In Part 1, we explored how TuxnDog, a Cane Corso trained on pet communication buzzers, began giving commands in English to another dog—our longtime friend and neighbor, Count Sydney. But behind the cuteness of two dogs “talking” to each other is something far more profound: an ongoing experiment to challenge the mainstream belief that dogs who count are simply reading their trainer’s facial expressions or body signals—not actually counting.
I’m Sky Blue Swain, the trainer of both TuxnDog and Count Sydney. For over a year, I trained Count Sydney to bark specific numbers on command—up to 10. Not through guesswork or hand signals, but through memorization of how many barks each number requires. Count Sydney’s understanding of the number prompts is precise. But for the experiment to be taken seriously, I needed to eliminate the human element—me.
That’s where TuxnDog came in.
Unlike Count Sydney, TuxnDog doesn’t know how to count. But TuxnDog does know how to communicate—in English. She uses pet communication buzzers, also known as talking dog buttons, to press phrases she’s learned and customized over time. TuxnDog began using these buttons naturally on her first day home, when she pressed “potty” within 12 hours, on her fifth trip outside. Since then, her vocabulary has grown to include favorites like “outside,” “want,” “Sydney,” and “if not now, when?”
Because TuxnDog had already been using the buttons for “Sydney” and “let’s play” anytime she saw Count Sydney outside, it felt like a natural evolution when she began pressing “Sydney,” “let’s count,” followed by a number. Not because TuxnDog knew what would happen next, but because she was delighted that she and Count Sydney were communicating—TuxnDog speaking in English through her buttons, and Count Sydney barking back in dog language every time she pressed a number button.
Interestingly, Count Sydney only barked when TuxnDog pressed a number button—never when she repeated “Sydney” or “let’s count.”
TuxnDog quickly caught on that Count Sydney only responded when she hit buttons she wasn’t familiar with yet—the ones labeled with numbers. TuxnDog began to favor those number buttons, testing them out to see what kind of reaction they triggered. And her reward, in this case, wasn’t a treat—it was Count Sydney barking joyfully at her.
And that’s the key.
TuxnDog doesn’t know what the correct answer is. She’s not reacting to Count Sydney’s performance. TuxnDog isn’t rewarding him or correcting him based on success. She simply presses another number button after he finishes barking—and Count Sydney gets a treat from me if he counted correctly.
The fact that TuxnDog has no concept of when Count Sydney is right or wrong is exactly what makes this experiment matter. If Count Sydney performs correctly even when TuxnDog is giving the commands—and I, the trainer, am completely still, even masked to hide all facial expressions—it helps disprove the common assumption that counting dogs are responding to subtle trainer cues.
So far, Count Sydney has only miscounted once—he barked three times when TuxnDog had pressed the “two” button. TuxnDog then barked twice herself and pressed “two” again. Count Sydney corrected himself and got it right on the second attempt. But let’s be clear: TuxnDog had no idea whether his original answer was right or wrong. Her response was purely coincidental. That moment is charming, even comical—but it’s not proof of understanding on her part.
The proof is in the structure.
This is a dog—Count Sydney—responding with precision to a numeric English command spoken through pet communication buzzers by another dog, TuxnDog, who cannot count. And all of this is happening while I remain physically still and uninvolved. That’s not a trick. That’s a controlled experiment with video evidence. And yes—it’s also extremely entertaining.
TuxnDog and Count Sydney may be an accidental comedy team. But the implications of their interactions are very real. They raise important questions about canine intelligence, communication, and independent memory recall—especially in inter-dog dynamics.
And the best part? We’re just getting started.