheERK


So, you're prepared to believe that humans have telepathy, but not that an another animal might?

I must say I disagree with ShowMe2. If anything I think that a parrot would be MORE likely to exhibit telepathy.

But this does go to show that believers don't always agree with each other, unlike skeptics. And that's despite the fact that we're so heavily outnumbered on here!

[Supercharts]

Paranormal Parrot Recipe
[Modified cornish hen recipe]

Crockpot Paranormal Parrot

Paranormal Parrot
Water
1 cup of butter, melted
Mix together:
tablespoon of pepper
tablespoon of nutmeg
tablespoon of ginger
tablespoon of ground thyme
tablespoon of cinnamon

Directions:

Invite the Paranormal Parrot to dinner.
[If you have a cat seat them next to one another]
Kill, defeather, gut and wash in warm water. Pat the ex-parrot dry.

Mix remaining ingredients together in butter.
Baste paranormal ex-parrot with mixture.
Fill crockpot with water until bottom is about 1/3d full. [Some people use chicken broth. Swanson's is very good. Use the low-sodium can]

Place paranormal ex-parrot inside crockpot. Spread remaining mixture over ex-parrot.
Turn on low.
Cook for eight hours or until done.

If you will be gone less than eight hours, you may turn crockpot on high. Paranormal ex-parrot will be done in four hours or so.

When done the paranormal ex-parrot will be very tender. The meat will fall from the bone.
Remove beak. Use as a paper-weight.
Feed rest to cat.

[CFLarsen]

Originally posted by Frederick Troteville
But this does go to show that believers don't always agree with each other, unlike skeptics.

This is a lie. Skeptics do disagree with each other.

Yes, it's a lie, because you know better. You have seen skeptics here disagree.

What is interesting is, skeptics seem much more likely to keep a civil tone among each other, and not resort to sewer talk, than believers.

Originally posted by Frederick Troteville
And that's despite the fact that we're so heavily outnumbered on here!

It's not a popularity contest, Ian. You can only win a debate by being right. Not by being in the majority, or - as you try - to be as rude as possible.

[Paul C. Anagnostopoulos]

I disagree with you Claus. Just as an example, of course.

~~ Paul

[Peter Morris]

Originally posted by Interesting Ian
Ian must be drunk again

So? Anything wrong with being drunk??

Ask a glass of water.

(Well someone had to say it)

[Beanbag]

Raw data and the methodology are required to make a valid judgement of the quality of the experiment and the findings.

There was a line used by a sugarless-gum manufacturer to hype their product on radio and TV: "Four out of five dentists surveyed recommend sugarless gum for their patients who chew gum."

Sounds reasonable, right? Well, at the time I was employed as a marketing research phone grunt, calling people at odd hours and asking the questions to complete a market survey. First, you had to "qualify" your potential survey-ee to make sure they fit the demographic you were targeting. After all, you really wouldn't want to ask a middle-aged housewife what product she used for jock itch, now would you?

Well, the way it worked for the gum-recommending dentists was that the surveyor first asked the qualifying question: "would you recommend that your patients chew gum?" If the dentist said no, they weren't qualified, and were dropped from the survey. If they said yes, then the surveyor proceeded to the actual (and only) survey question: "Given a choice, would you recommend that your patients chew sugared or sugarless gum?" Well, that's a no-brainer. Just about any dentist would recommend sugarless gum. It's like the often-quoted "fact" that doctors in particular are against handguns. Well of course they are -- if I spent all day in an emergency room trying to patch together shooting vivtims, I'd be a little bent against them myself.

Just from the brief amount of background Sheldrake provided about the experimental conditions, I'd say that too much data was being ignored, and only the hits were being scored, while most of the misses were being explained out of relevance.

Regards;
Beanbag

[S. D. Youngren]

So how about a proper retest? In addition to the standard safeguards, I'd be interested in an additional layer: Have a couple of researchers in the room with the parrot, listening to and watching him, and guessing from his comments what sort of picture he's describing.

I'd also like to know more about those "events of the day." "And then we went to Grandma's house! And then we had lunch! And then..."

Somehow I suspect they didn't quite sound like that.

[S. D. Youngren]

Great recipe, by the way.

[AlienX]

Originally posted by


It's not required for what? Interesting that you should state that compelling evidence destroying your view of the world, and how it works, is "not required".

>>It's not required for an interesting story. It does not require any discussion about telepathy to make it interesting.

Yes it would be. Any skeptic would state the same no matter how tight the experimental protocols were. And liikewise, a skeptic would view any experiment showing negative results as demonstrating an exemplary experimental protocol.

>>The controls need to be very tight indeed and repeatable for "proof" for such apparantly wild claims. A person would be much more valuable and controlable.


Pointless science?? How is smashing the materialist metaphysic and all the beliefs the pseudo-skeptics irrationally cling on to pointless science?? :rolleyes:

>>SEEMINGLY pointless science.. can you not undestand the difference?.

What are you talking about?? You don't believe in anomalous cognition, but if it does exist, human beings should be better at it than parrots and other non-human animals?? :rolleyes:
:rolleyes: :rolleyes:


>>I'm saying that using a parrot with a 950 word vocab is a very difficult scenario to prove anything.

Since you do not believe in telepathy in the first place, then how the f*ck do you come to this bizarre conclusion?? {SIGHS VERY HEAVILY}

>>How do you know I don't believe in telepathy, I think it's unlikely - i'm not certain coming to the conclusion that telepathy probably does not exist a £$%^ing bizare conclusion.

Are there any skeptics on this board who are not complete airheads?? :eek: :confused:

>>Hehe - nice one exacly what I thought when I seen this only with the roles reversed.
IMO you need to seriously chill out, just because I don't agree with you it doesn't mean im a bad person ;-).

Complete and total unbelievable a*seholes the lot of you.

[Goshawk]

Well, I have now read the official paper (here it is again so you don't have to go back to Page 1 to find it).

http://www.sheldrake.org/papers/Animals/parrot_telepathy.html

And I got as far as this, in the introduction--For each trial, the key word or words represented in the photograph were tabulated. Some images had only one key word, but others had two or more. For example, a picture of a couple hugging in a pool of water involved two key words, “water” and “hug”. --when I realized what they were saying was that, "Each photograph could count as both a hit and a miss".

And I stopped reading.
And I totally skimmed over all the statistical analysis mumbo-jumbo at the end.

Because it seems to me--even me, the math-challenged person--that if you're gonna set up your test so that the photographs involved can count as BOTH hits AND misses, then you ought to hand in your Junior Scientist badge and Secret Decoder Ring and go back to playing Everquest online.

Sheesh.

What a waste of time--they could have set up a real experiment, and really proved something. But no, because that would have proved that it didn't work, that Nkisi doesn't have telepathy, so they deliberately chose complex pictures that could do multiple duty (a photo of a couple on the beach could be practically anything).

And upon further perusal, I realized they also said this:Only trials in which N’kisi said one or more of the key words were included in the analysis of the data, because only in such trials could N'kisi have scored a hit or a miss. Trials in which he said nothing or used words that were not on the list of pre-specified key words were not included.

< snip >

In the remaining 60 trials, N’kisi either remained entirely silent, or said none of the 19 key words corresponding to the test images. Thus, in these trials, neither a “hit” nor a “miss” was scored, and they were irrelevant to the analysis. So, the trials in which he didn't say any of the 19 targeted key words, or trials in which he said nothing at all, were simply not included in all their fancy-shmancy statistical analysis.

Do they realize what they're saying? :rolleyes:

[Skeptical Greg]

Hmmmmm... Record the hits and ignore the misses..

Where have we heard that before?

[Interesting Ian]

Goshawk,

I have not as yet read the paper in question, and even when I do, I doubt that I will have the necessary understanding to judge the appropriateness of the experimental protocol. But I would just like to make a couple of comments.

First of all it was kind of inevitable that when you at last saw the paper you would have a contemptuously sneering attitude. Not just you but most other skeptics too. Could this be because all such research into this area is hopelessly flawed, or could it be because skeptics do not as a rule have an objective approach to this subject matter? In my opinion it would be wildly implausible to come to any other conclusion than the latter one. This is not to say that Sheldrake's experimental protocol might not be hopelessly flawed, it's just that the fact that you say it is, should carry very little weight for a person who is sincerely trying to find out the truth. And this conclusion is reinforced by your next comments:

And upon further perusal, I realized they also said this:So, the trials in which he didn't say any of the 19 targeted key words, or trials in which he said nothing at all, were simply not included in all their fancy-shmancy statistical analysis.

Do they realize what they're saying? :rolleyes: [/B]


Now I will need to leave aside the issue of the targeted key words until I've read the paper. But your suggestion that when the parrot says nothing at all, that this should be included in the statistical analysis, baffles me. If the parrot remains silent, then we simply do not know whether it was receiving any telepathic impressions or not. Therefore what on earth are you thinking of here? Sure, they could include it in the statistical analysis. But since we do not know whether or not the parrot received any telepathic impressions, what on earth would be the purpose?? Although I know nothing whatsoever about scientific experiments and statistical analysis, it seems to me that the occasions where the parrot is silent is simply irrelevant. Indeed, I would go further and say it could not possibly be relevant.

You try to convey the impression that you know what you're talking about. But I have to conclude I am far from impressed. A hysterical attitude towards any positive results for any aspect for the paranormal will not impress a genuine seeker of the truth. This is compounded when you make totally irrelevant criticisms.

Sorry.

[Interesting Ian]

Originally posted by Diogenes
Hmmmmm... Record the hits and ignore the misses..

Where have we heard that before?

Indeed this is bad if he has done this. If you don't me saying though, I find that skeptics often do not speak the truth when it comes to any evidence suggestive of anomalous abilities. I'll need to read the paper.

[Goshawk]

Okay, all you statisticians--analyze this. :D

Rearranging their list of "Hits" a bit. Pictures are in blue.
Data for trials in which N’kisi said at least one key word, according to the majority of scorers. Hits are...in bold type. The trials marked by asterisks involved pictures corresponding to key words that N’kisi rarely used, and never said in the entire series of trials. Some “key words said” were repeated during a given trial, and the number of times N’kisi said the key word are shown in parentheses.

16/1 feather car
22/2* TV glasses
5/1* teeth water(2), doctor(2)
18/1 cards water
16/2* key medicine, glasses
23/4* key glasses, feather
6/4* keys hug (3)
12/1* computer water, medicine(2)
15/3* computer phone, naked body
21/4* computer flower(4), books
31/5* computer hug, flower
4/1* CD hug, medicine, doctor(2), water
5/2* CD doctor
5/5* CD car, doctor
8/1* CD medicine
8/4* CD glasses, car, doctor(2)
12/3* CD phone
27/2* CD flower, doctor(3), naked body
19/1 car medicine
19/5 car flower
13/2 car car
28/1 car, book book, flower
11/4 book water, phone
20/3 books doctor
24/5 books glasses
19/4 flower flower, medicine
21/2 flower flower (10), doctor(2)
21/3 flower flower (7)
23/1 flower flower, doctor, glasses(2)
23/2 flower flower, car
24/2 flower naked body
27/1 flower flower
28/4 flower flower, naked body
28/5 flower flower (3), bottle
29/1 flower water
30/2 flower flower, bottle
11/3 flowers water, medicine
15/5 flowers car, water
16/3 flowers glasses(3)
30/4 flowers naked body, doctor
33/3 hug, flowers phone
21/5 flowers, hug flower (4)
6/2 hug , water hug (3), water
27/3 hug medicine, doctor(2)
8/2 hug, glasses glasses, medicine(2), doctor
28/3 hug, water, glasses flower(2)
5/4 water doctor(2)
15/4 water feathers
23/3 water flower
31/3 water flower, medicine
33/5 water phone(2)
9/1* fire medicine
21/1* fire flower(4)
8/5 fire, glasses hug, medicine
12/4 phone phone(2), medicine, water
18/4 phone glasses
19/3 phone doctor
29/3 phone flower, naked body, feathers
33/1 phone flower(3)
30/3 phone, glasses flower, bottle
4/2 medicine, bottle, doctor, glasses medicine(2)
5/3 bottle car(2), doctor(2)
8/3 bottle, medicine medicine(2), doctor(2)
18/2 water, hug, naked body water, glasses
24/3 naked body, water naked body
24/4 doctor, medicine, glasses medicine, feather, naked body, water
25/3 naked body, water naked body
25/5 naked body, water naked body(4)
28/2 bottle, medicine flower(4)
31/1 bottle book
31/4 naked body




Number of pictures altogether to which he said "flower" -- 23.
Number of pictures to which he said "flower" and it was applicable (pictures that were "hits" for "flower") -- 10

Number of pictures altogether to which he said "doctor" -- 16.
Number of pictures to which he said "doctor" and it was applicable (pictures that were "hits" for "doctor") -- 0.

Number of pictures altogether to which he said "car" -- 7
Number of pictures to which he said "car" and it was applicable (pictures that were "hits" for "car") -- 1

Number of pictures altogether to which he said "medicine" -- 16
Number of pictures to which he said "medicine" and it was applicable (pictures that were "hits" for "medicine") -- 3

Number of pictures altogether to which he said "water" -- 12
Number of pictures to which he said "water" and it was applicable (pictures that were "hits" for "water") -- 2

Number of pictures altogether to which he said "naked body" -- 11
Number of pictures to which he said "naked body" and it was applicable (pictures that were "hits" for "naked body") -- 3

Number of pictures altogether to which he said "glasses" -- 10
Number of pictures to which he said "glasses" and it was applicable (pictures that were "hits" for "glasses") -- 1

Number of pictures altogether to which he said "hug" -- 5
Number of pictures to which he said "hug" and it was applicable (pictures that were "hits" for "hug") -- 1

Number of pictures altogether to which he said "book/books" -- 3
Number of pictures to which he said "book/books" and it was applicable (pictures that were "hits" for "book/books") -- 1

Number of pictures altogether to which he said "phone" -- 6
Number of pictures to which he said "phone" and it was applicable (pictures that were "hits" for "phone") -- 1

Number of pictures altogether to which he said "feather" -- 4
Number of pictures to which he said "feather" and it was applicable (pictures that were "hits" for "feather") -- 0


So I make it:

113 times he said a word to a picture (or, rather, to the telepathically projected image of a picture).

