Evidence Discipline

Just Behaving · Rowley, Massachusetts

Version 1.1 · May 2026

Canonical meta-document on the program's epistemology. Public-facing reference for the practice of evidence discipline in the Just Behaving framework.

Evidence discipline is a daily practice. It is not a marketing posture, not a credential the program issues to itself, and not an ornamentation added to its documents to make them sound more rigorous than they are. It is the work that determines what the program is willing to claim about dogs, about its methods, and about why it does what it does. The discipline is held in specific operations that recur every day across the program's writing, decision-making, and engagement with new findings. This document explains those operations, demonstrates them with examples drawn from the program's recent canon, names what the discipline forbids, and argues that the practice is achievable, valuable, and increasingly necessary in a field where it remains uncommon.

A reader picking up this document should be able to follow it without having read any other Just Behaving material. Scientific terms are defined on first use. Internal terminology of the Just Behaving program is defined when it is introduced. Acronyms are expanded. Cross-references to other Just Behaving documents serve as worked examples, not as load-bearing arguments. The document stands on its own. A family considering the program, a veterinary professional auditing the framework, a graduate student evaluating the methodology, or a journalist writing about evidence-based dog raising should each be able to use the document for what they need.

The thesis the document is built around can be stated in a few sentences. The vast majority of claims made in the dog world about training, raising, behavior, breed character, evolution, and welfare are not evidence-based in any rigorous sense. They are anecdotal, traditional, intuitive, or commercially derived, and they are routinely presented at confidence levels the underlying evidence does not support. The Just Behaving program operates under a set of disciplines that govern what enters the program's canon and at what confidence level each claim is held. These disciplines are not unique to Just Behaving. Programs and practitioners in other parts of the dog world implement similar disciplines using their own specific tools. What this document offers is one program's articulation of its practice, in the hope that the articulation is useful both to readers evaluating Just Behaving's claims and to anyone in the field considering what comparable disciplines might look like in their own work.

The document moves through ten numbered sections. The first establishes the landscape of claims in the dog world, the mix of evidence quality that families and professionals encounter daily, the specific epistemological problems the field exhibits, and the case for program-level epistemology as one form of response. The second develops the five primitive evidence tags that constitute the foundational tool of the program's discipline: [Documented], [Observed - JB], [Estimated], [Heuristic], and [Ambiguous], and explains the two derived public badge and workflow composites that appear in the wiki. The third addresses the cross-species discipline, the way the program handles findings from research conducted in species other than dogs. The fourth describes the Scientific Claims Register (the SCR), the centerpiece tool that anchors evidence tags across the canon. The fifth treats slippage discipline, the single most common vulnerability in evidence-based writing across fields, and how the program audits for it. The sixth walks through worked examples drawn from the recent canon, showing the discipline in operation across four specific load-bearing claims. The seventh names the negative space (what the discipline forbids). The eighth describes the methodology of evidence maintenance over time. The ninth makes the case for evidence discipline as a service to the families the program serves. The tenth considers the wider implications of the practice for the field beyond Just Behaving specifically. A closing returns to the central thesis and ends on the document's landmark statement.

The voice throughout is intended to be accessible enough for a thoughtful family reader to follow while remaining rigorous enough that a board-certified veterinary behaviorist would not find the document careless. The document is making an argument for a way of practicing. Arguments have to be made carefully. The document earns its case by walking the reader through the practice, showing the operations the discipline requires, and being explicit about its own limits along the way. The reflexive challenge of writing a document about evidence discipline while observing the discipline runs through the work. The careful claims this document makes about the dog world's evidence quality and about the program's own contribution are themselves held at the confidence levels the underlying observations support, no higher. That challenge is part of the work. The work is what follows.

What This Page Explains

This page republishes the canonical evidence-discipline document for public wiki readers, researchers, journalists, veterinary professionals, graduate students, and AI or retrieval systems that need the full epistemology layer rather than a summary.

It explains the landscape of claims in the dog world, the five primitive evidence tags, the two derived public composites, cross-species discipline, the Scientific Claims Register, slippage discipline, worked examples from the canon, the negative space of what the discipline forbids, evidence maintenance over time, service to families, and the wider implications for the field.

The source document is intentionally long. The length is part of the function: a serious external reader should be able to audit the framework's method, examples, limits, and bibliography without needing access to private workspace files.

Core Explanation

Section 1: The Landscape of Claims in the Dog World

The case for evidence discipline begins with an honest description of where evidence quality currently sits across the field. The dog world is not uniform. Different communities within it (academic researchers, veterinary behaviorists, applied animal behaviorists, professional trainers, breeders, popular authors, equipment manufacturers, and the AI-amplified content layer that has emerged in the past several years) operate at different evidence standards. Families encountering the dog world do not typically have tools to distinguish between these communities or between the kinds of claims each produces. The structural problem this section names is not the malice of any particular community. It is the absence of consistent tools for distinguishing among the claim types that coexist in the field's everyday discourse.

The mix of evidence quality in the field

The dog world's claims come from a diverse set of sources, each operating at a different evidence standard. The diversity itself is not a problem. The problem is that the standards are not consistently distinguished from one another in the materials families and professionals typically encounter.

Peer-reviewed canine behavior science exists and is growing. Research groups at the Duke Canine Cognition Center, the Wolf Science Center in Ernstbrunn, Austria, the Family Dog Project in Budapest, the Indian Institute of Science Education and Research Kolkata, the Arizona Canine Cognition Center, the Working Dog Project at the University of Pennsylvania, and many others publish primary research in journals such as Animal Cognition, Applied Animal Behaviour Science, Behavioural Processes, Journal of Comparative Psychology, PLOS ONE, and Proceedings of the National Academy of Sciences. This body of research is the strongest evidence base available for understanding canine biology, cognition, and behavior. It is also relatively small compared to research in other species, with the canine literature still substantially thinner than the literature on rats, primates, and humans.

Clinical observation by veterinary behaviorists provides another evidence category. Diplomates of the American College of Veterinary Behaviorists (DACVBs) are veterinarians who have completed residency training in veterinary behavioral medicine and have passed a board-certification examination. The clinical observations they produce, drawing on years of case work with thousands of dogs across diverse conditions, are not peer-reviewed primary research, but they constitute structured professional observation by individuals with substantial training and accountability. Manual of Clinical Behavioral Medicine for Dogs and Cats by Karen Overall (2013), among the field's standard clinical references, exemplifies this kind of structured observation grounded in case experience and the broader literature.

Long-experienced breeder observation provides a third category. Breeders who have raised many dogs across many years, sometimes across decades and across multiple breeds, accumulate substantial experiential knowledge about temperament, development, breeding outcomes, and the long-arc behavioral signatures of dogs raised under different conditions. This knowledge is not peer-reviewed, is rarely systematically recorded, and is highly variable in quality depending on the individual breeder's attentiveness, skepticism, and willingness to update on new findings. At its best, breeder observation can be valuable; at its worst, it consists of strongly-held views that do not survive contact with rigorous evidence.

Cross-species inference from research in other mammals constitutes a fourth category. Much of the most relevant research on parenting, attachment, stress regulation, learning, and developmental neurochemistry has been conducted in rats, primates, and humans rather than in dogs. The inferences drawn from this research and applied to canine biology are sometimes well-grounded (the cross-mammalian convergence on certain features is robust), and sometimes loosely-grounded (a finding in rats does not necessarily transfer to dogs without methodological care).

Anecdotal experience reported by individual trainers, breeders, and owners constitutes a fifth category. Anecdote is the dominant evidence currency in much of the dog world's popular literature. A trainer who has worked with many dogs reports what has worked in those cases; an owner reports what worked for their dog; a breeder reports what they have seen in their lines. Anecdote can be useful for generating hypotheses or for identifying patterns worth studying more rigorously, but it is not sufficient evidence for confident claims about how dogs in general respond to particular interventions.

Traditional practices passed down through breed clubs, training schools, and breeding lines constitute a sixth category. Some of these traditions encode accumulated practical wisdom that has survived because it worked. Others encode beliefs that have persisted through cultural transmission long after the underlying scientific basis was withdrawn or never existed. Distinguishing these is not always easy from within a tradition.

Popular evolutionary narratives constitute a seventh category. The alpha framework derived from captive-wolf studies (Schenkel 1947), the pack-leader framing that dominated popular training discourse from the 1970s through the 2000s, and various "domestication tamed everything" framings have all had substantial cultural durability. Some of these narratives have been retracted by their original scientific sources without dislodging their popular use; the alpha framework was substantially retracted by L. David Mech in 1999, yet remains in popular circulation more than a quarter-century later.

Commercial marketing claims by training schools, breeding programs, food companies, and equipment manufacturers constitute an eighth category. Marketing claims are not inherently evidence-free; some companies make claims that are well-supported by research. But marketing has its own logic, with commercial pressure on claims that does not always align with rigorous evidence evaluation.

AI-amplified content of varying quality constitutes the ninth and newest category. Large-language-model-generated material on canine topics has proliferated rapidly. Some of it draws on rigorous sources; much of it draws on the broader mix of sources described above, with the AI layer flattening evidence quality distinctions in ways that can be difficult for readers to detect.

The point of this enumeration is not that any of these categories is uniformly bad. Each can produce useful information in its proper register. The point is that these categories coexist in the dog world, often appear together in the typical family's reading material, and are not consistently distinguished from one another. A family reading about how to socialize a puppy can encounter a peer-reviewed finding, a breeder anecdote, a commercial-marketing claim, and an AI-generated paragraph in the same hour, presented in formats that do not signal their differing evidence quality.

The specific epistemological problems

Several specific patterns recur across the field's evidence quality variation. Each contributes to the structural difficulty that families and professionals encounter when trying to evaluate competing claims.

The "evidence-based" label is widely used and seldom audited. Programs, schools, and authors describe their methods as evidence-based without making clear what evidence supports which specific claims at which confidence levels. A reader encountering the "evidence-based" label is often given no tools for verifying whether the label accurately describes the underlying material. The label has become, in many uses, a marketing term rather than a description of practice.

Cross-species inference is routinely laundered into canine-direct claims. A maternal-care study in rats produces findings about gene-environment interactions in offspring stress regulation; in popular writing the rat finding becomes "this is how attachment works" without the species qualification. A primate study produces findings about secure-base function; the finding becomes "this is how dogs form attachments" without the inference being marked. The laundering is rarely intentional; it is the natural drift of writing for readers who want the punchline. But the result is that readers encounter claims about canine biology that have not actually been tested in dogs.

