happy. ball. want. outside.
November 7, 2019 12:42 AM   Subscribe

Stella, an 18-month-old dog, can use a sound board to communicate using the 29 words she knows in short phrases. Christina Hunger, human and speech-language pathologist, documents her progress on instagram.

People.com (warning- autoplay video) interviewed Christina about Stella and how she can put words together to make simple sentences.

Stella is a Catahoula/Blue Heeler cross, and has been learning to 'talk' by pressing on buttons since she was 8 weeks old.
posted by freethefeet (74 comments total) 44 users marked this as a favorite
 
wow animals are getting smarter and smarter, don't you fell so?
posted by ElliotChang at 12:46 AM on November 7, 2019 [2 favorites]


I think we are only just realising how smart animals are.
posted by freethefeet at 12:55 AM on November 7, 2019 [9 favorites]


Oh, and here is Christina's blog- https://www.hungerforwords.com/. The first article I read made it sound like the instagram was the blog.
posted by freethefeet at 1:02 AM on November 7, 2019 [2 favorites]


me. good. Brent.
posted by CheesesOfBrazil at 2:21 AM on November 7, 2019 [44 favorites]


This story is fascinating...and obviously bumps up against the Clever Hans effect.

Going down the rabbit hole for this earlier in the week, I was surprised at the implications of the human/animal cue mechanic necessary to make any of this work. Just how important is the relationship between trainer and animal here? You can't just reproduce this experiment with another dog because the bond between trainer/animal is unique & critical. What does that mean exactly for the "validity" of what is happening? For example it's fact that these "cues" result in false positives with drug sniffing dogs (human trainer bias is passed through to the dog).

As always with these stories of animal intelligence, it's good to revisit Alex the parrot, which is another fascinating rabbit hole for anyone interested.

An additional question that never seems to come up in stories like this...is Stella or was Alex some sort of relatively one-off language genius? An avian/canine Shakespeare? And if they're orders of magnitude more intelligent than others of their kind, doesn't that make reproducing the results even more challenging?

Anyways loving this story and Christina is obviously a "good human" for putting in the time with Stella & having the perfect background/training and wherewithal to grind this all out.
posted by NervousVarun at 2:26 AM on November 7, 2019 [27 favorites]


Rats driving cars. Dogs using talking boards. We're doomed.
posted by tommasz at 3:13 AM on November 7, 2019 [2 favorites]


They taught horses to indicate if they wanted their blankets on or off. It took 11 days to teach 23 horses the symbols "blanket on" and "blanket off" and "no change". I am not saying horses are rocket scientists but they are certainly smart enough to know if they want their blankets on or off. Source.
posted by which_chick at 4:06 AM on November 7, 2019 [31 favorites]


is Stella or was Alex some sort of relatively one-off language genius?

I guess that would depend on how rigorous the selection process was. In these cases not very, so the likelihood that these results can be readily reproduced is probably pretty good.

(with the understanding that exceptionally stupid samples do pop up in every species on a regular basis)

What I'd like to see is this chatty puppers placed in a cohabitation environment with several other young clever dogs to see if they, upon seeing the play and snack benefits reaped from pawing the talking buttons, learn to communicate spontaneously without training.
posted by CynicalKnight at 4:06 AM on November 7, 2019 [8 favorites]


This is incredible.

From her blog: One afternoon shortly after the Daylight Savings time change, Stella said, “eat” repeatedly at about 3:00 PM. When Jake and I did not feed her dinner this early she said, “love you no” and walked into the other room.

I think if I helped my beloved pet to learn to express itself with language and it said "I don't love you," I would be forever broken.

So much for using this with my wife's cat.
posted by philotes at 4:07 AM on November 7, 2019 [71 favorites]


Just how important is the relationship between trainer and animal here? You can't just reproduce this experiment with another dog because the bond between trainer/animal is unique & critical.

I think the same can be said for the parent-caretaker / child bond, but neither scenario invalidates the innate capacity / intelligence of the learner. And as a parent, I can tell you there are plenty of very similar “false positives” with children as well.
posted by Silvery Fish at 4:09 AM on November 7, 2019 [5 favorites]


I think if I helped my beloved pet to learn to express itself with language and it said "I don't love you," I would be forever broken.

