Psychic parrot? What are the BBC thinking?

[CFLarsen]

CFL, I can't reply to your overly anal interpretation of my post. Sorry, that is the best way to describe it, I'm not intending to be insulting.

I am not interpreting your post in any way. I am going with what you claim.

I didn't make any "claim" in the context you see it in. I tried to recall what I had seen/read. You're taking that out of context and demanding further reply. It was a general comment about what I recalled that started this exchange, not a citation of a particular study.

Excuse me, but when you say that apes can communicate via sign language, then you - you - are making a claim. So, you - you - back it up with evidence.

So from there I tried to discuss the whole picture. You remain hung up on one detail of one study (or two). Had I presented any particular study as proof of xyz I might take the time to defend it. But I didn't.

Do you understand why I find this question so crucial?

As far as communicating with sign language with strangers, the researchers haven't claimed to have taught American Sign Language to primates. They have taught a modified version. As long as the stranger knows the same version, the answer is yes.

Evidence?

 

[Skeptic Ginger]

CFL, it is impossible to communicate with you. You are ignoring all the points of this discussion and demanding I provide ??????? What claim is it you think I must back up here? I made no specific claim. I said I recalled the researchers controlled for researcher cues. I will try one more time then I will not repeat myself again. Get as annoyed as you want to. If you don't make an effort to understand what I posted before and what I am posting now, I cannot respond to your focus on a side issue you have distorted my statement about.

I understand why the evidence is crucial. What I don't understand is why you can't see past your anal focus on a statement I made that was from my general recollection as an example of controlling for researcher cues. If you want to have this discussion, then discuss the issues, not what you think I claimed.

Take a deep breath, repeat after me, "Skeptigirl was speaking non-specifically. Skeptigirl was not making a claim of a single specific fact. Skeptigirl was giving an example from her recollection, not from a cited study. Skeptigirl recalls seeing a program on the research. In that program, Skeptigirl recalls seeing the researchers with masks on to prevent researchers cues other than the signing."

Now here's the important part, "Skeptigirl cannot recall each and every one of the methods used to control for researcher cues but she has the overall impression the research has been well done and the researchers have used proper methodology."

Now let's start from there. Go back again and I will not continue this exchange so you chose.

This is a very complicated subject. It takes hours of reading to understand it and to understand the controversy. We disagree on the conclusions. But so do many other people. There are not scientists on your side and only woo woos on mine. There are opposing scientific opinions as to what is being observed.

I posted a link to one very in depth analysis of the controversy and summarized what I thought was important to examine as factors which might be influencing the way scientists are drawing different conclusions based on the same evidence.


Now, I take it your issue is you believe ALL of the research was poorly done? You believe NONE of the research was done with proper controls for researcher cues? Is that the point you are trying to make by ignoring what I have posted in the last 3 posts and persisting on asking me to support a claim I did not make?

Why discuss a single study or something I vaguely recall? What I think you want is to say none of the evidence was properly collected and or evaluated. That I can discuss with you. (Maybe, you don't seem interested in anything other than one issue.)

And the last point? You want evidence they used a modified version of sign language? I don't get that one. Or are you also claiming I should personally provide evidence to you that an entire university research program validated their methodology tools?

 

[CFLarsen]

Look, it's perfectly simple.

I am asking if the apes could understand sign language - whatever type - from strangers.

If they can, then it is strong evidence that they do understand sign language. If they can't, then it is strong evidence that the handlers are over-interpreting - consciously or unconsciously, we don't know - what the apes are doing. And therefore, evidence that the apes do not understand sign language.

You said yes: They could understand strangers. That is a specific claim. So, I am asking for evidence of this.

Now, buried in all those words, I hear you say "I think I read it somewhere".

That's not evidence. That's your recollection. We can discuss that, of course, but it isn't evidence. You have to understand that.

Here's the crux of the matter: I am not commenting on the rest of the research, or how it was done. Because it doesn't matter.

I am solely interested in the evidence. The only thing that matters is, can the apes communicate with sign language? The test is to see if they can do it with strangers. That's called a control: Does it work, when you're not the one doing the test?

As long as we don't know that, it is futile to discuss what quality the rest of the research was done. All I want to see is the evidence. Then, we can look at how it was done, to see if the evidence is sound.

Now: Do you have any evidence that apes can understand sign language?

 

[Elaedith]

Clearly, from the links and quotes skeptigirl came up with, the apes clearly can see the people.

