Verified by 
"Joe recognizes another dog from afar and knows if they want to play or are not that playful and might be aggressive."
"Jasmine 'sees' other dogs from more than 200 feet and either tucks her tail as if she's submitting or runs toward them so she can play with them. She doesn't do this to any other animals and seems to know that whoever is out there is a dog and they're unfriendly or friendly."
Dogs seem to be able to recognize another dog at distances that cannot be accounted for what is known about their senses of smell, sight, or hearing. Their ability to do so remains a mystery, but I've seen this and have heard many stories, such as the ones above and below, about this cognitive capacity about which we know next to nothing.
Whenever I go to dog parks or interact with humans and their dogs at other locations, people often tell me something like, "My dog can recognize another dog at a distance of at least 100 feet" or "Harry knows when his best friend, Mallory, is here, across the dog park, but doesn't respond to any other dogs" (a distance of about 200 feet), or "Joe recognizes another dog from afar and knows if they want to play or are not that playful and might be aggressive." I've seen dogs who are sniffing here and there stop, look across a dog park at a distance of around 150 feet, and either run toward a dog who they don't know, clearly wanting to play, or, on the other hand, see an image at about the same distance and continue sniffing or walk in the opposite direction as if they've seen an other individual who might cause them harm. I honestly have no idea what's happening and that's why the question, "Do dogs recognize 'dog' and perhaps what they're feeling from afar?" is so intriguing.
The ethology of dogs' senses: Smell, sight, and sound
I'm writing this brief essay to get a discussion going about how dogs might be able to know "dog" and possibly assess what they're feeling from afar. Here are some quick facts about the senses of smell, sight, and hearing in dogs that might be used in this sort of discrimination. Dogs sense their world through smell, sight, sound, taste, and touch. Of course, like other animals, including humans, dogs often process a cocktail of stimuli coming in simultaneously and sequentially. Ethologists call these composite signals, and they usually contain more information than signals in a single sensory modality. The emerging and changing cacophony of sensory stimuli allows dogs to gather a good deal of detailed information about what’s happening in the moment. It might even tell them what happened in the past and what is likely to occur in the future. This information is vital in order for them to figure out what to do in any given situation.
Dogs' noses: Let's begin with dogs' noses, which are works of art. Dogs often like to sniff first and ask questions later. In Being a Dog: Following the Dog Into a World of Smell, Dr. Alexandra Horowitz calls dogs “nosed animals" — or, as Dr. Frans de Waal puts it, “a nose with a body attached" — and researchers refer to dogs as macrosmatic mammals because smell is so important, really essential, to their way of life. I always think that a dog without a working nose isn’t a dog. In fact, dogs also have what’s called a vomeronasal organ, also called Jacobson’s organ, that functions as a second nose. This is part of a dog’s accessory olfactory system, and it responds to stimuli that are liquid rather than volatile vapors.
You can read about a review of dogs' sense of smell in an interview I conducted with Dr. Frank Rosell, author of Secrets of the Snout: The Dog’s Incredible Nose, in an essay called "Secrets of the Snout: A Dog's Nose Is a Work of Art." I also review what's known about dogs' noses and other sense organs in Canine Confidential: Why Dogs Do What They Do. While dogs' noses are incredibly sensitive, estimated to be between 100,000 and 100 million times more sensitive than humans' noses, the distances at which dogs seem to be able to recognize "dog" and perhaps what they're feeling goes well beyond their olfactory acuity, even if they're downwind from another individual.
Dogs' eyes: Dogs clearly have a keen and highly evolved sense of smell. They also have a good set of eyes, which are also important for negotiating their social and physical worlds. I’m sure I’m not alone in having been stared down by a dog who has locked eyes and won’t let go. Dogs are not the only nonhumans to look people in the eye. I’ve also had similar stare downs with wild coyotes, black bears, and cougars around my mountain home.
People also are constantly telling me that their dog can “read” — visually discriminate — other dogs from afar and seem to be able to make reliable, long-distance assessments about whether another dog is friendly, wants to play, or is saying back off. However, dogs have a visual acuity of around 20/75 (called the Snellen fraction), which means that while we can see something from seventy-five feet away, a dog can only see it twenty feet. They’d do well to wear glasses! Using this method of measuring acuity, dogs have worse vision than humans. (For more discussion please see Dr. Stanley Coren's essay "Compared to Humans, How Good Is a Dog's Visual Acuity?" in which he concludes "Humans have better visual acuity than dogs under most light conditions.") However, it would be wrong to claim that dogs don’t see as well as humans, because the Snellen fraction only provides one small window into the larger sense of sight. For example, dogs are better adapted than humans for identifying movement in their peripheral vision. It would be more accurate to say that dogs and humans see the world differently, and their visual acuity evolved to meet their needs. Different doesn’t necessarily mean better or worse.