Retracted frameworks persist for decades after retraction. The most prominent example in the dog world is the alpha framework. L. David Mech's 1970 book The Wolf did more than any single work to popularize the framework, drawing on Rudolph Schenkel's 1947 observations of captive wolves at the Basel Zoo. Across the next three decades, Mech's field research on wild wolf populations produced data that did not fit the framework. The 1999 paper "Alpha status, dominance, and division of labor in wolf packs" in Canadian Journal of Zoology is the explicit on-the-record retraction. Wild wolf packs are not assemblies of unrelated competitors jockeying for dominance; they are families, with the "alpha" pair being the breeding parents and the rest of the pack being their offspring. The retraction has not, in the quarter-century since, dislodged the alpha framework from popular dog training discourse. The cultural durability of the framework has substantially outlasted its scientific support.

Heuristic and traditional practices are often defended in the rhetorical mode of settled science. A practice that has worked for a particular trainer over many cases is described not as "this is what has worked in my experience" but as "this is how dogs are." A traditional practice handed down through a breeding line is described not as "this is how we have done it" but as "this is the right way." The shift from heuristic claim to settled-science claim is rhetorical rather than evidentiary; the underlying support has not changed, only the language describing it.

Welfare-cost evidence and welfare-positive evidence are often presented selectively. The published literature on aversive training methods, including Ziv's 2017 Journal of Veterinary Behavior review, Vieira de Castro and colleagues' 2020 PLOS ONE paper, China and colleagues' 2020 Frontiers in Veterinary Science paper, and the AVSAB position statements, documents welfare costs from certain training methods. These costs are not always integrated into the materials that recommend the methods. Conversely, evidence supporting reward-based approaches is not always presented with appropriate caveats about effect sizes or contextual moderators. Selective presentation of welfare evidence occurs in both directions, and the family or professional reader has to do considerable work to recover the full picture.

New findings that complicate previous claims are sometimes ignored rather than integrated. A program that has built its identity around a particular position can find it difficult to update when new findings complicate that position. The Marshall-Pescini and colleagues 2017 Proceedings of the National Academy of Sciences paper documenting that wolves outperform dogs on conspecific cooperation tasks, for example, complicates simple "domestication made dogs better at everything social" framings; how this finding gets integrated into popular accounts of canine cognition varies considerably across sources.

The cost to families

The structural problems above translate into real consequences for the families that encounter the dog world's claims. A family bringing home a puppy faces decisions about socialization, training, environmental management, feeding, sleeping arrangements, exercise, social exposure, and dozens of other questions across the first months. These decisions are consequential. Developmental windows close. The relational architecture of the first year shapes the dog the family lives with for the next decade. Bad evidence produces bad decisions; good evidence produces decisions more likely to serve the dog over the long arc of the relationship.

The typical family does not have professional training in canine behavior science. Most families have not read peer-reviewed papers in Animal Cognition or Applied Animal Behaviour Science. Most families do not have a baseline for evaluating whether a claim in a popular book draws on rigorous research, on a particular trainer's anecdotal experience, on a traditional practice, or on commercial marketing. The decisions are made under emotional load: the family wants to do well by the puppy, and the desire to do well does not automatically come with tools for evaluating which advice to follow.

The result is that families often rely on whichever source feels most trustworthy, most credentialed-looking, or most aggressively marketed. The source that feels most authoritative is not necessarily the source with the strongest evidence. A confident, well-produced book by a charismatic trainer can outrank, in the family's perception, a more careful but less rhetorically-polished resource. A well-financed marketing campaign for a particular training method can outweigh, in the family's encounter, the published welfare-cost literature that complicates the marketing.

This is not the families' fault. The structural problem is upstream. The field's evidence quality variation creates conditions in which families cannot reliably evaluate claims, and the consequences fall on the dogs the families are trying to raise well.

The case for program-level epistemology

Individual practitioners cannot solve this problem alone. A single trainer, breeder, or veterinarian who personally maintains high evidence standards can serve the families and dogs they directly work with, but the structural problem remains. The field's overall evidence quality is not determined by the most rigorous individual practitioners; it is determined by the conditions under which claims circulate at scale.

A program-level commitment to evidence discipline can do something individual practice cannot. It can create a standard against which every claim in the program's material is audited. It can establish architectural choices (such as evidence tagging and a claims register) that enforce the discipline structurally rather than relying on individual writers' or reviewers' personal vigilance. It can make the discipline visible to outside readers, so the readers can verify the discipline's operation rather than trusting the program's assurance that the discipline operates. And it can demonstrate, by example, that this kind of rigor is achievable for working programs operating in real-world conditions.

The Just Behaving program's commitment to evidence discipline is one such program-level expression. The program does not claim to have invented evidence discipline, to be the only program practicing it, or to have arrived at a final or perfect form of the practice. The program practices a specific architecture (the five primitive evidence tags, the derived wiki-facing composites, the Scientific Claims Register, the slippage audits, the cross-species discipline, the negative-space prohibitions, the methodology of maintenance) that has emerged through years of work on the program's canon. The architecture is offered both as the program's working method and as one model that other programs or individual practitioners might find useful to consider. The discipline is what the next nine sections explain.

Section 2: The Five Evidence Tags

The foundational tool of the Just Behaving program's evidence discipline is its tagging system. Every empirical claim in a Just Behaving document carries a tag. The tag is not optional and is not decorative. It is the daily operational practice of distinguishing what is known from what is suspected, observed but not yet confirmed, or open in the literature. This section explains the system, walks through each of the five tags with examples, and concludes with the rationale for five tags rather than two.

The tagging system as the foundation

Tags appear in brackets within the prose of every Just Behaving document. A claim about canine social structure might be followed by [Documented - Dog], signaling that the claim is directly supported by peer-reviewed canine research. A claim drawing on cross-mammalian convergent findings might be followed by [Documented - Mammal]. A claim about the cross-species application of a finding from rats to dogs might receive dual tagging: [Documented - Rat] for the rodent finding, [Heuristic - Dog] for the canine application. A claim drawn from Just Behaving's own program experience might be followed by [Observed - JB]. A claim that the literature has not yet resolved might be marked [Ambiguous].

The brackets are visible to the reader. This visibility is itself part of the discipline. A reader can see, at the point of every empirical claim, what kind of evidence the program is offering for the claim. The visibility removes the need for the reader to do separate research to figure out whether a confidently-stated sentence is backed by peer-reviewed canine research or by a reasonable inference from related work. The work of distinguishing has been done up front, and the distinction is presented to the reader in the prose.

The discipline of tagging requires that the writer make a decision about every empirical claim. There is no way to write a Just Behaving document while skirting the question of what evidence supports each claim. The tag forces the decision. If the writer cannot identify a tag, the claim cannot enter the document. This forcing function is what makes the tagging system the foundation of the discipline. It is not a layer added to writing that would otherwise be unmarked; it is the structural requirement that shapes the writing from the beginning.

[Documented]: peer-reviewed research directly supports the claim

The strongest evidence level in the tagging system is [Documented]. A claim tagged [Documented] is one for which at least one peer-reviewed study directly supports the claim, with the species, method, and finding clearly mapping to the claim being made. The species qualifier is essential. Canine peer-reviewed research supports canine claims at [Documented - Dog]. Wolf research supports wolf claims at [Documented - Wolf]. Cross-mammalian convergent findings, drawing on research in multiple mammalian species with consistent results, can support [Documented - Mammal] claims. Human developmental research supports [Documented - Human] claims.

The careful operationalization of [Documented] is what gives the tag its meaning. [Documented] does not mean "an expert said this in a popular book." [Documented] does not mean "this is what trainers commonly say." [Documented] does not mean "this is consistent with broad scientific consensus on related topics." It means peer-reviewed research with directly applicable findings. The peer-reviewed standard is the standard at which research has been formally evaluated by other researchers in the field for methodological adequacy and clarity of finding. This standard is not perfect (peer review has well-known limitations and produces variable quality), but it is the standard at which the broader scientific community vets primary research, and the program treats it as the threshold for the strongest evidence tag.

A worked example helps. Consider the claim that adult dogs and puppies engage in observational learning from human and canine demonstrators. The Topál and colleagues 2006 paper in Animal Cognition, titled "Reproducing human actions and action sequences: 'Do as I Do!' in a dog," documented the capacity in pet dogs under controlled conditions. The Range and colleagues 2007 Current Biology paper extended the comparative picture. The Pongrácz and colleagues 2001 work on social learning in dogs provides earlier foundational evidence. A Just Behaving document making the claim that dogs engage in observational learning can tag it [Documented - Dog] with references to these specific peer-reviewed studies. The reader who wants to verify the claim has the references and can read the underlying research.

A related claim, that observational learning in dogs is mediated in part by attachment quality with the human demonstrator, requires more careful tagging. The general finding that secure attachment supports social learning has stronger evidence in human developmental research than in canine research. A claim that combines the canine observational learning capacity with the attachment-mediation hypothesis might require dual tagging: [Documented - Dog] for the basic observational learning capacity, [Heuristic - Dog] for the attachment-mediation extension. The dual tagging is what the cross-species discipline (developed in Section 3) requires.

The threshold for [Documented] is conservative by design. A claim near the edge of what the research supports is tagged at the next lower level rather than at [Documented]. The conservative threshold is part of what makes the tag meaningful when it appears. A reader encountering [Documented] in a Just Behaving document can rely on the tag because the tag is reserved for claims that meet the conservative standard, not stretched to cover claims that approximately meet it.

[Observed - JB]: drawn from direct Just Behaving program experience

The second tag is [Observed - JB]. This tag is used when the claim being made is supported by direct experiential evidence from the program's own operations. The program raises Golden Retrievers, places puppies in family homes, follows up with families across years, and accumulates observational data on placement outcomes, behavioral signatures of dogs raised under the program's specific architecture, and patterns that appear across many dogs and families over time. These observations are real evidence. They are also not equivalent to peer-reviewed research, and the tag is the program's honest acknowledgment of that difference.

The operationalization of [Observed - JB] requires holding two things together. The program does have genuine experiential evidence from its operations. A claim that puppies who have been through the program's specific socialization architecture tend to show a particular behavioral signature in their family homes is not a guess; it is the program's working observation of what happens across the dogs it has raised. The program is in a position to make this kind of claim with first-hand observational support.

At the same time, single-source observational evidence carries inferential limits. The program's observations come from one breeding line, one geographic location, one set of practices applied over a particular period of time. A claim supported by the program's observations is not a claim about all dogs everywhere or about what would happen under different conditions. It is a claim about what the program has seen in its specific operating context. The tag's "JB" qualifier is intended to signal exactly this limit. The reader is being told that the program is offering its observations, not that the program is offering a general scientific finding.

