Last week’s article on Flerkens spawned a discussion as they often do, correcting and expanding on a somewhat offhand comment about Flerken cognition. Specifically I asked, Are they sentient? That’s not what I should have asked, I was informed. The word I should have used was sapient.
I was actually riffing off the article before that, about Star Trek and Data’s cat, Spot. In the Star Trek universe, sentience is used in the sense of sapience. It goes beyond the processing of sensory data to the capacity for intelligence and self-awareness.
The essential episode in this context aired during Season 2, in 1989: “The Measure of a Man.” It was written by Melinda Snodgrass, and it’s considered to be one of the best episodes in the whole of the Trek Universe—some would say the best. It distills the philosophy of Trek into a single point: the nature of humanity, in the sense of both the human species and humane behavior. The question is whether Data is a person or a piece of property. Whether he belongs to himself or to Starfleet.
It’s a powerful episode on a number of levels. It’s strikingly apposite to questions we’re asking now about what we’re calling Artificial Intelligence. But what’s relevant here, in this series, is how it defines sentience (or sapience).
Dr. Maddox, the cyberneticist who wants to take Data apart and make a limitless number of copies, argues that Data is not a sentient being and is not entitled to the rights granted such a being in the Federation. The three criteria of sentence in this context are intelligence, self-awareness, and consciousness.
Intelligence, says Maddox, is the ability to learn and understand and to cope with new situations. Self-awareness means that “you are conscious of your existence and actions. You’re aware of yourself and your own ego.”
Consciousness is wrapped up in those two things. You know who you are, and that you are. You’re aware of your place in the universe.
In the episode, Captain Picard is able to convince both Maddox and the judge in the hearing that Data meets all three criteria and is therefore a sentient being. (Or sapient if you will.)
That’s the context in which Data composes the “Ode to Spot,” in which he says,
And though you are not sentient, Spot, and do not comprehend,
I nonetheless consider you a true and valued friend.
Spot of course is a sentient being. Merriam-Webster defines the word as “capable of sensing or feeling : conscious of or responsive to the sensations of seeing, hearing, feeling, tasting, or smelling.”
What she may not be is sapient: possessing or expressing great sagacity.
That is, wise in the sense of homo sapiens. Capable of higher cognitive functions. In Dr. Maddox’s terms: intelligent, self-aware, and conscious. Spot, Data believes, is not any of those things. She’s a creature of instinct. She doesn’t think. She’s not aware of herself.
That’s the traditional view of animals in Western culture. Man is unique among them all, raised up above them, ordained by heaven to rule over them by virtue of his superior intelligence. He is the only being with the ability to construct and use tools, and most of all, he is the only one who has developed language.
The conception of animal cognition has changed considerably since Next Generation first aired. One of the main criteria of human intelligence, tool use, turns out not to be unique to humans. Primates, birds, insects, and sea creatures have been observed using tools. There’s even a cow in Austria, who may or may not be a genius of her species.
Attempts to teach animals human language go back well before the original series. Increases in computing power and changes in how we perceive animals have given us the capacity to decode what may be language in sperm whales. It’s becoming clear that intelligence is not an on-off switch, or human-not human. Rather, it’s a spectrum, and many animals are a lot further along it than we used to think.
That brings us to a social-media phenomenon that’s been growing exponentially: animals pushing buttons equipped with human words, to communicate with their humans. Author Mary Robinette Kowal is part of the movement with her calico cat Elsie. There is a huge study, thousands strong, of button-pushing dogs, coordinated by Federico Rossano of the Comparative Cognitive Lab at the University of California in San Diego.
Rossano’s study and others like it are featured in a brand-new Nova documentary, “Can Dogs Talk?” It’s well worth watching if you’re interested in animal intelligence. It addresses the nature of language and asks whether dogs are actually using it to communicate, or if they’re applying more or less random associations to specific buttons, words, or sounds. Do they understand what Play or Outside or Beach actually means, as a word, or would any sound (or gesture or signal) do just as well to get them what they want? Are we training them to push a button to get a set response, or are they applying some form of reasoning to their choice of buttons?
Animal-communication studies prior to Rossano’s have tended toward very small samples, as small as one researcher focused on a single animal removed from its natural habitat. Under those conditions, it’s questionable as to whether the animal has learned language or if the researcher has trained them to respond in specific ways to specific stimuli.
The Rossano study encompasses ten thousand dogs in fifty countries. In its most basic form, each owner records button pushes on a simple spreadsheet. A more sophisticated version uses an app that records when the dog pushes a button, which button it pushes, and whether the button has been pushed by the dog or a human.
It’s the largest animal-cognition study that’s ever been done, and it’s ongoing. It’s meant to continue for years. The scale of it is immense. Rossano observes that in three months, the study records over a million button pushes.
Most of the communications between dogs and humans are transactional. Dog wants something, dog asks for it. Outside, Play, Food. This isn’t random, says Rossano. Dogs are stating their preferences. They’re letting their humans know what they want.
But are they “talking” through the buttons? That takes us into definitions of language, and the difference between action words and words for objects. A dog who can identify toys by name is rare compared to one who can tell their owner they need to go potty or they want to go to the beach.