How do you sign communicate with someone you can't see?

I'm not talking about anyone else's links. I'm talking about the video footage of studies that I have seen.

These studies didn't use sign language, but were on interpretation of verbal language, and use of artificial languages based on visual symbols. The use of visual symbols on a keyboard is preferred to sign language by some researchers and allows communication without anyone being present.

 

[Dave1001]

I'm not talking about anyone else's links. I'm talking about the video footage of studies that I have seen.

These studies didn't use sign language, but were on interpretation of verbal language, and use of artificial languages based on visual symbols. The use of visual symbols on a keyboard is preferred to sign language by some researchers and allows communication without anyone being present.

Visual symbols on a keyboard does seem like a superior way to control for Clever Hans effect. Also, using visual symbols on a keyboard it seems to me like it wouldn't be hard to create double blind experiments.

For example, the researcher wouldn't know if they were communicating with a chimpanzee or with a computer program generating random responses, or non-random responses that nevertheless aren't actual linguistic communications of the type the experimenters are looking for.

ETA: CHCI claims to do double blind studies with ASL, although it's hard for me to see how they can pull that off using ASL rather than visual symbols on a keyboard.

http://www.cwu.edu/~cwuchci/faq.html


Research
Q. How do the chimpanzees use ASL?

A. Under double-blind conditions, we have found that the chimpanzees communicate information in American Sign Language (ASL) to human observers. They use signs to refer to natural language categories: e.g. DOG for any dog, FLOWER for any flower, SHOE for any shoe, etc. The chimpanzees acquire and spontaneously use their signs to communicate with humans and each other about the normal course of surrounding events. They have demonstrated an ability to invent new signs or combine signs to metaphorically label a novel item, for example: calling a radish CRY HURT FOOD or referring to a watermelon as a DRINK FRUIT. In a double-blind condition, the chimpanzees can comprehend and produce novel prepositional phrases, understand vocal English words, translate words into their ASL glosses and even transmit their signing skills to the next generation without human intervention. Their play behavior has demonstrated that they use the same types of imaginary play as humans. It has also been demonstrated that they carry on chimpanzee-to-chimpanzee conversation and sign to themselves when alone. Conversational research shows the chimpanzees initiate and maintain conversations in ways that are like humans. The chimpanzees can repair a conversation if there is misunderstanding. They will also sign to themselves when alone and we have even observed them to sign in their sleep.

Q. What are some specific directions for future research?

A. One of our long-term research focuses is developing and evaluating enrichment methods. Enrichment serves to make life interesting for this family of chimpanzees and potentially help other captive chimpanzees as well. With regard to communication we have another long-term project that is studying gestural dialects used by this family of chimpanzees and free-living chimpanzees in Africa. In addition we still examine these chimpanzees' use of ASL with humans and each other.

It boggles my mind if this type stuff hasn't already been done experimentally. Has it? If so, the names of the researchers?

 

[CFLarsen]

I'm not talking about anyone else's links. I'm talking about the video footage of studies that I have seen.

These studies didn't use sign language, but were on interpretation of verbal language, and use of artificial languages based on visual symbols. The use of visual symbols on a keyboard is preferred to sign language by some researchers and allows communication without anyone being present.

You wouldn't happen to have a link to a video somewhere?

 

[Elaedith]

Visual symbols on a keyboard does seem like a superior way to control for Clever Hans effect. Also, using visual symbols on a keyboard it seems to me like it wouldn't be hard to create double blind experiments.

For example, the researcher wouldn't know if they were communicating with a chimpanzee or with a computer program generating random responses, or non-random responses that nevertheless aren't actual linguistic communications of the type the experimenters are looking for.

It boggles my mind if this type stuff hasn't already been done experimentally. Has it? If so, the names of the researchers?
[/i]

I don't know. Most of the research I have seen is quite old. For example the research with Lana involved a machine interpreting the commands and responding without any experimenter intervention. The machine would deliver different types of food, open and close blinds, play slide shows etc when the ape typed in grammatically correct 'sentences' in an artificial language. However, all these sentences could simply be rote learned and therefore would not count as spontaneous language use (although I believe Lana was supposed to have made up novel requests, such as asking the machine to tickle her).

I would not take this as evidence of language myself, but it is obviously not a clever Hans situation when a computer is 'interpreting' responses.

