Attachment, the Distress Triangle, and the Numbers

What Attachment Actually Means in a Dog

There is a folk theory, widespread among dog people, that separation anxiety is what happens when a dog is too attached to her person. The reasoning is intuitive: if your dog follows you everywhere, sleeps on your feet, can't bear to lose sight of you, and then loses her mind when you leave, the too-attached and the can't-handle-leaving must be different views of the same underlying problem. The intervention that follows from this folk theory is to make the dog less attached, to ignore her, to refuse to greet her, to enforce distance in the house, to send her to her bed when she tries to settle near you. The folk theory is wrong. The interventions that follow from it range from useless to actively harmful. To explain why requires a short detour through the science of attachment.

The framework was originally developed in the 1960s and 70s by John Bowlby and Mary Ainsworth to study the bond between human infants and their primary caregivers. The most famous tool to come out of that work is the Strange Situation Procedure (SSP), which Ainsworth designed in 1969 to classify how a one-year-old child responds to short separations from her mother. The SSP became the canonical assessment of attachment style in developmental psychology, secure, insecure-avoidant, insecure-ambivalent, and a fourth category, disorganized, added later.

In 1998, a Hungarian research group led by József Topál, Ádám Miklósi, Vilmos Csányi, and Antal Dóka at Eötvös Loránd University published a paper in the Journal of Comparative Psychology applying a modified SSP to fifty-one owner-dog pairs. They asked whether the framework that Ainsworth had built for human infants could meaningfully describe the bond between an adult dog and her owner. The answer was yes. Dogs exposed to a mild laboratory stressor in an unfamiliar room with an unfamiliar person reliably directed their proximity-seeking, contact-seeking, and stress-reduction behavior toward their owner, used the owner as a base from which to explore the room, and showed measurable physiological distress upon brief separation that resolved on reunion. The behavioral pattern was structurally analogous to what Ainsworth's infants did. The Topál 1998 paper is the foundational document of canine attachment science. Every modern study of dog-human bonding traces back to it.

Subsequent work has refined the picture. Solomon and colleagues 2019, in the journal Attachment & Human Development, classified fifty-nine pet dogs into Ainsworth's secure/insecure categories and found a distribution remarkably close to the human-toddler norm: roughly sixty per cent secure, the remainder distributed across the insecure categories. A 2026 Frontiers in Veterinary Science paper observing 180-plus dog-caregiver dyads found a similar pattern: about sixty per cent secure, twenty per cent ambivalent, seventeen per cent avoidant, three per cent disorganized. The Italian group at Pisa, the Hungarian group at ELTE, and the American work led by Solomon have converged on the same conclusion across two decades of independent work. Dogs form attachments to their humans that share the operational features and the statistical distribution of the bonds human children form with their parents. The Lagotto temperament, proximity-seeking, attention-tracking, return-on-call orientation, is, in attachment terms, the expression of a very securely-attached dog in an extreme breed. It is not pathology. It is competence.

The crucial paper, for our purposes, is Parthasarathy and Crowell-Davis 2006, published in the Journal of Veterinary Behavior. They compared thirty-two dogs with diagnosed separation anxiety against forty-three dogs without, looking at how much time each dog spent in close proximity to her owner during an attachment test. If the folk theory were correct, if separation anxiety were the consequence of more attachment behavior. The anxious dogs should have stayed closer. They did not. There was no significant difference in proximity-seeking between the two groups. Dogs with separation anxiety did not exhibit more attachment behavior during the test; they exhibited a different style of attachment, and they reacted differently when the owner left the room. The pathology is not in the strength of the bond. It is in the quality of the attachment relationship and the dog's ability to regulate her own arousal during the owner's absence.

What does this mean for the family raising a velcro-bred Lagotto?

It means, first, that the very ordinary Lagotto-following-you-around behavior you are seeing is most likely the expression of secure attachment, not impending pathology. Sixty per cent of the population sits there, in the secure-base zone, with their preferred distance from their person measured in inches. Your puppy following you to the bathroom is doing what well-attached human one-year-olds do at the playground when their mother is the parent on the bench: she is checking, briefly, where her person is, and then resuming what she was doing. This is the system working as designed.

Second, it means that the path away from separation problems is not the path of reducing the attachment. It is the path of building security within the attachment. Securely-attached dogs and securely-attached children both share a feature that anxiously-attached ones do not: they can be apart from their primary person without losing their composure, because they know that person reliably comes back. Reliability of reunion is the variable. A child who has been parented by an attuned adult develops the internal expectation that absences end. A puppy who is brought into a household that practices calm, sub-threshold separation from the first week develops, in functionally similar form, the same expectation. The Sixteen-Week Absence Staircase in Section XII is, in attachment-theory terms, a structured exposure program designed to build secure attachment, not to reduce attachment. The whole protocol works because the bond between you and your puppy is real.