23 times he got it right.

The seemingly uncannily high success rate of "flower" must of course be chalked up to the fact that obviously it's his favorite word--he said it to more pictures than any other word (a total of 55 times altogether by my count, and it was apparently the word most likely to be repeated multiple times). So just by dint of repeating it more often, it's more likely to be a "hit" a higher percentage of the time.

But even at that, it's still only the same as he would have scored if he had just chosen randomly each time, "say 'flower'/don't say 'flower' ".

[Goshawk]

You try to convey the impression that you know what you're talking about. But I have to conclude I am far from impressed. A hysterical attitude towards any positive results for any aspect for the paranormal will not impress a genuine seeker of the truth. This is compounded when you make totally irrelevant criticisms.Well, Ian, I'd be delighted to have you (or anybody else, actually) demonstrate exactly when and where I don't know what I'm talking about (not to mention my "hysteria" and "irrelevance"). I always welcome enlightenment, and if I'm wrong, I'm happy to admit it. But you have to show me I'm wrong, first. ;)

What we have here seems to be a simple difference of opinion--you think the parrot has ESP, I don't. And further, my opinion is that Sheldrake and Aimee haven't proved it; your opinion is that they have. I don't think that the situation calls for you accusing me of "not knowing what I'm doing." Because "expertise" isn't being called on here, for one thing, because nobody on these boards has any expertise at "parrots with ESP". So I'm not claiming any kind of "expertise" here, other than a sort of Alert Consumer expertise in knowing when someone's trying to pull a fast one by not telling me all the facts.

But your suggestion that when the parrot says nothing at all, that this should be included in the statistical analysis, baffles me.
Okay, here's an analogy.

You want to test a new drug, say a new antihistamine. You already know it works to dry up runny noses and stop the itching from hives, but you want to find out about side effects. You give it to 1,000 people. You tell them to record their reactions--did it make them drowsy, did it make their mouth dry, did it make them feel all jittery and nervous? You give them a three month's supply and a journal to write down their reactions.

Three months later, you go through the journals. You see the following reactions:

Some people said it made them drowsy.
Some people said it gave them a dry mouth.
Some people said it made them feel all jittery and nervous.

Some people said it didn't affect them at all.

Now, when you write up your findings, do you leave out the information that some people had no effects at all?

Of course not. That would be silly, and stupid, and unscientific. It's unscientific to leave out results just because they were null results. In the world of Science, even "no results" counts as "results". Because otherwise, you get skewed numbers. You can't do a statistical analysis of your results if you aren't working with ALL the results. And "the parrot was silent" counts as a "result".

So if Sheldrake and Aimee wanted to be "scientific", they would have gone ahead and reported their "null results", which would have been the numbers of times Aimee projected a telepathic image and the parrot just sat there and didn't say anything at all.

I'm sorry you think I'm "sneering" at them. I will admit I think the whole idea of parrot telepathy is ridiculous, but as I said on Page 1, I'd be delighted if they could prove to me that parrots had ESP. The thing is, they have to prove it, and so far, they haven't. And so far, their research methods are sadly lacking, IMO. For starters, as I mentioned, the images they chose were apparently so complex and ambiguous that they could be applied to any of a number of vocabulary words. For example, I find it VERY hard to believe that they couldn't find a simple picture of a "naked body"--that they had to depend on a picture of two people wearing swimsuits walking on a beach.

There's a reason why those Rhine cards are such big, simple symbols--it makes it easy to tell whether it's a "hit" or not.

[Interesting Ian]

Originally posted by Goshawk
What we have here seems to be a simple difference of opinion--you think the parrot has ESP, I don't.



Well, I read the paper. I found the anecdotes interesting. From what the owner of the parrot says these telepathic impressions occur under circumstances which I too have found to facilitate telepathic impressions.




And further, my opinion is that Sheldrake and Aimee haven't proved it; your opinion is that they have.



I most certainly have not stated this! With any scientific evidence for paranormal phenomena I think that other alternatives are possible. How plausible such alternatives are is open for debate.


I don't think that the situation calls for you accusing me of "not knowing what I'm doing." Because "expertise" isn't being called on here, for one thing, because nobody on these boards has any expertise at "parrots with ESP". So I'm not claiming any kind of "expertise" here, other than a sort of Alert Consumer expertise in knowing when someone's trying to pull a fast one by not telling me all the facts.


Well OK, let's see.



Ian

But your suggestion that when the parrot says nothing at all, that this should be included in the statistical analysis, baffles me.

Goshawk
Okay, here's an analogy.

You want to test a new drug, say a new antihistamine. You already know it works to dry up runny noses and stop the itching from hives, but you want to find out about side effects. You give it to 1,000 people. You tell them to record their reactions--did it make them drowsy, did it make their mouth dry, did it make them feel all jittery and nervous? You give them a three month's supply and a journal to write down their reactions.

Three months later, you go through the journals. You see the following reactions:

Some people said it made them drowsy.
Some people said it gave them a dry mouth.
Some people said it made them feel all jittery and nervous.

Some people said it didn't affect them at all.

Now, when you write up your findings, do you leave out the information that some people had no effects at all?



Of course you don't, but this is of course a blatently false analogy.



Of course not. That would be silly, and stupid, and unscientific. It's unscientific to leave out results just because they were null results. In the world of Science, even "no results" counts as "results". Because otherwise, you get skewed numbers. You can't do a statistical analysis of your results if you aren't working with ALL the results. And "the parrot was silent" counts as a "result".



But the parrot simply remaining silent cannot conceivably give either any evidence for or against telepathy. It might be receiving telepathic impressions, or it might not be doing so. We simply do not know because we cannot read the parrots mind! (even though it might be able to read ours). So sure, you can note the times when it's silent, but how could it possibly give any evidence for or against telepathy? The answer is it can't possibly. I really do not see how you can possibly argue against this.

And this is so even when the parrot does say something, but doesn't say one of the 19 specified words. For example Sheldrake says:

{quote}

Second, N’kisi could not have understood the need to limit his responses to the list of key words we had specified in advance. He has never been trained to produce specific words on demand; instead Aimée had always allowed him to use language as he pleases. In at least 10 trials he said words or phrases that corresponded to pictures Aimée was looking at but which did not involve prespecified key words. For example, in one trial Aimée was looking at a photograph of a stationary car whose driver had his head out of the window. This photograph had been chosen to represent the key word “car”. N’kisi did not say “car”. Instead he said, “Uh-oh, careful, you put your head out”, the moment Aimée noticed this unusual detail. This was not counted as a hit (or as a miss) because the phrase “head out” was not on the list of prespecified key words.

{/quote}

Now it seems very likely that N'kisa was receiving a telepathic impression here. But it can't be included because what he said wasn't on the list of predefined words. Likewise, if N'kisa had been saying another word repeatedly, standing for an object, but which wasn't the right object, but the word he kept repeating wasn't on the list, then this cannot provide any evidence for the notion that he wasn't getting any telepathic impressions.

I have to say though that I find the fact that the predefined word list was only 19 to be absolutely ludicrous. Doesn't the parrot have a vocabulary of about 900 words or something?? :confused:



So if Sheldrake and Aimee wanted to be "scientific", they would have gone ahead and reported their "null results", which would have been the numbers of times Aimee projected a telepathic image and the parrot just sat there and didn't say anything at all.



Yes, but such null results could not possibly alter the likelihood or unlikelihood that the parrot has genuine telepathic abilities. I find this absolutely clear and I am perplexed that you cannot understand this.




The thing is, they have to prove it, and so far, they haven't. And so far, their research methods are sadly lacking, IMO.



Well, I really don't know if they are or not. But since you're incorrect about the null results I don't feel confident about taking your word for it. Sorry! :)


For starters, as I mentioned, the images they chose were apparently so complex and ambiguous that they could be applied to any of a number of vocabulary words.


Well, it seems that it's going to be much more difficult testing for parrot telepathy than telepathy for human beings. This is unfortunate as I feel parrots are more likely to exhibit it on a more consistent basis. Oh well.



There's a reason why those Rhine cards are such big, simple symbols--it makes it easy to tell whether it's a "hit" or not.



Nah, they obviously wouldn't work.

[Meffy]

Originally posted by Interesting Ian


I have to say though that I find the fact that the predefined word list was only 19 to be absolutely ludicrous. Doesn't the parrot have a vocabulary of about 900 words or something?? :confused:


Just wondering. Do you find it absolutely ludicrous that Zener cards have but five symbols from which to choose? Doesn't the average human "psychic" have a vocabulary somewhat in excess of 900 words -- or, to make it proportional to the number of available choices, greater than 237 (= 900 /19 x 5) words?

Seems to me that a restricted vocabulary of symbols is pretty much mandatory unless dealing with a test subject capable of articulating very few distinct signals. But then I'm not :confused: -- that could account for the difference of opinion here.

Meffy

[Meffy]

Originally posted by Supercharts
Paranormal Parrot Recipe


What? No spam? :-)

Oops, sorry -- wrong Monty Python skit.

Meffy

[Zep]

Interesting Ian: But the parrot simply remaining silent cannot conceivably give either any evidence for or against telepathy. It might be receiving telepathic impressions, or it might not be doing so. We simply do not know because we cannot read the parrots mind! (even though it might be able to read ours). So sure, you can note the times when it's silent, but how could it possibly give any evidence for or against telepathy? The answer is it can't possibly. I really do not see how you can possibly argue against this.Absolutely true, and thereby you have raised yet ANOTHER reason why Sheldrake's claiming that the parrot is psychic is just so much dribble. The very fact that the parrot's speech is not conceivably any realistic measure or indicator of its psychic response means it is pointless taking ANY of its responses as meaning anything significant at all. For all we know, the parrot might indeed have been receiving the images, but it was just babbling its words while it did.

Honestly, what a total load of round things. Statistically flawed, methodologically flawed, original premise flawed, crackpots in charge of the research...what more do we need to see it is complete and utter sphericals???

[Interesting Ian]

Originally posted by Meffy
[B]

Just wondering. Do you find it absolutely ludicrous that Zener cards have but five symbols from which to choose?



Of course not. But this is of course a false analogy.




Seems to me that a restricted vocabulary of symbols is pretty much mandatory



Sure it is, but why not make it more than 19? Make it say 100, or if that's too difficult 50 or whatever.

[Interesting Ian]

Originally posted by Zep
Interesting Ian:

But the parrot simply remaining silent cannot conceivably give either any evidence for or against telepathy. It might be receiving telepathic impressions, or it might not be doing so. We simply do not know because we cannot read the parrots mind! (even though it might be able to read ours). So sure, you can note the times when it's silent, but how could it possibly give any evidence for or against telepathy? The answer is it can't possibly. I really do not see how you can possibly argue against this.
--------------------------------------------------------------------------------

Absolutely true, and thereby you have raised yet ANOTHER reason why Sheldrake's claiming that the parrot is psychic is just so much dribble. The very fact that the parrot's speech is not conceivably any realistic measure or indicator of its psychic response means it is pointless taking ANY of its responses as meaning anything significant at all.



Here you beg the question. Of course when it's silent there is no measure or indication of whether it is receiving any telepathic impressions. But that's why the silent times are not taken into account! When it is speaking, it might or might not give evidence of telepathic impressions. This is what the experiment is attempting to find out! On a purely emotional level, if you're the owner of the parrot and you find the parrot often makes comments appropriate to an object you are either looking at (even where the parrot cannot conceivably see), or thinking of, then the hypothesis of telepathy is surely a reasonable one. But to say that no matter what the parrot says, no matter how often it seems to be correct even though seemingly all sensory channels have been ruled out, it can never have anything to do with telepathy, is to rule out telepathy a priori. You're saying you will never believe no matter what. You are saying your prior beliefs trump empirical investigations!! :eek: :eek:




Honestly, what a total load of round things. Statistically flawed, methodologically flawed, original premise flawed, crackpots in charge of the research...what more do we need to see it is complete and utter sphericals???

You define anyone as a "crackpot" who wishes to research these things, or at least who research these things and don't come to a negative conclusion. This is begging the question in that you're not only assuming your worldview is correct, but very obviously correct. As for your charge that the experiment is hopelessly flawed, I cannot answer that. I don't have the appropriate expertise to judge. I'm afraid I'm not going to take your word that it is hopelessly flawed. For a kick off you would need to supply some sort of details as to why you think it is flawed. It may well be flawed, but no-one yet has demonstrated this to my satisfaction.

[Undodog]

Think of a card - any card.
OK now i'll randomly shout out things like "Ace" and "Red" and "Who's a pretty boy!" for 5 minutes.
Lets see how many hits i get..
More than 1 = evidence of psychic powers!

(If you're thinking of a Jack, "Who's a pretty boy!" counts as a hit.)

[Interesting Ian]

Originally posted by Undodog
Think of a card - any card.
OK now i'll randomly shout out things like "Ace" and "Red" and "Who's a pretty boy!" for 5 minutes.
Lets see how many hits i get..
More than 1 = evidence of psychic powers!

(If you're thinking of a Jack, "Who's a pretty boy!" counts as a hit.)

So far, although I know nothing whatsoever about scientific research or statistics, it is very clear to me that peoples criticisms are either hopelessly flawed or empty.

The images have to correlate with one of the 19 specified words. The misses and hits are both taken into account. You're just being silly.

[Undodog]

You mean, "This is of course a false analogy"

[Meffy]

Originally posted by Interesting Ian

Of course not. But this is of course a false analogy.

It's not an "analogy" at all, it's a simple matter of scaling symbol sets for testing to the size of the vocabulary of the creature being tested.

That's okay, I've seen that you use "false analogy" to mean "I can't refute that, so will handwave the issue and pretend I won." Nice meeting you. :-)

Meffy

[Meffy]

Originally posted by Interesting Ian


So far, although I know nothing whatsoever about scientific research or statistics, it is very clear to me that peoples criticisms are either hopelessly flawed or empty.

