By Sam Basso
OVERVIEW: “Resource guarding” in dogs has been used to describe behavior around something a dog may value. But, there are problems with this term. For example, the way it is tossed around does not identify the cause or tell us how to treat “it”. To be fair to a dog, an assessment must examine an entire interaction, not lump it into a large basket of different causes and effects, and give it all one name.
I don’t diagnose dogs as “resource guarders,” and I think other dog trainers should stop using that label as though it were a diagnosis or a treatment plan. It tells us that a dog did something around something supposedly valuable. It does not tell us why the dog did it, what the other individual did, or what needs to change.
That distinction matters. When people believe they have named a problem, they may stop investigating it.
What is wrong with calling a dog a resource guarder?
The term gathers too many potentially different behaviors under one heading. A dog growling beside a food bowl, moving away with a stolen object, protesting when someone reaches for its collar on a couch, and threatening another dog near a person may all be described as “resource guarding.” But the apparent similarity does not demonstrate a shared cause.
There is a second problem with the word resource. Food, toys, beds, furniture, people, and even access to a doorway can be called resources. If almost anything can qualify, finding a resource near an incident tells us very little. It can become circular: the object is assumed to be valuable because the dog reacted, and the reaction is then explained as the dog protecting something valuable.
Researchers have wrestled with this problem. In a 2018 study, Jacobs and colleagues asked canine behavior experts how to define resource guarding. Most preferred that name, but they did not agree completely on its scope. Their proposed definition concerned retaining control of food and other physical items; the researchers left people and larger areas out because those situations might involve different processes (Jacobs et al., 2018).
That is a useful descriptive definition for research. It is not a diagnosis of an individual dog’s emotional state, motivation, or prognosis. I would rather describe the behavior precisely than attach a label that seems to explain more than we know.
If a dog guards food, will it also guard the couch or its owner?
No. A response in one setting does not establish that the same response will occur with every item, individual, or place. In fact, if a dog growls around a particular chew but shares other objects peacefully, the difference is evidence we ought to investigate, not an inconvenience to ignore.
The same caution applies when someone tells me a dog “guards its owner.” Did the dog prevent another dog from approaching the person? Did the dogs collide while rushing to greet someone? Was the dog reacting to being crowded, restrained, or touched? Did the other dog threaten first? An owner standing nearby is an observation. Guarding the owner is an interpretation.
I am not saying dogs cannot compete over food or defend possession of something. Of course they can. I am saying that naming the object does not identify the behavioral mechanism. Competition for access, retaining possession, resisting unwanted contact, and responding to perceived danger are different questions.
What if the dog growls when someone approaches its place on the couch?
A couch incident can have several explanations, and the fact that the dog likes the couch does not decide among them. Perhaps the dog wants to remain in a comfortable resting spot. Perhaps it anticipates being grabbed or pushed off. Perhaps the approaching dog blocks its way out. Pain, disturbed sleep, previous confrontations, and learned expectations might also matter.
Think about a person sitting down to a meal. If a stranger interferes with the plate, the response could be annoyance at the intrusion rather than fear of losing the food. That human comparison helps expose an assumption, but it cannot tell us how a particular dog feels. Nor does the opposite assumption, that every growl is motivated by fear, fare any better.
What I would ask is exactly what preceded the growl. Was the dog resting? Did someone lean over it, reach for its collar, or insist it move? Could it leave? What happened when the person or other dog stopped approaching? Were there uneventful encounters at the same spot? Without those details, “couch guarding” risks becoming the answer before the examination has begun.
Why does the other dog or person matter so much?
Because aggression is often part of an interaction, not an isolated performance. Suppose Dog A has a chew and growls when Dog B comes near. If B backs away and A resumes chewing, we have observed one sequence. If B keeps approaching and A retreats with the chew, we have observed another. If B withdraws but A pursues and attacks, the sequence is different again.
The behavior of B changes the encounter, sometimes dramatically. Humans do the same. A person may withdraw, take the item, restrain the dog, punish the growl, offer a trade, or continue moving closer. Those actions can alter what happens immediately and may influence later encounters. But seeing one outcome does not establish what changed the response, or what emotion produced it.
This is why I want both sides of the interaction recorded: who approached, who moved away, what signals appeared, whether contact occurred, and whether the confrontation ended after withdrawal or continued into pursuit. I also want to know what happened in similar situations when nobody fought. A history consisting only of the worst incidents can be misleading.
