Telepathy is not proven…not remotely.
Communication is (as this thread proves).
I provided a shred of evidence. More than a shred in fact. Evidence that was done by scientists and peer reviewed that suggested dogs understand human language to some extent.
What hasn’t happen is you providing any proof beyond your opinion.
Of the two of us it is you lacking evidence for his argument.
Consider herding dogs.
Words are not used but words are not the only form of communication (again, ask a deaf human about this).
The herding dog when “told” to do “X” has a concept of what “X” means. The herd it is moving is not the same from day-to-day or even moment to moment. The dog understands the larger concept of what it is meant to do. Given command “X” the dog achieves the desired result and not merely “run there and sit”.
Admittedly there are limits to what the dog understands of what the human wants it to do but watching them in action there is little doubt they grasp a bigger picture of what is intended than merely “run right” or “run left”.
A more skeptical alternative is that Rico’s abilities have nothing to do with human word learning. For a child, words are symbols that refer to categories and individuals in the external world. Even oneyear- olds appreciate the referential nature of words. When children learn a word such as “sock,” they do not interpret it as “bring the- sock” or “go-to-the-sock,” and they do not merely associate it with socks. They appreciate that the word refers to a category, and thereby can be used to request a sock, or point out a sock, or comment on the absence of one.
Does Rico understand reference? It is not clear. In the experiments, Rico’s abilities are limited to specific routines. All new items are learned in the course of fetching, and Rico’s understand-ing of these items is tested in this context as well. Also, it is always Rico’s owner who is communicating with him. These experiments are carefully designed, and so there is no worry about problems akin to those of Clever Hans (a horse that seemed to have mastered arithmetic but was actually responding to subtle cues by its owner). Yet, if Rico really is learning sound-meaning relations, as Kaminski et al. maintain, it should not matter who the speaker is.
Further experiments can help to resolve this issue. Can Rico learn a new word by being shown an object and hearing a person name it? Can he learn a word for something other than a small fetchable object? Can he display knowledge of a word in some way other than fetching? (Kaminski et al. note that there is anecdotal evidence that he can—this issue is worth pursing experimentally.) Can Rico follow an instruction not to fetch an item, just as one can tell a child not to touch something? Rico’s abilities are fascinating, but until we have answers to these sorts of questions, it is too early to give up on the view that babies learn words and dogs do not.
Bloom, 2004 “Can a Dog Learn a Word?” Science 304
knowledge of words learned by exclusion. These findings are less convincing than they appear at first sight. Although the follow-up comprehension results look impressive, they could be attributable to the reinforcement Rico received after initially retrieving the novel object (J. Fischer, pers. commun.). Would long-term learning result from reasoning via exclusion, with no positive feedback given to Rico? The presence of feedback does not, of course, explain how Rico arrived at the correct mapping in the first place. On this point, however, another potential problem with Kaminski et al.’s exclusion task is that, unlike studies with children, this study did not include a control for baseline novelty preference. If Rico tends to pay more attention to novel objects than familiar ones, this tendency could be solving the identification problem for him. Comparing performance on this task with a no-label control where Rico is asked simply to ‘fetch’ from a set that contains a novel object among familiar ones could rule out a novelty bias. In summary, Rico could have initially retrieved the novel object because it was more salient or interesting but he was then reinforced for his selection and so retained the object–word mapping in the follow-up test. At present, the question of whether Rico achieved word learning by exclusion remains open
Markman and Abelev 2004 “Word learning in dogs?” TRENDS in Cognitive Sciences
All these criticisms were published within months of the Rico story being published. Yet as far as I can tell no further results from Rico have been published. Since it would be so easy to do the experiments suggested to establish whether Rico is using language, that fact that nobody has published for 7 years suggests that the follow up experiments invalidated the original claims.
At the very least, the results from Rico have been seriously called into question, and nobody has bothered to address those questions, so the evidence is of little value. More’s the pity.
Telepathy is communication.
No, you did not, the researchers themselves specifically stated that it did not establish that dogs understand human language.
As I said, science does not deal in proofs. Learn it.
Yes the multiple references i have provided are not in fact there. It is all a conspiracy.
Herd dogs? Seriously? Read post number 73. I have no intention of repeating the exact same post with the example of herd dogs rather than he last example you used.
Do you have any evidence whatsoever that telepathy exists?
If it does not exist it is not communication (it is not anything).
Define “understand”.
Can a dog discuss Plato with you? Of course not.
What is your definition of understanding because the way I read it the researchers felt the dog was grasping language beyond a mere conditioned response.
And you still have no answer for how a social creature manages its social society without communication.
Social constructs require more than pre-programmed responses to work.
Dogs are eminently social creatures. Wild wolves are. Fido is.
Do you have any evidence that dogs understanding human language exists?
If it does not exist it is not communication (it is not anything).
I have done so at least three times in this thread.
You quote where they say that and we can discuss it.
And you still have no answer for how dogs manage to recite Shakespeare, yet nobody ever hears them.