Oh, dear.

Don't you think it would be wise to learn a little about scientific research and statistics before making sweeping, grandiose proclamations concerning scientific research and statistics?

Just an idea, you don't have to if you don't want to. Might lend some credence to your opinions. More likely, it might change them. :-)

Meffy

[Interesting Ian]

Originally posted by Meffy


It's not an "analogy" at all, it's a simple matter of scaling symbol sets for testing to the size of the vocabulary of the creature being tested.

That's okay, I've seen that you use "false analogy" to mean "I can't refute that, so will handwave the issue and pretend I won." Nice meeting you. :-)

Meffy

I believe the parrot has a vocabulary of about 900 words. Why restrict the predefined words to just 19? Read the paper to see all the problems this causes. This is not remotely comparable to zenor cards wether it is an analogy or whatever.

[Interesting Ian]

Originally posted by Meffy


Oh, dear.

Don't you think it would be wise to learn a little about scientific research and statistics before making sweeping, grandiose proclamations concerning scientific research and statistics?

Just an idea, you don't have to if you don't want to. Might lend some credence to your opinions. More likely, it might change them. :-)

Meffy

Where have I made grandiose proclamations concerning scientific research and statistics?

If anyone were to claim that the parrot saying nothing whatsoever gives evidence for or against telepathy, then it is clear they are talking absolute nonsense. If any statistics class I attended attempted to teach me otherwise, I would simply walk out of the class or lecture (I've done this before a couple of times at University).

It seems to me that people put far too much faith in education and not enough on raw intelligence.

[Meffy]

First you say you have no understanding of scientific research or statistics, but then claim your "raw intelligence" enables you to discard any portion of mathematics you don't agree with? Mhm.

Ian, it's been nice chatting with you. Have a pleasant discussion.

Meffy

[T'ai Chi]

Originally posted by Goshawk
Because it seems to me--even me, the math-challenged person--that if you're gonna set up your test so that the photographs involved can count as BOTH hits AND misses, then you ought to hand in your Junior Scientist badge and Secret Decoder Ring and go back to playing Everquest online.

Sheesh.


That's not really a problem I don't think, since the photo itself isn't counted as a hit or a miss but rather what the parrot says while a photo is being thought upon, and if it matches the content of the photo, is what is counted as a hit or a miss.

I read the paper, and have a few comments on the statistics involved. This might be a typo, but this doesn't make any sense (bold mine):


Statistical Analysis
The data were analysed independently by Jan van Bolhuis, assistant professor of statistics at the Free University of Amsterdam, Holland. He used three methods. The first two ignored repetitions of key words by N’kisi during a given trial. The second two took repetitions into account.


3 methods, but we have the first two, and the last two, which gives 4 methods if I am reading it correctly. :) Or is it that van Bolhuis used 3 methods, but Sheldrake used 1 method in addition, and that is where the total of 4 is coming from?

With that said, I personally have school and consulting experience with permutation tests, bootstrap analysis, and Fisher's Exact test, and can see no flaws in description of the choice and carring out of these statistical tests. If there are flaws, they seem reside elsewhere.

I would personally have liked to test the hypothesis that the parrot would score higher if the parrot was actually shown the pictures, and see what resulted from that.

Goshawk,


The seemingly uncannily high success rate of "flower" must of course be chalked up to the fact that obviously it's his favorite word--he said it to more pictures than any other word (a total of 55 times altogether by my count, and it was apparently the word most likely to be repeated multiple times). So just by dint of repeating it more often, it's more likely to be a "hit" a higher percentage of the time.


Which is why I believe they did the analysis also after eliminating "flower" and still obtained a highly significant p-value of .006.

[!Xx+-Rational-+xX!]

For legal reasons TBK can’t tell us what happened to the last believer who came to his door promoting religion! I’ll just say at the time his Thai Boxing came in handy as it does here!

[S. D. Youngren]

I'm still wondering why Aimee found it necessary to teach her parrot to say "naked body."

[S. D. Youngren]

If this bird is Really psychic, they ought to be able to explain the experiment to him so that he can respond more intelligently.

Just a thought.

[Interesting Ian]

Originally posted by S. D. Youngren
If this bird is Really psychic, they ought to be able to explain the experiment to him so that he can respond more intelligently.

Just a thought.

I guess the parrot, although extremely intelligent, is not that intelligent.

[T'ai Chi]

Originally posted by !Xx+-Rational-+xX!
For legal reasons TBK can’t tell us what happened to the last believer who came to his door promoting religion! I’ll just say at the time his Thai Boxing came in handy as it does here!

LOL, maybe he did his ritual dance (the one thai boxers do before actually fighting) to get help from the spirits, wearing his superstitious headband, as well as buddha amulets in his arm bands, all the time yelling 'superstition is religion, and both are for the ignorant!'

[T'ai Chi]

Originally posted by S. D. Youngren
If this bird is Really psychic, they ought to be able to explain the experiment to him so that he can respond more intelligently.

Just a thought.

Uh, strawman?

I'm sure you have an idea of what psychic means, and the researchers have a wohle other idea, one that was tested in a scientific manner.

[S. D. Youngren]

Well, if the experiment was supposed to determine whether the bird can understand his owner's thoughts, then maybe he should be able to have the business explained to him. Or at least to understand that she wanted him to talk; that he shouldn't be sitting silently through any of the trials.

I didn't mean to imply that his inability to do so proved he couldn't "get" the images. I just don't think much of the experiment as explained, and, well...if they really had an amazing and useful psychic link, he might have understood what she wanted of him.

[Interesting Ian]

Originally posted by S. D. Youngren
Well, if the experiment was supposed to determine whether the bird can understand his owner's thoughts, then maybe he should be able to have the business explained to him. Or at least to understand that she wanted him to talk; that he shouldn't be sitting silently through any of the trials.

I didn't mean to imply that his inability to do so proved he couldn't "get" the images. I just don't think much of the experiment as explained, and, well...if they really had an amazing and useful psychic link, he might have understood what she wanted of him.

I'm unclear what you're saying. Are you saying that the owner should explain to the parrot what is required, or are you saying that what is required should be telepathically conveyed?

Also, how would you improve the experiment?

[TheBoyPaj]

I think there have already been a few suggestions on ways to improve the experiment:

Have more than 19 keywords. The parrot has a vocabulary of 900, so why pick such a small set (with a likelihood that the parrot's favourite words have been selected)?

Do not reject results simply because the parrot did not perform as you wish him to perform.

At the experiment stands the potential for errors is huge and as such removes any possible worth it may have had.

[Goshawk]

And, I would add, use pictures that are simple--for "car", have a picture of a "car", nothing else.

Etc.

[AlienX]

Ian can you please explain how you can comment so strongly upon the validity of a statistically based test when you say you have no understanding of stats?

How do you know that the silence periods should not be taken into account?, how do you know that these phases are not statistically important?. Simple answer is you don't, stats isn't just about number crunching it's about interpretation and understanding the significance of sampling etc.

The silence periods fall into this area yet you claim they are not relevent in the statistical analysis. Surely you must relent upon this specific issue as you have an acknowledged lack of understanding in this field and you simply don't know.

It's easy to see patterns in stats and draw false conclusions - especially in marginal cases.

Personally i'm sick of seeing - "statistically significant" used when what they really mean to say is it was about what you would expect only slightly +ve and well within the envelope of possibility, essentially the result means little in isolation but it's waved about as proof - not suprisingly it's by a "scientist" and not a statistician ;-).

Stats is a whole field of mathematics on it's own, just because someone can number crunch well does not mean they are a good statistician - see the bible code stuff for an example of a great pure maths specialist trying to do stats and then one of the worlds top statisticians pointing out the fundamental mistakes he's made - not in the adding up etc just the methodology of stats.

I notice you (Ian) do focus on some of the anecdotes and not the stats which is fair enough, but the real point is the stats.. really what matters here is actual discussion of the stats and methodology.
You also point at people discussing the stats and questioning their interpretation.. obviously i'm confused as to how you can do this without the tools to do so?

AX

[Interesting Ian]

Originally posted by AlienX
[B]Ian can you please explain how you can comment so strongly upon the validity of a statistically based test when you say you have no understanding of stats?

How do you know that the silence periods should not be taken into account?, how do you know that these phases are not statistically important?. Simple answer is you don't, stats isn't just about number crunching it's about interpretation and understanding the significance of sampling etc.


If you don't understand from what I've already said, then there seems little else I can say. One does not need to know anything about statistics to understand that the silent periods are irrelevant to establishing whether or not the parrot has telepathic abilities.

Can other people understand this? Or am I the only one? :confused:

[Paul C. Anagnostopoulos]

Well folks, I'm coming in late on this one. I only just read the paper last evening. The objections raised already in this thread are certainly concerns.

I have a simple question about Table 3. Column E is supposed to give the probability of N'kisi saying a given word by chance on any single trial. It is calculated by dividing the number of trials in which the word was said by the number of words said (ignoring repeats). Does it make sense to divide trials by words? I would think he should divide trials by trials.

~~ Paul

[Paul C. Anagnostopoulos]

To continue on: Imagine a set of trials similar to those reported, but where N'kisi also said many other keywords on each trial. Now take a look at the statistics. The number of hits could stay the same, but the probabilities in Table 3 column E would be lower and thus the overall expected hit rate would be lower. The more N'kisi chatters, the better he does.

Consider also the photographs chosen. For these statistics to be valid, there must be no correlation between the content of the photos and the frequencies of N'kisi's word usage. But there is no way to know if this is the case without some baseline word usage statistics. In fact, the keywords were chosen based on the ability to represent those words in pictures! Even given this problem, why aren't there the same number of photos for each word, to prevent accidental correlation between photo frequency and word frequency?

And, of course, the fact that they discarded trials in which N'kisi failed to say any keywords is really quite absurd.

~~ Paul

[CFLarsen]

Originally posted by Paul C. Anagnostopoulos
And, of course, the fact that they discarded trials in which N'kisi failed to say any keywords is really quite absurd.

Not absurd at all, if you look at - or remote view, if you like - it from the right perspective. It is called the "file drawer effect", a well-known and much used trick in the field of paranormal research to make the results look better than they actually do.

I cannot imagine that Sheldrake is not aware that this trick is a no-no in serious research.

I repeat:

I cannot imagine that Sheldrake is not aware that this trick is a no-no in serious research.

[Interesting Ian]

Paul C. Anagnostopoulos
And, of course, the fact that they discarded trials in which N'kisi failed to say any keywords is really quite absurd.
--------------------------------------------------------------------------------



Paul, for God's sake man, read my comments upon this!


Larsen
Not absurd at all, if you look at - or remote view, if you like - it from the right perspective. It is called the "file drawer effect",



No, this is certainly not the file drawer effect. The file drawer effect is ignoring negative results. One might as well accuse Sheldrake of ignoring positive results. The times when the parrot doesn't say a key word is neither a negative result nor a positive result.



a well-known and much used trick in the field of paranormal research to make the results look better than they actually do.



I believe that parapsychology is the only branch of science which actually has a policy of publishing all results, whether positive or negative.

[Paul C. Anagnostopoulos]

Ian said:
No, this is certainly not the file drawer effect. The file drawer effect is ignoring negative results. One might as well accuse Sheldrake of ignoring positive results. The times when the parrot doesn't say a key word is neither a negative result nor a positive result.
It is a trial in which N'kisi did not manage to match the stimulus. This is not an experiment to determine whether, when N'kisi says something, is it relevant to the stimulus. It is an experiment to determine whether N'kisi can match the stimulus.

~~ Paul

[Interesting Ian]

Originally posted by Paul C. Anagnostopoulos

It is a trial in which N'kisi did not manage to match the stimulus. This is not an experiment to determine whether, when N'kisi says something, is it relevant to the stimulus. It is an experiment to determine whether N'kisi can match the stimulus.

~~ Paul

The times when N'kisis fails to say anything cannot be deemed to provide evidence that he doesn't have any telepathic ability anymore than it can be deemed to provide evidence that he does have telepathic ability.

Are you truly unable to understand this? :eek:

[Paul C. Anagnostopoulos]

Carrying on, I recalculated Table 3 under the following assumption: Suppose N'kisi said "cards" on every trial, in addition to the other words he said. Then the overall expected hit rate would drop to 4.926, making the results appear even more significant. The table is incredibly sensitive to the number of words N'kisi says. I repeat my assertion that column E is bogus.

On another note, imagine a Zener card test where the circle was presented far more often than the other cards. Now there is some chance that the subject's predilections match the skewed stimulus probabilities. I think we would agree to discard such an experiment. The N'kisi experiments have this problem in spades.

~~ Paul

[CFLarsen]

Originally posted by Interesting Ian
No, this is certainly not the file drawer effect. The file drawer effect is ignoring negative results. One might as well accuse Sheldrake of ignoring positive results. The times when the parrot doesn't say a key word is neither a negative result nor a positive result.

In what case do we have a "negative" result, then?

Originally posted by Interesting Ian
I believe that parapsychology is the only branch of science which actually has a policy of publishing all results, whether positive or negative.

There is a difference between having a policy and living up to it.

Can I see this policy described somewhere?

Originally posted by Interesting Ian
The times when N'kisis fails to say anything cannot be deemed to provide evidence that he doesn't have any telepathic ability anymore than it can be deemed to provide evidence that he does have telepathic ability.

Now, me thinks you are moving the goal posts:

Originally posted by Interesting Ian
The times when the parrot doesn't say a key word is neither a negative result nor a positive result.

"No key words" or "nothing at all"?

[Paul C. Anagnostopoulos]

Ian said:
The times when N'kisis fails to say anything cannot be deemed to provide evidence that he doesn't have any telepathic ability anymore than it can be deemed to provide evidence that he does have telepathic ability.
Why would you assert this and not also assert it for trials in which N'kisi says inccorect keywords? After all, N'kisi does not know the set of keywords. Therefore, there is no difference between an absence of chatter, chatter of non-keywords, and chatter of incorrect keywords.