Is the behavior defensive, frustrated, competitive, or something else?
It could involve any of these processes, singly or together, and we should not pretend the label decides the matter. A dog trying to obtain a toy held by another dog is in a different position from a dog already holding the toy. A dog resisting interference is not necessarily experiencing the same thing as a dog confronting an animal it perceives as dangerous. A dog trying to escape uncomfortable handling may have no interest in keeping the couch at all.
A 2026 scholarly paper by Wilson, Soulsbury, and Mills argues that canine behavior science needs clearer distinctions between fear and frustration. The authors propose that interference with obtaining or retaining a desired outcome may involve frustration rather than automatically being described as fear. That is an important theoretical proposal, not a test that can tell us what an individual dog was feeling at the time (Wilson et al., 2026).
Learning research also complicates the picture. In a detailed functional analysis of one pet dog displaying human-directed food guarding, Mehrkam and colleagues found that the behavior was maintained by more than one type of consequence, involving escape, attention, and access to items. The investigators designed and tested different treatments for those functions (Mehrkam et al., 2020). We cannot transfer that case’s findings to every dog, but it demonstrates why a label cannot replace an analysis.
Even research outside dogs gives us a reason for caution. In a classic experiment, pigeons sometimes attacked another pigeon when expected food reinforcement stopped. The study did not require an animal to be defending a bowl or toy to produce aggression (Azrin et al., 1966). It does not establish the cause of canine aggression; it illustrates why aggression around rewards need not have one explanation.
Could the dog’s response change as another dog grows up?
Yes. Social encounters have histories, and development can change them. In research on free-ranging mother dogs, food competition with puppies changed across weaning and differed with the kind of food offered and maternal condition (Paul et al., 2015). These findings concern mother-offspring interactions in a particular population, not a universal rule about pet dogs or adult males.
Still, they expose another weakness of the label. An adult may permit a very young puppy to feed nearby and respond differently when that puppy is older and more persistent. The food may be the same. The puppy’s behavior, the relationship, and the history of encounters are not. We should not immediately diagnose a newly revealed “guarding problem,” nor claim that the adult intended to teach a social lesson. We need to observe what both animals did and what followed.
Is there a passive form and an active form?
Possibly, but I would not announce two new diagnoses on that basis. There are at least two separate questions: did the dog initiate the encounter or respond after an approach, and did it accommodate, avoid, threaten, displace, or pursue the other individual?
A dog that moves away with a chew is actively moving, but may be avoiding confrontation. A dog that stands motionless is not necessarily relaxed. A dog that growls while remaining on a bed may be responding to another dog’s approach. A dog that leaves the bed to displace another dog has initiated a different interaction. None of those observations alone establishes confidence, fear, or a fixed personality trait.
The active-versus-passive idea is worth investigating, but the distinctions must be defined before they can be tested. Otherwise we will simply replace one vague label with two more.
Why does the label lead people toward the wrong treatment?
Because a generic label invites a generic remedy. People may be told to make the dog surrender its bowl, demonstrate who is in charge, punish a growl, or repeatedly approach with treats, without first determining what is happening. Some training procedures can be useful under carefully chosen conditions. None is appropriate simply because someone said “resource guarding.”
Does a dog that supposedly “guards” something mean its owner has failed to become a strong leader or establish themselves as the “alpha”? No. That conclusion substitutes one unproven explanation for another. A dog that growls when someone approaches its food, takes away a toy, or tries to move it off the couch is not necessarily challenging anyone’s authority or attempting to achieve a higher social rank. The behavior may involve competition, frustration, anticipated loss, unwanted handling, pain, previous learning, or something we have not yet identified. Dominance has a legitimate, more specific meaning in animal behavior research, but it does not explain every conflict between dogs and people. Worse, telling an owner to demonstrate leadership by forcibly taking possessions, confronting the dog, or punishing its warnings may create additional, or even dangerous, problems without addressing the original cause. As I explained in my article, Does Your Dog Need a Strong Leader? What Science Says About Leadership, Obedience, and Behavior, responsible leadership involves guidance, appropriate boundaries, effective teaching, and attention to the dog’s welfare, not proving who is boss.
