Key Takeaways
Judge a working dog by the partnership, not by supposed magic
A dog needs a good team. Cat Warren expected Solo, her German shepherd puppy, to be a calm companion. Instead, the singleton pup was unruly, intensely scent-focused, and difficult around other dogs. Trainer Nancy Hook gave him a job: searching for human remains. The shift did not turn Solo into a wonder dog. It gave his energy a purpose and taught Warren to handle him differently.
The human matters. A working dog detects things its handler cannot, but the handler must train the dog, read its behavior, account for terrain, and interpret the result. Warren’s central discovery is that a dog’s success depends on the quality of this partnership, not simply on a remarkable nose. The work can be extraordinary, but neither dog nor handler is infallible.
What's striking is how the partnership divides labor without making either partner a mere tool. Solo gathers information through smell; Warren supplies context, judgment, and care. This resembles distributed cognition, where a team solves problems by combining different abilities rather than expecting one mind to do everything. The analogy also clarifies the risk: a handler can misread a signal or overrule it. Calling a dog a teammate is useful only when it comes with accountability for the human's part.
Turn a dog's energy into expertise through rewarding, independent search
Start with motivation. Working-dog trainers call a dog's intense readiness to pursue a task drive. Solo had plenty of it, though Warren initially mistook his frantic energy for bad behavior. His early lessons paired the scent of human remains with a toy reward. The game taught him what smell mattered and gave him a reason to keep searching through difficult terrain.
Make the dog think. Trainer Kevin George used a shell-game exercise to teach dogs not to rely on their handlers' gestures. A handler appeared to hide a toy in one box, then secretly moved it to another. After being fooled, dogs began searching with their noses rather than waiting for a cue. Skill grew through staged challenges, timely rewards, and repetition, not through elaborate equipment or obedience alone.
The exercise is a vivid example of shaping: a trainer arranges small successes and errors so the learner discovers a useful strategy. In education and motor learning, practice becomes more productive when feedback is immediate and the challenge is demanding but manageable. The dogs' independence is not pure autonomy, however. People choose the target, control the reward, and define success. The lesson is less that dogs become self-directed than that good training gives them room to use their own sensory strengths.
Train the nose, not the myth of a universally superior breed
Smell rankings mislead. Popular accounts assign dogs noses ten, a hundred, or thousands of times better than humans'. Warren finds no sound universal ranking. Smell depends on the odor, concentration, task, and experience. In one experiment, blindfolded Berkeley students tracked chocolate scent across grass; on a repeat attempt, they cut their tracking time by nearly two thirds. Practice changed their performance.
Training changes capacity. Research on scent-detection dogs found that trained German shepherds could distinguish between identical twins who lived together and ate the same food, while an earlier test with untrained dogs had found them unable to do so. The comparison is not simply one breed against another, or one species against another. A dog's usefulness comes from the combination of individual ability, training, and a well-defined task.
This is a useful correction to the habit of treating sensory ability as a fixed biological score. A pilot's performance cannot be inferred from an untrained person's flight-simulator attempt, and the same caution applies to dogs. The point also reaches beyond smell: expertise is often a learned way of noticing relevant signals. Still, training does not erase biological differences. Breed, health, and individual temperament matter, but receptor counts alone cannot predict who will do a particular job well.
Use blind tests to catch the handler's expectations before they become evidence
Handlers can cue dogs. In a study described by Warren, researchers placed red paper markers where handlers were told drugs or explosives could be found. No target material was present. Handlers were more likely to report an alert near those markers, suggesting that expectation can shape what people think they see in a dog's behavior.
Make the test fair. In a blind problem, the handler does not know where the target is. In a double-blind problem, the person supervising the search does not know either. These tests reduce the chance of unconscious cues, while careful records make false alerts visible instead of explaining them away. Warren connects this discipline to court cases in which handlers exaggerated dogs' tracking abilities, contributing to unreliable evidence and wrongful convictions.