While we're here, why did they reject "camera" as a keyword, but none of the others? The reason given is that there was a camera in the room with N'kisi, so he said the word often. Were there any flowers, or pictures or flowers, or flowers yesterday, or is flower one of his favorite words? If so, it should have been eliminated, too.

~~ Paul

[TheBoyPaj]

Here is some number crunching I performed, based on Table 2 (the selected results). It gives a very different story. I am not a professional statistician, so I am more than willing to be corrected.

Taking the case of the word "flower" on its own:

Number of times the word was spoken = 54 (including repetitions)
Number of keywords spoken in total = 185

Probability of parrot saying "flower" at any time: 54/185 = 0.292 (not 0.197 as claimed in table 3)

Average number of keywords spoken per trial: 185/71 = 2.606 (they missed this bit out of their calculations)

Therefore, the chance of saying "flower" in any one trial: 0.292 * 2.606 = 0.761

Expected number of hits = 0.761 * 17 (no. of flower images) = 12.937

Compare this to their own expected frequency of 3.341, and I believe that they vastly underestimated the chances of this happening randomly. Nearly 13 hits from just one word. From this perspective their claim of 26 hits using 19 words doesn't look so good.

[TheBoyPaj]

Also, isn't it a bit odd that there were more than double the amount of "flower" images than of any other key word? And it just happens to be his favourite word?

Of course, it doesn't say anywhere that these pictures were randomly picked from the source library.

[Paul C. Anagnostopoulos]

Very interesting, Paj. Is there a statistician in the house? We need a statistician!

~~ Paul

[Paul C. Anagnostopoulos]

There should have been the same number of images for each keyword.

Edited to add: And perhaps they should have been selected with replacement?

~~ Paul

[TheBoyPaj]

More parrot stuff (http://www.aouc84.dsl.pipex.com/parrot.htm)

I'm bored so I went and looked at the whole table. I came up with an expected total of 21 hits. Sounds about right, doesn't it?

Now, this was working with a simplified model which has 91 pictures, each with a single keyword. In actual fact there were only 71 pictures used, but many had multiple keywords. I'm sure this will act to increase the likely number of hits.

In a word.. crackers.

[Paul C. Anagnostopoulos]

Paj, you're a man with time on your hands.

I'm tempted to vouch for your statistical method, but I probably don't know any more statistics than you do.

My kingdom for a statistician!

~~ Paul

[CFLarsen]

Originally posted by TheBoyPaj
In a word.. crackers.

Well....(wait for it....).....Polly have a cracker.... :D

[TheBoyPaj]

OK, enough now. This is my second analysis. (http://www.aouc84.dsl.pipex.com/parrot2.htm).

This time I wanted to take into account the fact that many pictures were associated with multiple keywords, thus making it much easier to achieve a hit.

So, for each picture I worked out the likelihood that the bird would say each key word (based on the parrot's own preferred frequency distribution), added them up, then multiplied this by the number of attempts the bird had for that picture.

Result: 24 hits by random probability.

He may be a pretty boy, but he's not a miracle.

[S. D. Youngren]

Originally posted by Interesting Ian


I believe that parapsychology is the only branch of science which actually has a policy of publishing all results, whether positive or negative.

Do you mean to tell me that you have NEVER seen or read of a medical study in which such-and-such was NOT found to cause a Dread Disease? Or to cure one? NEVER???

Or at least, not every day or so?

[Goshawk]

You can do a Google search for 'study "does not cause" ' and ' study "found no correlation" ' and come up with a whole slew of hits.

Testosterone replacement does not cause prostate cancer: Study (http://in.news.yahoo.com/040129/139/2b6zk.html)

Study: Birth control pill does not cause weight gain (http://www.healthypages.net/news.asp?newsid=4022)

Folic acid 'does not cause miscarriage' (http://news.bbc.co.uk/1/hi/health/1528244.stm)

And my personal favorite...

...two British scientists have laid to rest the idea that a man's shoe size is in any way correlated to the size of his penis. Their scientific study found no link between the two variables, according to a report in the British Journal of Urology International... (http://www.drbobmartin.com/2002k_09_30news04.html)

[Interesting Ian]

Originally posted by S. D. Youngren


Do you mean to tell me that you have NEVER seen or read of a medical study in which such-and-such was NOT found to cause a Dread Disease? Or to cure one? NEVER???

Or at least, not every day or so?

No, read what I said.

[Interesting Ian]

Originally posted by TheBoyPaj
[B]Here is some number crunching I performed, based on Table 2 (the selected results). It gives a very different story. I am not a professional statistician, so I am more than willing to be corrected.

Taking the case of the word "flower" on its own:

Number of times the word was spoken = 54 (including repetitions)
Number of keywords spoken in total = 185


Not 185 but 177



Probability of parrot saying "flower" at any time: 54/185 = 0.292 (not 0.197 as claimed in table 3)



They're not including repetitions. So their figure is correct.




Average number of keywords spoken per trial: 185/71 = 2.606 (they missed this bit out of their calculations)

Therefore, the chance of saying "flower" in any one trial: 0.292 * 2.606 = 0.761



This is hopelessly wrong because the parrot has the propensity to keep repeating words. It would only be correct if the probability of the parrot saying a key word was independent of any other times the word was uttered during the trial concerned. But in one trial the parrot said "flower" 10 times!! Thus your calculation must surely be hopelessly out.


Expected number of hits = 0.761 * 17 (no. of flower images) = 12.937

Compare this to their own expected frequency of 3.341, and I believe that they vastly underestimated the chances of this happening randomly.



I have to confess I have far more faith in their figures! LOL

Besides, the parrot saying flower frequently was looked at. Sheldrake has this to say:


One of the reviewers of this paper pointed out that much of N'kisi's success hinged on the frequency of his hits with the word "flower", which was both the commonest key word he said during the series of tests, and was also represented by the largest number of images. If this word were to be excluded from the analysis, the statistical significance of N'kisi's success would be lower. This is true, but post hoc. In any experiment, if the most obvious evidence is arbitrarily removed afterwards, the significance of the results will be reduced.

Nevertheless, to examine this argument more closely, we carried out a statistical analysis eliminating "flower" both as a target and as a response. Using the data from the majority scoring method, as shown in Table 2, following the BRA procedure with 20,000 random permutations, the results excluding "flower" were still strikingly significant (p = 0.006).


Well, I think we can say your argument has comprehensively been refuted!

I rather expected this though. I just wish I wasn't wholly ignorant of statistics so I could do a more in depth analysis. Oh well.

[Interesting Ian]

Originally posted by TheBoyPaj
Also, isn't it a bit odd that there were more than double the amount of "flower" images than of any other key word? And it just happens to be his favourite word?

Of course, it doesn't say anywhere that these pictures were randomly picked from the source library.

Where are you getting the information that flower is his favorite word??

[Jeff Corey]

Originally posted by Interesting Ian

I just wish I wasn't wholly ignorant of statistics so I could do a more in depth analysis. Oh well.
'Nuff said, there.

[TheBoyPaj]

I am stunned and still reeling from the revelation that Interesting Ian is not prepared to accept my analysis! ;)

I have corrected those miscounting errors (177 instead of 185) on the web page, but cannot alter the post. The web page is correct, as far as I know.

Ignoring the repeated words is NOT acceptable. How do you know they were all said in a string? The experimenters have discarded all the words which were spoken in between these keywords. It may have been telling us about its breakfast in between the repeated uses of the word "flower". If the experimenters had been a little more honest with their raw data maybe we would be able to verify if this happened. But they weren't.

Oh, and if you doubt the accuracy of my own calculations (again, I am not a professional but I seem to have a greater understanding or the principles than you), then I would also point out that the source material's mathematics were not without errors.
(Table 3, Column B should read 91, not 90)

[TheBoyPaj]

Yes, the flower issue was "looked at", and promptly discounted. I wonder why?

My little girl says the name of her grandma's dog, "Harley", an awful lot. It drives us crazy. If I get a load of photos and carefully ensure that there are more pictures of this dog than of any other (more than double the amount), I can be sure that she will get loads of hits. THAT is why it cannot be ignored, and that is why the selection of pictures in this "experiment" stinks to high heaven. Nevertheless...

Originally posted by Interesting Ian (regarding Sheldrake's claim that, when flower is removed, the results are still significant)
Well, I think we can say your argument has comprehensively been refuted!

If you remove data for all the "flower" targets then I am calculating an expected 8 hits whereas the parrot got 12.

However, this brings the number of tests down to 52. I doubt you will find a statistician (or at least an unbiased one) who would be happy to declare significance with that few trials. It's working out at roughly a one in six chance of getting a hit, Ian. The same as rolling a dice. Would you get excited if you rolled 12 sixes in 52 attempts?

If you took it upon yourself to check before making such gra... ah, what's the point?

[Paul C. Anagnostopoulos]

Ian, give it up.

Paj, I've been thinking about this since yesterday. I think your method is correct. Sheldrake's column E is flawed. Shall we tell him?

I think you should present your analysis as you did the first time, in a table that corresponds to Table 3 of the paper. This makes it clearer that you are using the overall probability of each keyword to determine the expected hit rate.

In any event, you might want to make two changes to your second table. Specify how column E is calculated. And change the heading of column F to "Probability of a Hit". The word "cumulative" implies that the column would have a running total of some sort.

~~ Paul

[Paul C. Anagnostopoulos]

I was explaining this thread to my wife this morning, telling her about finding flaws in the paper and Paj's reanalysis. She rolled her eyes, grabbed a piece of paper, and presented me with her analysis of parrot telepathy:

It is stoopid.

She thinks we're spending an inordinate amount of time on this topic. :D

~~ Paul

[TheBoyPaj]

Table 2 (http://www.aouc84.dsl.pipex.com/parrot2.htm) has been amended following Paul's suggestion.

I have also made another concession in favour of the believers. I have capped the expected frequency column for each test to "1" (after all, even if something is dead certain to happen it still only counts as one hit). It's still coming out as 19 hits from 71 tests, comparable to the Amazing Birdi's performance.

Really, I can see that the only way one can refute this conclusion is to really hold onto the idea that repetitions should not be counted. However, my point is to demonstrate that similar results would be produced by a totally random selection, where there IS no link between the words that are spoken.

Ian, I agree that parrots do seem to have a tendency to repeat themselves, but this only serves to question the merit of the test even further! The 2 minute trials heppened one after another. The testers admit that there was no way to tell the parrot that one trial has ended and another has begun. How do we know that, for some of the hits, the bird was simply repeating a thought from a previous test?

Trial 20, for example. The parrot said "phone" and it was counted as a hit. But the bird started saying "phone" in the previous test. Surely it was simply repeating its prior thought? Unless you are now claiming precognition?!

Really, if you start assuming that the bird knows what it is doing and the words it speaks are NOT random then one must also look at all the other words which that parrot spoke but which were disregarded. Why should repetitions of a keyword be taken as attempts to identify the picture but repetitions of "cracker" to be meaningless chatter?

[TheBoyPaj]

Originally posted by Paul C. Anagnostopoulos
Paj, I've been thinking about this since yesterday. I think your method is correct. Sheldrake's column E is flawed. Shall we tell him?

I expect he knows his experiment is rubbish. He doesn't strike me as the sort of person who cares.

I was explaining this thread to my wife this morning, telling her about finding flaws in the paper and Paj's reanalysis. She rolled her eyes, grabbed a piece of paper, and presented me with her analysis of parrot telepathy

I received a similar response from my own wife. How's that for a repeatable phenomenon? :D

[Paul C. Anagnostopoulos]

Sorry Paj, I didn't make myself clear. I think you should present the data as you did in your first table, one row per keyword, in the same format as Table 3 in the paper. This makes it clear that you are calculating the probability of saying each keyword by dividing the number of times that keyword was spoken by the total number of keywords spoken.

The way the data is presented in your second table makes it appear as if you calculate the probability of each keyword independently on each trial. I realize this is pedantic, but we might as well be clear.

But the real question is: Are you going to tell Sheldrake or am I? The first thing we should ask him is whether someone has already pointed this out. No reason to waste our time. :D

~~ Paul

[Paul C. Anagnostopoulos]

Paj said:
I expect he knows his experiment is rubbish. He doesn't strike me as the sort of person who cares.
Well, he payed attention when I talked to him about a flaw in his telephone telepathy experiments. You may still be right, of course. He may just go through the motions.

~~ Paul

[CFLarsen]

Originally posted by TheBoyPaj
Ian, I agree that parrots do seem to have a tendency to repeat themselves

Really????

Webster
Main Entry: parrot
Function: transitive verb
: to repeat by rote

;)


Originally posted by TheBoyPaj
but this only serves to question the merit of the test even further! The 2 minute trials heppened one after another. The testers admit that there was no way to tell the parrot that one trial has ended and another has begun. How do we know that, for some of the hits, the bird was simply repeating a thought from a previous test?

Yup.

[TheBoyPaj]

Originally posted by Paul C. Anagnostopoulos
Sorry Paj, I didn't make myself clear. I think you should present the data as you did in your first table, one row per keyword, in the same format as Table 3 in the paper. This makes it clear that you are calculating the probability of saying each keyword by dividing the number of times that keyword was spoken by the total number of keywords spoken.


The problem I had with the original format is that it does not make allowances for the fact that each picture has multiple keywords, and a hit is therefore much more likely.


But the real question is: Are you going to tell Sheldrake or am I? The first thing we should ask him is whether someone has already pointed this out. No reason to waste our time.

I will email him and point him towards this thread. Maybe he will make friends with Ian!

[Interesting Ian]

Originally posted by TheBoyPaj
I am stunned and still reeling from the revelation that Interesting Ian is not prepared to accept my analysis! ;)

I have corrected those miscounting errors (177 instead of 185) on the web page, but cannot alter the post. The web page is correct, as far as I know.

Ignoring the repeated words is NOT acceptable.


