Training A Dog To “Speak” With A Sound Board

The field of Augmentative and Alternative Communciation (AAC) covers communication methods used by those who are unable to otherwise produce or comprehend spoken or written language. Many will be familiar with the speech synthesizer used by Stephen Hawking as just one such example of AAC technology. [Christina Hunger] is a speech language pathologist, and is intimately familiar with such tools. She decided to use these techniques to teach her dog, Stella, to talk.

[Christina] began her project by implementing a button board which triggers various speech samples when triggered. There are plenty of typical words that a dog may wish to use, like beach, park, and ball – as well as words describing concepts, such as where, later, and come. Over time, she has observed Stella using the button board in various ways, that she claims indicate a deeper understanding and use of language than would normally be ascribed to a dog.

From the outset, [Christina] has been intentional in her methods, being sure to only demonstrate the use of the board to Stella, rather than simply pressing the buttons for her. The experiment has many similarities to the case of Koko the gorilla, known for learning symbols from American Sign Language. The project is also documented on Instagram, where she films Stella using the device and gives interpretations of the meaning of Stella’s button pressing.

Attemping to communicate on a higher level with animals has long been a mysterious and complex pursuit; one which we’re sure to see more of as various technologies continue to improve. We’d love to see a broader scientific study on the use of AAC tools to “talk” to animals. In such matters, context and interpretation play a large role, and thus it’s difficult to truly gauge the quality of understanding an animal may actually have. More research would be great to shed light on these techniques. Video after the break.

[Thanks to SoggyWaffles1984 for the tip!]

37 thoughts on “Training A Dog To “Speak” With A Sound Board

    1. Color me skeptical. How many times has the dog also said “play doorknob toaster” in a similar situation? It’s the whole extraordinary claims needing extraordinary proof situation.

        1. Yeah. Dogs learn tricks to please owners, and they learn fast. The dog learns your triggers, so it can read you like a book from the slightest expression.

          You think you’re not influencing the dog, but your expressions tell the dog “warmer” and “colder” all the time. You think the dog wants to go to the beach so the dog takes the cue and goes to press the “beach” button, but oh, it’s broken, so it peeks back at you and tries to guess which buttons you want it to press by how excited you get.

          https://en.wikipedia.org/wiki/Clever_Hans
          > “the horse was responding directly to involuntary cues in the body language of the human trainer, who had the faculties to solve each problem. The trainer was entirely unaware that he was providing such cues.”

          1. Most importantly:

            “After Pfungst had become adept at giving Hans performances himself, and was fully aware of the subtle cues which made them possible, he discovered that he would produce these cues involuntarily regardless of whether he wished to exhibit or suppress them.”

        2. Environment is also another key word. There was the one guy who wrote a whole long scientific paper on how to make rats get lost in mazes. Turns out they’re really good at leaving clues for themselves and picking up things that the researchers don’t notice, like the arrangement of ceiling lights above the maze.

  1. I taught my husky x dingos a fairly complex sign “language” for hunting

    it worked really well, getting them to encircle a herd of goats, then drive them into a kill box was something we did all the time

    for those that think dogs don’t “understand” english, my mate has a Japanese wife and a few Japanese friends

    when people were speaking Japanese, you could see it confused the hell out of the dawgs

    so I got them to teach the dawgs some Japanese

    in the end mah dawgs understood english, german, spanish, japanese, a bit of greek and italian

    which is way better than I ever managed

    1. Shepherds use Border Collies to assist them. The dogs are often trained in different languages (English, German, French) so that the shepherd can control one dog at a time just by using the commands in the right language

  2. Dogs, as well as other animals, have their own language, that is not restricted to words. There is also scope, semantics, and it is tailored to neural processes that are specific of their brains. In other words animals speak at a different domain than humans and there is no 1:1 relationship between such domains. Nevertheless we can understand some concepts of the animal language and I am sure that they can sometimes understand part of ours, and even build some tool to help on this process, like the board from this article.

    BTW I hope that the aliens bring a human version of such boardvin their flying saucer on the day of our first contact.

    1. As I’ve understood the research into language in other animals it seems that their languages are each fixed, without variation in the same species. A chimpanzee that has never met another chimpanzee will understand and be understood by other chimpanzees. So it’s really interesting how well we can teach animals to understand and even communicate in other languages.

      1. I think humans have a base level language also. Toddlers “speak” it. It’s often described as a body language, but some gestures are more abstract than the obvious “I’m sitting in something wet” discomfort expression. There are a handful of actual sign language signs that match up with it, drinking and eating for example.

    1. I suppose it all depends on what the buttons are labeled then. I might think this could be a tool for finding the limitation of animal comprehension. If animals use instinct more so than human beings, this is like talking in emoji. Lets have a cow or a pig press the buttons. Although very interesting, there may be some unforseen ramifications. Worked in a college basement doing some ventilation work for their labs, where the monkey’s were being experimented on in the name of science, maybe give them some buttons. Just not sure about this.

  3. I recall reading from somewhere that when they had these experiments with chimps, they never initiated any conversation with people or each other for that matter – they used the symbol boards only when they were asked something by the handlers or they wanted something from the handlers. It’s still a bit controversial whether what they actually learned was just responding to cues or performing simple conditioned tricks, like “point at symbol, receive banana”, and the rest is just selective observation.