A worked example. Consider the claim that puppies raised through the Just Behaving program's extended breeder-home period (twelve to sixteen weeks with continued exposure to calm adult dogs) display a characteristic settled baseline in their family homes during the first month after placement. This is a claim about a pattern the program has seen across many placements. The program can make the claim, tagged [Observed - JB], with the program's experience as the supporting evidence. A peer-reviewed controlled comparison documenting that the extended breeder-home architecture produces this specific outcome does not exist. The program is offering its observation. The reader is being told that the observation is real but is single-source and does not have the evidentiary weight a peer-reviewed controlled comparison would provide.

[Observed - JB] is, in the program's tagging system, evidence at a level above pure inference and below peer-reviewed research. It is honest reporting of what one program has seen. The program uses the tag where it is appropriate and does not use the tag to launder weaker claims into apparent evidence support. A claim that the program cannot honestly report from its own observation does not receive the [Observed - JB] tag, even if the program would prefer to support the claim that way.

[Estimated]: derived inference, biologically plausible, but not directly tested

The third tag is [Estimated]. This tag is used when a claim is derived from related evidence (theoretical frameworks, adjacent research, biological reasoning) rather than directly supported by a study of the specific claim being made. The claim is the program's best inference from the available evidence, presented as inference rather than as a documented finding.

A worked example illustrates the distinction. Consider the claim that puppies separated from their natal canine social environment at six to eight weeks experience the transition as a substantial developmental discontinuity. The general biological reasoning is sound. Puppies in the natal environment have continuous exposure to adult canine social signaling, the dam's calibrated correction patterns, and the broader group's social dynamics. Removing the puppy from this environment introduces a substantial change in the puppy's social context. A direct controlled comparison documenting the specific developmental cost of early versus extended natal-environment exposure, with longitudinal outcome measures across the dog's life, would be difficult to conduct and has not been published in a form that directly supports the specific claim.

The program might still want to make a version of this claim, anchored to the biological reasoning and to the adjacent research on social development in canids and other mammals. The tag [Estimated] signals that the claim is inferred from related evidence rather than directly supported by a study of this specific question. The reader is being told that the program considers the claim biologically plausible but is not claiming direct documentation. The honesty of the tag protects the reader from the impression that more direct evidence exists than actually does.

The boundary between [Estimated] and [Heuristic] is sometimes fine and requires editorial judgment. [Estimated] leans toward claims derived from theoretical frameworks or adjacent research with reasoned inference; [Heuristic] leans toward claims supported by convergent observational evidence across many practitioners and biological plausibility but without controlled study. The distinction is operational; both tags signal that the claim is reasonable but not directly tested.

[Heuristic]: patterns observed that are consistent with biology and make good practical sense but have not been through controlled study

The fourth tag is [Heuristic]. This tag is used for claims supported by patterns the program and adjacent practitioners have observed across many dogs, families, and contexts, where the claim is consistent with biological plausibility but where formal controlled-comparison evidence is absent.

[Heuristic] claims are the workhorse of practical canine behavior knowledge. Many of the operational protocols that guide skilled trainers and breeders are heuristic in this sense. The protocols emerged from accumulated practice, were refined through trial and revision, and now constitute a working knowledge base that is not formally tested but is widely held and consistent with what we understand of canine biology.

A worked example. The relational-context firewall claim that the Just Behaving framework develops in JB_Indirect_Correction.md (and that the Section 6 worked example in this document examines in detail) is held at [Heuristic - Dog]. The claim is that the developmental outcome of a correction depends substantially on the relational context within which the correction is delivered, not only on the technical operant classification of the correction. The biological plausibility is strong: cross-species literature on caregiver-buffered stress response, the secure-attachment literature, and the broader mammalian parental-care literature all converge on the importance of the caregiver-offspring relational frame. The clinical resonance is strong: skilled handlers across multiple traditions consistently report different outcomes from corrections delivered within versus outside secure relationships. The direct controlled-comparison evidence in dogs (a randomized study comparing developmental outcomes for puppies receiving corrections in matched operational form but different relational contexts) does not exist. The program holds the claim at [Heuristic - Dog], signaling that the claim is the program's working position based on convergent observation and biological plausibility, not a documented finding.

The honesty of [Heuristic] is that the tag does not pretend the claim is more strongly supported than it is. The claim is held as the program's working position. The program would update the claim if controlled evidence contradicted it. The program would upgrade the tag if controlled evidence supported it at the higher confidence level. In the absence of such evidence, the program presents the claim at the level the evidence currently supports.

[Heuristic] claims are not weak. They are working knowledge held at the appropriate confidence level. The program operates substantially on heuristic claims because much of what skilled practitioners know about dogs is heuristic in this sense. The discipline is in holding the claims at heuristic confidence rather than rhetorically promoting them to settled-science confidence.

[Ambiguous]: the literature is genuinely contested or insufficient

The fifth tag is [Ambiguous]. This tag is used when the relevant scientific literature does not provide a clear answer, when the literature is genuinely contested, when methodologies differ in ways that affect findings, or when the evidence base is insufficient to support a confident position. The tag is the program's acknowledgment that some questions have not been resolved by the field.

A worked example. The specific heritability of various behavioral traits in dogs is, in the published literature, a contested topic. Different studies using different methodologies on different populations produce different heritability estimates. Some traits are estimated as having high heritability in some studies and substantially lower heritability in others. The methodological choices (the breed populations sampled, the behavioral phenotyping protocols, the genetic markers used) substantially affect the findings. A claim about the heritability of a specific behavioral trait is, in the current state of the literature, a claim into a genuinely contested area.

The program might make a version of such a claim with the tag [Ambiguous], signaling that the claim is one position in an active scientific debate rather than a settled finding. The reader is being told that the program is naming uncertainty rather than asserting a resolved position. The discipline is in resisting the temptation to flatten contested questions into apparently-settled positions because the program would prefer to have a clear answer.

[Ambiguous] is the tag that the program uses sparingly but uses honestly. Many claims in the canon do not require this tag because the evidence is clearer one way or the other. But for the claims that are in contested territory, the tag protects both the program and the reader from false confidence.

The rationale for five tags rather than two

A binary distinction (evidence-based versus not evidence-based) is too coarse to capture the actual structure of knowledge claims. Real knowledge is graded. A claim supported by extensive peer-reviewed research is in a different evidence position from a claim supported by convergent observation across many practitioners but without controlled study. A claim that the literature has not resolved is in a different position from a claim that has been clearly supported by recent work.

A two-tag system invites two failure modes. The first is overclaiming. Everything that does not meet the strict "evidence-based" threshold gets dropped, with the program losing the ability to discuss heuristic knowledge or to acknowledge biologically plausible inferences. The second is underclaiming. Everything that does not meet the strict threshold is presented as guesswork, with the program losing the ability to distinguish between a heuristic claim grounded in convergent observation and biology and a pure inference with little support. Neither failure mode serves the reader.

The five-tag system lets the program hold claims at their actual evidence levels. A [Documented] claim is presented as documented. A [Heuristic] claim is presented as heuristic. The reader sees both, knows which is which, and can evaluate each accordingly. The program retains the ability to discuss working knowledge without falsely promoting it; the program retains the ability to acknowledge contested questions without retreating from them entirely.

The design follows the general epistemological principle that evidence is graded rather than binary. The GRADE (Grading of Recommendations Assessment, Development and Evaluation) Working Group, which has developed the dominant evidence-grading methodology in clinical medicine, uses a four-level system (High, Moderate, Low, Very Low certainty) with explicit operationalization of what factors raise or lower the rating. The Cochrane systematic reviews use comparable graded evidence assessment. The principle that intellectual honesty requires marking the gradations rather than collapsing them is well-established in evidence-based practice across disciplines (GRADE Working Group; Cochrane methodology).

The Just Behaving five-tag system is the program's specific adaptation of this general principle to the dog-behavior context. The tags are not identical to the GRADE levels, because the canine behavior context is different from the clinical-medicine context that GRADE was developed for. But the underlying epistemological commitment is the same. Evidence is graded; the gradations matter for the reader's decisions; the discipline is in marking the gradations.

The cost of the five-tag system is that the writing becomes slightly more visually busy. A document with tags is not as smooth as a document without them. The program accepts this cost. The cost of unmarked writing is that the reader has no tools for distinguishing what is documented from what is inferred. The trade-off favors the marking, because the marking serves the reader. The system is the practice in operation.

A Note on Display and Workflow Composites

The five evidence tags above are the program's primitive epistemic categories. They identify the kind of confidence a claim carries and the kind of source that supports it. Two additional tags appear on entry-level public badges and in the wiki's frontmatter system, and they require explicit framing to avoid confusion.

Documented-Cross-Species is a notational shorthand applied at the public-badge level when an entry carries a documented finding from another species being applied to dogs at heuristic confidence. The full epistemic state is dual-tagged at the claim level: [Documented - Species] for the original finding, [Heuristic - Dog] for the canine application. The composite badge compresses these two tags into a single visible label so the reader does not need to parse two primitives when scanning the entry. The dual-tag underlying state is the primitive truth; the composite is a display affordance.

Mixed Evidence is a page-level rollup applied when an entry carries multiple claims at different evidence weights. A page on Calmness, for example, may carry one claim at [Documented - Human], another at [Heuristic - Dog], and a third at [Observed - JB]. Rather than display a single primitive tag that misrepresents the page's full evidence state, the badge displays Mixed Evidence and the SCR panel exposes the underlying claim-level primitive tags. Mixed Evidence is a workflow rollup, not an independent epistemic category. A page tagged Mixed Evidence is not less rigorous than a page tagged Documented; it is heterogeneous, and the heterogeneity is the relevant signal.

The five primitives remain the foundational taxonomy of the program's evidence discipline. The two composites exist because public-facing pages need readable badges and dashboard surfaces need rollup labels, not because the program has seven independent evidence categories. When in doubt about how a claim should be tagged, consult the five primitives; the composites derive from them.

Section 3: Cross-Species Discipline

Cross-species generalization is where pseudo-scientific claims most often enter the dog world's discourse. Research conducted on rats becomes "this is how attachment works." Research on primates becomes "this is how dogs form bonds." Research on captive wolves becomes "this is how dogs naturally relate to other dogs." The laundering is rarely intentional; it is the natural drift of writing for readers who want the punchline. But the result is a body of popular and even semi-professional material that overstates what is actually known about canine biology. The Just Behaving program's cross-species discipline is the architecture that prevents this drift within the program's canon. This section explains the challenge, names the temptation, develops the dual-tagging discipline, and works through several specific examples.