Well first of all, I have to say this is funny. Since the parrot repeats the word flower (10 times, 7 times, 4 times) precisely when those words were a hit, how are you implying that this somehow gives evidence against telepathy?? One would imagine the skeptic should be arguing quite the converse of you, and wanting only a statistical analysis omitting repeating words. Otherwise you are getting 10 hits for flower for just the one image! :eek: And this is precisely what Sheldrake did, otherwise the statistical significance would obviously be that much more impressive! But yet, despite this, your statistical analysis using repeated words is much less impressive than Sheldrakes analysis omitting repeated words. There's something deeply wrong here! :eek:


How do you know they were all said in a string? The experimenters have discarded all the words which were spoken in between these keywords. It may have been telling us about its breakfast in between the repeated uses of the word "flower". If the experimenters had been a little more honest with their raw data maybe we would be able to verify if this happened. But they weren't.


I repeat, the point is that repeated words are not independent of that same word being said before. Once the parrot latches onto a word there is a good chance it will have a psychological propensity to keep repeating it. Hell, us human beings are the same! And yet you want only a statistical analysis done incorporating repeated words. :confused:

Also you'll note that the word "flower" wasn't spoken until quite a fair few number of trials had been undergone, then after the owner of the parrot had started seeing images of flowers, this seems to have precipitated the word flower from the parrot. You allege that the parrot loves the word flower. It seems to me that a far more plausible hypothesis was that the parrot suddenly started to continually say the word flower because of the cluster of images depicting flowers at that point! That is it supports the telepathy hypothesis.


Please note that I am not saying the experimental protocol was perfect. I have already talked at length for example about my dissatisfaction with only having 19 key words. But you're doing a disservice to everyone presenting a hopelessly flawed analysis.

[Interesting Ian]

Originally posted by Paul C. Anagnostopoulos
Ian, give it up.


LOL
Doesn't it say something when a person who knows absolutely nothing about statistics blows holes in the chief skeptics statistical analysis? ;)



Paj, I've been thinking about this since yesterday. I think your method is correct. Sheldrake's column E is flawed. Shall we tell him?



Yes, you tell him. It will be interesting to see his response :)

[TheBoyPaj]

Ian, if someone who knows what they are talking about chooses to label my analysis as flawed, I'll accept it. Not from you.

[Interesting Ian]

Originally posted by Paul C. Anagnostopoulos
I was explaining this thread to my wife this morning, telling her about finding flaws in the paper and Paj's reanalysis. She rolled her eyes, grabbed a piece of paper, and presented me with her analysis of parrot telepathy:



She thinks we're spending an inordinate amount of time on this topic. :D

~~ Paul

You mean she doesn't believe in it??

[Interesting Ian]

Originally posted by TheBoyPaj
Ian, if someone who knows what they are talking about chooses to label my analysis as flawed, I'll accept it. Not from you.

Read my argument! :rolleyes:

[Paul C. Anagnostopoulos]

Paj said:
The problem I had with the original format is that it does not make allowances for the fact that each picture has multiple keywords, and a hit is therefore much more likely.
I think it does, no? Notice how the total number of images in Table 3 is 90, even though there are only 71 trials.

~~ Paul

[TheBoyPaj]

Ian. I am fully aware of your argument. I think you are mistaken and that you do not understand what this data means.

But just to show willing:

If one discards every occasion where the parrot simply repeated a word which has gone before (since we can assume he was "latched on" in that quirky parrot way), you end up with only 9 trials. Those 9 tests are the only times where a new word was used. The parrot scored hits on two of these: Trial 11 (glasses) and trial 33 (flower).

This is slightly above chance, but we're talking really, really small and insignificant numbers of tests here. Bring that to a statisticians' party and you wouldn't even get through the front door.

[Paul C. Anagnostopoulos]

Ian, take a look at column E of table 3 in the paper. This column is supposed to be the probability of saying the keyword in any given trial. Now look at how it is calculated:

(trials in which word was said) / (total nonduplicate words said)

Think about this. What does it mean to divide a number of trials by a number of words? It's like dividing kilometers by grams. At the very least, he should divide by the total number of trials, so he's dividing trials by trials. But the big problem here is that he is using the actual outcomes of the trials to calculate the expected outcomes of the trials. It's meaningless.

Paj's method is straightforward. The probability of saying a keyword in any given trial is:

P(utterance = keyword) * (mean number of utterances per trial)

The probability of an utterance being the particular keyword is quite simply the number of times that keyword was uttered divided by the total number of utterances.

The fact that the parrot repeats keywords increases his chances of uttering that keyword appropriately. This must be taken into consideration. And you should agree, because you think the opposite should be taken into consideration: If he doesn't utter any keywords, the trial should be discarded.

~~ Paul

[Interesting Ian]

Originally posted by TheBoyPaj


Really, I can see that the only way one can refute this conclusion is to really hold onto the idea that repetitions should not be counted.



They shouldn't be. And if you had the remotest sense you would see that not doing so is actually advantageous to the skeptical position.



Ian, I agree that parrots do seem to have a tendency to repeat themselves, but this only serves to question the merit of the test even further! The 2 minute trials heppened one after another. The testers admit that there was no way to tell the parrot that one trial has ended and another has begun. How do we know that, for some of the hits, the bird was simply repeating a thought from a previous test?



Yes absolutely. Many "misses" appeared to be the parrot saying words corresponding to images seen 2 minutes earlier.



Trial 20, for example. The parrot said "phone" and it was counted as a hit. But the bird started saying "phone" in the previous test. Surely it was simply repeating its prior thought?



Yes, this seems likely. But it happens more the other way around therefore favouring telepathy over the skeptical hypothesis. Therefore I rather think it's unwise of you to complain! ;)



Unless you are now claiming precognition?!



{shrugs} It could well be. But we can't of course do such a post-hoc analysis of the data. Is that not called data-mining? (Hmmm, this is a miss, how can we interpret it as a hit LOL).


Really, if you start assuming that the bird knows what it is doing and the words it speaks are NOT random then one must also look at all the other words which that parrot spoke but which were disregarded. Why should repetitions of a keyword be taken as attempts to identify the picture but repetitions of "cracker" to be meaningless chatter?

Well, it might well be getting an impression of a cracker, but we don't know do we? Likewise if it says words clearly associated with the image, but not on the list of 19 words. So it works both ways. If you're saying the list should consist of more than 19 words then I absolutely agree. But are you saying that the list of words should be vastly expanded without having any images associated with most of these words? That seems highly unfair to me.

[Interesting Ian]

Originally posted by Paul C. Anagnostopoulos
[B]Sorry Paj, I didn't make myself clear. I think you should present the data as you did in your first table, one row per keyword, in the same format as Table 3 in the paper. This makes it clear that you are calculating the probability of saying each keyword by dividing the number of times that keyword was spoken by the total number of keywords spoken.

The way the data is presented in your second table makes it appear as if you calculate the probability of each keyword independently on each trial. I realize this is pedantic, but we might as well be clear.

But the real question is: Are you going to tell Sheldrake or am I? The first thing we should ask him is whether someone has already pointed this out. No reason to waste our time. :D



I might email him and simply refer him to this thread.

[Interesting Ian]

Originally posted by CFLarsen
Originally posted by TheBoyPaj
but this only serves to question the merit of the test even further! The 2 minute trials heppened one after another. The testers admit that there was no way to tell the parrot that one trial has ended and another has begun. How do we know that, for some of the hits, the bird was simply repeating a thought from a previous test?



Yup.

[/B]

As I pointed out, I believe this is disadvantageous to the skeptical stance.

[TheBoyPaj]

Originally posted by Interesting Ian
But are you saying that the list of words should be vastly expanded without having any images associated with most of these words? That seems highly unfair to me.

Imagine you have two sisters. You challenge me to guess their names.

There are just two keywords: the girls' names.

I reel off about a hundred names (using a book of girls names) and eventually get hits on both of them.

Wouldn't you want the 98 other guesses to count?

[TheBoyPaj]

Originally posted by Interesting Ian
I might email him and simply refer him to this thread.

Already done this.

[TheBoyPaj]

Originally posted by Interesting Ian
As I pointed out, I believe this is disadvantageous to the skeptical stance.

Nope. Read my post above ("Just to show willing:")

[Paul C. Anagnostopoulos]

Paj, I took another look at your first table. You calculated the average number of keywords uttered per trial by dividing by the "number of images." But sometimes there are multiple images in one photo. I think you want to divide by the number of trials, which gives 2.49 utterances per trial and raises the expected hits to 21 or so.

~~ Paul

[Darat]

Originally posted by TheBoyPaj


Imagine you have two sisters. You challenge me to guess their names.

There are just two keywords: the girls' names.

I reel off about a hundred names (using a book of girls names) and eventually get hits on both of them.

Wouldn't you want the 98 other guesses to count?

If Ian thought that somehow it supported his monomania about his faith then his answer would be "Are you stupid that you don't see that the answer is ‘yes’?" ;)

[Paul C. Anagnostopoulos]

Paj, I suggest you also send him a concise statement of the problem. Not sure he'll want to read this thread.

~~ Paul

[Paul C. Anagnostopoulos]

Paj said:
Imagine you have two sisters. You challenge me to guess their names.

Easy pie: Interesting Ione and Interesting Irene.

~~ Paul

[Interesting Ian]

Originally posted by Paul C. Anagnostopoulos
Ian, take a look at column E of table 3 in the paper. This column is supposed to be the probability of saying the keyword in any given trial. Now look at how it is calculated:

(trials in which word was said) / (total nonduplicate words said)

Think about this. What does it mean to divide a number of trials by a number of words? It's like dividing kilometers by grams. At the very least, he should divide by the total number of trials, so he's dividing trials by trials. But the big problem here is that he is using the actual outcomes of the trials to calculate the expected outcomes of the trials. It's meaningless.



Paul, I haven't given this sufficient thought yet. I would need to read through the paper carefully and look at your argument here. I'm not sure if I'll understand it though, but I could have a shot.

My point here though is that we have a prior problem. TheBoyPaj is claiming that there is less statistical significance when including repeated words than without. But this clearly cannot be the case as the parrot continually repeats the word flower when it actually is a hit! But anyway, it seems to me that we should certainly not consider repeated words anyway. And BTW "TheBoyPaj", I mean repeated words on a given trial as you're probably very well aware!

[Interesting Ian]

Originally posted by Interesting Ian


I might email him and simply refer him to this thread.

Theboypaj? Are you doing this then? If so I won't bother.

[TheBoyPaj]

The parrot doesn't know where one trial ends and another starts!

How can you say that when a parrot repeats a word in a single trial it is nothing to be concerned about, but when it repeats a word from the last trial, or one a few trials back, it is making a whole new connection? And how far back is too far?

That is data mining at its worst, I'm afraid.

[CFLarsen]

Let's just make sure that ONE person does it. I'm not sure Sheldrake will appreciate 25 people emailing him about the very same topic....

Paj should do it. It's his analysis.

[Paul C. Anagnostopoulos]

Ian said:
My point here though is that we have a prior problem. TheBoyPaj is claiming that there is less statistical significance when including repeated words than without. But this clearly cannot be the case as the parrot continually repeats the word flower when it actually is a hit! But anyway, it seems to me that we should certainly not consider repeated words anyway. And BTW "TheBoyPaj", I mean repeated words on a given trial as you're probably very well aware!
Ian, you need to think of the experiment as a whole. The more keywords the parrot says, the greater his chances for a hit. The entire experiment rests on him chattering keywords as much as possible. The probability of uttering any given keyword in a trial must take into consideration his overall chattering level, both the probabilities of each keyword and the number said each trial.

The idea that he knows when one trial ends and the next begins, and emphasizes words within one trial by repeating them, is ludicrous.

~~ Paul

[Interesting Ian]

Originally posted by TheBoyPaj
Originally posted by Interesting Ian
But are you saying that the list of words should be vastly expanded without having any images associated with most of these words? That seems highly unfair to me.


TheBoyPaj
Imagine you have two sisters. You challenge me to guess their names.

There are just two keywords: the girls' names.

I reel off about a hundred names (using a book of girls names) and eventually get hits on both of them.

Wouldn't you want the 98 other guesses to count? [/B]

Yes of course, and no-one has suggested otherwise. But we need a predefined list, otherwise how are we able to calculate the statistical probability??

So lets say there is a 100 predefined names. By chance hits will occur 2% of the time, and misses will occur 98% of the time. So anything above 2% of hits which is of statistically significant are suggestive of telepathy. Yes??

As long as you don't expect to get hits consistently above 50% which would be clearly absurd.

[TheBoyPaj]

Originally posted by CFLarsen
Let's just make sure that ONE person does it. I'm not sure Sheldrake will appreciate 25 people emailing him about the very same topic....

Paj should do it. It's his analysis.

Well, I have emailed the guy and asked him to take a look at the thread. I just said that there is a discussion going on and that we have some specific concerns about his method (since he is no doubt unwilling to listen to people who reject the very idea outright).

Like I said, these are just my thoughts. I'm not claiming to be an expert analyst and I'd rather not go pounding on his door unless I get some confirmation that I haven't got it all terribly wrong.

[TheBoyPaj]

Originally posted by Interesting Ian


Yes of course, and no-one has suggested otherwise. But we need a predefined list, otherwise how are we able to calculate the statistical probability??

So lets say there is a 100 predefined names. By chance hits will occur 2% of the time, and misses will occur 98% of the time. So anything above 2% of hits which is of statistically significant are suggestive of telepathy. Yes??

As long as you don't expect to get hits consistently above 50% which would be clearly absurd.

No, you're not getting it. There are only two keywords. Just like there are only 19 keywords in the parrot trial. I am using my vocabulary of 100 words, just like the bird is using his vocabulary of 900 words.

But by your logic you can ONLY count a hit or a miss when I use the keywords. All those 98 other guesses should be discarded because they are "irrelevant".

[Paul C. Anagnostopoulos]

Ian said:
Yes of course, and no-one has suggested otherwise. But we need a predefined list, otherwise how are we able to calculate the statistical probability??
There is a predefined list: The book of names. In N'kisi's case, it's his complete vocabulary list, since he does not know the list of keywords.

It doesn't matter, though. Even discarding nonkeyword trials, Paj shows that the number of hits was expected.