If the problem involves painful handling, the first priority may be veterinary assessment and changing the way the dog is approached. If two dogs repeatedly collide in a narrow passage, managing movement and access may be more relevant than working around a food dish. If a dog fears losing a chew, predictable exchanges may eventually be helpful, but the dog’s responses and the risk of injury should guide the plan. If a dog pursues another after it retreats, that requires particular caution; ending the initial contest has not ended the danger.
Above all, do not provoke an encounter to see whether the dog is “really guarding.” Do not punish warning signals to make the dog appear quiet. When bites, escalating threats, children, or repeated dog fights are involved, use safe separation and qualified veterinary and behavior assessment. Preventing another injury is a valid immediate result, but it does not by itself prove the underlying relationship is repaired.
What should we say instead of “resource guarding”?
Describe what was seen, and keep the explanation separate. “The dog stiffened and growled when another dog approached its food bowl” tells me more than “He’s a resource guarder.” So does “She moved away with a chew when someone came near,” or “He snapped when a person reached for his collar while he was resting.” Those descriptions are useful even if we still do not know why the behavior happened.
Then ask what the other participant did, whether the dog could withdraw, what happened next, and how the interaction differs in other circumstances. The goal is to discover a workable and humane way forward, not to win an argument over vocabulary.
I am asking researchers to erase the term from published studies. When a study defines a term, it is appropriate to start reporting that definition accurately to explain what is happening in that study. What I want professionals to stop doing is presenting “resource guarding” as if it were an explanation, a diagnosis of the dog, or an instruction manual for treatment. It is otherwise no better than using the term, “reactive”, which also is tossed around and practically meaningless for any kind of serious research or application.
If your dog has growled, snapped, or bitten around food, toys, a resting place, or another individual, begin with safety and a careful account of the encounter. Tell me what happened between the participants. Do not tell me the label and assume the work is done.
Glossary
Resource guarding: A useless research and training term for certain behaviors associated with retaining access to items; definitions vary, and the term alone does not establish a cause.
Antecedent: A condition or event occurring before a behavior, relevant to examining when it occurs.
Behavioral sequence: The order of actions by all participants before, during, and after an encounter.
Competition: Interactions in which individuals seek access to an opportunity that cannot be used equally or simultaneously under the circumstances.
Possession retention: Continued access to an object or opportunity already held; an observational category, not an emotion.
Defensive behavior: Behavior associated with responding to an actual or perceived threat; it does not establish a particular conscious feeling.
Frustration: A proposed emotional process associated with blocked or unsuccessful motivated activity; its presence cannot be read directly from a bite or growl.
Operant function: The effect of consequences on the future occurrence of behavior under specified conditions; identifying it requires more than noticing what happened once.
Disengagement: A participant ending or reducing its involvement in an interaction, for example by moving away.
Generalization: Occurrence of a learned or established response in circumstances different from those in which it previously occurred.
Scholarly bibliography
Azrin, N. H., Hutchinson, R. R., & Hake, D. F. (1966). Extinction-induced aggression. Journal of the Experimental Analysis of Behavior, 9(3), 191-204. https://doi.org/10.1901/jeab.1966.9-191
Jacobs, J. A., Coe, J. B., Widowski, T. M., Pearl, D. L., & Niel, L. (2018). Defining and clarifying the terms canine possessive aggression and resource guarding: A study of expert opinion. Frontiers in Veterinary Science, 5, 115. https://doi.org/10.3389/fvets.2018.00115
Mehrkam, L. R., Perez, B. C., Self, V. N., Vollmer, T. R., & Dorey, N. R. (2020). Functional analysis and operant treatment of food guarding in a pet dog. Journal of Applied Behavior Analysis, 53(4), 2139-2150. https://doi.org/10.1002/jaba.720
Paul, M., Sen Majumder, S., Nandi, A. K., & Bhadra, A. (2015). Selfish mothers indeed! Resource-dependent conflict over extended parental care in free-ranging dogs. Royal Society Open Science, 2(12), 150580. https://doi.org/10.1098/rsos.150580
Wilson, B. M., Soulsbury, C. D., & Mills, D. S. (2026). Development of a psychobiological nosology for fear and frustration in clinical animal behavior using dogs as a model species. Frontiers in Psychology, 17, 1888439. https://doi.org/10.3389/fpsyg.2026.1888439