The core issue is not whether dogs are honest, but whether the observation system is protected from bias. This is a standard challenge in medicine, psychology, and forensic science: expectations can influence both behavior and interpretation. Blind testing is valuable because it does not require handlers to become perfectly objective; it designs around predictable human limits. A single controlled study cannot settle every operational question, but refusing such tests leaves confidence resting on anecdotes rather than evidence.
Practice finding nothing, because most searches end without a recovery
Empty searches are real work. Trainer Andy Rebmann estimates that more than nine out of ten searches find nothing. A negative search means a trained team checks an area and finds no target. It can help investigators rule out a location, even when no dramatic discovery follows.
Train against the expected story. At a seminar, Roy and Suzie Ferguson built two convincing-looking mounds of dirt to resemble graves, then added animal odors to distract the dogs. Several teams alerted on the false clues. June Bug, a young pit bull, did not. In another exercise, handlers searched within marked boundaries while scent came from beyond them. Such scenarios train teams to resist assumptions about where evidence ought to be, and to recognize when a dog is signaling that the search area should change.
This is a lesson in base rates as much as canine skill. When true finds are rare, even a sensitive search can produce many misleading leads if training rewards every suspicious-looking anomaly. Negative results also have value, but only if the search method and coverage are documented well enough to support them. A simulation cannot reproduce every pressure of a real case, yet deliberately practicing uncertainty is better preparation than rehearsing only the satisfying moment of discovery.
Treat scent as a moving plume, not a dot on a map
Scent travels through a landscape. Wind, temperature, humidity, vegetation, water, and ground shape where odor collects. A dog's alert marks where scent is detectable, not necessarily where a body lies. Experienced handler Deborah Palman describes scent gathering against terrain much as debris gathers in a stream; in one search, dogs alerted in several places before a victim was found elsewhere by chance.
Solo's swamp search shows the stakes. A man had fled into flooded woods and was presumed missing. Solo first followed signs of live human scent, then worked into a shaded grove where the air held the smell of a body. He located the victim, while Warren had to interpret the dog's changing behavior and guide the search. The discovery depended on the dog's nose, but also on understanding how scent and terrain interact.
The plume model makes search work an inverse problem: teams observe odor at one location and infer where its source may be. Similar reasoning appears in environmental monitoring, where a pollutant can be measured far from its release point. That means alerts should guide investigation, not dictate excavation or conclusions. The model also explains why a search can miss a body even when a skilled dog is present: scent may not reach the dog, and the team may not cover the right ground.
On water, let the dog's changing signals steer the search boat
Water makes scent three-dimensional. A submerged body releases gases and oils that rise through the water and spread across its surface. Currents, depth, and temperature can carry that scent away from the body, making the search a problem of direction and depth as well as distance. The dog cannot simply run to the strongest odor; it must signal while confined to a moving boat.
A search can narrow a vast area. In one lake case, two dogs alerted at a dock where a woman was last seen. After extensive sonar, diving, and dragging had failed, the dogs' repeated signals helped narrow the search to a small area. A submersible camera found her about thirty feet from the dock and 230 feet underwater. The handlers had to read each dog's behavior and translate it into a route for divers and boat operators.
Water work highlights a general feature of human-machine systems: sensing is not the same as acting. The dog detects; the handler interprets; the boat operator changes position; divers take on the physical risk. Each link can introduce error, so clear signals and coordination are essential. The dogs' contribution is remarkable, but an alert alone does not make a recovery safe. This chapter's strongest implication is that a useful sensor must be integrated into a competent response system.
Recover the dead without hiding the politics of who gets remembered
Body recovery has a history. In the 1970s, researchers and police began developing dogs specifically to find human remains, extending scent work beyond finding the living. The role is not morally simple: dogs have been used to track enslaved people and control populations, as well as to find missing people and victims of disasters. Their purpose reflects human choices, not canine virtue.
The missing include the forgotten. In 2003, Rhode Island trooper Matt Zarrella took two German shepherds to Vietnam to search for American service members missing since the war. At a site connected to Captain David Phillips, the dogs showed interest in an area away from the reported burial. Excavators found artifacts and possible bone fragments there; Phillips's remains were identified in 2004. Dogs have also helped locate possible historic burial sites. Their alerts can direct attention, but records, excavation, and historical context are needed to establish what lies beneath.