The challenge: most directly applicable research is not on dogs

The canine peer-reviewed literature is substantial and growing, but it remains thinner than the literature on several other species whose findings inform our understanding of dogs. Mammalian parenting research is heavily concentrated in primates (especially rhesus macaques and chimpanzees), in rodents (especially rats and mice), and in humans. Attachment theory was developed by John Bowlby and Mary Ainsworth in human infants and only later extended to dogs through the foundational Topál and colleagues 1998 Journal of Comparative Psychology paper, with the canine attachment literature still much smaller than the parallel literature in humans. Learning theory was developed in rats and pigeons (the Skinnerian tradition), with applications to dogs being a later development. Developmental neurochemistry research is heavily concentrated in rodents (the Meaney and Champagne lab tradition) and primates. Stress biology research is heavily concentrated in humans, rats, and primates.

The consequence for canine behavior writing is direct. A writer who wants to discuss canine parenting, attachment, learning, or stress regulation will find that much of the most directly relevant research has been conducted in other species. The writer has two options. The first is to limit claims strictly to what has been documented in dogs, accepting that some interesting and biologically plausible claims cannot be made because the canine research has not yet been done. The second is to draw on cross-species research with appropriate methodological care, marking the cross-species inference and holding the canine application at the appropriate confidence level.

The Just Behaving program takes the second approach, with explicit discipline. The cross-species research is part of the evidence base, but it is part of the evidence base for what it actually shows, not for what it might suggest about dogs by uncritical extrapolation. The dual-tagging discipline developed below is the operational form of this approach.

The temptation: apply findings directly to dogs

The temptation to apply cross-species findings directly to dogs without marking the inference is strong, for several reasons. The findings are often dramatic. A maternal-care study in rats showing that variation in licking and grooming behavior produces measurable epigenetic differences in the offspring's stress regulation systems is genuinely striking. The implication for parenting practice (that early maternal care matters in ways that persist into adulthood) is the kind of finding that translates well into popular writing. The writer wants to tell the reader about it, and the cleanest way to do that rhetorically is to write "maternal care affects offspring stress regulation across the mammalian class," dropping the species qualifier.

A primate attachment study showing the secure-base function of the mother for infant exploration translates similarly. The finding has clear implications for how attachment functions in mammalian development generally. The writer wants to apply the finding to the context the reader cares about, which is the dog. The cleanest rhetorical move is to write "this is how dogs form attachments," dropping the species qualifier.

These moves are not malicious. They are the natural drift of writing for an audience that wants the punchline rather than the methodological caveat. The audience may not particularly want to read about rats or primates when they have come for information about dogs. The temptation is to give the audience what they want by translating the finding without marking the translation.

The cost of yielding to the temptation is that the resulting material misrepresents what is known about dogs. The rat finding is a rat finding. The primate finding is a primate finding. The canine application is an inference, and inferences from one species to another vary in their reliability depending on the specific question, the specific biological mechanism, and the specific cross-species comparison. The discipline of marking the inference is what allows readers to know that an inference has been made, so they can evaluate it accordingly. Without the marking, readers receive an apparent finding about dogs that was never actually demonstrated in dogs.

The Just Behaving discipline: dual-tagging cross-species claims

When a claim in a Just Behaving document depends on cross-species research, the program tags the species studied and the canine application separately. The dual-tag pattern is the operational form of the cross-species discipline.

A typical dual-tagged sentence looks something like this: "Maternal caregiving variation in rats has been documented to produce epigenetic differences in offspring methylation patterns and adult stress responses [Documented - Rat] (Weaver et al. 2004; Champagne 2008). The convergence with the broader mammalian-parenting pattern across species suggests that the canine application is biologically plausible, though direct controlled comparisons of maternal-care variation effects in dogs have not been conducted at the same level of detail [Heuristic - Dog]."

Several things are accomplished by this dual-tagging. The reader is told that the original research was conducted in rats. The reader is told that the canine application is an inference rather than a documented canine finding. The reader is given the citations for the rat work and can read the underlying research if they want to evaluate the inference themselves. The reader understands that the program is offering the cross-species translation as a heuristic claim, not as a documented canine finding.

The dual-tagging pattern is the architecture that prevents cross-species laundering. A document that consistently uses dual tagging at every cross-species translation cannot produce sentences that read as canine findings while actually being rat findings. The architecture forces the marking. The marking is what allows the reader to know what kind of evidence supports each claim.

The dual-tagging pattern is applied at every cross-species translation in the program's canon, not just at the first mention in a document. This is part of the slippage discipline (developed in Section 5). A claim introduced with dual tagging in a document's opening section does not lose the dual tagging when the same claim is referenced in the document's later sections. The cross-species discipline is maintained throughout, not just at introduction.

Worked examples

Three specific examples from the program's recent canon illustrate the dual-tagging discipline in operation.

The Baumrindian parenting typology applied to dogs. Diana Baumrind's developmental psychology research, developed at the University of California, Berkeley from the 1960s through the subsequent decades, established the parenting-typology framework that distinguishes authoritative (high warmth, high demandingness), authoritarian (low warmth, high demandingness), permissive (high warmth, low demandingness), and uninvolved (low both) parenting styles, with developmental outcomes for children that vary substantially by style. The framework was developed and extensively tested in human populations, with the Eleanor Maccoby and John Martin 1983 extension formalizing the two-axis structure and the four-style model. Subsequent decades of research, including Martin Pinquart's 2017 meta-analyses in Developmental Psychology and Educational Psychology Review, have refined the operational definitions and documented the developmental outcomes across populations. The framework is [Documented - Human] at the level of the human developmental research.

The Just Behaving program's JB_Structured_Leadership.md (one of the program's recent canonical satellite documents) applies the Baumrindian framework to dog raising, with the claim that authoritative caregiving as understood in human developmental psychology has recognizable analogues in mammalian parenting more broadly and is reasonably extended to the dog-raising context. The canine application is not directly tested in the controlled-comparison sense. The application is biologically plausible: the cross-mammalian convergence on certain parental-care patterns is well-documented, the canine handler-effects literature is consistent with the framework, and skilled practitioners across many traditions recognize something like authoritative caregiving as a working pattern. But the framework itself has not been formally tested in dogs in the way it has been tested in humans.

The program tags this carefully. The Baumrindian framework itself is [Documented - Human]. The cross-mammalian convergence is [Heuristic - Mammal]. The canine application is [Heuristic - Dog]. The triple-tagging pattern preserves the reader's ability to distinguish the human-developmental evidence (which is strong) from the cross-mammalian inference (which is plausible but not formally tested across the relevant range of species) from the canine application (which is the program's working position for the dog-raising context). The cross-species discipline is held at every mention of the framework throughout the JB_Structured_Leadership.md document.

Maternal caregiving methylation in rats applied to puppies. The Meaney and Champagne lab tradition has documented that variation in maternal licking and grooming behavior in rats produces measurable differences in offspring methylation patterns, particularly in genes involved in stress regulation, and that these differences persist into adulthood and affect the offspring's behavior. The Weaver and colleagues 2004 Nature Neuroscience paper is a foundational reference. The Champagne 2008 Frontiers in Neuroendocrinology review extends the picture. The findings are [Documented - Rat].

The Just Behaving program's canon discusses the cross-species relevance of these findings to canine development carefully. The biological plausibility of an analogous effect in dogs is real. The mammalian stress-regulation systems are conserved across the class. Maternal care variation in dogs is real and observable. A reasonable inference is that maternal care variation in dogs may produce comparable epigenetic effects on offspring stress regulation, though the specific methylation patterns and behavioral correlates would need to be documented in dogs to support the canine application at the same confidence level as the rat findings.

The program tags this carefully. The rat findings are [Documented - Rat]. The cross-species mammalian convergence is [Heuristic - Mammal]. The canine application is [Heuristic - Dog]. The reader is told that the dramatic rat findings do not directly translate into documented canine findings. The translation is the program's reasoned inference, held at heuristic confidence.

Wolf social structure applied to dog training. The most prominent failure mode of cross-species generalization in the dog world's history is the alpha framework's translation from captive-wolf observations to dog training. Rudolph Schenkel's 1947 observations of captive wolves at the Basel Zoo produced a dominance-hierarchy framework that L. David Mech's 1970 book The Wolf brought to wide audiences. The framework was then applied to dogs through popular dog-training books in the 1970s and 1980s and through subsequent television and popular media. The framework's empirical premise was a particular reading of wild wolf social structure based on captive-wolf observations.

The framework's empirical premise was substantially wrong. Mech's own 1999 paper in Canadian Journal of Zoology retracted the framework as a description of wild wolf social structure: wild wolf packs are families, with the breeding pair as the alpha pair, and the rest of the pack as their offspring. The captive-wolf hierarchies that Schenkel observed were artifacts of the captive context, not features of natural wolf social biology. The cross-species translation that had built dog training methods on the captive-wolf framework rested on a wolf-empirical claim that the field's primary scientific source publicly retracted.

The program's JB_Structured_Leadership.md Section 3 develops the dominance critique at length, with the careful handling that the empirical literature on free-ranging dog social structure also reveals: free-ranging dogs do form hierarchies, but the hierarchies are age-graded and tolerance-based rather than the captive-wolf alpha pattern (Bonanni et al. 2017; Cafazzo et al. 2010). The Bonanni and Cafazzo findings are tagged [Documented - Dog]. The Mech 1999 retraction is [Documented - Wolf]. The Schenkel 1947 captive-wolf observation is acknowledged as documented in its own original context but retracted as a description of wild wolf social structure and not relevant to dog social structure in any case.

This example illustrates the cross-species discipline in inverse form: not the laundering of a finding from one species to another, but the careful unwinding of a cross-species translation that had been doing widespread damage in the field. The discipline is applied to both kinds of cases. The same tagging architecture that prevents inappropriate laundering of new findings also supports the careful retraction of inappropriate laundering from earlier eras.

Why this matters

Cross-species laundering is one of the most common epistemological failure modes in the dog world. A program that does not maintain the cross-species discipline produces material that overstates what is known about dogs. The reader of such material encounters claims that read as canine findings but are actually inferences from other species, often without realizing the difference. The reader's confidence in the material is then misaligned with the actual evidence base supporting it.

A program that maintains the discipline produces material that is more credible to professional readers and more useful to family readers making real decisions. The professional reader can audit the cross-species translations and evaluate them on their merits. The family reader does not need to do this audit; the dual-tagging tells them what species each claim is anchored to and what confidence the canine application carries. The discipline serves both audiences by giving them the information they need at the level of detail they need it.