No, they do not. There is a reason why bees are referred to as social insects.
I think the dog’s “understanding” and the human’s understanding stem from similar places. However, I don’t think they fall apart at some fundamental as vaguely definable as abstract thought. I think it falls apart at something more similar to “tokenization.” For an animal, an entire phrase, complete with inflection is tokenized to be associated with an event, along with everything that event entails. If you say things with similar inflections, due to lack of an ability to tokenize, these all, independently, will be related to these circumstances. With a human, however, our brain finds patterns, since the data is stored relationally we end up taking what amount to probabilistic averages of the phrase given.
Over time, the phrases become pattern matched, separated, and stored until they give the structure of grammar and words. The more times you say a certain phrase, the more we look for patterns in the situation, cross check it with other phrases we know, and then mentally note the similarities and differences, both between the situations and the overlap point between the phrases (all, of course, in the matter of a couple seconds). Do this enough times, and we have enough relational data to learn that “toilet” means that thing in the bathroom. And of course, since it’s probabilistic based on your percepts, this allows for things like homophones.
The problem with animals doesn’t come in that they can’t think, so much as the fact that their pattern recognition qualities aren’t quite as advanced. It’s debatable exactly where this ends: inflection? Syllables? Time taken to utter the phrase? Since you don’t have to be dead precise and they respond to different voices, there’s clearly at least some relational matching going on, but it’s clearly not as advanced as a human’s.
As such, I think both a human’s knowledge and an animal’s knowledge are both, fundamentally, operant conditioning, but the human’s ability to parse and store the information in increasingly more discrete chunks causes a little bit of a snarl that manifests as “understanding.”
If you think that then you are wrong. It is trivially easy to falsify any claim that human language comprehension hinges upon operant conditioning. I have given several examples in this thread already.
Can you define exactly what you mean by the conditioning them? I’m probably just stupid and don’t understand it correctly. The way I learned operant conditioning, it was necessary for any semi-thinking entity to even function, and that everything in the universe, from breathing to doing math was fundamentally driven via it.
I cited dogs understanding in Post #59
I’m dumb and missed it. Give me the short version.
Post #59. To wit: “Fischer and colleagues set up experiments to test the dog, and are satisfied that he understands the words.”
What?
Bees communicate with each other too inrather complex ways at that.
No, you didn’t.
Language is more than sounds.
Not a quote and not made by a researcher.
What?
That is my point.
Only if you consider breathing to be volountary.
I think the thing I don’t grasp is: does a reward/punishment stimulus have to be external? Does it have any meaning for a brain to condition itself? If so, that’s what I was referring to with the operant conditioning. The brain ITSELF gets a kick out of pattern recognition, and rewards itself on a job well done innately, which reinforces the learning behaviour.
Oh Og, that is whole other can of worms that we really should not open in this thread.
The short answer is that in theory the brain can condition itself.
The complication is that once you allow that you have a theory that predicts everything, and hence nothing. If the-brain-in-the-rat-in-the-box can condition itself then the box itself becomes a worthless observational tool. Every possible response is predicted because the system is now self-referential. If you feed the rat and it avoids repeating the action, you have to conclude that the rat’s brain conditioned itself aginst the stimulus or against the reward, rather than being able to conclude that the hypothesis itself has been falsified.
What has been cited is the common experience that variations in inflection are ignored by many dogs. I know my dogs seem to have some words that are clearly understood by inflection, and even some that are only associated with the speaker. But there are certain words that they respond to no matter what the context.
I have a feeling you are going to tell me this is insufficient for language, and that is what’s frustrating. You gave us something we could directly talk about, we show you a contradiction, and then you change the parameters. Even if you (as I assume) meant to be saying the same thing throughout the entire thread, I do not believe you have been effectively communicating that.
And, BTW, I think the Chinese box is very relevant to this thread, especially if your interpretation is that anything that appears to be language and reasoning really is (a common philosophy in AI). That basically undermines the entire distinction of language and communication.
As for my opinion (as if that matters in GQ), dogs are really good at pretending they understand. Seeing as “fake it till you make it” works in so many other learning situations, it wouldn’t surprise me if some dogs have a form of sentience that we haven’t been able to test for, yet.
No, what has been cited is anecdote. We actually have *more *anecdotes saying precisely the opposite, and they are also worthless.
That is certainly your anecdote.
No, you have given nothing that we can talk about. You give anecdotes, I acknowlegde that as being your anecdote. There is nothing there for us to talk about.
No, I do not.
Since I say exactly the same thing in every post, I can only conclude that that you are reading effectively at this level.
Since I have never said anything remotely like that this, or even anything that could be twisted into this meaning, your comment requires no further discussion.
What you wrote is very largely true, but too absolute. When dogs are conditioned to a stimuli, they clearly CAN respond to that stimuli in a variety of contexts. Nowhere near to the degree that humans can, true, but still – the context of the stimuli does not have to be exactly the same every time.