So lets say there is a 100 predefined names. By chance hits will occur 2% of the time, and misses will occur 98% of the time. So anything above 2% of hits which is of statistically significant are suggestive of telepathy. Yes??
You're oversimplifying. The expected number hits depends on a lot of things, such as the frequency of occurrence of the targets. It also depends on the number of keywords uttered per trial, which Sheldrake does not even consider.

~~ Paul

[Paul C. Anagnostopoulos]

Crossposting at its finest, folks. Don't you just love an active thread? I do.

~~ Paul

[Interesting Ian]

Originally posted by Paul C. Anagnostopoulos
Crossposting at its finest, folks. Don't you just love an active thread? I do.

~~ Paul

Well I'll have to come back to this thread later. I don't have anymore time at the moment.

[CFLarsen]

Originally posted by Paul C. Anagnostopoulos
In N'kisi's case, it's his complete vocabulary list, since he does not know the list of keywords.

He doesn't? Isn't he telepathic? :D

[Interesting Ian]

Originally posted by Paul C. Anagnostopoulos
[B]
There is a predefined list: The book of names. In N'kisi's case, it's his complete vocabulary list, since he does not know the list of keywords.


Well we could include his whole vocabulary consisting of things that one can have images of. But some of his vocabulary the parrot scarcely uses. So would this be fair?


You're oversimplifying. The expected number hits depends on a lot of things, such as the frequency of occurrence of the targets. It also depends on the number of keywords uttered per trial, which Sheldrake does not even consider.


But if the number of keywords uttered is increased, the number of misses uttered should increase proportionatly to the number of hits, should there be no telepathy involved.

[Interesting Ian]

Yes of course, and no-one has suggested otherwise. But we need a predefined list, otherwise how are we able to calculate the statistical probability??

So lets say there is a 100 predefined names. By chance hits will occur 2% of the time, and misses will occur 98% of the time. So anything above 2% of hits which is of statistically significant are suggestive of telepathy. Yes??

As long as you don't expect to get hits consistently above 50% which would be clearly absurd.

TheBoyPaj

No, you're not getting it. There are only two keywords. Just like there are only 19 keywords in the parrot trial. I am using my vocabulary of 100 words, just like the bird is using his vocabulary of 900 words.

But by your logic you can ONLY count a hit or a miss when I use the keywords. All those 98 other guesses should be discarded because they are "irrelevant". [/B]

But in the sisters names example you must therefore necessarily either get it right or the test is discarded. With the parrot test there are 19 key words of which the image normally corresponds to only 1 of those words. So if the parrot is just chattering randomly it should get 18 times more misses than hits. Right?

[Interesting Ian]

Originally posted by TheBoyPaj


Nope. Read my post above ("Just to show willing:")

I have. You have not justified your position.

[Paul C. Anagnostopoulos]

Ian said:
Well we could include his whole vocabulary consisting of things that one can have images of. But some of his vocabulary the parrot scarcely uses. So would this be fair?
I'm not saying this to suggest a change in the protocol. I'm merely pointing out that he has his entire vocabulary available during the trials, not just the keywords. Therefore, a trial in which he utters nonkeywords should be counted as trial with no hit.

But if the number of keywords uttered is increased, the number of misses uttered should increase proportionatly to the number of hits, should there be no telepathy involved.
Perhaps, but we're not calculating that proportion.

~~ Paul

[Paul C. Anagnostopoulos]

Ian said:
But in the sisters names example you must therefore necessarily either get it right or the test is discarded. With the parrot test there are 19 key words of which the image normally corresponds to only 1 of those words. So if the parrot is just chattering randomly it should get 18 times more misses than hits. Right?
If there was only one image per trial, and if he was somehow forced to utter only keywords, and if he only uttered one per trial, and if he uttered them at random, and if the images were chosen equally among the keywords, then yes.

~~ Paul

[Paul C. Anagnostopoulos]

This experiment highlights the problem with psi experiments, namely, that they are about the statistics, not about the phenomenon.

When we use the actual counts of keywords uttered to determine the expected number of hits, we are still using the actual results to determine the expected results. Once could argue that we should simply use .107 as the probability of utterance of each keyword, that is, 19 / 177. We should assume the parrot utters keywords at random.

After all, imagine he is psychic. Then we have used the correlation between his utterances and the photos, which follows from his psychic ability, to determine the expected hit rates, thus rendering the hits meaningless.

But there is no reason at all to think that he would utter keywords at random if he were not psychic. First, We don't really know how the keywords were chosen (are there really only 30 words out of 700 that represent concrete objects?). Second, I very much doubt he says all his 700 words with the same frequency.

On the other hand, look at the ratios of keywords uttered to number of images with that keyword (Table 3, column C / B). The ratios are all over the place. If he was psychic, you'd think he would utter keywords in proportion to the number of images with those keywords, more or less.

So, as usual, without a model of the process, we don't really know how to calculate the expected number of hits. And that's at the heart of the matter. It's a hopeless mess.

~~ Paul

[TheBoyPaj]

Originally posted by Interesting Ian
I have. You have not justified your position.

Right. How about the following (taking a little bit of your ideas and mine concerning repetition):

a)Imagine we discard all repetitions of a word within a trial AND if they appear to be repetitions of the trial which occurred immediately before that one. That way we can at least try to rule out the possibility that the parrot was simply stuck on its last thought, but we allow for another image of the same type appearing later.

b)And, since that would make it impossible to score a hit if a similar image appeared twice consecutively, those tests would have to be discarded too.

c)And, in case that doesn't do it for you, we allow for a little bit of "delay" factor. We give another hit if the parrot said the matching keyword in the trial immediately after a certain image.

What would you say to an analysis based on these principles?

[Interesting Ian]

Originally posted by TheBoyPaj


Right. How about the following (taking a little bit of your ideas and mine concerning repetition):

a)Imagine we discard all repetitions of a word within a trial AND if they appear to be repetitions of the trial which occurred immediately before that one.


Well we can't do this. This makes it far far too subjective. Much better would be if the trials were seperated in time to a reasonable degree.




b)And, since that would make it impossible to score a hit if a similar image appeared twice consecutively, those tests would have to be discarded too.



See my above comment. The trials need to be separated in time to a certain extent.



c)And, in case that doesn't do it for you, we allow for a little bit of "delay" factor. We give another hit if the parrot said the matching keyword in the trial immediately after a certain image.

What would you say to an analysis based on these principles?

It doesn't seem to me that you have understood my criticisms. How often must I repeat myself?? Allow me to do so again. Let's see if you understand this time! ;)

Please note that I am not saying the experimental protocol was perfect. Indeed it might well be hopelessly flawed! Nevertheless, it doesn't appear to be hopeless flawed for the reasons you mention. I have already talked at length, for example, about my dissatisfaction with only having 19 key words. But you're doing a disservice to everyone by presenting a hopelessly flawed analysis.

So to repeat myself . . .


You are claiming there is less statistical significance when including repeated words than if we do not. But this seems to me to be implausible due to the fact the parrot continually repeats the word flower when it actually is a hit, but doesn't do so when it is a miss! For example the parrot repeats the word flower (10 times, 7 times, 4 times) precisely when those words were a hit, but does not do so when they are a miss. The crucial quation here then is how on earth can this somehow give any evidence against telepathy?? In short it makes me highly suspicious of your statistical analysis. This is apart from the question of why you wish to take into account repeated words anyway. I suspect it's because you mistakenly believe it favours skeptical beliefs. So you're making a double error. The first one is that one ought not to take into account repeated words for a reliable statistical analysis. And secondly, it seems to militate against your beliefs in anycase! LOL

I reiterate the point that repeated words spoken by the parrot are not independent of that same word being said before. Once the parrot latches onto a word, there is a good chance it will have a psychological propensity to keep repeating it. Hell, us human beings are the same! And yet you want only a statistical analysis done incorporating repeated words. :confused:

Also you'll note that the word "flower" wasn't spoken until quite a fair few number of trials had been undergone; then after the owner of the parrot had started seeing images of flowers, this seems to have precipitated the word flower from the parrot. You allege that the parrot simply likes the word flower and says it often as a matter of course. It seems to me that a far more plausible hypothesis was that the parrot suddenly started to continually say the word flower because of the cluster of images depicting flowers at that point! That is it supports the telepathy hypothesis.

[Paul C. Anagnostopoulos]

Why don't we back up a minute and ask ourselves: How can we calculate the baseline expected hit rate for each of the keywords? We need this number in order to compare the actual hit rate to see if it is signficantly higher.

I claim that Sheldrake's method is flawed. Paj's method is straightforward, yet still relies on the actual count of spoken words, thus making the expected hit rate a function of the actual hit rate.

What we really need is a few days of typical N'kisi chatter to analyze his word counts. If we did, do you think that more than 50% of his little conversations would involve only 19 words, as was the case with the trials? He supposedly has a vocabulary of 700 words! I wonder if there is a way for his owner to get him to focus attention on specific words?

Bottom line: It's hopeless. But I'll willing to continue chatting about it.

By the way, Ian, he does repeat a keyword when it's a miss. It happens with flower quite often.

~~ Paul

[TheBoyPaj]

Ian, I perfectly understand all your opinions on this subject. I just think they're all wrong.

[Stimpson J. Cat]

I think that Paul is absolutely correct that the analysis is hopelessly naive. Even if each trial was limited to one word, and separated in time, statistical analysis could only be done by determining the relative frequencies of the various keywords in the bird's speech. With multiple words per trial, you would also have to take into account how many words are spoken, as well as conditional probabilities. Not only do you have to take into account the fact that, for example, the word flower might be more likely to be used multiple times in a trial than what would be expected from the frequency estimates alone, but you also have to adjust the frequency estimates to take into account the probability of a given word occurring given the condition that at least one keyword has been used.

If this experiment were done correctly, the analysis would have to be incredibly complicated. This is mostly because the entire experimental concept is horribly flawed. If the researchers were interested in actually scientifically testing their hypothesis, they should have spent more time trying to come up with a practical way to test it. The experiment they chose, and the analysis methods they used, clearly indicate either gross incompetence, or a deliberate attempt to use unreliable methods to produce (apparently) positive results. I am inclined to suspect that they simply don't know what they are doing.


Dr. Stupid

[Ed]

Originally posted by Stimpson J. Cat
If this experiment were done correctly...

Why is this such a refrain in the swirling, ectoplasmic draped world of paranormal research?

The experiment they chose, and the analysis methods they used, clearly indicate either gross incompetence, or a deliberate attempt to use unreliable methods to produce (apparently) positive results.

Why is this such a refrain in the swirling, ectoplasmic draped world of paranormal research?

I am inclined to suspect that they simply don't know what they are doing.

Really? Given the patterns of deceit in the world of etc. why would you assume this?


Dr. Stupid [/B]

[Paul C. Anagnostopoulos]

In spite of the hopelessness of determining expected hit counts in this experiment, particularly in light of what Stimpy just posted, I feel obliged to note that Table 3, column E, of Sheldrake's paper is not as meaningless as I thought. After staring at it for awhile longer, I realized that he is computing, the following, ignoring word repetitions within a trial:

(number of times word used in all trials) / (count of words used in all trials)

This does produce some kind of probability of the word being used in any given trial, but it's horribly dependent on the actual outcomes of the trials. Paj's method is better, because it considers overall word usage across all trials. However, even Paj's method derives expected hits from word usage within the experiments.

~~ Paul

[Goshawk]

Also you'll note that the word "flower" wasn't spoken until quite a fair few number of trials had been undergone; then after the owner of the parrot had started seeing images of flowers, this seems to have precipitated the word flower from the parrot.
The first picture of "flowers" occurs on 11/3.
Subsequent pictures of "flowers" occur on 15/5,16/3, and 19/4. On 19/4, the 33rd trial in this list, he started saying "flower".

So according to you, it took a while for the picture of "flowers" to sink in telepathically, after which time the parrot started saying "flower".

However, before he started saying "flower", he said "doctor" and "water" and "medicine" a LOT more. Out of the first 33 trials listed, he said "doctor" in 10 of them, and "water" in 10 of them, and "medicine" in 12 of them.

Now, you would probably say that that was in response to the second trial's picture of "medicine, bottle, doctor, glasses".

But, three questions: why would he say "doctor" and "medicine" so many times in response to that picture, and not "bottle" or "glasses" (he says "glasses" only 6 times in the first 33 trials, and he doesn't say "bottle" at all in the first 33 trials, not until 28/5)?

And, why would he say "water" at all in response to that picture, even if it "took a while to sink in telepathically"?. There's no water in the picture.

And, the first time he said "medicine", "doctor", and "water" was before the picture of "medicine, bottle, doctor, glasses" was presented--the picture during the first trial was of a CD. A picture of "water" doesn't occur until 5/4, the sixth trial on the list, by which time he has already said "water" during two preceding trials.

And, by the 19/4 trial, the CD picture had been presented 6 times, more than any other picture. So, by your way of thinking, he should have eventually latched onto "CD" telepathically and started repeating that, since that was the thought that would have been in Aimee's mind the most. But he didn't--he stuck to "doctor", "water", "medicine", and eventually added "flower". Not "CD".

Answer: He has favorite words he just plain likes to say: on this particular day they happened to be "flower", "water", "doctor", and "medicine". And absent any other stimulus (remember, during the trials he's sitting there in a room all by himself with nothing to look at but a camera, hence his repetitions of the word "camera" which were discounted by the researchers), he'll say his current favorite words aloud, to amuse himself.

And remember, there are data missing. The table of trials is only a partial list. They didn't give the results for the trials when he said none of the words on their arbitrarily chosen list of keywords. So, he sat there and chattered away to himself, certainly saying many more of his favorite words besides "flower", "doctor", "water", and "medicine", but because some of the words he was saying weren't on the keywords list, the researchers aren't telling us about them, or what they were.

Repeating link to paper so latecomers won't have to hunt through all four pages of the thread to find it.
http://www.sheldrake.org/papers/Animals/parrot_telepathy.html

[Stimpson J. Cat]

Ed,

I am inclined to suspect that they simply don't know what they are doing.

Really? Given the patterns of deceit in the world of etc. why would you assume this?