The cross-species discipline is, more broadly, part of what makes evidence-based practice possible in fields that draw on multiple species of research. Comparative biology, veterinary medicine, and animal behavior research all routinely draw on cross-species findings. The discipline of marking the cross-species inference is not unique to Just Behaving or to the dog world; it is a general methodological commitment that the program has adapted to its specific context. The principle holds: the species in which a finding was documented matters, and the cross-species inference is an additional epistemic step that should be made visible to the reader rather than hidden in confident-sounding prose.

Section 4: The Scientific Claims Register

The Scientific Claims Register, abbreviated as the SCR throughout the Just Behaving canon, is the program's centerpiece tool for maintaining evidence discipline across the canon. The SCR is the authoritative reference for every empirical claim the program makes. This section explains the SCR's architecture, its role as sole authority, how claims enter the register, the rhetorical-ceiling enforcement that the SCR makes possible, the versioning practice, and the rationale for maintaining a register rather than just citations.

The architecture

The SCR is a numbered list of empirical claims the Just Behaving program treats as part of its working knowledge base. Each entry has several components. The claim itself, stated precisely. A confidence level, drawn from the five evidence tags developed in Section 2. The supporting sources, with citations to peer-reviewed research where applicable and explicit acknowledgment when the support is something other than peer-reviewed research. Rhetorical-ceiling notes that govern how the claim can be presented in the canon's prose. Cross-references to other SCR entries that bear on related claims.

Entries are numbered for stable reference. An entry might be SCR-005 on relational-context modulation, SCR-019 on authoritative caregiving, or SCR-029 on canine social correction patterns. The numbers do not encode importance or priority; they are administrative identifiers that allow documents across the canon to reference the same entry consistently. A document that wants to invoke a particular claim cites the SCR number, which directs the reader to the register entry where the full evidence base for the claim is documented.

Entries are organized by topical area for navigation, but the navigation structure is operational rather than conceptual. The conceptual organization of the canon (the Five Pillars, the empirical anchors, the operational documents) is established in the canon's framework documents; the SCR is the evidence-tracking infrastructure that supports the conceptual organization rather than substituting for it.

The register is versioned. As of this writing the SCR is at version 3.32, with a changelog tracking the revisions across versions. New entries are added as new claims enter the canon. Existing entries are revised as evidence develops. Some entries are retired when the evidence base no longer supports the claim at any confidence level. The version history is itself part of the register's value as a record of the program's intellectual development over time.

The role: sole authority for evidence tags

The SCR is not one source among many. It is the sole authority on what confidence level each claim in the program's canon can be presented at. A claim in any Just Behaving document cannot be tagged at a higher confidence level than its SCR entry allows. The principle is the operational core of the program's evidence discipline.

The principle's mechanism: rhetorical ceilings are enforced. An SCR entry at [Heuristic - Dog] sets the rhetorical ceiling for any document discussing the corresponding claim. The document may use the claim at the heuristic level, with appropriate hedging and acknowledgment of the limits, but cannot exceed that level. A document that finds the claim compelling and wants to present it more confidently than the SCR ceiling allows is signaled to either revise the prose to stay within the ceiling or to propose an SCR update with primary-source verification. The ceiling is binding across the canon.

This principle is what makes the SCR more than a citations list. A citations list points the reader to sources but does not constrain how the claim is presented in the prose. The SCR's rhetorical-ceiling enforcement constrains the prose. The constraint is what produces consistency: a claim referenced in five different Just Behaving documents carries the same confidence level in each, because each document is bound by the same SCR entry. The reader who encounters the claim in multiple documents finds it presented at the same evidence level each time, with the same hedging, the same acknowledgment of limits. The consistency is not the result of careful copy-editing across documents; it is the structural consequence of the SCR's authority.

The principle also matters for the program's own internal accountability. A writer working on a Just Behaving document who finds a claim's SCR ceiling constraining is forced to confront the constraint rather than working around it. The writer either accepts the constraint and revises the prose, or the writer proposes an SCR update, in which case the proposal goes through the verification process described below. There is no third option of writing more confidently than the SCR allows on the basis of the writer's personal conviction that the claim deserves more confidence. The structural constraint prevents personal-conviction overriding from entering the canon undetected.

How claims enter the SCR

A claim enters the SCR when the program needs to make the claim in its canon and when the supporting evidence has been evaluated against the program's standards. The evaluation has specific components.

Primary-source verification is required. The cited research is retrieved and read. The species in which the study was conducted is checked. The methodology is evaluated for relevance to the claim being made. The finding itself is read in the original rather than summarized from secondary sources. If the supporting research is a meta-analysis or systematic review, the underlying studies are spot-checked when feasible. The species, method, finding, and applicability to the proposed claim are all evaluated.

The evidence tag is set conservatively. If the supporting research clearly documents the claim, the tag is [Documented] with appropriate species qualification. If the supporting research is cross-species and the canine application is an inference, the tag is dual: documented for the original species and heuristic or estimated for the canine application. If the claim is the program's own observational evidence, the tag is [Observed - JB]. If the supporting evidence is heuristic, the tag is [Heuristic]. If the literature is contested or insufficient, the tag is [Ambiguous]. The principle: when in doubt, the lower confidence tag is used. The burden of proof for higher confidence is on the writer proposing the entry, not on the reviewer.

Rhetorical-ceiling notes are added. These notes specify how the claim can and cannot be presented in the canon's prose. A claim at [Heuristic - Dog] might carry a note specifying that the claim cannot be presented as a documented canine finding, must be hedged at each mention, and should reference the convergent observational and biological-plausibility evidence rather than claiming direct controlled study. The notes are not aesthetic preferences; they are the operational instructions for how the claim is to be handled in the canon.

AI assistants are part of the workflow. Large-language-model-based tools are used to surface relevant literature, format citations, draft documents, and propose SCR entries. The AI assistants accelerate the work substantially. But AI assistants do not independently upgrade confidence levels. The discipline: AI-proposed evidence claims are reviewed against primary sources before they enter the canon at any confidence level. An AI assistant might propose a [Documented - Dog] tag based on its training data; the proposal does not become the SCR's position until a human reviewer has verified the underlying research, checked the species and methodology, and confirmed the tag's accuracy. The discipline cannot be automated because evidence evaluation is judgment work. AI assistants enhance the work; they do not replace the judgment.

The rhetorical-ceiling enforcement

The rhetorical-ceiling enforcement is the operational manifestation of slippage discipline applied at the structural level. Section 5 develops slippage discipline at length; here it is enough to note that the ceiling enforcement is what gives slippage discipline a structural rather than an editorial-only foundation.

An SCR entry at [Heuristic] sets the ceiling. A document writing about the corresponding claim may use heuristic language, may acknowledge the limits, may hedge appropriately. The document cannot present the claim as [Documented] or with the rhetorical confidence of documented findings, regardless of how compelling the claim feels to the writer.

An SCR entry at [Documented] permits stronger language but does not require it. A document may present a [Documented] claim with documented-level confidence, with appropriate citation and context. The document is not required to over-hedge a claim that the SCR has determined to be documented; the documented tag is meaningful and the prose can rely on it.

The ceiling enforcement creates a specific operational pattern across the canon. When a writer is drafting a document and encounters a claim that they want to present confidently, they check the SCR. The SCR's ceiling governs the prose. If the writer finds the ceiling constraining and believes the claim deserves higher confidence, the writer must propose an SCR update with primary-source verification rather than working around the ceiling in the prose. The structural constraint is what prevents the conviction-based overriding that produces overclaiming.

A writer who finds an SCR ceiling constraining without being able to provide primary-source verification for a higher confidence level is signaled, by the structural constraint, that the claim does not deserve higher confidence in the canon. The writer may be right that the claim is true; the writer may be wrong; the discipline does not depend on resolving the writer's intuition. The discipline depends on the structural requirement that confidence be earned through primary-source verification rather than through prose strength.

The versioning practice

The SCR is versioned because claims are revised over time. The version history is not a defect of the register; it is part of the register's value.

Some claims are upgraded. A claim that was initially tagged [Heuristic - Dog] may be upgraded to [Documented - Dog] when new peer-reviewed canine research directly supports the claim at the higher confidence level. The upgrade requires primary-source verification of the new research, evaluation of methodology, and explicit recording in the version history. The upgrade is conservative; the burden is on the new evidence to clearly support the higher confidence level.

Some claims are downgraded. A claim that was tagged at one confidence level may be downgraded when new evidence complicates the original support, when methodological criticism of the original research surfaces, or when subsequent studies fail to replicate the original finding. The downgrade is also recorded in the version history.

Some claims are retired. A claim that does not survive scrutiny is removed from the register. The retirement is recorded in the version history with the reason. Documents in the canon that had referenced the retired claim are flagged for revision.

The version history is, at the level of the program's intellectual practice, the record of the program's willingness to change. A program that has never revised a claim in its register is either not engaging seriously with new evidence or is too young to have encountered the need. The version history's existence is evidence that the program does engage with new findings and does update positions when the evidence supports updating.

A worked example of revision: the Marshall-Pescini and colleagues 2017 Proceedings of the National Academy of Sciences paper documenting that wolves outperform dogs on conspecific cooperation tasks complicated the simple "domestication made dogs superior at social cooperation" framing that some canine cognition writing had implied. The Just Behaving program integrated this finding into JB_Free_Ranging_Dogs.md Section 4 rather than ignoring it. The integration required acknowledging that the cross-species cognition picture is more nuanced than earlier framings had suggested, with both species sharing substantial cognitive infrastructure and differing in specific domains tracking their socioecologies. The integration was a small revision to the framing of comparative cognition claims, recorded in the relevant SCR entries.

Why a register rather than just citations

A document with citations is doing part of the work of evidence discipline. A register is doing the whole work. The distinction is structural.

Citations tell the reader where a claim comes from. The reader can follow the citation to the underlying source and evaluate the source on its merits. This is valuable, but it is also limited. The reader has to do the work of following each citation, evaluating each source, and figuring out whether the claim is presented at the appropriate confidence level. A document with extensive citations can still overstate what the citations support, if the writer has interpreted the citations more confidently than warranted. The reader's only protection is doing the work of checking each citation independently.

A register tells the reader what confidence the claim carries. The reader does not need to evaluate each citation independently to know that a claim is being presented at [Heuristic] rather than [Documented]. The register has done that work and made its judgment visible. The reader can choose to dig into the underlying citations if they want, but they do not need to dig in to evaluate the program's epistemic position on the claim.

Citations can be selectively presented. A writer can cite the studies that support a preferred position and omit the studies that complicate it. The reader who has not done independent research on the topic may not realize that the cited studies are not the complete picture. A register, by existing in one place and being audited as a whole, makes selective presentation harder. The register has to account for the literature on the topic, including findings that complicate the program's positions. The audit is at the level of the register, not at the level of individual documents.