Third, it means that the rhetoric of "ignoring your dog to make her independent" is exactly the wrong intervention. An anxiously-attached dog needs more reliable presence, not less. She needs you to come back when you said you would, every time, until she stops needing to wonder whether you will. The dog who has been emotionally rationed by a well-meaning owner trying to make her tougher is the dog you will see clinging hardest at thirty months. Security is built; it is not refused into existence.

The Lagotto temperament is not pathology. It is competence. The pathology is not in the strength of the bond. It is in the dog's ability to regulate her own arousal during the owner's absence.

The Distress Triangle: Boredom, Frustration, Panic

One reason the welfare literature has been slow to progress on this topic is that three very different underlying states can produce almost identical surface behaviors. A dog who chews a pillow while the family is at work may be understimulated, she may be confinement-frustrated, or she may be in full physiological panic. These three conditions have different causes, different timelines, different welfare implications, and completely different interventions. Mistaking one for another is the single most common diagnostic error made by well-meaning owners and, frankly, by general-practice veterinarians who have not trained specifically in behavior.

Here is the distinction, laid out plainly.

The clinical and practical difference between these three states is enormous, and the interventions are almost opposite. A bored dog needs more enrichment. A frustrated dog needs patient, stepwise habituation to confinement without being rescued in the middle of a tantrum. A dog in panic needs the absence itself to stop happening at its current duration, immediately, while a proper desensitization program is begun from a sub-threshold starting point, often with veterinary-behaviorist oversight and, in more severe cases, medication to break the conditioning cycle long enough for learning to take place.

If you try to treat panic with enrichment toys, you will fail, and you may make it worse. If you try to treat boredom with anxiety medication, you will also fail. The first diagnostic step in anything that looks like a separation problem is asking which of these three categories you are actually dealing with. Video evidence helps. A baby-monitor camera pointed at the dog while you are out of the house for thirty minutes will tell you almost everything you need to know. The behavior itself looks similar across the three. The physiology and the trajectory over time do not.

The Numbers: Salonen, Flannigan, Generation Pup, and the Pandemic

The best large-scale dataset on pet-dog behavior anywhere in the world was collected by Hannes Lohi's research group at the University of Helsinki and published in Scientific Reports in 2020 by Milla Salonen and colleagues. The study surveyed the owners of 13,715 Finnish pet dogs across 264 breeds about seven categories of anxiety-related behavior: noise sensitivity, fearfulness, fear of surfaces and heights, inattention and impulsivity, compulsion, separation-related behavior, and aggression. For anyone trying to understand how common any of these traits actually is, the Salonen dataset is the closest thing we have to a reliable baseline.

Here is what it found.

The pandemic cohort changed the picture. In the Royal Veterinary College's longitudinal study of 985 dogs acquired as puppies during UK lockdowns in 2020, followed up at 21 months of age, 31% displayed owner-reported separation-related behavior, a more than six-fold increase over Salonen's pre-pandemic baseline. The same study found that 97% of these dogs exhibited at least one problem behavior, and 82% of owners had tried to address problems using aversive training methods (physical punishment, shock, intimidation) of the kind that a substantial body of research has shown to be worse than useless and sometimes to create new behavioral symptoms.Brand 2024

These two data points, 5% pre-pandemic baseline and 31% in the pandemic cohort, are, in my view, the most important numbers in this whole essay. They tell you two things. First, separation-related behavior is not a rare outlier problem. Even at baseline, it affects one dog in twenty. It is, and has long been, among the most common behavior complaints presented to general-practice veterinary clinics and a leading reason for dogs being surrendered to rescue organizations in their first two years of life.

Second, and more consequentially: it is spectacularly sensitive to environmental conditions. When the pandemic confined dogs and owners in the same small spaces for twelve to eighteen months, and those owners then returned to offices, the rate of observable separation-related behavior in the affected cohort jumped by more than six-fold. This is not a subtle effect. It is a screaming signal that the environmental inputs during a dog's early life, specifically, how much independent alone-time she experiences in the first four months, determine the adult outcome to a degree almost no other behavioral variable does.

The Generation Pup Findings

A second longitudinal study, smaller than the RVC Pandemic Puppies but more focused on the early-life developmental window, published its first major findings in late 2024. Fiona Dale, Charlotte Burn, Jane Murray, and Rachel Casey at the Royal Veterinary College, working through the Dogs Trust-funded Generation Pup cohort, tracked 145 puppies in the UK and Republic of Ireland through their first six months of life. The headline number was sobering: 46.9% of the puppies showed at least one separation-related behavior by six months of age. The most common behaviors expressed only when left alone were pacing (14.5%), whining (7.6%), and spinning (6.9%), the classic signature of low-grade separation distress at the puppy stage.