If it's anything like with kids, the first time feels like that, and by the tenth time you just tell "that's too bad, but you still can't have Halloween candy for breakfast."
posted by Bulgaroktonos at 4:48 AM on November 7, 2019 [83 favorites]


All right, be honest, how many people commenting here are actually dogs?
posted by Acey at 4:54 AM on November 7, 2019 [62 favorites]


On the Internet, nobody knows you're a dog.
posted by briank at 4:57 AM on November 7, 2019 [31 favorites]


"On the Internet, nobody knows you're a . . ." oh, never mind.
posted by dannyboybell at 4:58 AM on November 7, 2019 [9 favorites]


All right, be honest, how many people commenting here are actually dogs?

There is a suspicious amount of cat slander from some people here.
posted by Bee'sWing at 5:07 AM on November 7, 2019 [21 favorites]


No dog

Cat where chase

No dog
posted by ianso at 5:35 AM on November 7, 2019 [10 favorites]


Who's a good Mefite?
posted by Reverend John at 5:41 AM on November 7, 2019 [10 favorites]


It just occurred to me while watching the video that the dog is not only communicating using the device but is either able to 'read' the labels or memorize the exact location of the appropriate device for the word she wants to say?
posted by Karaage at 5:59 AM on November 7, 2019 [5 favorites]


They're all good Mefites, Reverend John.
posted by Quindar Beep at 6:13 AM on November 7, 2019 [9 favorites]


Watching the Instagram videos I really cannot tell if the dog is communicating at all or just repeating behaviors. When we were training our dog and teaching her a new trick, if she didn't understand the command yet she would just cycle through all the tricks she knew in no particular order, hoping that would get her a treat and praise. This is a dog that has been trained to press buttons that make sounds, but I don't believe she is aware of much more than the reactions the sounds have gotten from her trainer in the past and the ability to read cues and guess what which sound is likely to elicit a reward.

I don't think this person is lying, but so much of the significance of what Stella does comes from the way her trainer frames the videos in the comments. Without seeing the whole training process here it's hard to say anything much is going on besides a dog trained to push buttons.
posted by dis_integration at 6:14 AM on November 7, 2019 [19 favorites]


I knew a dog (husky) who could absolutely read analog clocks and understood when people would say things like "we'll go outside in ten minutes" or "you'll get dinner at seven thirty" - she'd listen intently, look at the clock, go back to her nap or toy or to bother the other dog, and come back at exactly the described time. She also recognized the logo for Petsmart and would get all wiggly in the car on road trips when passing one, even an unfamiliar one. So, some dogs can be wicked smart, and learn really complex things. Some dogs are also highly adept at using what they have to communicate what they want.

I am inclined to believe that this (very good) dog is communicating with her person. But I also see a lot of confirmation bias in it, one of those "everything's a nail if all you have is a hammer" situations. Reading the descriptions and some of the blog shows a lot of comparison to human children, and that can be a great point of reference to begin with and as a familiar touchpoint with fans online, but it also seems kind of oblique to the actual situation. What I'm really curious about is what Christina is learning about how Stella, a dog, appears to understand about the meanings of words that are not the human understandings of them. We've evolved alongside dogs but they are still experiencing the world very differently from us. Apart from the very cute dog, that's what I'm most interested in, here. It's not going to magically result in talking dogs like in Up, and the closer it looks like that the less inclined I am to accept Christina's interpretations of meaning. I want insights into what she's learned about Stella's perceptions and experiences.
posted by Mizu at 6:37 AM on November 7, 2019 [11 favorites]


I think I've told the story here before, but my dog trained me how to communicate. It took me a while to catch on, as I thought she was just playing at first, but if she brings me an empty water bottle or starts pawing at the case under the kitchen table, I know her water bowl needs refilling.

(Yes, we use bottled water. Yes, the pets get bottled water, too. We need a whole house water filter, ok? Don't judge.)
posted by Ruki at 6:41 AM on November 7, 2019 [6 favorites]


Stella is a very good and very smart girl, but isn't this a Chinese-room issue? She has learned that certain buttons pressed in a certain way can yield desired results, plus they make her people very happy, which is always a desired result for a dog. The fact that the bells are attached to words is meaningful to us, not to her. She is combining the bells in sophisticated ways because she is a clever pup and wants to try anything that might seem right.
posted by Countess Elena at 6:48 AM on November 7, 2019 [8 favorites]


Not seeing a button for "squirrel."
posted by doctornemo at 6:59 AM on November 7, 2019 [6 favorites]


isn't this a Chinese-room issue?