Based on what I have read, if they were being deliberately deceitful, I think they could have done a much better job of it. Of course, there is also the possibility that they were being deliberately deceitful, and are just incompetent at that, too.


Dr. Stupid

[CFLarsen]

Originally posted by Ed
Why is this such a refrain in the swirling, ectoplasmic draped world of paranormal research?
...
Really? Given the patterns of deceit in the world of etc. why would you assume this?


It seems, that once otherwise perfectly intelligent people delve into the field of parapsychology, they discover that they have a choice:

Either:

Accept that there simply isn't evidence of any paranormal phenomenon - and they can't find it themselves, either,

or

Lower the scientific standards so much that their research becomes so tainted that it is a travesty,

or...

Resort to fraudulous means to produce the most coveted results.


When people begin to believe in paranormal phenomena, something happens to them. They get increasingly credulous and those with scientific integrity seem to completely forget whatever they learned. Willingly.

Belief is a very strong phenomenon.

[TheBoyPaj]

In the end, I'm with Stimpson on this one. It's just not possible to draw any meaning from this due to the profound errors in the design of the experiment. On that we all seem to agree (even Ian).

So, bearing this in mind, what would we suggest as a far more meaningful exercise?

Are we agreed that a more sensible protocol might be:

Prior to the tests, spend a good deal of time monitoring the parrot's speech, recording the normal frequencies of the words so that probabilities can be calculated.

Expand the list of keywords to encompass as much of the parrot's vocabulary as possible. Obvious exclusions would be words like "you", "is", "and" etc. But any nouns and verbs that can be represented by images should be left in the list. The owners of the parrot would have to attest that the parrot would know one of these things if he saw a picture of it. (ie. if it can say "swim" but cannot pick out a picture of someone swimming then it's no good)

Many, many more trials. The numbers of tests being performed here are too low to be reliably analysed.

Separate the tests with breaks, to reduce the chance of crossover between images.

When analysing data, take into account the number of keywords attached to each image, and the number of keywords spoken per trial.

Does anyone else have recommendations?

[Ed]

Originally posted by CFLarsen


It seems, that once otherwise perfectly intelligent people delve into the field of parapsychology, they discover that they have a choice:

Either:

Accept that there simply isn't evidence of any paranormal phenomenon - and they can't find it themselves, either,

or

Lower the scientific standards so much that their research becomes so tainted that it is a travesty,

or...

Resort to fraudulous means to produce the most coveted results.


When people begin to believe in paranormal phenomena, something happens to them. They get increasingly credulous and those with scientific integrity seem to completely forget whatever they learned. Willingly.

Belief is a very strong phenomenon.

Yes and particularly in religion. What do you think would happen if the Curia in the Vatican had absolute, accurate, iron clad proof that Jesus lived out his days operating a Bordello in Caesaria? Do you think for one moment that they would admit to it?

Same thing here. For whatever reasons people seem to get tainted by the paranormal. Whether a nutsie creationist saying "there are no geologic columns" in spite of voluminous evidence to the contrary or people who insist that "there is ample proof" citing silly and discredited studies (remember the late lamented Keen's exemplar cases?). Most recently we have seen, again, the bald faced assertion that the PEAR guys actually found something. It's wild.

[Paul C. Anagnostopoulos]

I wonder if a better protocol might involve fewer words, a single photo of each, and selection with replacement on each trial. This is the way Zener cards are used. Count a hit if the parrot says the appropriate word on the trial, a miss of not.

A problem with spacing out the trials is that the experiment would take forever.

But the real problem is selecting the 10 or so words to use. Either they have to be words that the parrot uses with equal frequency, or the photos have to be selected with a probability matching his frequency of use.

Also, the owner cannot have any contact with the parrot from the beginning of the experiment to the end. Otherwise she can coach him between trial groups.

I wonder if a better experiment would be Autoganzfeld with the parrot?

~~ Paul

[alfaniner]

Originally posted by Interesting Ian


...
Please note that I am not saying the experimental protocol was perfect. Indeed it might well be hopelessly flawed!

...



Then why are you arguing for its validity at all?

[alfaniner]

Originally posted by Interesting Ian


...
Please note that I am not saying the experimental protocol was perfect. Indeed it might well be hopelessly flawed!

...



Then why are you arguing for its validity at all?

Well, tried to delete this duplicate post, but it won't let me! Sure would help the forum database size if I could do that now, wouldn't it?

[T'ai Chi]

Originally posted by TheBoyPaj
Bring that to a statisticians' party and you wouldn't even get through the front door.

We don't have parties.







;)

[Hellbound]

Hmmm, a thought occured to me as I was reading this thread, how's this for an idea on testing:

Many parrots are "trained" to respond to certain words, items, or pictures with a word they've learned. For example, a parrot we had at work would say "ball" when he saw his ball, "keys" if you held keys up, and "hello" if you opened a cell phone.

If one could find the items that illicit these responses (which would be much more reliable a list of keywords) then those items or images could be used a the girl's focal points.

It would seem relatively simple then to determine, for example, the frequency with which he would say "cracker" when shown a cracker. Then, during the trials, it would make sense that non-keywords are ignored (parrots will chatter), the simple test would be that if he says the appropriate keyword, it's a hit, if not, it's a miss. Seems rather simple to me, but I'm sure there's a problem in it somewhere. Anyway, does this sound reasonable?

[Paul C. Anagnostopoulos]

Sounds like a fine idea, Huntsman. The stimuli would be things that we know the parrot normally responds to.

You know, folks, perhaps the parrot isn't telepathic at all. Perhaps it's remote viewing the stimuli. Let's take his owner out of the experiment altogether and find out.

~~ Paul

[TheBoyPaj]

Originally posted by Huntsman
Many parrots are "trained" to respond to certain words, items, or pictures with a word they've learned. For example, a parrot we had at work would say "ball" when he saw his ball, "keys" if you held keys up, and "hello" if you opened a cell phone.

I don't understand how this is different from the original experiment, except that they are real objects rather than pictures. It's still using a limited number of keywords.

Presumably the original keywords were chosen because the parrot's owner thought he would respond to those pictures.

[Hellbound]

Paj:

Yes, I see what you're saying. However, thinking about my idea and reflecting on it I can see what I was getting at, I just didn't realize the difference at first :)

Anyway, you can do the same thing with the pictures. But, the idea would be to do a control experiment as well, where the parrot was simply shown the pictures and his keyword responses were recorded. That would give a "baseline" or control reading of how well the parrot reacts when it sees the images. Then the "telepathic" viewings would at least have something to compare to. One would know which keywords each picture evoked, and with what frequency.

The only advantage to using objects would be that they leave less "wiggle room" in declaring a hit. For example, there are only a few words that might be considered a hit when the object is a ball (ball, play, perhaps a few others). Compare this to a picture of a doctor in a lab, where there are dozens, possibly hundreds of possibilities. Even if these other items are not keywords for the parrot, it would seem that using simpler images and/or singular items would, at the least, be easier for the parrot to understand.

Heck, just find, say, a dozen or two dozen items that the parrot can identify. Use these in the test. If the parrot correctly identifies the item, it's a hit...if it says nothing or does not correctly identify the item, it's a miss. I'm no statistician or experimentalist (whatever term one wishes to use), but it would seem to be a simpler protocol than that used, and less prone to error and interpretation.

[Ed]

I'm sort of surprised no one mentioned this:

http://dogtraining.co.uk/hans.htm

There will alway be woo's.

[Ed]

Parrot Stew

Recipe By : Key Gourmet CD Rom
Serving Size : 4 Preparation Time :0:00
Categories : Soups & Stews

Amount Measure Ingredient -- Preparation Method
-------- ------------ --------------------------------
Parrots
1 large can tomatoes -- mashed
1 onion -- chopped
Curry powder -- to taste

Parboil parrot, cut in pieces. Drain and dry. Then brown the pieces with salt and pepper to taste. Next prepare mixture using tomatoes, onions, curry powder, salt and pepper. Add browned parrot pieces to mixture and stew at low temperature for 3 to 4 hours. Thicken juice with flour and water for the gravy. Meat will fall off the bones.

[Interesting Ian]

Originally posted by Ed
I'm sort of surprised no one mentioned this:

http://dogtraining.co.uk/hans.htm

There will alway be woo's.

Yup, this is another example which skeptics absolutely love. Together with the birthday example. How on earth this supports their beliefs is never explained :rolleyes: A few questions. Is doubt being cast upon this parrots ability to converse intelligently as well as his alleged telepathic abilities? Why should non-human animals be so very different from humans? Are non-human animals actually conscious? Sorry, just trying to get into this mysterious mindset skeptics have about the world.

[Interesting Ian]

Originally posted by Ed
Parrot Stew

Recipe By : Key Gourmet CD Rom
Serving Size : 4 Preparation Time :0:00
Categories : Soups & Stews

Amount Measure Ingredient -- Preparation Method
-------- ------------ --------------------------------
Parrots
1 large can tomatoes -- mashed
1 onion -- chopped
Curry powder -- to taste

Parboil parrot, cut in pieces. Drain and dry. Then brown the pieces with salt and pepper to taste. Next prepare mixture using tomatoes, onions, curry powder, salt and pepper. Add browned parrot pieces to mixture and stew at low temperature for 3 to 4 hours. Thicken juice with flour and water for the gravy. Meat will fall off the bones.

My God, what a very sad despicable person you are.

[Ed]

Originally posted by Interesting Ian


Why should non-human animals be so very different from humans?


[shrugs, fart, scratches, rolls eyes, gives a nervous twitch, adjusts personal bits in underpants] Cuz it's called "Parrot Stew, not "Human Stew".

[Ed]

Originally posted by Interesting Ian


My God, what a very sad despicable person you are.

[burps, scratches butt, mops Parrot juice with last bit of bread]
Yes, but full.

[Paul C. Anagnostopoulos]

Ian said:
Is doubt being cast upon this parrots ability to converse intelligently as well as his alleged telepathic abilities?
I have no idea. Not seen any papers on that, or any evidence. Has someone claimed he can have an intelligent conversation? Is there a transcript?

I still don't understand how they managed only 30 concrete nouns out of 700 words. What, does he say things like "free will" and "ontology"?

~~ Paul

[Interesting Ian]

Originally posted by Paul C. Anagnostopoulos

I have no idea. Not seen any papers on that, or any evidence. Has someone claimed he can have an intelligent conversation? Is there a transcript?

I still don't understand how they managed only 30 concrete nouns out of 700 words. What, does he say things like "free will" and "ontology"?

~~ Paul

Don't be silly. You know what I mean.

http://www.animalsentience.com/news/2004-01-26.htm

[Paul C. Anagnostopoulos]

And here's the conversation after they went for a ride in a boat:

N'Kisi: "Remember, we went in a boat"
Aimee: "Yes! Did you like it?"
N'Kisi: "I like that - wanna go out in the boat"
Aimee: "We can't, wo don't have a boat now"
N'Kisi: "Wanna go in a boat right now"
Aimee: "I'm sorry, we can't right now - maybe we can go again later"
N'kisi: "Why can't I go in a boat now?"
Aimee: "Because we don't have one"
N'kisi: "Let's get a boat"
Aimee: "No Kisi, we can't get a boat now"
N'kisi: "I want a boat"
Aimee: "I'm sorry, baby, not today"
N'kisi: "Hurry up, wanna go in a boat. Remember? We were in a boat"

Love to hear a recording of a conversation. Something just doesn't feel right.

~~ Paul

[Mike D.]

Originally posted by Paul C. Anagnostopoulos
Love to hear a recording of a conversation. Something just doesn't feel right.

~~ Paul

Paul,

You can hear a recording of a conversation here:

http://www.sheldrake.org/nkisi/nkisi1_text.html

Mike

[AlienX]

Well i've just been through this entire thread as i've been busy of late and my initial conclusion that this test does not demonstrate either a +ve or -ve result for telepathy has not changed. So ultimately the test conclusion is wrong.

Initially I thought a test with the parrot was difficult and why not do something really crazy like use a person in place of the parrot?

Ian seems to be saying that certain concerns raised do not show that telepathy does not exist.. that's not what I think the situation is here. Were looking at this data in the light that people say it proves telepathy. Unfortunately there are serious concerns over the methodolgy and data manipulation. This data is saying that telepathy exists and that it proves it somehow, yet it seems that it does not show anything. So I don;t think people are trying to show that the parrot is not telepathic simply that the test does not show anything either way.

Use a person next time - use pictures that only one word can be applied to and allow the "telepath" to say only one word in reference to a specific period.... ohh yeah it's been done already and i've not seen anything yet which proves telepathy conclusivly to me yet.

I see no problem in investigating this field but at least make some effort and do it right, stuff like this only gets in the way of finding out if there is any truth to the belief of telepathy.

Looks like we have yet another inconclusive test on our hands.... how you can conclude this as proof is beyond me, I think the whole test is irrelevent due to the fundamental method design flaws. Clearly its going to take a well thought out test to show anything even using human subjects. It's poor "scientists" like these demonstrating questionable science methodology which hold everyone else back in the end - if they keep producing this crap and if they do get meaningful data at some point people may be inclined to ignore it out of hand - The boy who cried wolf so to speak ;-)

AX

[Paul C. Anagnostopoulos]

Interesting audio recording of N'kisi. Something is still bothering me, but I'm not sure what it is.

~~ Paul

[AlienX]

Yep it's an amazing animal even without the telepathy claims.

[Goshawk]

I would bet that what's bugging you, Paul, is the fact that that transcript isn't a "conversation"--it's simply a record of a trained animal responding to certain visual and sound cues with the appropriate sounds.

Here's a list of the words and phrases that Nkisi has apparently learned are appropriate to playing with this particular toy.

There's a square.
There's a square.
There's a square.
There's a square.

There's a little square.

Look at the square.
Look at the square.

Look, there's a square!

Look!

Star.

Wow! there's a...
Wow

Lemme
Lemme try
Lemme write
Lemme show you.

That's so cool!
That's so cool!

Listen!

Can you see that?

Can ya hear
Can ya hear

You pushed the button.
Could push the button.
You could...

Triangle.And he sings "Old McDonald", and whistles, and mews like a cat.