    Dogs are excellent at guessing what humans want and reading subtle body cues as well. It’s too easy to get yourself into a Clever Hans situation, which is why the real test is whether the dogs would try to communicate with anyone else – whether they’ve learned that their owners respond to certain actions in a certain way and don’t have a clue what the words mean.

    1. “whether they’ve learned that their owners respond to certain actions in a certain way and don’t have a clue what the words mean.”

      That raises the question of meaning in an interesting way.
      If the dog learns “go” “outside” makes the owner open the door and let them out… did they learn that “go outside” means they get to go outside?

      I guess where I’m trying to go with this is that if you learn a given symbol always produces the same effect, have you learned the meaning of the symbol?

      1. My dog lets me know he wants to “go outside” by ringing the bell we attached to the front door.
        (Of course, as we were house training him, we’d ring the bell whenever we took him out to do his business.)

        B^)

      2. My friend had a hunting dog trained to fetch birds, and he would demonstrate to me that his dog knows three different languages by giving the command in different words but the same voice and posture.

        The dog never listened to the words at all.

      3. > did they learn that “go outside” means they get to go outside?

        They learn that some specific actions means they get to go outside, but whether they associate that with the words, the tone of the words, the circumstances of the words, the action of pressing certain clickers…

        The most plausible explanation is probably the simplest: the dog learns an association between an action and an event, rather than figuring out the meaning of words. If the “outside” button is always at a certain corner of the mat, or always a particular color, shape, smell, etc. they are most likely to simply associate the button with the event: click on a particular button and hope that the desired event happens. If it doesn’t, try again or give up. It doesn’t matter whether the button makes a noise, a light show, or shoots up fireworks.

        Put yourself in the position of the dog: You press a button and the owner responds in a positive way. You’re a dog: your needs are simple. You want to go outside, play, eat or drink. Pressing a random sequence of buttons ending with any of these options will have a great chance of doing what you want anyways, and if it’s the time of the day you normally take walkies anyways then it’s just a ritual to press the “outside” button before you get to go.

        1. That is still communication. My dog has a squeak toy that initiates communication. If it wants outside it takes the toy to the door and squeaks it. If its water bowl is dry it squeaks it at the water bowl, etc…

          If this dog wants to go outside it hits (Some button) and the human understands that it wants to go outside.

          1. Though on the other hand, “communication” pre-supposes that the dog understands there is another mind receiving the message. The dog may understand its owner more like a force of nature. It doesn’t necessarily have a theory of other minds.

            Suppose there’s a button on the wall and when you press it, the door opens. Is this an automated door, or does someone have to open the door? Who cares – the door opens anyways. In one case, you’re just using a mechanism, in the other you are communicating your wish to open the door to another person. As long as you don’t know or understand which case it is, it’s meaningless to say that you are communicating.

          2. Based on personal experience, it is clear that dogs DO have a concept of other minds. If I drive an RC car around and bug my border collie with it, he doesn’t attack the toy. No, he ignores the toy, looks right at me and barks in a disaproving tone. I also remember another dog my family had when I was a boy. He wasn’t supposed to run off and if my family was coming home in the car he would notice our arrival from a distance run as fast as he could to get to the garage and then come out slowly, stretch, and yawn so as to give us the impression that he had been resting peacefully in the garage while we were away, the whole performance for the purpose of putting a false idea in someone else’s mind. No, people who deny that dogs have a sense of other minds are simply wrong. The daily struggles of will between me and my also stubborn border collie reinforces that perspective. Especially with these smarter, strong willed dogs, you definitely don’t own them, you have to establish a working relationship and friendship but more or less as equals, soul to soul. (Shadowfax?) I’m sure elephant caretakers get the same impression….horses, donkeys, domesticated (or wild) foxes, etc….any of the higher functioning smarter animals, especilly those that are wired to be social. These creatures clearly are people with the ability to establish relationships with other people, not the kind of things we should consider killing, eating, or making sellable products out of.

          3. >”it is clear that dogs DO have a concept of other minds.”

            It’s more complicated than that. For example, autistic people have theory of other minds, but theirs is often incomplete or immature to the point that they can’t understand and predict other people. This shows up as behavioral quirks, like assuming that other people understand you want the door opened without you telling them.

            It’s not an on-off deal. Almost all animals have some limited understanding of other minds because they need it to predict their actions. What parts of other minds they understand depends on what actions they are trying to predict. For example, a dog may understand the act of pointing at an object, but not understand communication as an exchange of thoughts – so they may not understand that by performing an action they are telling you to do stuff. They just observe that certain actions result in certain outcomes as if by magic and just accept that as a rule of the universe.

          4. > the whole performance for the purpose of putting a false idea in someone else’s mind

            That’s too easy to anthropomorphize. The dog doesn’t need to understand the trick to perform it. A simple observation is enough: “If I yawn, the owners will act happy. If I don’t they will scold me for some reason I don’t understand.”

