When my dog, Fauci, wants to go out to play, he drops a toy at my foot and paws my knee. Stella, a three-year-old Catahoula Blue heeler mix in Chicago, is more articulate. When she's ready to run, she says, "Park, outside, come."

Stella's no Martha who Speaks or Scooby-Doo. She's a real, living animal — and she "talks" by pawing buttons that activate prerecorded words.

Stella is an internet sensation. (No surprise.) She has nearly 800,000 followers on Instagram and is now the star of a book, "How Stella Learned to Talk," written by her owner, Christina Hunger. Online, millions of viewers can — and do — watch Stella type out sentences and "converse" with Hunger, a 27-year-old speech therapist. Stella can now tap 48 words and a variety of phrases. She has become both an inspiration for dog lovers and the subject of debate about how much canines really comprehend. Stella's accomplishments also raise questions about how we define speech.

Given the title of her book, it's clear where Hunger stands on the matter. Hunger, whose day job is working with children with language development delays, is a passionate advocate of Augmentative and Alternative Communication (AAC), expression through nonverbal means (think of Stephen Hawking's eye-activated device), and she used her experience with children to teach Stella.

Early in Stella's puppyhood, Hunger said in an interview, she noticed "these strikingly similar language milestones that kids show right before they are able to say words, and I had this lightbulb moment, 'Wait a second. Dogs can understand words, shouldn't they have an opportunity to say words, just in a different way — not with verbal speech?'" So she set out using the same tools she used with kids — modeling words by, for example, narrating an action ("look outside!") and then pushing a button on a communication board that said the word "outside."

She repeated the process for a few hours a day, until little by little, Stella began associating an action with a word. It took about a month for Stella to learn her first word: "outside."Over time, Stella began to put words together in phrases, even tapping out emotions like "love" and "mad." Hunger believes she is the first person to teach a dog to "talk" using augmentative communication.

Hunger's book does not read as scholarship but rather as the memoir of an experiment and a relationship. It also includes tips, and if you're serious, you can buy communication-board starter sets on her website (also T-shirts that say "Believe in Potential"); a portion of the proceeds will go toward the purchase of communication devices for children.

Q: Congratulations on teaching Stella to "talk." Some people might say that barking and other canine behaviors are also speech. What are your thoughts on that?

A: That's definitely a form of language. Communicating with gestures and with vocalizations like barking are all language milestones that happen before words. That's something many people wouldn't necessarily know unless they're in the field of speech therapy: There are so many language milestones that happen before words, and there are so many ways to express language — you can use a gesture to express a concept, you can write down a word, you can verbally say a word. It's all still language.

Q: That raises a kind of philosophical question: What is speech?

A: In my field, I would say verbal speech is how you and I are communicating now. We say talking is any way that someone is using words — it could be talking through a communication device, talking through sign language, talking through verbal speech.

Q: Are you working with cognitive scientists or animal behaviorists?

A: I'm not personally working with anyone, but research has begun because of what I've done with Stella. It's amazing to see all these other professionals interested and just taking what I've done and running with it.

Q: One of the more intriguing aspects of Stella's "speech" is her expression of feelings and emotions. How did you get her to do that?

A: I decided to add "happy" and "mad" when I noticed how she was communicating her emotions nonverbally. When I saw how strongly Stella had these nonverbal communications — whining, twirling her tail — for these emotional states, all I did was add a word to it, so anytime she was showing those emotions through her body gestures, I would model a word. When wagging her tail and smiling, I would model the word "happy, happy" — to put a word to the gestures she was already showing.

Q: How do we know what Stella really knows?

A: It's the same with children as they are developing language. It's all through patterns that kids use in their language and in their speech and knowing the context and the environment. When a toddler says, "Daddy monkey!" Does it mean, "Daddy, look at the monkey"? Or: "Daddy, let's sing the monkey song"? Or: "This is the daddy monkey"? You don't know unless you know the child and understand what the child has seen and heard through the day and how you interact together. It takes a lot of knowing the situation and knowing the patterns of communication.

Q: What are some of the misconceptions about your work?

A: I think one is that since I am giving Stella a way to say words I am taking away from her natural forms of expression. But that's just completely the opposite of what's happening. Using the words is just one of the ways for her to express herself. She is incredibly expressive through all of her natural forms — and now in words, too. She also expresses herself through, for instance, barking, tail wagging, pawing at things and turning her ears. It's really interesting, too, because with humans, research has shown that using AAC devices doesn't take away from verbal speech development, if anything it helps verbal speech development. People were once scared that introducing an AAC device would keep kids from talking. But we've found that when kids have success communicating in one area it goes to all areas. And that's something I'm seeing with Stella. She has so much success and so many ways to communicate, that all forms of communication are amplified.

Q: One critique I have seen is that Stella might be pressing buttons because she knows something good might happen if she presses the "right" one. How do you respond to that?

A: I've never given her any treats for pressing buttons. My only response is to show her the natural meaning of the word she presses. It goes back to understanding her patterns. When she says "eat" in the morning around her mealtime and in the evening again at mealtime, that makes sense — she's not using it constantly, randomly throughout the day.

Q: What would you advise for dog owners who want to try this at home?

A: The first step is noticing how your dog is already communicating. So when your dog is gesturing for something, you can narrate what they're gesturing for with the word. Before introducing buttons, dogs need to hear words and understand them before they "say" them.