You may be right but the same argument could apply at scale to you and me. Philosophically I find the arguments for and against evaluating degree of consciousness based on behavior to be uninteresting.
posted by hypnogogue at 7:00 AM on November 7, 2019 [6 favorites]




I apologize. Will do.
posted by Countess Elena at 7:08 AM on November 7, 2019 [18 favorites]


I mean, if three humans were kidnapped by intelligent aliens whose language required seeing in infrared and generating sounds that humans are entirely incapable of producing, and the aliens were incapable of understanding our language and then the aliens spent time with one of the humans teaching them to use a series of buttons to communicate intent and reaction, while the others were merely fed and tickled, it would be a strikingly similar scenario. The aliens would wonder: are these creatures really intelligent and capable of language, or did this one human just learn that pressing buttons elicited actions?
posted by grumpybear69 at 7:20 AM on November 7, 2019 [6 favorites]


It is literally a Searle's Room problem. But more importantly, using individual words isn't really human language. Human language includes the ability to use grammar to string together words in novel sentences. Some of the videos here seem to show a simple version of this ("come" "outside" "look") but I'm gonna remain skeptical until there's some evidence more substantive than Instagram videos.
posted by Nelson at 7:20 AM on November 7, 2019 [4 favorites]


And of course dogs aren't going to have human intelligence, because they are not human, but that does not mean they cannot engage in meaningful communication using an intermediary translation system like, say, the button matrix in the video.

In order to draw any sort of meaningful conclusion the same educational process (which is what that is) would need to be applied to a large cohort of dogs by a variety of trainers and the results analyzed.

And, come on, it is not a Searle's Room problem because dogs are animals and not computers. The whole concept of "if not human, then automaton" is tiresome.
posted by grumpybear69 at 7:25 AM on November 7, 2019 [8 favorites]


She has learned that certain buttons pressed in a certain way can yield desired results

And isn't this exactly what communication is? When proto-Humans first learned to talk to each other, it wasn't more complicated than this dog's word board.
posted by Xyanthilous P. Harrierstick at 7:29 AM on November 7, 2019 [15 favorites]


The fact that the bells are attached to words is meaningful to us, not to her.

I think they should add buttons like 'annoy cat' , 'give me better food that doesn't come out of the can shaped like the can' and 'eat humans' to really test them.
posted by The_Vegetables at 7:32 AM on November 7, 2019 [5 favorites]


If someone wrote an algorithm to activate three or four of those buttons at random, I wonder how many of the resulting word groups could be charitably interpreted as sentences.
posted by Flexagon at 7:34 AM on November 7, 2019 [3 favorites]


NervousVarun:
With dogs it’s probably a mixture of classical and operant conditioning, clever Hans, forced evolution of social instincts, and the fact that most central nervous systems above a certain threshold of complexity experience emotions. Regional/global weighted modifiers on neurotransmitter activity are a lot more common in the animal kingdom than compartmentalization and mapping of internal states (read: passing the mirror test, step one of several for self-awareness/recursive self-as-agent modeling). Dogs do not possess the latter capability, nor the capacity for contextual anguish it enables humans to subjectively experience. That said, emotional pain, suffering and even prolonged sadness (along with other more positive emotions) doesn’t require there to be anybody at home. Basic emotions are more universal than identity, particularly in mammals.

I’m sorry, I know people hate hearing the above because they love their dogs - as I’ve loved all of mine - but it doesn’t change the facts and the incessant anthropomorphism is beyond frustrating. You love them, they love you: sure. Absolutely true. They feel loved in a vague, ambient warmth/fulfilled pack instinct way: probably true. They have a discrete sense of self, an identity, a mind’s “I”, and contextualize those experiences like a person (*I* am loved, am valued as an individual): sorry, no. Utterly false. Exhaustive testing forces us to conclude that they simply don’t meet the most basic requirement for that to be possible.

Nor do cats, but cats wouldn’t care even if they had the cognitive resources to appreciate it. Dogs, I am certain, would be sad if they knew.

For about 30 seconds.