But the value of Generation Pup is not in the prevalence figure. It is in the four specific protective factors the analysis identified.

The recommendation from the Generation Pup authors is calibrated and simple: when you walk in, regardless of what you find, enter the house calmly, do not address the damage in the puppy's presence, clean it up later without comment. The reaction is the training signal. Calm reaction trains calm expectation.

If you take nothing else from this essay, take that. Separation-related behavior is not mostly a matter of what breed you chose or what genes the dog was dealt. It is mostly a matter of what happened, or did not happen, during a specific window of the puppy's early life. That is simultaneously a heavy responsibility and a piece of extraordinarily good news: the same variable that drove a six-fold increase during lockdowns is, in the other direction, the thing you can control.

The Pandemic Legacy

It is worth pausing on what the pandemic cohort actually experienced, because the particulars matter and the lessons generalize.

In 2020, under lockdown conditions in the United Kingdom, a puppy acquired in, say, May of that year met a household in which one or both adults were working from home, rarely traveling further than the end of the street, and reliably available around the clock. The puppy learned, in the critical socialization window between seven and sixteen weeks, that her primary people were continuously present. She never learned what absence was. When she was placed in a crate for a nap, her person was twelve feet away at the kitchen table; when she woke, her person was there. The background baseline of her world was we are always together.

Then, over the next twelve to eighteen months, the adults began returning to offices. The transition, for many of these dogs, was not graduated. It was abrupt. One Monday the baseline changed, and the puppy, who by then had developed into an adolescent dog with fully-consolidated expectations about what normal life looked like, was suddenly left alone for eight hours, with no preparation, no graduated exposure, and no behavioral capacity to regulate her own arousal in the face of her world turning inside out.

What the RVC longitudinal study captured, at 21 months of age, was the downstream consequence of that abrupt transition. Thirty-one per cent of the cohort displayed observable separation-related behavior. Ninety-seven per cent displayed at least one problem behavior. Eighty-two per cent of the owners had tried to address the emerging problems using physical punishment, intimidation, or training tools designed to cause discomfort, and in the measurable aggregate, those owners got worse outcomes, not better, because aversive methods are astonishingly poor at treating anxiety-driven behavior and frequently manufacture new symptoms on top of the old ones.

The lesson from the pandemic cohort is not that dogs acquired in 2020 were somehow defective. They were not. They were ordinary dogs, in ordinary homes, who had been quietly and inadvertently trained, during their most malleable weeks, to expect continuous human presence as the baseline state of the world, and who then had that baseline revoked without warning. The same thing happens, in smaller doses, to ordinary puppies in ordinary years whenever a family takes two weeks of vacation to welcome the new arrival and then goes back to work on Monday of week three. The pandemic was just a natural experiment that made the effect, normally distributed across thousands of households, visible at scale.

The takeaway, for any family reading this in any year: the baseline your puppy experiences during her first sixteen weeks is what she will expect for the rest of her life. If your working pattern involves absences. Most do, then those absences need to be introduced early, deliberately, and at durations the puppy can tolerate, from the first week she is in your home. Not the week before you go back to work.

Is My Dog Struggling? A Self-Check

If you are reading this essay because you have a dog already, perhaps an adolescent or adult Lagotto, perhaps a puppy whose first weeks did not go the way this essay describes. You may be wondering whether what you have been seeing at home is ordinary velcro behavior or something more concerning. The following self-check is based on the behavioral features that Flannigan and Dodman, and subsequent researchers, identified as distinguishing clinical separation-related problems from normal Lagotto attachment.

It is not a diagnosis. A real diagnosis involves a qualified professional, ideally with video evidence of the dog during an actual absence. But it is a reasonable first-pass tool for deciding whether professional input is warranted, and if so, how urgently.

A few notes on how to read your own result. The weighting is not uniform: item eight (self-injury during escape attempts) counts as two items rather than one, because self-injury signals a welfare issue that rises above the threshold for professional consultation regardless of other scores. Any tick on item eight warrants a call to a veterinary behaviorist, full stop.

Also worth saying: the self-check is designed to catch both ordinary hyperattachment (normal for the breed, manageable at home) and clinical separation-related behavior (not manageable at home without help). It does not catch boredom-driven destruction, because boredom and separation distress look different in the time course, not in the single-snapshot symptom list. For that distinction, return to the Distress Triangle in Section V and, ideally, point a camera at your dog during an actual absence. The video will make the pattern unambiguous.