Recovery is both forensic work and a decision about whose absence receives public attention. The book's examples connect present-day search practice with slavery, war, and neglected local histories, showing that technical tools operate inside systems of memory and power. A dog's alert can make an overlooked site visible, but it cannot identify a person or tell a community what the discovery means. That distinction protects both scientific rigor and the dignity of people whose remains are being sought.
Keep scent machines as partners, not replacements for working dogs
Technology promises more than it delivers. Researchers have proposed electronic noses, insects, trained bees, and even genetically altered plants as alternatives to detection dogs. Each has limitations: some respond too slowly, fail outside controlled conditions, or cannot move through difficult terrain. The book cites military assessments in which technology found improvised explosive devices at about 50 percent, while dog-handler teams reached about 80 percent.
Combine tools instead. Chemist Arpad Vass developed LABRADOR, an instrument intended to detect odors associated with buried remains. It was designed to complement dogs, investigators, and ground-penetrating radar, not replace them. Dogs are mobile, adaptable, and able to communicate a change in behavior quickly, but their noses are not infallible. Machines can add measurements and consistency; dogs can help locate targets where instruments are less effective.
The useful comparison is not dog versus machine, but system versus system. A machine may offer repeatable measurements; a dog can move through irregular environments and respond to complex mixtures. The book is persuasive about the value of that flexibility, though its cited performance figures should not be treated as universal rankings across every device, setting, and task. Complementarity is a stronger design principle than replacement: different sensors can check one another's blind spots.
Prepare for a working dog's decline before loyalty makes denial easier
Skill does not stop aging. Solo reached his later working years with arthritis and a wobbling leg, even while experience helped him search efficiently. Working dogs can lose physical capacity, and their sense of smell may also change with age or illness. Their careers depend on health, not devotion alone. Warren's difficult searches also leave another uncertainty: a dog may miss scent, and a team may never know unless remains are found elsewhere.
A successor is a new partnership. Trainer Nancy Hook urged Warren to start another dog while Solo could still teach the puppy. Warren chose Coda, a high-drive German shepherd from Kathy Holbert's kennel. The decision meant beginning again, with no guarantee of another dog as capable as Solo. The book closes not with a perfected method but with an aging dog meeting a young one, and a handler accepting both continuity and risk.
This ending resists the fantasy that dedication can suspend biological limits. In care professions, succession is often treated as administration; here it is intimate, involving grief, training, and responsibility to an individual animal. A new dog cannot simply inherit the old dog's expertise, and a handler's experience does not eliminate mistakes. The practical challenge is to plan early enough that retirement protects the dog's welfare without pretending that the bond can be replaced.
Analysis
What the Dog Knows is at once memoir, working-dog primer, history, and an argument for disciplined skepticism. Its narrative engine is the changing relationship between Warren and Solo. The unruly singleton puppy who seems to disrupt the author's life becomes a partner in searches for the dead, and the work reshapes how she sees scent, terrain, training, and grief. The transformation is not that a difficult dog becomes obedient in the ordinary sense. Rather, the handler learns to make useful room for the dog's drive and independent perception.
The book's strongest intellectual move is to hold wonder and uncertainty together. Canine scent detection is impressive, but there is no single number that makes a dog's nose universally superior, no fully settled account of what decomposition smells like to a dog, and no basis for treating alerts as proof. Warren repeatedly makes the handler part of the measurement problem. Her accounts of false alerts, blind searches, and fraudulent testimony show that a dog's ability can be compromised by the human who interprets it. This attention to error makes the praise of working dogs more credible, not less.
The historical material widens the subject beyond a sentimental portrait of the loyal animal. Dogs have helped find children, disaster victims, missing soldiers, and ancient graves, but they have also served coercive and violent human purposes. The question is not whether dogs are good or bad tools. It is who decides how they are used, what evidence their work supports, and which dead receive resources and remembrance.