Interestingly, dogs do practice basic forms of deceit: deliberately framing other dogs for stealing food, that sort of thing. Which implies modeling their master’s mental state despite not having the capacity to distinguish their own as separate from the world. From a cognitive perspective that’s a gross distortion of pack instincts on par with the most grotesque examples of “purebred” traits, and almost certainly humanity’s fault.

Cats and dogs aside, whether any of that matters to you is your business, but I deeply wish people would stop insisting that there is literally a face in that cloud, a man in that moon. Your brain is heavily predisposed towards recognizing the patterns of faces and agent-intent, and it drives me absolutely crazy when that bias isn’t taken into account.

Gorillas are a lot more interesting (to me) because they collectively rub right up against the border of a number of cognitive inflection points, including self-recognition, with select individuals falling on either side of the mirror test, or facility with language, etc. Gorillas are endlessly fascinating in the way that twin studies or exotic brain injuries are: there’s so much we can learn from observing how specific individuals perform intuitive leaps or feats of directed modeling/problem-solving, while others simply never develop similar abilitIes no matter how you train them.

If we ever crack minds in the general case - as an abstract model of which humans are but one example - I believe it will be because we had access to gorillas. Dogs will remain dogs unless (probably until, damn us) we decide to get a lot more...hands-on with their cognitive abilities than mere tens of thousands of years of artificial selection.

In conclusion: every dog is a good boy and what we have collectively done to them is a horrific moral crime against the whole of dogmanity. Thank you for coming to my TED talk.
posted by Ryvar at 7:35 AM on November 7, 2019 [21 favorites]


if she didn't understand the command yet she would just cycle through all the tricks she knew in no particular order, hoping that would get her a treat and praise

I've had co-workers like this, with graduate degrees.
posted by paper chromatographologist at 7:43 AM on November 7, 2019 [74 favorites]


Read an old scifi story where a group of explorers on a planet had a disaster happened that left them naked and stranded. Guess what? *Another* group of aliens was visiting that planet and grabbed them as specimens => kind of a zoo thing.

They were wondering these same questions then one day they were recognized as "intelligent". How? Why? Because they had caught a couple of mice like critters that were in their pen and were feeding them as "pets".
posted by aleph at 7:52 AM on November 7, 2019


It might be interesting to add buttons that corresponded to concrete, meaningful, but uninteresting concepts. Like household objects that Stella is familiar with but does not interact with (e.g. a broom, perhaps). That might help disprove or affirm the "push buttons at random until something good happens" hypothesis.

For example, if Stella says things like "love broom" or "eat broom" then that suggests she doesn't understand "broom" (or is a very strange dog). But if she basically ignores it then that could be evidence in favor of comprehension.

Given that dogs probably have a relatively limited vocabulary capacity, adding uninteresting words now runs the risk that the words are ignored because Stella is at her limit for distinguishing sounds or learning new vocabulary. To counter that one would also need to then add some additional interesting words and observe that she added those to her vocabulary.

Basically: interesting concept! But more testable hypotheses, please.
posted by jedicus at 7:52 AM on November 7, 2019 [2 favorites]


This is a good stopgap until the eyebrow translator is perfected.

Also: can he start fires with his mind?
posted by BeeDo at 8:05 AM on November 7, 2019 [1 favorite]


I'm disappointed at the unfounded skepticism in here. (the founded skepticism is fine). they've already done research that dogs understand the difference between actual words and sounds. From that, it's not hard to believe that they can repeat what they are hearing and assign a concept to it.
posted by FirstMateKate at 8:10 AM on November 7, 2019 [12 favorites]


OK, but perhaps we should be talking about the border collie who digitized the speech samples, fixed a small bug in the fast fourier transform algorithm in passing, and wired up the soundboard.
posted by Wolfdog at 8:10 AM on November 7, 2019 [29 favorites]


So the headline of the article on People.com is Dog Learning to Talk By Using a Custom Soundboard to Speak: 'I'm in Constant Amazement' . And when I scrolled past it on Facebook a few days ago, I genuinely thought that that’s what the dog had expressed. Because like have you ever met a dog?
posted by 2or3whiskeysodas at 8:17 AM on November 7, 2019 [51 favorites]


Not sure the dog's at it's vocab limit— wasn't there a doctor that trained their dog with near/over 1000 retrieval commands? I thought most dogs can tell ~150 commands/words pretty easily, the majority of owners just never bother.