For instance, if the dog is conditioned to find the socks in pile and return them to his master, the stuffed-gorilla test posited above will be no problem for the dog. The socks have not only visual cues, but oflactory and tactile ones. The gorilla-with-socks is nothing more than a little “pile of stuff” – it doesn’t really matter that the socks are placed on the gorilla’s foot as opposed to laying loose on top of the gorialla. Either way, the dog will have no problem identifying those socks as the ones he’s been fetching in prior training. And if you change the socks out … the dog will still be able to identify “socks” by their attributes.
Passing the stuffed-gorilla test does not mean that the dog fully comprehends the word “socks”, of course. The test as you proposed it, however, can show that dogs can respond appropriately to more contextually-diverse stimuli than you seemed to be allowing in prior posts.
A better test of word comprehension in canines is one modelled after comprehension tests given to small children. Use a photograph or drawing of the socks, and see if the dog can identify that in response to the command “socks”. The dog will not be able to do so. Once such transference (aka displacement) is introduced in the stimuli, the dog will not respond to it. This is one of the principal ways in which human speech differs from animal communication.
Examining some common canine behaviors can lend greater understanding to the inability of dogs to deal with another type of displacement – that of time. A dog may grab his leash and whine at the door to make an immediate request: “go for walk now”. A dog will never grab the leash, walk up to you, and then attempt to communicate “don’t want to walk now, want to walk later”. Similarly, even if your dog can communicate to you that (a) it wants to go for a walk now, and (b) it wants to be fed now, it will never communicate the two in sequence within the same “utterance”. A dog will never communicate that it “want to eat now, then walk later”. The temporal window in a dog’s mind is very short. Their storage/recall of stimuli/responses is pretty good among the main run of mammals, but nowhere near enough to lend dogs the kind of displacement in communication that humans deal with routinely.
Blake, this would all be a lot more interesting if anyone here had ever claimed that dogs can understand language like humans can. Saying ‘they can recognize their human names’ does not mean ‘they identify it as a proper noun and can speak English pretty damn well.’ It means what you were saying: ‘they can recognise their name via learning to respond to that sound.’
Where some differ is that they claim the tone of voice is needed too; someone above made a good point about human tonal languages. Dogs need context too - so do humans. I don’t respond to ‘sam’ if I hear someone say ‘I was dancing the samba.’
But if you think it has to be said in the exact same tone of voice and same context every time then you haven’t spent much time with dogs, or had to learn to never use the word ‘walkies.’ Claiming that dogs can learn to recognise the sequence of syllables in their name - which are said repeatedly and for great reward early on in the dog’s life - is just saying that they can learn to respond to those sounds, not claiming they learn actual language.
So, anyway, which other animals can learn to recognise the sound that makes up what humans would recognise as a name?
Hmm, thread died. Anyway, I went off to check some of the primary literature to, you know, see what the actual latest theoretical underpinnings are in the field of animal behavior and cognition and language use. In the Journal of Comparative Psychology, I found a paper from Couchman, Coutinho, & Beran 2010 titled “Beyond Stimulus Cues and Reinforcement Signals: A New Approach to Animal Metacognition”, which, while it did not directly address language usage, did investigate where conditioning leaves off and cognition picks up.
I found the following quotes quite relevant to the current thread, and rather telling:
“Comparative researchers naturally proceed cautiously in attributing
metacognitive capacities to animals. Indeed, comparative
psychology’s tradition of parsimony, as exemplified by Morgan’s
Canon (1906, p. 53), demands a search for an explanation of the
animal data patterns that relies on the lowest-level psychological
capabilities possible. This is why the appropriate psychological
interpretation of animals’ uncertainty responses has been a source
of ongoing theoretical discussion.”
“This caution highlights a central theme of our research, which is
that the field should avoid an all-or-none approach toward animal
metacognition. For that approach could distance researchers from
the theoretically fertile middle ground wherein one grants animals
a fairly sophisticated uncertainty-monitoring capacity without
overinterpreting what they do. We believe that in this middle
ground lies the phylogenetic emergence of human metacognition,
and probably also the ontogenetic emergence of metacognition in
human development. Therefore, we believe that it is critical that
the animal-metacognition literature not take an all-or-none approach
toward a focal construct that itself is not all-or-none.”
Note the first quote indicates that there is still quite a bit of uncertainty involved and that the subject most certainly is not closed and settled. The 2nd quote is notable in that it strongly suggests that simplistic all-or-none approach (which, in the case of this thread takes the form of “either animals grasp the essence of the words communicated to them” vs. “they are just conditioned automatons with absolutely no deeper comprehension”) may not be optimal and may end up leaving a lot out of the explanation derived.
You see, I tend to react strongly to absolutist language (where words and phrases like “none”, “never”, “100% of all researchers”, “not even a shred” and “none of it indicates in any way”); it usually is a sign that a dogmatist (heh) is on the loose. [I’m also suspicious of absolutist reductionistic thinking too, FWIW]
So care to relax your 100% claims on what the experts currently think, Blake?
Good grief!
Does this mean my weekend experiment is cancelled?