A register enables cross-document consistency. The same claim cited in five different documents carries the same confidence level in each, because each document references the same SCR entry. The reader who reads multiple Just Behaving documents finds the program presenting the same claim with the same confidence in each, with the same hedging, the same acknowledgment of limits. The consistency is not editorial happenstance; it is structural.

A register makes the program's evidence base visible to external readers as an integrated whole. The reader who wants to understand the program's epistemological position on a topic can read the relevant SCR entries directly. The reader who wants to audit the program's claims across the canon can do so by examining the register. The visibility is what allows external scrutiny. Without the register, the evidence base is distributed across documents and not easily audited as a system.

The register is, in operational terms, what makes the rest of the evidence discipline structural rather than personality-dependent. A program that relies on individual writers' and reviewers' personal vigilance to maintain evidence discipline is vulnerable to the inevitable variations in individual attention. A program that has a structural register, with rhetorical-ceiling enforcement and versioning, has built the discipline into the architecture in a way that does not depend on any individual writer or reviewer being uniformly vigilant. The structure carries the discipline. The structure is what makes the discipline durable across the canon over time.

Section 5: Slippage Discipline

The single most common vulnerability in evidence-based writing across fields is slippage. The pattern is recognizable in scientific writing, in popular science writing, in policy documents, in clinical guidelines, in any genre where evidence is meant to anchor claims. It is also recognizable in the dog world's writing about training, behavior, and breeding. The Just Behaving program treats slippage as a standing audit target rather than as a one-time editorial check. This section names the pattern, explains why it happens, describes how to detect it, walks through how to fix it, illustrates the discipline with worked examples, and concludes by treating slippage discipline as an ongoing rather than completed practice.

What slippage is

Slippage is the pattern in which a claim is properly hedged in an evidence-grounding section of a document but presented as settled science in a later philosophical, operational, or rhetorical section of the same document. The first introduction of the claim is careful; subsequent mentions soften the hedging or drop it entirely. By the document's closing, the claim reads as established even though the underlying evidence base has not changed.

A worked illustration of the pattern. A document introduces a claim early on with appropriate care: "The cross-mammalian convergence on certain features of authoritative caregiving is consistent with the canine application, though direct controlled comparison of authoritative caregiving in dogs has not been conducted at the level the human-developmental literature has tested the framework [Documented - Human] [Heuristic - Dog]." The hedge is explicit, the dual tagging is in place, the reader is told that the canine application is the program's working position rather than a documented finding.

Several pages later, the same document discusses how the framework operates in practice. The writer, now in operational rather than evidence-grounding mode, writes: "Authoritative caregiving produces the developmental outcomes the dog needs." The hedge has disappeared. The dual tag is gone. The claim now reads as if the canine application were a documented finding, when the underlying evidence base has not changed since the careful introduction earlier in the document.

This is slippage. The same claim has slid from a properly-hedged presentation to a confidently-stated presentation across the document. A reader who reads only the first introduction encounters the careful hedge. A reader who reads only the operational section encounters the confident statement. A reader who reads both encounters two versions of the same claim, with the second version presenting confidence the underlying evidence does not support.

Slippage is not lying. Most slippage is unintentional. Writers who hedge in evidence-grounding sections often forget to re-hedge when they return to the same claim in operational or philosophical sections. The writing styles appropriate to evidence-grounding sections and to operational sections are different. The evidence-grounding style is naturally hedged and tentative; the operational style is naturally confident and direct. The shift between styles is what produces slippage, even in writers who are not trying to overclaim.

The consequence is that slippage produces material that, on careful reading, contains internal inconsistency. The same claim is presented at two different confidence levels in the same document. The reader who notices this can recover the actual evidence base by re-reading the evidence-grounding section, but the reader who reads sequentially or who focuses on the operational sections receives the more confident version without the corresponding evidence.

Why slippage happens

Several factors contribute to the prevalence of slippage in evidence-based writing.

Confidence is rhetorically natural. Direct sentences read more clearly than hedged sentences. The writer who wants the prose to be readable, accessible, and persuasive naturally drifts toward direct statement. The hedge that was carefully constructed in the evidence-grounding section feels rhetorically awkward when carried into later sections. The writer's editorial instinct is to smooth the prose, which often means smoothing out the hedges.

Writers who hedge in the evidence-grounding section often forget to re-hedge in later sections. This is not malicious; it is the natural pattern of attention. The evidence-grounding section is where the writer's attention is focused on the evidence base. The operational section is where the writer's attention shifts to application. When attention shifts, the careful hedging from the evidence section is not always carried forward with the rest of the writer's work.

Readers and editors expect different rhetorical registers in different sections of a document. Evidence-grounding sections are expected to be hedged and tentative; operational sections are expected to be confident and direct. The mismatch between the expected register and the discipline of consistent hedging creates pressure toward slippage. A document that is fully hedged throughout reads, to many readers, as less confident or less useful than a document that is hedged where evidence is presented but direct where application is described.

The slippage pattern is rarely intentional. It is the natural drift of careful writing if no audit catches it. The discipline of slippage detection is not a discipline of preventing slippage in the first place (which is difficult and may not be fully achievable); it is the discipline of catching slippage during review and correcting it before publication.

How to detect it

Slippage detection is editorial work. It requires re-reading the document with specific attention to the consistency of confidence levels across mentions of the same claim.

For each load-bearing claim in a document, the reviewer checks the tag at the first mention and then checks the rhetorical register of every subsequent mention. The reviewer looks for specific shifts. A shift from hedged language to declarative language. A shift from "appears to" to "is." A shift from "this suggests" to "this proves." A shift from "the literature documents" to "we know." A shift from acknowledging the limits to omitting them. Each of these shifts is a signature of slippage.

The reviewer also checks the closing sections of the document with particular attention. Closings tend to be where slippage is most pronounced, because closings are written in a summative and confident register that pulls toward declarative statement. A claim that was hedged in the document's body but is presented declaratively in the closing has slid across the document. The closing is the natural concentration point for slippage and the natural focus point for slippage audit.

The reviewer also checks transitions between evidence-grounding and operational sections. The transition is the moment when the writer's attention shifts and the hedging is most likely to fall away. A reviewer who watches the transitions closely will catch the slippage that emerges at the boundary between the evidence section and what comes after.

The reviewer also checks the document against the SCR. The SCR's rhetorical-ceiling notes for each claim specify how the claim is to be presented. If the document presents the claim more confidently than the SCR ceiling allows, the slippage is structural and the prose needs revision regardless of how natural the prose reads.

How to fix it

Slippage is fixed by holding the confidence level at every mention of the claim, not just at the first introduction.

A [Heuristic] claim remains a [Heuristic] claim in the closing paragraph. The hedging that was constructed in the evidence-grounding section is preserved in the operational and philosophical sections. The reader encounters the claim with consistent confidence across the document, with the SCR's rhetorical ceiling visible in every relevant passage.

A [Documented - Rat] [Heuristic - Dog] claim remains dual-tagged in every mention. The canine application remains at heuristic confidence regardless of how compelling the rat finding is. The dual tagging is the architectural prevention of cross-species laundering, applied at every mention, not just at the first.

Specific sentences that drift toward higher confidence than the tag supports are rewritten. The drift might be obvious ("this is how attachment works") or subtle ("attachment functions through the secure-base architecture"). In either case the rewrite preserves the hedge: "in human developmental research, attachment functions through the secure-base architecture, with biologically plausible canine application that has not been formally tested in the controlled-comparison sense [Documented - Human] [Heuristic - Dog]." The rewritten sentence is longer and less rhetorically smooth than the drifted sentence. That is the cost of the discipline. The cost is accepted because the discipline serves the reader.

This work is editorial. It cannot be automated. AI tools can flag potential slippage by identifying confidence-language shifts across a document, but the judgment of whether a particular passage is slipping requires reading the passage in context and evaluating it against the SCR's ceiling. The editorial judgment is the work; the AI tools accelerate the work but do not replace it.

Worked examples

Two specific examples from the program's recent canon illustrate the slippage discipline in operation.

The relational-context firewall claim from JB_Indirect_Correction.md Section 4. The relational-context firewall is the program's claim that the developmental outcome of a correction depends substantially on the relational context within which the correction is delivered, not only on the technical operant classification of the correction. The mechanical operations of indirect correction (body pressure, calm vocal markers, deliberate gaze, quiet disengagement) can be classified in operant-conditioning terms as positive punishment, negative punishment, and negative reinforcement contingencies. The program accepts these classifications. The program's claim is that the developmental outcome of corrections delivered within a secure relationship is categorically different from the developmental outcome of corrections delivered outside a secure relationship, despite the operant classification being the same.

This claim is tagged [Heuristic - Dog] with SCR-005 (the relational-context modulation entry). The hedging is substantial. The published welfare-cost literature on aversive training methods documents costs that the program does not dispute (Ziv 2017; Vieira de Castro et al. 2020; AVSAB 2018). The program's claim is not that these costs do not exist; the claim is that the operational features of indirect correction within secure relationships (brief duration, proportional intensity, immediate re-engagement, embedded in caregiver-attachment) are categorically distinct from the operational features of the methods the welfare-cost literature studies, and that the developmental outcomes differ accordingly.

The slippage discipline as applied to this claim: at first introduction in JB_Indirect_Correction.md Section 4, the claim is presented with explicit acknowledgment of the welfare-cost literature, with the dual nature of the operant analysis explicitly developed, with the firewall claim held at [Heuristic - Dog]. The heuristic level is held at every subsequent mention in the section. The section's closing does not soften into "this is settled science." The acknowledgment that a board-certified veterinary behaviorist could read the section carefully and still disagree with the program's position, and that the disagreement would be reasonable, is preserved through the section. The reader encounters the firewall claim at the heuristic level in every mention, with the exposed flank named rather than hidden.

This is what the slippage discipline looks like in operation on a contested claim. The claim is not promoted in the prose to a confidence level the SCR does not support. The hedge that is constructed at first introduction is preserved through the section. The reader who reads the section can evaluate the claim at the actual confidence level the program is offering.

The modern-to-ancestral continuity claim from JB_Free_Ranging_Dogs.md Section 7. The modern-to-ancestral continuity claim is the program's claim that modern free-ranging dogs occupy a niche plausibly continuous with the ancestral commensal niche of the species, and that their behavior therefore provides indirect evidence about what the species evolved to experience. The claim grounds the broader commensal-pathway thesis that animates the program's Origins documents.