            Distinguishing between genuine understanding and a simple conditioned response is extremely difficult since we’re biased towards believing the former instead of the latter. You’d have to ask the dog.

        2. Can you show that what you’ve learned is any different? You only know you comprehend because you can see inside your head. But humans learn the exact same way, conditioned responses.

          Look at toddlers learning. There’s no comprehension. Or people who have been abused. They might know what the word love means, but that perceived abstraction is usually secondary to the way that they were conditioned to use it.

          Is this a qualitative difference or a quantitative difference?

          1. I appreciate your comparison as quantitative vs qualitative.

            On the one hand I think the level of self awareness in the conscious mind (I’m not only aware of myself and aware of others, I am also aware that I am aware and aware that awareness is a concept in my mind) in humans is so far removed from that in animals that it constitutes a qualitative difference.

            On the other hand, a large majority of human behavior (how we act, think, and feel) is driven by the subconscious. What portion is debatable but could be up to 95% according to some. That subconscious mind is a lot easier to think of as a quantitative difference from animals, even if there is a large difference in complexity.

            On the comprehension/communication point, I have always thought that the association is most often direct rather than indirect. For example, when my dog hears the command to get off the couch, I don’t think he really understands the command. But subconsciously it reminds him of times when he responded incorrectly and had negative reinforcement—so he is inclined not to react that way—and of times when he responded correctly and had positive reinforcement—so he gets off the couch. It’s almost as automatic as eating because you’re hungry. He’s getting off the couch because he has a vague feeling of not wanting to be on the couch anymore—but only I understand what caused him to feel that way.

            And after saying all of that, I think the most surprising thing is that dogs are so loving and lovable. I honestly think that dogs are proof that god loves people.

          2. I think it’s a qualitive difference. I believe there are two types of processes going on in anyone’s mind – automatic conditioned, and deliberative understanding. This is identified as system I and II thinking in people.

            Going from simpler animals towards more complex animals, the amount and scope of the two types of thinking shifts. Not all animals have the same cognitive powers – they haven’t evolved to spend as much energy on running a brain as we have – so they’re taking more shortcuts and relying more on heuristics than understanding. There is inevitably a cross-over point where there’s not enough system II thinking to support for example an ego or the notion of a self, so you cannot meaningfully talk of things like language. The difficulty is the gray area between, because animals do not necessarily distribute system I and II thinking the same way we do. They may be great at reading body language, but rubbish at understanding even the simplest concepts of spoke language.

            It is like the situation where a color-blind person tries to figure out between a green and a red traffic light in the dark when they can’t see which one of the bulbs is on top and bottom – it’s impossible because they simply lack the faculties for it. If other cues are present, it may look like they understand color, but that’s merely an illusion. Likewise, the animal relies on other cues and the owners are tricked into thinking that their dog is thinking on the same terms as they do.

    2. There was some suggestion that apes trained in sign language or other communication methods do not seem to be aware that any other ape or human could know anything more than they know. Therefore simple conversational things like “Was it raining when you were outside?” don’t come into play because if subject doesn’t know if it was raining or not, how could any other.

      There may have been disbelief in the early human mind of the concept of originality, original thought, ideas. New ideas, ways to do things were attributed to gods. That’s as recent as 10,000 years ago, after the basic human genotype was here for a few hundred thou before that. Anyway, if you have no notion that you yourself are capable of original thought, why should you expect it out of others, hence why talk, they’re all thinking what you are right?

      However in humans, language has developed beyond the necessities as maybe a replacement for social grooming. In animals where social grooming and close contact body languages exist, there’s no need for vocal “idle chatter” it’s happening another way.

  4. I like the idea a lot but I am still very skeptical of this story (seen it before) and would like to see the idea studied more rigorously. I have a mil. spec. dog so she is very smart but I am not sure she would communicate in that manner, she has very different ways of letting you know what she is thinking or what she wants. It is more a case of learning to understand her rather than teach her to say it the way that suits you.

  5. My experience is it depends how you raise the dog, just like if you ignored a child in it’s formative years it would grow up struggling to communicate, many (often loving) dog owners “ignore” their dogs as puppies because they assume they can’t learn the same way.

    You look at all the endless gooing and gahing from parents endless pointing and peekabooing with children, fo that with a smart pup, and be consistent with language and actions and see how they grow up.

    Given humans treat other humans from just a few centuries as “unsophisticated” and a few thousand years ago as unintelligent, similarly many humans “talk down” to dogs, and get the results accordingly.

    This is the opposite of excessively anthromorphasising the dog. Its not human, but it’s not dumb. It can communicate in a very sophisticated manner to operate with others as a pack, it can do the same with a human, maybe better.

    Humans often over rate themselves, “hey, I can use a computer , I am the smartest animal to ever live”. Not.

  6. I have a very food drivin chocolate lab, I’m pretty sure once she figured out what button was “Feed me” or “treats” or “cookie” those are the ones she’d press all the time….

    1. The dog just pokes me when he wants something. It used to be obvious, he’d poke and then flip on his side, wanting a rubbing. But after being in the hospital last year for six months, he seems to be poking for food, and maybe to go outside. He’s really good at telling me he wants something, not so good at being specific.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.