Hardly a conversation.

[CFLarsen]

Originally posted by Mike D.
You can hear a recording of a conversation here:

http://www.sheldrake.org/nkisi/nkisi1_text.html

Pardon me, but where in that recording does N'Kisi do anything else but - at most - repeat what the toy "said"?

Aimee holds the toy to N'Kisi: Subtle guiding? You bet.

Aimee is the one reacting to what N'Kisi says.

Aimee should be removed from the experiment.

[S. D. Youngren]

Removed altogether? Ah, but if she's the one person he has a psychic link to...

I do like the suggestion about getting a baseline on his utterances. This should be done out of Aimee's presence, just like the trials.

I do have a very good statistician on the premises here: my husband. Let's see if he can squeeze this parrot thing into his schedule.

[Darat]

Originally posted by Interesting Ian


My God, what a very sad despicable person you are.

What a very strange reaction, I haven’t noticed you react like this to other meat based recipes here.

I've eaten parrot, any reason why I shouldn't have? It's just another bird like chicken, ostrich, turkey and duck and so on.

[Undodog]

Originally posted by Darat


What a very strange reaction, I haven’t noticed you react like this to other meat based recipes here.

I've eaten parrot, any reason why I shouldn't have? It's just another bird like chicken, ostrich, turkey and duck and so on.

Quick, Ian, pretend you're deeply offended and make out that your the victim of torment by the evil skeptics.

---

I think Aimee was the parrot in that 'conversation'.
It was no better than me talking to my cat:
"What are you doing?"
"-meow-"
"I see. And when will you be finished?"
"-meow-"
"OK. See you later"
"-meow-"
...etc

(actually that would be paranormal as my cat died months ago)

[CFLarsen]

Originally posted by S. D. Youngren
Removed altogether? Ah, but if she's the one person he has a psychic link to...

Distance is not a factor. :)

[TheBoyPaj]

Originally posted by Goshawk
I would bet that what's bugging you, Paul, is the fact that that transcript isn't a "conversation"--it's simply a record of a trained animal responding to certain visual and sound cues with the appropriate sounds.

And doesn't it sound spookily like Aimee herself? I'm not suggesting that its not the parrot speaking, I'm not familiar with any speaking parrots so I don't know if they mimic their owners so exactly. But if I owned that thing it would freak me out.

[Undodog]

I'm no expert either but Im quite sure that they mimic the sound.
It doesnt understand what it's saying or if it's saying one word or two or three.
It also says 'square' exactly like the toy.

[Darat]

Originally posted by TheBoyPaj


And doesn't it sound spookily like Aimee herself? I'm not suggesting that its not the parrot speaking, I'm not familiar with any speaking parrots so I don't know if they mimic their owners so exactly. But if I owned that thing it would freak me out.

Two places near me have parrots, one the local pet shop and they have a parrot that "pings" like the till being opened and in the barbers I use they have a parrot that "trills" like the door opening.

I think they do just mimic the sounds, so I wouldn't be surprised if they mimic the sound of the owner's voice.

[Ed]

Originally posted by Darat


What a very strange reaction, I haven’t noticed you react like this to other meat based recipes here.

I've eaten parrot, any reason why I shouldn't have? It's just another bird like chicken, ostrich, turkey and duck and so on.

It is understandable, Darat. He thinks he is a Pirate.

[Darat]

Originally posted by Ed


It is understandable, Darat. He thinks he is a Pirate.

And before anyone asks me "Yes it did taste a bit like chicken".

[Interesting Ian]

Would you also eat dolphins Darat?

[Darat]

Originally posted by Interesting Ian
Would you also eat dolphins Darat?

Well tell me first what is so terrible about eating a parrot compared to say a chicken.

[Interesting Ian]

Originally posted by Darat


Well tell me first what is so terrible about eating a parrot compared to say a chicken.

You would make no moral distinction between eating a parrot who arguably is capable of rudimentary conversation, and a chicken?

[Ed]

Ian?
http://home.aaris.net/~nce/Captain_Chris_the_Pirate.jpg

[CFLarsen]

Originally posted by Interesting Ian
You would make no moral distinction between eating a parrot who arguably is capable of rudimentary conversation, and a chicken?

Conversation? Depends on what you mean by that. But, let's just assume that a parrot indeed can learn sounds that - to us - are meaningful, and that it understands, e.g. "Yes" and "No".

Ian: "Polly want a cracker?"
Polly: "Yes".

OK, that's a conversation. Not a Socrates-level one, but nevertheless: Communication occurred. I'm sure you can recognize the level.

What about a chicken? A chicken (like a lot of animals) can be taught to do the same: Pick one lever for food, another for nothing.

The only difference here is, that the parrot communicates by way of it's ability to mimick sound. The chicken does it by its beak.

Ian, aren't you merely attaching human attributes to an animal, just because it is good at mimicking sound?

[Darat]

Originally posted by Interesting Ian


You would make no moral distinction between eating a parrot who arguably is capable of rudimentary conversation, and a chicken?

It's a question I can't really just give a yes or no answer because of how you've framed it.

"Conversation" to mean means much more then just communication, to use an online definition (Encarta) :

Conversation: 1. casual talk - an informal talk with somebody, especially about opinions, ideas, feelings, or everyday matters.

So as soon as you say conversation in the above question it loads it with too many assumptions for a quick an easy answer.

One way I can answer it is that I don’t consider a parrot anymore a person i.e. a human being then I do a chicken, despite the parrot being able to mimic human vocal sounds*, therefore my conclusion based on that premise is that I don’t think there is any moral distinction between eating a parrot or chicken.


(*Remember parrots can mimic a huge range of sounds, from doorbells to cash registers and just because they can make those sounds I don’t assume a parrot has all the other characteristics of a cash register.)

[Interesting Ian]

Originally posted by CFLarsen


Conversation? Depends on what you mean by that. But, let's just assume that a parrot indeed can learn sounds that - to us - are meaningful, and that it understands, e.g. "Yes" and "No".

Ian: "Polly want a cracker?"
Polly: "Yes".

OK, that's a conversation. Not a Socrates-level one, but nevertheless: Communication occurred. I'm sure you can recognize the level.

What about a chicken? A chicken (like a lot of animals) can be taught to do the same: Pick one lever for food, another for nothing.


The only difference here is, that the parrot communicates by way of it's ability to mimick sound. The chicken does it by its beak.

Ian, aren't you merely attaching human attributes to an animal, just because it is good at mimicking sound?

You need to be careful here. The parrot might well be responding "yes" because s/he understands the meaning of a word and wishes to convey this message. Alternatively, it could be saying "yes", not because it understands the meaning of "yes", but because it precipitates a desired response or event. The latter is what the chicken is doing (if indeed chickens can do this). The former seems to be what the parrot is doing (i.e it can understand the abstract concept of "yes"). All the difference in the world.

BTW, delighted you're not ignoring me like others are. :) (see my feelings thread)

[Paul C. Anagnostopoulos]

Darat said:
Remember parrots can mimic a huge range of sounds, from doorbells to cash registers and just because they can make those sounds I don’t assume a parrot has all the other characteristics of a cash register.
This bears repeating, regardless of your culinary taste for parrots.

Perhaps we could teach N'kisi about ten philosophy words. He could carry on a perfectly good philosophy conversation then, I bet.

And we could teach him ten Freudian words and we'd have an expert psychoanalyst.

~~ Paul

[Darat]

Originally posted by Paul C. Anagnostopoulos

This bears repeating, regardless of your culinary taste for parrots.

...snip...

~~ Paul

Can I just make it clear I have had "parrot stew" once, I do not habitually eat parrot, I do not crave the taste of dead parrots.

Goodness me I'll be castigated next for saying I've eaten dog...

[Ed]

Originally posted by Darat


Can I just make it clear I have had "parrot stew" once, I do not habitually eat parrot, I do not crave the taste of dead parrots.

Goodness me I'll be castigated next for saying I've eaten dog...

Querry: Was the parrot dead or sleeping? What species, Norwegian Blue?

Darat, are YOU a Pirate? Why this fascination with parrots? Do you wear a feather boa?

[alfaniner]

Originally posted by Paul C. Anagnostopoulos

...
Perhaps we could teach N'kisi about ten philosophy words. He could carry on a perfectly good philosophy conversation then, I bet.

...
~~ Paul

Well, with Ian anyway.

[CFLarsen]

Originally posted by Interesting Ian
You need to be careful here. The parrot might well be responding "yes" because s/he understands the meaning of a word and wishes to convey this message. Alternatively, it could be saying "yes", not because it understands the meaning of "yes", but because it precipitates a desired response or event. The latter is what the chicken is doing (if indeed chickens can do this). The former seems to be what the parrot is doing (i.e it can understand the abstract concept of "yes"). All the difference in the world.

There are a lot of "might", "could be", "seems to be" here. Not much to build on.

It seems to me that you need to be careful.

[Interesting Ian]

Originally posted by CFLarsen


There are a lot of "might", "could be", "seems to be" here. Not much to build on.

It seems to me that you need to be careful.

WOW, what a comprehensive refutation of what I said! :eek: Does your intellect know of no limits Claus?? :eek:

[Darat]

Originally posted by Ed


Querry: Was the parrot dead or sleeping? What species, Norwegian Blue?

Darat, are YOU a Pirate? Why this fascination with parrots? Do you wear a feather boa?

The parrot was no more! He had ceased to be! 'E's expired and gone to meet 'is maker! '... THIS WAS AN EX-PARROT!!

Bloody hell you mention you've had one bloody munch on one blasted parrot and you're never allowed to live it down...

[CFLarsen]

Originally posted by Interesting Ian
WOW, what a comprehensive refutation of what I said! :eek:

I merely pointed out that there are a lot of unknowns.

Not everything you have to say need to be "refuted", Ian. You shouldn't be so confrontational all the time.

Originally posted by Interesting Ian
Does your intellect know of no limits Claus?? :eek:

And you really wonder why people find you rude?

[Interesting Ian]

Originally posted by CFLarsen
Originally posted by Interesting Ian
Does your intellect know of no limits Claus??


Claus
And you really wonder why people find you rude? [/B]

Where did I wonder that?

[CFLarsen]

Originally posted by Interesting Ian
Where did I wonder that?

It was sarcasm, Ian.

[Ed]

Originally posted by Darat


The parrot was no more! He had ceased to be! 'E's expired and gone to meet 'is maker! '... THIS WAS AN EX-PARROT!!

Bloody hell you mention you've had one bloody munch on one blasted parrot and you're never allowed to live it down...

Hmmmmm....

Darat "Two Parrots" Jackson:D

[Goshawk]

And doesn't it sound spookily like Aimee herself? I'm not suggesting that its not the parrot speaking, I'm not familiar with any speaking parrots so I don't know if they mimic their owners so exactly.
I thought the two identical voices was startling, too, at first, until I realized what was happening. Parrots will also mimic doorbells, telephones, microwave ovens, dogs barking, really, any sound that they're exposed to.

And African greys have long been legendary even among parrots for their mimicry skills. The other famous talking parrot, Alex, is also an African grey.

http://www.alexfoundation.org/research/index.html

[Darat]

Originally posted by Goshawk

...snip...

And African greys have long been legendary even among parrots for their mimicry skills. The other famous talking parrot, Alex, is also an African grey.

http://www.alexfoundation.org/research/index.html

I wonder if you could train one to recite a recipe for parrot stew, that would be useful.

Imagine the cooking instructions, it wouldn’t be the standard “see over for cooking instructions”, it would be “listen to the birdie”, no need to scramble around looking for your reading glasses.

(Edited to add: And no I'm not obsessed with eating Polly Ed.)

[Paul C. Anagnostopoulos]

Get your parrot here! Come get your parrot! Makes a great pet. Will repeat just about anything you tell it, or anything it hears, man or machine. Get your personal Polly here! Tired of it? No worries, mate! Hold up a cleaver in front of it and it's gay-ron-teed to repeat a marvelous parrot stew recipe. It's the world famous Darat-Parrot stew recipe. Once it completes the recitation, you're ready to go. Get your parrot here! On sale now!

~~ Paul

[S. D. Youngren]

Originally posted by CFLarsen


Distance is not a factor. :)

So they say. All I was saying is that my idea of a nice tight experiment would involve getting a baseline under the same conditions as the experiment itself, including the removal of the visual stimulus of the bird's owner (as she is removed for the experiment). She would remain the person whose mind the bird is supposed to read because the claim is that he can read her mind. And, after all, let's be sports and give him a chance.

[S. D. Youngren]

Hope that doesn't come off sounding humor-impaired. But it seemed to me I might do well to explain.

[TheBoyPaj]

I just thought I'd post Rupert Sheldrake's recent reply to my email:

------------------------------------------------------------

Thanks for telling me about this. I've only just received it because I've been away. I'm impressed by the detail you have gone into in this discussion.

In the JSE where it was published, they also published the referees comments, which cover some of the ground in the online discussion. I don't know if those remarks are online or not.

Strangely enough the discussion does not cover the randomized permutation analysis, which is the most accurate detailed way of looking at the results. This was carried out independently by Jan van Bolhuis in Amsterdam, and one of the referees did another version, taking into account all sessions where nkisi was silent, with very similar results.


Best wishes
Rupert Sheldrake

------------------------------------------------------------

The permutation analysis referred to is the assertion that, when the order of the pictures are rearranged, there no longer appears to be a correlation. This is contrary to what one might expect if the parrot was speaking randomly.

Of course, we only have Mr Sheldrake's word for this, as the report does not give the actual results, just a summary of them. Given that we have found what we feel are some pretty deep oversights in the analysis as published, we cannot tell if there are similar problems with the permutation analyses.

[Paul C. Anagnostopoulos]

I believe the permutation analyses also involves ignoring repeated words on each trial. Does that make sense for a permutation analysis?

I can feel my headache returning . . .

~~ Paul

[TheBoyPaj]

Don't get stressed again. I think we have laid this particular parrot to rest, to my satisfaction anyway.

Rupert, if you're reading this, maybe another experiment could be designed which takes into account some of our suggestions?