Langauge

Discussion in 'Behavior and Intelligence' started by hlywkar, Aug 27, 2009.

  1. hlywkar

    hlywkar O. vulgaris Registered

    Joined:
    Jul 18, 2009
    Messages:
    81
    Likes Received:
    0
    any one ever try to teach their octopus any type of language?

    for example, sign language, or a picture based language?

    These creatures are supposed to be one of the smartest creatures on the planet. you'd think if a dog can learn to differentiate language (including hand signals) than an octopus could as well.

    that being said.... what about a writing implement?

    octopuses are extremely dexterous, with practice they could probably learn to manipulate either a stick in sand... or pull switches/push buttons.

    if a chicken can learn to play tic tac toe, what about an octopus?

    not sure how safe it would be to all electronics to an aqaurium... I am sure there could be a way.

    just curious if anyone, or any studies have been done in this manner?

    My degrees in psych. so this is of particular interest to me.
     
  2. monty

    monty Colossal Squid Staff Member Supporter

    Joined:
    Mar 8, 2004
    Messages:
    4,887
    Likes Received:
    11
    My impression is that octos are somewhat unfortunate subjects for this sort of thing for the same reason cats are: they may be able to do tasks, but they're largely unmotivated to be trained to do them on request. The closest ones I'm aware of are discrimination tasks regarding a food reward. These were described in Wells' book Octopus, but I don't remember if Wells was the experimenter. They could discriminate based on tactile or visual cues in those experiments. I don't know of any manipulation tasks more complicated than picking up or attacking objects, but I don't remember how the tactile experiments were carried out, just that the objects were cylinders with various weights and textures. I suspect Lene may have more up-to-date knowledge on this topic. You might want to look up papers by Jennifer Mather as well.
     
  3. DWhatley

    DWhatley Cthulhu Staff Member Moderator

    Joined:
    Sep 4, 2006
    Messages:
    19,076
    Likes Received:
    1,123
    Location:
    Gainesville, GA USA

    I think you may be over impressed with an octopuses stated intelligence. Smartest mollusk, yes. One of the smartest animials on the planet, no.
     
  4. ckeiser

    ckeiser GPO Supporter

    Joined:
    Oct 16, 2008
    Messages:
    120
    Likes Received:
    1
    Communication, being the transmission of information from one organism (the sender) to another (the receiver), seems like it would only be achieved from a ceph towards a human for reward. Teaching an octopod or cuttlefish sign language would be more operant conditioning than interspecific communication.

    What kind of information would a coleoid have to transmit towards a human?

    Great inquiry though, hlywkar.
    Cheers.
     
  5. ceph

    ceph Wonderpus Staff Member Moderator

    Joined:
    Dec 20, 2002
    Messages:
    198
    Likes Received:
    18
    Location:
    West Palm Beach, Florida
    "or pull switches/push buttons." - I suspect this would be the easiest. Most octopuses learn very quickly when food is involved and tasks are broken down into steps.

    What is the difference between an animal learning commands (Sit, stay, fetch me a beer) and a language?

    For communication studies, I have focused on cephs that live in groups such as cuttlefish or squid. That work so far has been communication between other members of the same species and to predators.
     
  6. ckeiser

    ckeiser GPO Supporter

    Joined:
    Oct 16, 2008
    Messages:
    120
    Likes Received:
    1
    I think the difference lies in the receiver's reaction to a command.
    A dog may understand that "fetch me a beer" means "If I go to the kitchen and get my human a beer, he or she will reward me", which is definitely communication. However, to make it a shared language, the dog would have to be able to transmit that same signal back to the human, or to other dogs, and understand the communicative factors that created the signal.
     
  7. ceph

    ceph Wonderpus Staff Member Moderator

    Joined:
    Dec 20, 2002
    Messages:
    198
    Likes Received:
    18
    Location:
    West Palm Beach, Florida
    Good answer.

    What if you trained a horse to do math and communicate the correct results? Then that horse must understand math and communication right?
    http://www.skepdic.com/cleverhans.html

    What about observational learning in octopuses? Is there another explanation?

    The next question is a bit harder: How do you test the dogs understanding of the communicative factors that created the signal and separate that out from "If I do XXX behavior, I will receive positive feedback"?

    This is where designing an experiment which has results that can clearly distinguish between these two alternatives becomes a challenge.


    On the much easier and more fun per effort side, I'd like to try to train an octopus and really see just how much one can get them to learn in their short lifespan. TONMO members with octopuses can try this too.
     
  8. DWhatley

    DWhatley Cthulhu Staff Member Moderator

    Joined:
    Sep 4, 2006
    Messages:
    19,076
    Likes Received:
    1,123
    Location:
    Gainesville, GA USA
    Ceph,
    There are often requests for ideas. If you came up with something simple and more importantly, simple to observe and retest in a consistent way that might make an interesting thread.
     