 

[Dave1001]

I don't know. Most of the research I have seen is quite old. For example the research with Lana involved a machine interpreting the commands and responding without any experimenter intervention. The machine would deliver different types of food, open and close blinds, play slide shows etc when the ape typed in grammatically correct 'sentences' in an artificial language. However, all these sentences could simply be rote learned and therefore would not count as spontaneous language use (although I believe Lana was supposed to have made up novel requests, such as asking the machine to tickle her).

I would not take this as evidence of language myself, but it is obviously not a clever Hans situation when a computer is 'interpreting' responses.

Right, but there's probably a threshhold of the sentences a machine could interpret, that doesn't delve much into grammatical complexity.

I thinking something like a chimp turing test could be useful. I'd really like to see this done with deliberately-stunted, non-linguistic computer generated answers vs. chimp generated answers and see if researchers could tell the difference.

 

[Elaedith]

Right, but there's probably a threshhold of the sentences a machine could interpret, that doesn't delve much into grammatical complexity.

I thinking something like a chimp turing test could be useful. I'd really like to see this done with deliberately-stunted, non-linguistic computer generated answers vs. chimp generated answers and see if researchers could tell the difference.

People can interpret sentences without much grammatical complexity also. Most parents don't have too much difficulty working out what a toddler at the 'two-word utterance' stage means to say without grammar.

I don't know how you would test a chimp against a computer using 'deliberately-stunted' answers, because the most you could show is that the chimp is worse at grammar than a deliberately stunted computer. Its just a question of how stunted the computer has to be before you can't distinguish it from a chimp.

 

[Dave1001]

People can interpret sentences without much grammatical complexity also. Most parents don't have too much difficulty working out what a toddler at the 'two-word utterance' stage means to say without grammar.

I don't know how you would test a chimp against a computer using 'deliberately-stunted' answers, because the most you could show is that the chimp is worse at grammar than a deliberately stunted computer. Its just a question of how stunted the computer has to be before you can't distinguish it from a chimp.

Right. And that provides us with very useful information, that makes it easier to avoid reading things into chimp communication that aren't there. For example, I think it might sort out whether chimps are making novel syntactical patterns or are just randomly punching different key combinations -if researchers consistently also interpret occasional computer randomness and novel syntactical patterns.

 

[Elaedith]

You wouldn't happen to have a link to a video somewhere?

No. I don't know of any videos online.

 

[Amapola]

.....snip.......

I am asking if the apes could understand sign language - whatever type - from strangers.

If they can, then it is strong evidence that they do understand sign language. If they can't, then it is strong evidence that the handlers are over-interpreting - consciously or unconsciously, we don't know - what the apes are doing. And therefore, evidence that the apes do not understand sign language.

.....snip.......

OK, I thought I would be reading about parrots and whether or not they are psychic (seems pretty unlikely to me) but instead this is about animal communication. All right!

I thought the above quote was a little strong of an assumption, if by "stranger" it is taken to mean anyone at all. I have friends with very young children, and while the parents can understand and talk to them just fine, I can't understand them and they apparently can't understand me. (I used to be particularly puzzled by the phone "conversations" I had with my young nieces!) I don't therefore conclude that the child can't use or understand language; I simply conclude that I don't know the child well enough to communicate with it at that time. With children they usually grow up and become more skillful with language; that's not going to happen with an animal.

I think by "stranger" you can't just pick some person off the street and have them attempt to communicate with an animal, with any hope of actual good results. I think the person has to be someone with training in the behaviour of that particular animal. Otherwise, you could run the risk of (a) the person can't understand the animal at all, even though the animal is genuinely communicating or (b) the person misinterprets a normal behaviour of the animal as an attempt to communicate, when it actually is a random normal action that is common for that type of animal. Someone with real training and skill in the behaviour of that particular animal might not seem like a true double-blind test, but it seems to me that would be the only way to avoid getting wrong results.

I've never worked with chimps or apes so I don't know how likely it is that they can develop language. I have found Skeptigirl's links really interesting reading though, so thanks, Skeptigirl! It's an interesting subject but I doubt we can settle it one way or the other here.

 

[CFLarsen]

I think by "stranger" you can't just pick some person off the street and have them attempt to communicate with an animal, with any hope of actual good results. I think the person has to be someone with training in the behaviour of that particular animal. Otherwise, you could run the risk of (a) the person can't understand the animal at all, even though the animal is genuinely communicating or (b) the person misinterprets a normal behaviour of the animal as an attempt to communicate, when it actually is a random normal action that is common for that type of animal. Someone with real training and skill in the behaviour of that particular animal might not seem like a true double-blind test, but it seems to me that would be the only way to avoid getting wrong results.