The claim is tagged [Heuristic - Dog] with SCR-001 (the commensal pathway entry). The hedging is explicit: modern free-ranging dogs are not the proto-dogs of fifteen thousand to thirty thousand years ago; they are post-domestic dogs in a niche plausibly continuous with the ancestral one but not identical with it; the continuity is biological and ecological rather than historical-direct; the genetic-history literature shows that modern dog populations have been substantially modified by subsequent selection, including breeding-program selection in the past several centuries.

The slippage discipline applied to this claim: JB_Free_Ranging_Dogs.md Section 7 ("Limits and Cautions") is specifically constructed to hold the heuristic level explicit. The section's "Gap between modern free-ranging dogs and ancestral commensal proto-dogs" subsection names the limit directly. The heuristic tagging is preserved in every section of the document that invokes the claim. The closing landmark statement ("They didn't learn to live with humans by us following them. They followed us.") is anchored in the empirical foundation the document has developed, but is presented in the closing in the same heuristic register the rest of the document has maintained.

The slippage that the document specifically resists: the temptation to translate the heuristic continuity claim into a documented claim about the ancestral process in the document's closing. The closing could have said "this is how dogs evolved." The discipline kept the closing at the level the underlying evidence supports: this is the program's working interpretation of what the modern evidence converges on, held at heuristic confidence, with the gap to ancestral observation explicitly acknowledged.

Slippage as a standing audit target

The Just Behaving program treats slippage as an ongoing audit concern rather than a one-time editorial check. New documents are audited for slippage during review. External reviewers are specifically asked to flag slippage they detect. The discipline is honest about the difficulty: even disciplined writers do this; the discipline is the catch, not the prevention.

The standing audit is part of what makes the discipline durable. Slippage is the natural drift of careful writing if no audit catches it. The audit is the mechanism that catches the drift. Without the audit, the discipline is one-time editorial work that does not survive the program's growth and the continued production of new documents. With the audit, the discipline is sustained across the canon as new documents enter and as existing documents are revised.

The audit is also a contribution to the program's intellectual humility. The acknowledgment that even disciplined writers slip is the acknowledgment that the discipline is hard and ongoing rather than a state the program has achieved. The discipline is the practice of catching the slippage that emerges despite the writer's best intentions. The catching is the work. The work continues.

Section 6: What This Looks Like in the Canon

The discipline developed in Sections 1 through 5 is abstract until it is made concrete in the actual practice of producing documents. This section walks through four worked examples drawn from the recent canonical satellite documents in the Just Behaving canon. Each example takes a specific load-bearing claim, walks through how the evidence discipline operates on the claim, names the exposed flank that the discipline acknowledges, and shows what the discipline produces in the actual prose the reader encounters.

The worked examples are not the complete inventory of the discipline's operation. Every Just Behaving document contains many empirical claims, each of which has been through the discipline. The four examples in this section are selected because they are load-bearing claims in their respective documents, because they illustrate different aspects of the discipline (heuristic claims, cross-species translation, ancestral inference, framework synthesis), and because the worked examples make the discipline visible in operation rather than in abstract.

Example 1: The Relational-Context Firewall

The first worked example is the relational-context firewall claim from JB_Indirect_Correction.md Section 4. The document, one of the program's recent canonical Pillar satellites, develops the Indirect Correction pillar's evidence base, signal vocabulary, and operational practice. Section 4 of that document takes up the operant-conditioning analysis of indirect correction, which is the section's most exposed flank.

The claim. Indirect Correction's mechanics can be described in operant-conditioning terms. A body block that decreases the future likelihood of a behavior is positive punishment in the operant sense. Quiet disengagement that decreases a behavior through the removal of attention is negative punishment. The release of pressure when the dog adjusts is negative reinforcement. These operant classifications are accurate as descriptions of the local mechanics. The Just Behaving claim is that the operant classification, while accurate for the local mechanics, does not determine the developmental outcome, because the developmental outcome depends substantially on the relational context within which the correction is delivered. The same operant operation, delivered within a secure caregiver-attachment relationship in brief, proportional, immediately-re-engaged form, produces a categorically different developmental outcome from the same operant operation delivered outside such a relationship, in extended, escalated, relationally-cold form.

The evidence base. Three strands of evidence anchor the claim. First, the published welfare-cost literature on aversive training methods documents costs that the program does not dispute. Ziv's 2017 Journal of Veterinary Behavior review synthesized seventeen studies on the welfare impact of aversive techniques. Vieira de Castro and colleagues' 2020 PLOS ONE paper documented elevated stress-related behaviors in dogs trained at schools using aversive methods. China and colleagues' 2020 Frontiers in Veterinary Science paper compared e-collar training against reward-based controls and found welfare costs without efficacy advantages. The AVSAB 2018 position statement synthesizes the field's clinical position. These findings are [Documented - Dog] and the program accepts them.

Second, the cross-mammalian literature on attachment-buffered stress response is substantial. Caregiver presence reduces offspring stress response across many mammalian species studied. Myron Hofer's 1994 work on hidden regulators in attachment, Regina Sullivan's 2017 review on attachment figure regulation of infant brain and behavior, and the broader mammalian-parenting literature converge on the importance of the caregiver relational frame in offspring stress regulation. The findings are [Documented - Mammal] across the broader mammalian literature.

Third, the canine-specific literature on attachment, caregiver-buffered stress response, and handler effects is consistent with the claim. The Topál and colleagues 1998 Journal of Comparative Psychology paper established canine attachment behavior with the Strange Situation procedure. The Schöberl and colleagues 2017 work documented owner presence as a buffer of dog stress response. The Sundman and colleagues 2019 Scientific Reports paper on hair cortisol coupling documented long-arc physiological synchrony between dogs and owners. The Cimarelli and colleagues 2016 work documented that owner interaction styles correlate with dog responses to social threat. These findings are [Documented - Dog].

The tagging. The firewall claim itself, that the operant classification does not determine the developmental outcome because the relational context matters, is tagged [Heuristic - Dog] with SCR-005 (the relational-context modulation entry). The component findings are documented at their respective species levels. The synthesis claim that the relational context modulates the developmental outcome of operationally-equivalent corrections is the heuristic claim that the program holds at heuristic confidence.

The discipline in operation. The firewall claim is presented at first introduction with explicit acknowledgment of the welfare-cost literature. The acceptance of the operant classifications is explicit. The dual nature of the analysis (the mechanics are correctly classified; the outcome depends on more than the mechanics) is developed with care. The heuristic level is held at every subsequent mention in the section. The closing of the section explicitly acknowledges that a board-certified veterinary behaviorist could read the section carefully and still disagree with the program's position, and that the disagreement would be reasonable. The acknowledgment is what the discipline produces: not the assertion that the program has resolved the question, but the honest acknowledgment that the question is contested and the program is offering its position with appropriate hedging.

The exposed flank. The flank the discipline names: a controlled comparison documenting that operationally-equivalent corrections produce different developmental outcomes in dogs when delivered within versus outside secure relationships does not exist in the literature. The relational-context firewall is the program's heuristic synthesis grounded in convergent observational evidence, cross-mammalian biological plausibility, and clinical resonance, but the direct experimental test in dogs has not been conducted. The discipline names this flank explicitly rather than hiding it.

What the discipline produces in practice. The reader of JB_Indirect_Correction.md Section 4 encounters the firewall claim as the program's heuristic position, with the supporting evidence and the limits both visible. The reader can choose to accept the program's position, to remain skeptical, or to develop their own intermediate position. The discipline makes the choice available to the reader rather than presenting the position as a foregone conclusion.

Example 2: The Cross-Species Translation of Baumrindian Categories

The second worked example is the cross-species translation of Baumrindian parenting categories from JB_Structured_Leadership.md Section 1. The document, one of the program's recent canonical Pillar satellites, develops the Structured Leadership pillar's evidence base and operational practice. Section 1 of that document opens with the parenting-typology framework that anchors the pillar's interpretive claims.

The claim. Diana Baumrind's developmental psychology research, developed at Berkeley from the 1960s through the subsequent decades and extended by Eleanor Maccoby and John Martin's 1983 work, established the parenting-typology framework that distinguishes authoritative (high warmth, high demandingness), authoritarian (low warmth, high demandingness), permissive (high warmth, low demandingness), and uninvolved (low both) parenting styles. The framework has held up across decades of research in human populations, with consistent findings on developmental outcomes for offspring raised under each style. The Just Behaving claim is that the framework, originally developed and tested in humans, is reasonably extended to dog raising through the cross-mammalian convergence on parental-care patterns and the canine handler-effects literature consistent with the framework. Authoritative caregiving as understood in the framework is the program's translation of the framework into the canine context.

The evidence base. The human-developmental research on the Baumrindian framework is extensive. Baumrind's 1966 Child Development paper, the 1971 monograph in Developmental Psychology Monographs, and the subsequent decades of replication and refinement establish the framework as one of the most rigorously-tested in developmental psychology. The Maccoby and Martin 1983 chapter in the Handbook of Child Psychology formalized the two-axis structure. The Steinberg and colleagues 1994 Child Development paper, the Lamborn and colleagues 1991 work, and the broader literature have documented the developmental outcomes across many populations. Martin Pinquart's 2017 meta-analyses in Developmental Psychology and Educational Psychology Review synthesize the cross-cultural evidence. The framework is [Documented - Human] at the level of the human research base.

The cross-mammalian convergence on parental-care patterns has its own substantial literature. Hofer's 1994 work on hidden regulators in attachment, Sullivan's 2017 review on attachment figure regulation, the Suomi tradition of rhesus macaque maternal-style research, and the rat maternal-care work of Meaney and Champagne all document within-species variation in maternal-care quality with developmental consequences for offspring that parallel the human findings in important respects. The cross-mammalian convergence is [Documented - Mammal] with the appropriate caveat that the specific operationalizations differ across species.

The canine-specific evidence base, while thinner, is consistent with the framework's plausible extension. The canine attachment literature (Topál et al. 1998 and subsequent work) documents that dogs form attachment bonds with their humans that parallel the infant-caregiver dynamic. The canine handler-effects literature (Cimarelli et al. 2016; Dodman et al. 2018) documents that owner caregiving practices shape dog behavior across measurable dimensions. The autonomic synchrony work (Sundman et al. 2019) documents long-arc physiological coupling between dogs and owners. These findings are [Documented - Dog] at the level of the specific empirical findings.

The tagging. The framework itself is [Documented - Human]. The cross-mammalian convergence is [Heuristic - Mammal] with the operationalization caveat. The canine application is [Heuristic - Dog] with SCR-019 (the authoritative caregiving entry). The triple-tagging pattern preserves the reader's ability to distinguish the human-developmental evidence (strong) from the cross-mammalian inference (plausible but not formally operationalized across the relevant range of species) from the canine application (the program's working position for the dog-raising context).