  9. ceph

    ceph Wonderpus Staff Member Moderator

    Joined:
    Dec 20, 2002
    Messages:
    198
    Likes Received:
    18
    Location:
    West Palm Beach, Florida
    There are lots of thing hobbiests can do to with learning and training - opening jars is the classic. Once your octopus is trained to open things you could time pet octopuses speed with new puzzel boxes/containers that it hasn't encountered before. If you want to know exactly when your octopus has opened a small container but can't see it underneath its web add a drop of food coloring in with the food. Idealy these experiments are done behind a blind since the octopus will very likely watch you in case you give it more puzzel boxes with more food. . . If your just doing this for fun, ditch the blind. You could train your octopus to pull a lever to get fed. . . You could train it to pull the white level and not the black one, or vice versa. Mazes are an option if you like working with plexiglass. One could have a lot of fun with this.

    As for how you test an animals understanding of the communicative factors that created the signal and separate that out from "If I do XXX behavior, I will receive positive feedback"? I don't have any good answers and therefor don't do that kind of behaviour research. Often these sorts of interesting but grey areas boil down to semantic arguments about interpretations of results. Since we can never know what an animal is thinking, it is difficult for scientists to clearly show that a behavior is clearly one of these two explanations and not the other one. For those who do want to take this farther, one place to start looking for ideas to test is in primate research. . .
     
  10. ckeiser

    ckeiser GPO Supporter

    Joined:
    Oct 16, 2008
    Messages:
    120
    Likes Received:
    1
    Yes, applying the research methods and results from communication studies across taxa, I think, will be a great advantage in understanding animal communication as an abstract phenomenon.

    Another difficulty lies in the field of human observer limitation, as the media of communication in non-human animals often lies outside of our own observable and manipulative methods (technologies obviously aid in this). Progress in the study of multi-modal signaling I think gives us a much better understanding of whats going on in animal communication channels. However, ceph, the point you bring up about arguing semantics over results is a definite difficulty. I'll continue working on it :read:
     
  11. hlywkar

    hlywkar O. vulgaris Registered

    Joined:
    Jul 18, 2009
    Messages:
    81
    Likes Received:
    0
    look to humans

    well let's look at the brain.

    we know that there are areas. these areas tend to be used for functions... I say tend to, because sometimes they are not used.

    example. say brain has areas a + b.
    a is for action x
    B is for action y.

    cut out a.

    at times B is able to learn the new function to later be developed to perform X and Y.

    why is that important...

    that suggests that designated brain area may not be necessary to be able to perform a function.

    we are constantly comparing tests from rats and other animals then using these to equate back to us.

    why?

    Because maybe if we find it happening in animals, it can happen in humans too.

    now... turn it around... if it can happen in humans, can it happen in animals?

    maybe... who knows.

    some issues come up with are they speaking, or is it just reinforcement?

    well, one could argue that just because it's learned through reinforcement, does not mean it's not language.

    I am a teacher... sometimes the only way kids learn is through reinforcement.

    does that mean they are not speaking english?

    I love my wife... but I only know a little korean... largely due to the fact that I do not have enough reinforcement. I KNOW there is a long term reinforcement available. But I still don't study. I need short term. And I am an adult, human, with a university education.

    I can say hello. goodbye, thankyou, this, that, how much, like, hate, love, left, right , straight, give me, foreigner, teacher, park.

    maybe a few others... but that's about it.

    am I speaking korean when I use this? technically.. yes.
    am I speaking a language? by some of the standards we judge animal communication on... no, I am not.

    some people demand that abstract thought and others be included... here's a list from wikipedia

    those are argued... and some say outdated.

    also... most kids below the age of 5, or maybe higher could not pass this test. (I'm guessing here)

    Why is that important?

    well maybe our tests are too strong.

    maybe we are too afraid. what would happen if someone created a way to test all animals... teach them, and talk with them... do they get more rights? or the same?

    some say language can be a measure of intelligence... but is it.
    I can't speak korean. if english is my native tongue, and I go meet some strange koreans in a far off land... does that mean I am not smart? no... because I am. the test is wrong.

    what about that girl that was locked in a basement and not talked to?

    she came out and had no language what so ever. her intelligence was stunted by it.

    what does that suggest?
    it suggests that language helps build intelligence.

    so - language helps build intelligence.
    minimal language may not be a measure of existent intelligence.
    minimal language use, to later gain more language can further increase intelligence.

    immersion is an issue. so is reinforcements.

    I live in korea. but I have other sources. I also don't have reinforcements.

    If I had no Need for language I would not learn it.

    how can we solve this and use it with animals?

    use a language with animals constantly. not just in lessons, but with game time as well. not just when we are with them, but when others are with them... not just when people are with them... but also when they are alone.

    I have tv. I have books. I have games. I have friends. I have parents. I have teachers.

    create a form of language that can be recreated by an animal. if it barks. use barks... or use something that can be manipulated by the animal.

    what about tv?... translate it. I can watch korean tv with subtitles... create tv with subtitles.

    what sort of subtitles?... well what language are they learning to speak? sign? buttons? barks?

    use the same...

    sure reward is necessary... but so is immersion.

    every little detail should be worked out before hand... if you want to have an octopus speak via buttons... 1st create a language that I can speak to you via buttons.

    obviously a keyboard will not work. but buttons for phrases or concepts may or even words.
    or even a switch. think of Stephen Hawkins. He has only one switch to use and a computer screen.
     

Share This Page