Even rarer is the dog who is capable of what’s called language productivity. Given a limited number of words or buttons, if the dog wants or needs to convey a concept that isn’t in the existing vocabulary, they will combine two or three words into a new super-word. Some 800 dogs in Rosano’s study have been doing this.
One for example lost her button for Beach—it broke. On her own, without prompting, she pressed Water and Outside instead.
One of the animal-cognition experts in the documentary warns that one has to be skeptical. The substitute buttons are close by the broken one, so maybe she just pressed the nearest ones. They can’t test it because once she’s used those buttons, they mean Beach. A test would mean moving the buttons and seeing if she still used the same ones.
But there are other dogs who have combined other buttons into new configurations, and those don’t seem to be random. The one that convinced Rossano to do the study despite his reluctance to risk his career on it was a dog who went through a progression of buttons: Mad, Ouch, Stranger, Paw. The owner at first didn’t make sense of it, till she asked the dog for her paw—and found the “stranger,” a foxtail stuck in the tender web beween toes.
That’s communication. It gets even more amazing in a smaller study based in Hungary, which looks at what canine cognition expert Claudia Fugazza calls gifted word-learner dogs. There are only about fifty in the world, that she has found so far. These dogs can learn multiple words for objects—as many as a thousand—and are also capable of sorting them into categories. Tug toy versus fetch toy, for example.
Can these or any dogs compose complex sentences with correct grammar? As far as we know, no. But that they can learn and remember words, and acquire a considerable vocabulary, yes. And they can create new words by combining those they already know.
Some appear not just to ask for things, or name things, but to narrate actions. One might press the button for Settle, without being asked or rewarded for it, then put herself to bed. She’s just talking to herself, apparently. Using her word.
Is she intelligent? She seems to be. Is she self-aware? She may be. Is she conscious? It’s possible. The more we learn, the more data we gather, the more we’ll understand about how, and whether, and to what extent dogs think.
Cats, too. Eventually. One hopes.
The idea of dogs or cats having a range of both capability and desire to learn buttons is most interesting. One cat may take to buttons very quickly in order to better communicate with their human but another might never learn anything beyond “treat.” Is that because he lacks the capability to do the reasoning required or is he confident that cats ways of communicating will be sufficient?
The sentient/sapient confusion which dogs SF is very annoying.
I think the big complication is the cat brain which is more specialized, smaller, and less complex than other animals such as dogs. Cats do a surprising amount of visual processing directly in their optic nerve which reports directly to their spine and muscles so that the cat’s brain may not be aware of what is happening when the cat chases the mouse.
In my experience, that depends on the cat… ;)
But that would explain tail-chasing behavior, I guess. And I think maybe it’s specious to define the brain and the nervous system as separate entities rather than a continuous whole. There are other species that decentralize their processing, like cephalopods, whose arms can act independently even when severed.
I have not looked for similar studies of cats. It’s on my list. A big problem with them of course is that dogs are eager to please, but cats, generally, are not. Even if the cat hears you, she may not bother to acknowledge you.
I was reading about the challenges of working with the cats who played Spot. Brent Spiner was not exactly thrilled.
Completely off topic, but our hostess may appreciate this.
“How a Horse Whinnies: With a Whistle and a Song”
https://www.nytimes.com/2026/02/23/science/horses-whinny-sounds.html?unlocked_article_code=1.OVA.0ObV.-oxasbKiTPVl&smid=url-share
Very interesting, thank you! I used to say my late stallion was a Heldentenor. He had a song he would sing when a mare was ready for him. It was enough to split an eardrum, and there was a definite melody to it. (Coincidentally, I just watched a Reel from a whale watch with a male humpback singing nearby. They had to turn the underwater-cam volume down to 35% to keep from blasting viewers’ eardrums. Gorgeous song, but talk about power.)
This is fascinating, and I have been floored by some of the things MRK has described her cats doing/saying with the buttons. I do think it may be worth being careful about assertions of “without prompting” – I would prefer “without deliberate prompting,” because it’s hard to rule out animals picking up on non-verbal unintended cues when they look like they’re about to interact with the “right” button. (I have been conducting my own involuntarily longitudinal anecdotal study into “nonverbal clues we are going to the vet”)
The documentary shows tests of this exact thing, after several observations/warnings about overinterpreting the animals’ actions. They tested whether the dog would use the same buttons for someone other than the owner, and whether a human using the word with tone, intonation, body language, etc. would get the same response as the button (which always has the same tone and there’s no body language involved). In one test, the owner had her back to the dog, with headphones on so she couldn’t hear which buttons the dog pressed. Very interesting to see it in action.
Can dogs compose complex sentences with correct grammar?
In GERMAN?
(Or, say, Latin?)
Have mercy on the poor critters.
But I have two photos of my beloved Elf, who went to the Otherworld last month, which prove conclusively that she, at least, knows about her existence in the world and her (and my) place in it.
In the first, she is staring intently at the top of my bureau, where the treat boxes are.
In the second, taken less than two seconds later, she is staring just as intently at me. If that wasn’t an attempt at reasoned communication, I don’t know what is.