But the point is that the animal is communicating via sign language - not that it is interacting with people it knows, or exhibiting normal behavior.

 

[Amapola]

But the point is that the animal is communicating via sign language - not that it is interacting with people it knows, or exhibiting normal behavior.

All right. But I don't understand why you assume an animal would cease exhibiting normal behaviour because it was supposedly communicating via sign language? I think the animal would continue to behave in a normal fashion and the sign language would be an added behaviour - I don't think the new behaviour would cause all the other normal behaviours to cease. For example the animal might wave away a fly, which could be interpreted as a "sign" when all it amounts to is the animal made the motion for another reason which has to do with its normal behaviour. A trained person would know that that was not a sign at all, and could discount it.

This is all hypothetical to me because I have no experience with sign language. I don't really know if young children who are learning to sign are more difficult to understand than adults who are adept at signing, as in language use in young children, or if young children are immediately and universally understood by anyone who can sign. Do you know anyone who could answer that?

 

[Miss Whiplash]

All right. But I don't understand why you assume an animal would cease exhibiting normal behaviour because it was supposedly communicating via sign language? I think the animal would continue to behave in a normal fashion and the sign language would be an added behaviour - I don't think the new behaviour would cause all the other normal behaviours to cease. For example the animal might wave away a fly, which could be interpreted as a "sign" when all it amounts to is the animal made the motion for another reason which has to do with its normal behaviour. A trained person would know that that was not a sign at all, and could discount it.

This is all hypothetical to me because I have no experience with sign language. I don't really know if young children who are learning to sign are more difficult to understand than adults who are adept at signing, as in language use in young children, or if young children are immediately and universally understood by anyone who can sign. Do you know anyone who could answer that?

Young children learning to sign is quite a bit like young children learning to speak. Hearing children of deaf parents who use sign language will sign before speaking. Whether it's universally understood, I can't say. But sign language is not universal in itself. Sign language can differ in meaning from location to location. It contains colloquialisms the same as human speech.

I doubt this will add anything to add to this discussion, but I do know a few things about sign language (ASL) and the deaf culture.

 

[CFLarsen]

All right. But I don't understand why you assume an animal would cease exhibiting normal behaviour because it was supposedly communicating via sign language? I think the animal would continue to behave in a normal fashion and the sign language would be an added behaviour - I don't think the new behaviour would cause all the other normal behaviours to cease. For example the animal might wave away a fly, which could be interpreted as a "sign" when all it amounts to is the animal made the motion for another reason which has to do with its normal behaviour. A trained person would know that that was not a sign at all, and could discount it.

This is all hypothetical to me because I have no experience with sign language. I don't really know if young children who are learning to sign are more difficult to understand than adults who are adept at signing, as in language use in young children, or if young children are immediately and universally understood by anyone who can sign. Do you know anyone who could answer that?

Nope.

I'm not saying that the ape must cease to exhibit normal behaviour. What I'm saying is that the ape must be able to communicate with sign language to a degree that anyone unknown, but who understands sign language, is able to understand what the ape is communicating. And vice versa.

 

[Skeptic Ginger]

Look, it's perfectly simple.

I am asking if the apes could understand sign language - whatever type - from strangers.

If they can, then it is strong evidence that they do understand sign language. If they can't, then it is strong evidence that the handlers are over-interpreting - consciously or unconsciously, we don't know - what the apes are doing. And therefore, evidence that the apes do not understand sign language.

You said yes: They could understand strangers. That is a specific claim. So, I am asking for evidence of this.

Now, buried in all those words, I hear you say "I think I read it somewhere".

That's not evidence. That's your recollection. We can discuss that, of course, but it isn't evidence. You have to understand that.

Here's the crux of the matter: I am not commenting on the rest of the research, or how it was done. Because it doesn't matter.

Finally, a little headway. You are now mixing up two different things in this exchange, but at least you are talking about the basis of the issue. What I said about how the experimenters controlled for researcher cues was from general recollection and non-specific. That is what you have said over and over I claimed was a fact, and demanded I 'prove'. Serendipity, I did find the evidence below as to how researcher cues were completely removed. See the video.

I am solely interested in the evidence. The only thing that matters is, can the apes communicate with sign language? The test is to see if they can do it with strangers. That's called a control: Does it work, when you're not the one doing the test?