The evidence is necessarily uneven. The author describes many deployments and experiments, but relatively few controlled studies of cadaver dogs. A memorable recovery demonstrates possibility, not a general success rate. Warren's insistence on training records, realistic scenarios, and corroborating tools is therefore central to the book's value. Its final lesson is not that dogs can solve every mystery. It is that their extraordinary perception becomes dependable only inside a careful, accountable human partnership, one that recognizes both the dog's gifts and its limits.
Review Summary
What the Dog Knows is highly praised for its engaging blend of science, memoir, and history surrounding cadaver dogs. Readers appreciate Warren's well-researched approach, compelling storytelling, and insights into the fascinating world of working dogs. Many found the book both educational and entertaining, with a perfect balance of personal experiences and scientific information. The author's journey with her German Shepherd, Solo, resonated strongly with dog lovers and those interested in canine behavior. While some readers wanted more focus on Solo, overall the book was widely recommended for its unique perspective on an intriguing subject.
People Also Read
Glossary
drive
task-directed motivationIn working-dog training, drive is the intense motivation that keeps a dog pursuing a task or reward. Warren describes Solo's energy as essential raw material, not automatically good behavior. Training channels that motivation toward scent work, search persistence, and clear signals rather than trying to suppress the dog's vitality.
canine ergonomics
dog-environment interactionA term used by cognitive psychologist William Helton for the study of how a working dog functions in its environment. It considers the physical and sensory demands of a task, including balance, movement, scent detection, obstacles, and communication with a handler. The concept treats expertise as learned performance in real working conditions.
final indication
trained target signalThe distinctive behavior a detection dog learns to show when it has located the target odor. Depending on the dog and task, the signal may be a sit, a down, a focused stare, a bark, or another consistent action. Trainers shape an indication that handlers can recognize and describe, then reward it to reinforce accurate alerts.
blind problem
handler unaware of targetA training search in which the handler does not know where the scent target is located. Because the handler cannot knowingly direct the dog toward a hidden sample, the exercise helps reveal whether the dog is locating odor independently or responding to human cues. It is one safeguard against expectation shaping the result.
double-blind problem
trainer and handler unawareA search exercise in which neither the handler nor the person supervising the search knows whether or where a target has been placed. With both people unaware, the opportunity for conscious or unconscious cueing is reduced. The method is particularly valuable when assessing whether a detection dog's alerts are reliable.
negative search
search with no targetA training or operational search in which no target is present in the designated area. The dog is expected to search without producing a false alert. Negative searches help teams practice clearing locations and teach handlers not to assume that every exercise or deployment must end in a find.
Winthrop Point
durable burial landmarkA stable feature in the landscape that may help a person who buried a body remember its location. As described by trainer Lisa Higgins, it should be unlikely to disappear or change, unlike a tree or a road. Recognizing such landmarks can help investigators and handlers reason about where a clandestine burial may be located.
body dog
human-remains search dogAn early term for a trained dog used to locate human remains, including bodies, body parts, and traces associated with decomposition. The work developed from experiments in scent detection and became a practical tool for police investigations and disaster recovery. The book also uses the later terms cadaver dog and human-remains detection dog.
sagdid
Zoroastrian dog ritualA Zoroastrian funerary rite whose name means seen by the dog. A specially selected dog was brought to gaze upon a deceased person, a ritual understood to help repel forces associated with death and putrefaction. Warren discusses the practice as an example of the long and varied cultural relationship between dogs and the dead.
LABRADOR
buried-remains odor analyzerAn acronym for Lightweight Analyzer for Buried Remains and Decomposition Odor Recognition. Developed by forensic scientist Arpad Vass and colleagues, the instrument was intended to detect odors associated with buried human remains. It was presented as a complementary tool for investigators, alongside dogs, ground-penetrating radar, and other methods.
Clever Hans
unconscious cueing exampleA German horse famous for apparently solving arithmetic questions by tapping his hoof. Psychologist Oskar Pfungst showed that Hans responded to subtle cues from people who knew the answer. Warren uses the case to explain how handlers' expectations can shape a dog's behavior or their interpretation of an alert, even without deliberate deception.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.