A few studies show that dogs can recognize photographs of other dogs. C. Claiborne Ray, discussing a study done by Dominique Autier-Dérian and her colleagues called "Visual discrimination of species in dogs (Canis familiaris)" says: “Ranging in size from a tiny Maltese to a giant St. Bernard, and showing myriad differences in coats, snouts, ears, tails and bone structure, dogs might not always appear to belong to one species. Yet other dogs recognize them easily, even in the absence of clues like odor, movement and vocalizations.” This study looked at the ability of nine dogs to discriminate two-dimensional pictures that were presented to them using a simultaneous discrimination paradigm. Nonetheless, as with dogs' sense of smell, their ability to recognize "dog" and perhaps mood goes well beyond their visual acuity when one considers the distances at which some dogs seem to recognize "dog."1 While tails might be important in detecting "dog" and perhaps mood, many stories I've heard about the ability of a dog to recognize "dog" at a distance include dogs with short tails or with no visible tail other than a stub.
Dogs' ears: Turning now to dogs' ears, it's well known that they come in a wide variety of shapes and sizes: from long and floppy to short and erect. But whatever the shape, they hear sounds of which humans are totally unaware. Their ears are rather mobile and capable of turret-like movements, which allow them to more precisely locate a sound. Depending on the breed and age, dogs can hear in frequencies ranging from around forty to sixty thousand hertz (one hertz equals one cycle per second). Humans can hear sounds around twelve to twenty thousand hertz. Dog whistles produce a sound that’s usually in the range of twenty-three to fifty-four thousand hertz. Dogs have more than eighteen muscles that control their flexible pinna (the external part of the ears).
Overall, dogs perceive frequencies approximately twice that of humans, and it’s been widely reported that they can detect and distinguish sounds about four times as far as humans. This means that what a human can hear at twenty feet a dog can hear at roughly eighty feet. However, even given dogs' highly evolved sense of hearing, their ability to recognize "dog" and perhaps mood, goes well beyond their auditory acuity when one considers the distances at which dogs recognize "dog," even in the presence of a great deal of ambient noise from other dogs, humans, and traffic, for example. For this discussion, I'm not considering situations where a dog is barking loudly and can be heard for hundreds of feet.
Do dogs know there's a dog "out there" and read their mood from afar? The possible role of composite signals and a "dog gestalt"
Given that each of the three senses that dogs might use to detect "dog" at a distance alone don't readily account for their apparent ability to do so, it's possible that composite signals consisting of input from smell, sight, and sound, or from some combination other than those including loud barking, might help them discriminate "dog" and perhaps mood. However, this doesn't seem likely under the conditions in which it's been reported, "My dog knows 'dog' from afar."
Given what we now know about the sensory systems of dogs and the information that might be used for them to be able to know "dog" and perhaps assess mood from afar, I wonder if there's some sort of "dog-gestalt" that dogs are able to use, a composite signal perhaps, but we don't have any data that support this possibility.
Along these lines, some people have suggested that gait is a key variable, so sight might be paramount in combination with input from other senses. While I also lean toward sight being a key variable, we don't know if this is so. A challenge for future research with dogs is to learn not only how each sense works on its own, but also how dogs combine and use the input from multiple senses -- how they use composite signals -- to understand the world and to make the best decisions they can. For instance, one study by dog researcher Ludwig Huber showed that captive dogs could integrate information from sight and sound to identify other dog breeds correctly. In this study, dogs matched a projected visual image of dogs of different sizes with the vocalization that is usually made by dogs of each size.
Where to from here?
I wrote this essay to get the ball rolling on generating more systematic and controlled research focused on the apparent ability for some dogs to know there's a dog "out there" and possibly to read their mood from afar. "Out there" refers to distances beyond which their senses of smell, sight, and hearing are thought to work. There really are two parts to studying this cognitive capacity, the first being the ability to recognize "dog" and the second being the ability to assess their mood. While scientific data are lacking for either ability, many anecdotes and a good deal of citizen science strongly suggest that some dogs are able to recognize "dog" and in some cases to assess the dog's intentions (for more discussion please see "Citizen Science as a New Tool in Dog Cognition Research"). So, it's high time to do some detailed research on these questions.
In addition to collecting data with real dogs, another way to study whether and how dogs recognize "dog" would be to use a "robotic" dog on which its body morphology and tail, ears, and gait can be varied alone or in combination with one another to assess the possibility that composite signals are important in recognizing "dog" and perhaps mood. These studies could shed some light on whether or not there's a generalized "dog gestalt," for example, and which variables are important in recognizing "dog" from afar. Where there's a will there's a way, and there are good reasons to pursue these investigations.
It's surely an exciting time to study the cognitive and emotional lives of dogs and other animals. I look forward to more discussions on this aspect of dog behavior and many others for which systematic data are lacking. Please stay tuned for such exchanges.
Note
1Many people report that, on their dogs’ first encounter with other dogs, members of the same breed prefer one another and that dogs treat breed members differently from individuals of other breeds. Is this odor based, as is kin recognition in some rodents? While dogs know what they themselves smell like, they don’t necessarily know what they look like -- or might they? Research done on birds in the 1960s suggests that they might learn their own color from reflections in water.