As long as we don't know that, it is futile to discuss what quality the rest of the research was done. All I want to see is the evidence. Then, we can look at how it was done, to see if the evidence is sound.

Now: Do you have any evidence that apes can understand sign language?

But here you are asking for research where the apes understood strangers. That I got from the summary reference in my previous post.

Finally, Woodruff and Premack are reported to have devised a cuing-free experiment in which chimpanzees indicated by gesture the presence of food in a container to human participants who did not know its location. They would correctly direct "friendly" humans who would then share the food with them, but would <mislead> "unfriendly" humans who would not share the food _ since the animals were then permitted to get the food for themselves.15

And here I may have made the wrong assumption about "unfriendly" vs "stranger". It turns out chimps are not communicative with strangers. But there are more ways than that test to show the validity of the research conclusions.


It is hard to find a lot of these studies as I don't know any data bases to search such as pubmed for medical research.

I did find this fantastic recorded event:

Watch this video from 1977 research with "Lana". If you can't play the video, I'll write out the transcript of the exchange, which was conducted strictly via a keyboard of lexicons with no visual connection between primate and researcher.

At first it looks like an animal trained to push buttons for a reward. Scoff as you will. But watch the next exchange which is shown! I rest my case, but you may rebut away.

I did gather a lot more and will post it as I get a bit more time. I'll also move it to the other thread and put a link here. There are many more issues to consider such as the limiting factors of age language may be acquired and the problems with interacting with strangers which chimpanzees don't do well with and imprisonment all of which affects the research.

 

[CFLarsen]

I have always asked if apes understood strangers: That is the litmus test, to see if someone who hasn't been involved with the previous results can replicate them independently. Blondlot discovered that the hard way.

I see several problems with the LANA project. First, this is very reminiscent of Facilitated Communication. It is obvious, when the clip of a retarded person is shown: The person sitting next to her is leading her hand to the correct button. Atsa no good, paisan.

I am very suspicious of how Lana constructed sentences: She did it basically in the same fashion as humans do. But that would mean that either she has been taught that a specific string of buttons results in a specific response, or - which would be truly amazing - apes understand the structure of language without having a spoken language. In other words, they understand the written word before they have evolved speech.

Again, a simple test would reveal if she understands the symbols: Were the buttons moved around from time to time, so she had to seek out the correct symbols, or did she merely repeat a pattern? Think about when you stand at an ATM: You have done it so many times, that your fingers move in a pattern, but you don't really watch if the numbers are in the correct place. (It's actually a good way of remembering passwords, as long as the keyboard stays the same!)

Finally, this is 30 years ago. I haven't heard of any replications. Given the fantastic nature of this experiment, I would have expected enough replications to have become pretty much accepted by science.

 

[Skeptic Ginger]

CFL, you are in denial.

There was no contact with the human except via the keyboard.

They showed you the exact symbols Lana chose.

You are saying that she merely had a memorized sequence here?

6-11-75 exchange after putting cabbage (disliked) into the container that opened when Lana requested Monky Chow (liked).

Lana: Question you put chow in machine

Researcher: Chow in machine

repeated 4 times

Lana: Question chow in machine

Researcher: yes

Lana: No chow in machine

Researcher: Question what in machine

no response

Researcher: Question what in machine

Lana: cabbage in machine

Researcher: yes cabbage in machine

Lana: question you move cabbage out [of] machine

Researcher: yes

(cabbage removed)

If you claim this only appeared to be communication and was essentially just some button pushing for a reward, you are in denial.

Just the use of the symbol for "question" is strong evidence of language beyond merely memorizing reward sequences.

They needed to move the lexigram symbols around why?

How could this possibly have just been some stimulus response for a reward?

So taking this and all the other volumes of research, I suggest you are just going to have to bite the bullet. The evidence is pretty darn incontrovertible.


As for the stranger communication, there was a study which had observers document and rate the signing. One deaf native signer was critical but all the other observers correctly identified the signs the chimps were using. I'll have to find it. Again, these things are cited in the papers discussing the issue but only available for money if you want to see the research paper.

 

[CFLarsen]

You are not paying any attention to what I write.

The point is, is Lana merely repeating a sequence of buttons in the anticipated positions? Or is she actually reading the symbols?

They need to move the symbols around to avoid Lana remembering the position of the symbols.

If you use an ATM without really looking at the keys, and they switch the keys around, you would not punch the right code.

Do you understand now?