Our little nugget surprised our friends with her non-verbal communication skills. They were looking after her for two days, and about sundown on day one she started going up to them, nudging the back of their leg with her nose, then as they turned around, she backed off outside head-pat distance. When they took a step towards her she'd back up again, rinse-- repeat.

Only after a minute or two did they realize she was bridging them to the fridge, as it was now... three minutes past her the time we would usually feed her. (She has not enjoyed the daylight savings shift either!)
posted by Static Vagabond at 8:22 AM on November 7, 2019 [4 favorites]


I'm disappointed at the unfounded skepticism in here. (the founded skepticism is fine). they've already done research that dogs understand the difference between actual words and sounds. From that, it's not hard to believe that they can repeat what they are hearing and assign a concept to it.

Oh, I have no doubt that dogs can distinguish between tone and word and distinguish one word from another. My dog knows "Go see dis_integration" vs "Go see mrs. dis_integration" and will reliably run to either of us no matter who is making the command. She knows the difference between toys based on name "ball" "rope" and will go get those specific things.

I can easily believe they could repeat commands expressed to them and learn that the "outside" sound often means they get to go outside.

But "no no help no no" or "no love you", or for that matter anything that seems syntatically correct or expresses complex concepts? Skeptical! Very! And is this "language" or does it not go beyond the Wittgensteinian "Slab!" thought experiment?
posted by dis_integration at 8:24 AM on November 7, 2019 [4 favorites]


FWIW the skepticism here is not new. The debate re: animal language use is massive and many a linguistics PhD has been earned from work on the topic. This one situation seems to suggest that the dog is using productive language which I have serious doubts about.
posted by lazaruslong at 8:31 AM on November 7, 2019 [3 favorites]


I have to admit that mood-verb-object is pretty much how I would expect dog syntax to work. But the claims about the use of negation restore my skepticism. Humans have been bred for language skills for far longer than dogs have been bred for people skills, yet even humans have an awfully hard time with negation.
posted by Not A Thing at 8:41 AM on November 7, 2019 [4 favorites]


For example, if Stella says things like "love broom" or "eat broom" then that suggests she doesn't understand "broom" (or is a very strange dog).

Insert joke about humans' failure to understand the meaning of "eggplant" here.
posted by Not A Thing at 8:45 AM on November 7, 2019 [2 favorites]


My dog (a golden retriever) taught herself how to use reflections. She spent about 45 minutes at a time, batting a ball around and watching what it did in the reflection until she understood what was going on. I have video of her doing it. I can now give her hand commands in the mirror with me standing behind her and have her do them. I can hold a toy over her head where she can only see them in the reflection where she will invariably turn the right direction to where she can grab them. She uses a mirror in the hall to locate me if I hide around the corner. According to Scientific American, she is an exceptional dog.
posted by Xoc at 8:46 AM on November 7, 2019 [11 favorites]


mood-verb-object is how cat syntax works, too, it's just that the mood, verb, and object are invariably "furious-fuck-you"
posted by Wolfdog at 8:50 AM on November 7, 2019 [11 favorites]


It is very easy to be skeptical of canine intelligence until you have seen a dog concoct a deceitful plan and see it through over the course of six to eight hours.
posted by wierdo at 8:53 AM on November 7, 2019 [8 favorites]


Nah, Wolfdog, you obviously haven't met the millions of lovely cats out there. Both of mine are as affectionate as any dog.
posted by agregoli at 8:55 AM on November 7, 2019 [3 favorites]


Of course, watching my other dog wrap himself and his lead around a tree or post or something and seeing the surprise on his face when it keeps getting shorter can easily restore one's skepticism, plus some.

