If This Then Dog
Dogs, algorithms, and black boxes
Dogs are pattern recognition machines.
Fuzzy pattern recognition machines.
And dogs’ lives depend on this capability.
So I shouldn’t have been surprised when we adopted our dog and named him (perhaps suitably), Emoji, that he would be no exception. At the age of 4 1/2, he’s a master reader of his human algorithms.
You’ve probably heard of If This Then That (IFTTT), the digital platform that allows people to plug their various digital and physical apps and devices together to do things like copy your Instagram photos to Dropbox, or get Alexa to turn down the heat on your Nest.
In our house, we have If This Then Dog, or IFTTD: the platform of our dog.
Emoji is extraordinary at gaining data from us to figure out what’ll happen next. We’re the inputs of a magical algorithm, and if he watches or sniffs closely enough, then he’ll be able to figure out the outputs—which mostly have to do with him being fed, walked, or with us going to work and leaving him at home (his least favorite outcome).
Some of the many functions of If This Then Dog
- Sniff pant legs while human gets dressed -> Determine whether they’re hanging out for a day of reading or putting on the go-to-work clothes.
- Human scrapes bottom of yogurt container -> Dog stands up in anticipation of getting the yogurt container to lick clean
- Garage door squeaks -> Human comes to back door
- Human runs bathtub -> Bad news
Many IFTTD functions revolve around Emoji’s concern that we’re leaving the house without him. Some of these functions are pretty rudimentary. Emoji knows, of course, that putting on our coats without his leash means that we’re going somewhere and he’s not coming. But he notices things that we don’t, like:
- Close laptop -> Dog jumps up because human is changing to new activity. Possibly one without him.
- Human wears shoes in house -> Humans about to leave; dog about to be left alone.
And I’m not sure why he sometimes knows we’re about to leave for work when we go out:
- Human takes dog for a walk -> Dog slows down on walk
Maybe he notices that we’re more hurried. Or maybe it’s the way we tied our shoes. But somehow, just around the corner from the back gate, he’ll do that gravity thing dogs do, you know:
- Human put dog car, take to vet
-> Dog invokes superpower of suddenly weighing 100 times more, as if on surface of Jupiter
If Dog Then Human
Emoji tries to reverse-engineer the algorithm, to greater or lesser effect. It doesn’t usually work around us leaving for work. For instance:
- Dog slow down walk -> Maybe human not leave
- Dog sneak up to Simon’s armchair: -> Maybe human not leave
- Dog snuggle up in bed in dining room: -> Maybe human not leave
But it works really well for other things:
- Dog stare at human: -> Maybe human give dog ears a nice scritch-a-scratch
- Dog stare at human: -> Maybe human give dog tasty treat
- Dog nudge human calf with wet nose: -> Maybe human give dog scritch-a-scratch or tasty treat
We are black boxes to our dogs.
We are black boxes to our dogs. The world around them is sensory input and activities to be experienced. Perhaps they think that if they could just figure out those clues, they’ll know what’s going to happen next. For a rescue dog like Emoji, this capability helped keep him alive when he was roaming Milwaukee’s streets with a pack of small dogs. But maybe they don’t need to figure out what’s going on inside of us: they understand us by our inputs and outputs. For all of our black boxes, they’ve got black, wet noses. And they’re at doing what they do—and what they want us to do.
And they’ll hope it’s nice walks outside, tasty treats, and long days at home snuggling with their humans.
Molly Wright Steenson is the co-owner of Emoji, a 4 1/2 year old bichon frisé-poodle-chihuahua rescue, along with her husband Simon. She’s the author of Architectural Intelligence: How Designers and Architects Created the Digital Landscape, a history of artificial intelligence in architecture and design.