I think the smarter members of the species are more capable than we generally give them credit for, though largely hobbled by what we would call severe ADHD in humans.
posted by wierdo at 9:03 AM on November 7, 2019 [4 favorites]


It's not a Searle's Room problem because there's no unboundedness being demonstrated here. Thud the particular setup for the dog is effectively unfalsifiable, which means if you are concluding that this dog has linguistic intelligence, then you are using an unfalsifiable subtheory

Which is not even science
posted by polymodus at 9:44 AM on November 7, 2019 [3 favorites]


For example, if Stella says things like "love broom" or "eat broom"

i love lamp
posted by ricochet biscuit at 10:10 AM on November 7, 2019 [9 favorites]


That is because you are crazy! Lamp has no feelings.
posted by praemunire at 10:22 AM on November 7, 2019 [10 favorites]


if Stella says things like "love broom" or "eat broom" then that suggests she doesn't understand "broom" (or is a very strange dog)

You clearly have never swept a floor with a dog in the room.
posted by dchase at 10:57 AM on November 7, 2019 [12 favorites]


Your dog wants steak.
posted by NedKoppel at 11:37 AM on November 7, 2019


Being skeptical that Stella is demonstration human language (which I am) is not the same thing as being skeptical that dogs are intelligent (which they are), or that dogs can communicate through a variety of means (which they do).

Hunger's youtube videos showing training (mostly cuing) are few and Instagram videos are frustrating because they are short and without visual context. If you watch all four of Stella's paws there are some interesting behaviors that are hard to examine due to the format (for example in this video Stella paws at the "Eat" button before the behavior that Hunger highlights, but it's completely unremarked.)
posted by muddgirl at 12:39 PM on November 7, 2019 [1 favorite]


what i want to know is can dogs understand memes
posted by poffin boffin at 12:40 PM on November 7, 2019 [2 favorites]


Xoc:
For whatever it's worth your dog sounds extremely intelligent and I would've assumed golden retriever even if you hadn't specified. Pigs are another case where specific individuals demonstrate understanding of how mirrors work, and some mixed success at tests of object permanence. Like dogs they consistently fail at self-recognition (I really appreciate you linking an article that leads with that disclaimer), but it's clear there's more general awareness in play than your average non-primate mammal.

For anybody interested in reading more on this topic, AnimalCognition.org appears to have a number of excellent writeups and interviews with current researchers about their ongoing work, written for general audiences. I think a lot of people reading this thread would enjoy it.
posted by Ryvar at 1:32 PM on November 7, 2019 [6 favorites]


The one that gets me is that one day a large and scary package arrived to the house, and Stella spontaneously pressed Help and No repeatedly to indicate her displeasure. That's both a novel way of communicating her emotions, and they're negative emotions - she doesn't just press buttons to receive good things.
posted by nakedmolerats at 1:48 PM on November 7, 2019 [3 favorites]


I think that's a good example of why programs trying to teach animals to speak human language are unintentionally deceptive, actually. Hunger is the one who is interpreting for us that "help no" means that Stella is distressed (rather than Stella is refusing help), yet "bye no" means that Stella doesn't want her owners to leave.
posted by muddgirl at 1:58 PM on November 7, 2019 [2 favorites]


Of course this is consistent if "no" isn't a negation at all but an expression of negative emotion, which again is communication but not necessarily language.
posted by muddgirl at 1:59 PM on November 7, 2019 [3 favorites]


I wouldn't go so far as to call it human-level language, either, but it is pretty impressive and (as far as I knew) unprecedented for a dog to put together novel strings on their own.
posted by nakedmolerats at 2:01 PM on November 7, 2019 [2 favorites]


As a very new speech-language pathologist who is working with a LOT of students learning to use their Augmentative and Alternative Communication (AAC) devices, I love all of the very important messages that Christina includes in her Instagram posts - like her initial comment in this one:

"Give your AAC user time to learn, lots of opportunity to practice, and encouragement through all of their attempts. Your belief in their potential makes all of the difference." Convincing people of this has been one of the most challenging aspects of my job thus far.

True Language or "just" communication, whatever Stella and Christina are doing together, if it encourages one teacher, aide, or family member to give their AAC user the time and support they need, I am 100% for it.
posted by DingoMutt at 2:20 PM on November 7, 2019 [2 favorites]


I went back and looked at that post and Hunger doesn't say it was a novel string - as far as I can tell she doesn't film her training sessions, just refers to them incidentally so it's impossible to say. Reacting to novel situations with a learned response is called "generalization" in dog training.

The most interesting thing about that post, to me, is Stella's body language. Her ears are up and engaged, and her tail is up and engaged as well. She doesn't look like a frightened dog, she looks like a dog who is focused or what we call "in the game" in positive reinforcement training. This is pretty much the same body language she has in every video.
posted by muddgirl at 2:25 PM on November 7, 2019 [3 favorites]


What if you got this and your dog turned out to be kind of a dick?
posted by schadenfrau at 3:41 PM on November 7, 2019 [10 favorites]


What if you got this and your dog turned out to be kind of a dick?

JET
FUEL
NO
MELT
STEEL
BEAMS
posted by CynicalKnight at 7:05 PM on November 7, 2019 [17 favorites]


This seems like a good place to mention Chaser, the border collie who died earlier this year and knew the names of over 1000 fetch toys, plus other words. The second link is from a Nova segment with NDGT where he meets Chaser and tests his skills. He picks 7-8 unique toys from a pile of 1000 and puts them behind the couch, then says "Chaser, fetch Inky" and Chaser runs behidn the couch and picks out Inky the squid. The best part is when NDGT adds a new toy ("Darwin") and Chaser picks it out by recognizing it as the only one in the pile he doesn't know.
posted by freecellwizard at 6:42 AM on November 8, 2019 [4 favorites]


All right, be honest, how many people commenting here are actually dogs?

*nervous panting*
posted by Tehhund at 9:17 AM on November 8, 2019 [6 favorites]


Like DingoMutt I find the fact that she's an SLP is central to this whole endeavor, as AAC (Which is what this is. Side note: this set up must be significantly expensive) is a core competency for speechies.

What that means is that they spend a lot of time (and, as an OT, I'm privy to a decent amount of this) teaching humans with wildly varying levels of cognitive and physical ability to use devices like this, trouble-shooting set-ups, and assessing what's going on with what they're seeing.

We have kids who use AAC who are absolutely not thinking linguistically, but are using their systems to meaningfully communicate their wants and needs, which significantly improves their quality of life and ability to connect with the people around them. Looking for "scrolling" behavior (hit buttons till something happens) and watching for generalization and spontaneous innovation are things everyone on the team look for (and not all kids come out on the "good sign" side of those indicators). I'm sure, if she chose to, she'd be able to use those SLP tools to write up something a lot more rigorous (though not necessarily in the same language/framework as the cognitive/neuro piece people here have been using here).

So, yeah, for me, that's the other half of the puzzle. Half of it is "look how smart the dog is" but also "look how much is possible with the right AAC set-up" in a way that's novel and less soft-focus inspirational than a lot of other media about AAC. I recognize that the substitution of a doggo for a disabled person in the set-up warrants careful handling in how it's framed, but I hope other SLP's and others can use this to help facilitate understanding.
posted by DebetEsse at 2:05 PM on November 8, 2019 [10 favorites]


I would be very interested to see how a native ASL speaker would set this up and teach their dog. ASL doesn't translate word-for-word like writing does.
posted by FirstMateKate at 7:11 AM on November 11, 2019


What if you got this and your dog turned out to be kind of a dick?

If a lion could speak, we could not understand it.
posted by jackbishop at 9:53 AM on November 11, 2019


It seems to me like questions of facilitator-led interpretation like we see with Hunger and Stella are even more important when we are talking about AAC devices for people who are disabled. It's not like the field has been immune from these kinds of problems in the past.

ASL is a different language with different grammar from English. Gorillas who learned ASL hand signs never seemed to pick up grammar, just as chimps who learned to use AAC-type devices never learned (or developed their own) grammar either. Grammar is not superfluous to (our human conception of) language - it's a huge part of how we interpret meaning. "Man bites dog" vs. "Dog bites man" is a classic example in english.
posted by muddgirl at 4:34 PM on November 11, 2019


It is very easy to be skeptical of canine intelligence until you have seen a dog concoct a deceitful plan and see it through over the course of six to eight hours.

I've had a few smart dogs over the years. When I was kid we had a border collie who could herd chickens in a open field. All you had to do was define a perimeter by walking it with her and then she'd keep a flock of chickens within the confines of the perimeter until you told her to stop.

I also had a dog growing up who was the deceitful type. Would frame the cat for stealing food/getting on the furniture. Would fake a limp for attention. Was an escape and break in artist and would regularly go on unauthorized walkabout to a former owner (several kilometres and one river crossing away) because said owner would send him home in a taxi and he loved car rides.
posted by Mitheral at 8:48 PM on November 13, 2019 [3 favorites]


« Older Scrollin' scrollin' scrollin', keep those pages...   |   The Great British Dildo Heist Newer »


This thread has been archived and is closed to new comments