The discipline in operation. JB_Structured_Leadership.md Section 1 establishes the dual-tagging discipline explicitly at the section's close: "The Just Behaving position holds this cross-species translation as a [Heuristic - Mammal] / [Heuristic - Dog] claim, not a [Documented - Dog] claim. Baumrind's framework is documented in humans. The mammalian convergence is documented across multiple non-human mammalian species. The canine application is a plausible and clinically useful extension supported by adjacent canine evidence, but the framework has not been directly validated in the controlled-comparison sense for the canine population."

The dual-tagging discipline is then held throughout the rest of the document. Section 2 (the attachment-theory substrate) maintains the tagging through the canine attachment literature. Section 4 (the authoritative posture in practice) translates the framework into operational practice while preserving the heuristic confidence for the canine application. Section 6 (the twin functions) develops the secure-base and safe-haven functions at the operational level with the same tagging discipline. Section 8 (the life arc) traces the practice across the dog's life with the same hedging maintained.

The exposed flank. The flank: a controlled-comparison study formally operationalizing Baumrindian categories in dog raising and documenting developmental outcomes has not been conducted. The cross-mammalian convergence is documented across multiple species, but the specific operationalization of authoritative-versus-authoritarian-versus-permissive parenting in dogs has not been published in a form that would support the canine application at [Documented - Dog] confidence. The framework's canine extension is the program's principled position grounded in convergent evidence, not an empirically-validated finding at the controlled-comparison level.

What the discipline produces. The reader of JB_Structured_Leadership.md Section 1 encounters the framework as the program's principled cross-species translation. The triple tagging is visible. The reader can verify the human-developmental evidence by consulting the cited research. The reader can evaluate the cross-mammalian convergence on its own merits. The reader can decide whether the canine extension at heuristic confidence is reasonable enough to accept as a working position. The discipline gives the reader the information needed to make this evaluation rather than presenting the framework as a foregone conclusion.

Example 3: The Modern-to-Ancestral Continuity

The third worked example is the modern-to-ancestral continuity claim from JB_Free_Ranging_Dogs.md Section 7. The document, the program's recent canonical empirical-anchor document on free-ranging dogs, synthesizes the published literature on free-ranging domestic dogs across multiple research populations. Section 7 of that document is the limits-and-cautions section that the program describes as the document's most important credibility-builder.

The claim. Modern free-ranging dogs occupy a niche plausibly continuous with the ancestral commensal niche of the species, and their behavior therefore provides indirect evidence about the species's evolved profile. The claim grounds the broader commensal-pathway thesis that animates the program's Origins documents.

The evidence base. Several strands of evidence converge on the claim. The free-ranging dog literature documents the modern niche in detail. The Bonanni and colleagues 2017 Behavioral Ecology paper on age-graded dominance hierarchies, the Cafazzo and colleagues 2010 work on dominance signaling, the Pal 2005 and 2008 work on parental care and developmental behavior, the Bhadra group's work on Indian street dog cognition and social behavior, and the broader free-ranging dog literature establish what modern free-ranging dogs do in human-adjacent contexts. These findings are [Documented - Dog]. The Vanak and Gompper 2009 Mammal Review paper on dogs as commensal carnivores characterizes the ecological niche in detail. The Hughes and Macdonald 2013 Biological Conservation paper on global free-ranging dog distribution documents the demographic scale.

The genetic-history literature documents some continuity in modern dog populations across the longer timescales. The Botigué and colleagues 2017 Nature Communications paper on ancient European dog genomes documents genetic continuity in European dog populations across the Neolithic period. The Bergström and colleagues 2020 Science paper extends the picture across broader temporal and geographic ranges. The Frantz and Wang lab tradition has documented the longer-arc canine genetic history. These findings are [Documented - Dog] at the level of the specific genetic findings.

The tagging. The modern free-ranging dog findings are [Documented - Dog]. The genetic-history findings are [Documented - Dog]. The continuity claim from modern free-ranging dogs to ancestral commensal proto-dogs is [Heuristic - Dog] with SCR-001 (the commensal pathway entry).

The discipline in operation. Section 7 of JB_Free_Ranging_Dogs.md names the continuity claim's status explicitly. The section's "Gap between modern free-ranging dogs and ancestral commensal proto-dogs" subsection directly addresses the limit: "The free-ranging dogs of today are not the proto-dogs of fifteen thousand to thirty thousand years ago. This bears stating directly because the heuristic translation from modern free-ranging dog evidence to claims about the ancestral domestication process is one of the document's most exposed flanks."

The discipline is then held through the rest of the document. Section 6 (the synthesis section) presents the continuity claim at heuristic confidence with the limit named. The Closing landmark statement ("They didn't learn to live with humans by us following them. They followed us.") is anchored to the empirical foundation the document has developed, but is presented in the closing in the same heuristic register the rest of the document has maintained. The closing does not promote the heuristic claim to a documented one.

The exposed flank. The flank: the ancestral commensal proto-dogs of fifteen thousand to thirty thousand years ago cannot be directly observed. The continuity from modern free-ranging dogs to ancestral proto-dogs is an inference from convergent biology and ecology, not a direct observation. Modern free-ranging dogs are selected by the modern niche, which is similar to but not identical with the ancestral one. The documentation of continuity is genuinely heuristic.

What the discipline produces. The reader of JB_Free_Ranging_Dogs.md encounters the modern free-ranging dog evidence at documented confidence and the continuity-to-ancestral claim at heuristic confidence. The distinction is preserved through the document. The closing statement on origins lands as the program's working interpretation grounded in the convergent evidence, not as a documented claim about the ancestral process. The reader can accept, evaluate, or remain skeptical about the continuity inference, with the evidence base for both the documented and the heuristic components visible.

Example 4: The Commensal Pathway Itself

The fourth worked example is the commensal-pathway claim itself, which threads through the Just Behaving canon as the program's interpretive thesis about the species's evolutionary history.

The claim. Domesticated dogs evolved through a commensal pathway in which proto-dogs self-selected toward human-adjacent niches across the long domestication period, with the Five Pillars of the Just Behaving framework (Mentorship, Calmness, Structured Leadership, Prevention, Indirect Correction) describing the selection pressures that operated during this process. The framework is, on this interpretation, not a set of techniques layered onto an arbitrary species but a named description of how the species's social biology was actually shaped by its evolutionary history.

The evidence base. Multiple strands of evidence converge on the claim. The broader self-domestication literature (the Coppinger and Coppinger 2001 framework, the Brian Hare and colleagues work on the self-domestication hypothesis, and the broader genetic-history work) provides the framing. The free-ranging dog literature (documented in JB_Free_Ranging_Dogs.md) provides the modern empirical foundation. The comparative wolf-dog cognition literature (Marshall-Pescini and Range tradition, including the 2017 PNAS finding that wolves outperform dogs on conspecific cooperation, and the broader Wolf Science Center work) characterizes the cognitive divergence. The mammalian parenting literature provides the cross-species context. The findings are [Documented - Dog] for the modern free-ranging dog evidence, [Documented - Wolf] and [Documented - Dog and Wolf] for the comparative cognition work, and [Documented - Mammal] for the cross-mammalian parenting work.

The Just Behaving interpretive claim, that the Five Pillars describe the selection pressures that operated during the commensal-pathway evolution, is the program's distinctive interpretive contribution. The interpretive claim integrates the convergent evidence into a particular reading of what the species evolved into and what raising a dog within an evolutionary-attentive framework looks like. The interpretive claim is [Heuristic - Dog].

The tagging. The empirical components are [Documented] at their respective species levels. The interpretive synthesis (the Five Pillars as the program's reading of the selection pressures) is [Heuristic - Dog].

The discipline in operation. The commensal-pathway claim is developed across multiple documents in the canon, with the empirical foundation in JB_Free_Ranging_Dogs.md, the philosophical-historical treatment in JB_Origins_2_0.md and adjacent Origins documents, and the framework-application in the Pillar satellite documents. At each invocation, the empirical components are presented at their documented confidence and the interpretive synthesis is presented at heuristic confidence. The cross-document consistency is enforced by the SCR's ceiling on the interpretive claim, which keeps the interpretation at heuristic confidence across the canon.

The slippage discipline applies particularly to this claim because the claim is intellectually compelling to the program and is at the center of the program's framework. The temptation to write the interpretive synthesis with the rhetorical confidence of a documented finding is real. The discipline of holding the claim at heuristic confidence in every mention, in every document, requires editorial attention because the natural drift of the prose is toward stronger confidence than the SCR ceiling allows.

The exposed flank. The flank: the JB-specific interpretive claim that the Five Pillars describe the selection pressures that operated during commensal-pathway evolution is the program's distinctive interpretive contribution, and it is also where the program's evidence discipline is most demanded. The claim is biologically plausible, internally coherent, and consistent with the convergent evidence, but it is the program's interpretation rather than a documented finding. The discipline names this rather than hiding it.

What the discipline produces. The reader of the Just Behaving canon encounters the commensal-pathway thesis at the appropriate confidence level: the empirical foundation at documented confidence, the interpretive synthesis at heuristic confidence. The program is offering its working interpretation grounded in the convergent evidence, not claiming that the interpretation has been empirically validated as the only defensible reading of the evidence. The reader can accept the interpretation, evaluate it, or develop their own alternative reading. The discipline makes the program's interpretive contribution visible as interpretation rather than as documented science.

What this shows

The four worked examples together illustrate the discipline in operation across different kinds of claims. The relational-context firewall is a heuristic synthesis grounded in convergent evidence. The Baumrindian translation is a cross-species inference handled with dual tagging. The modern-to-ancestral continuity is a heuristic inference with explicit limit-acknowledgment. The commensal-pathway thesis is an interpretive synthesis held at heuristic confidence across the canon.

The discipline produces, in each case, prose that the reader can audit. The evidence tags are visible. The supporting research is cited. The limits are named. The exposed flanks are acknowledged rather than hidden. The reader can choose how to engage with each claim, with the information needed to make the choice provided in the prose rather than hidden behind confident rhetoric.

The discipline is what produces the prose. Each Just Behaving document operates under the same discipline applied to its specific claims. The reader can audit any Just Behaving document by reading the evidence tags, checking the SCR references, and tracking the consistency of confidence levels across mentions. The discipline is transparent and verifiable. The transparency is the practice in operation.

Section 7: The Negative Space

A discipline is defined as much by its prohibitions as by its commitments. The Just Behaving program's evidence discipline forbids specific kinds of writing, specific kinds of claims, and specific kinds of presentation. The prohibitions are not optional features of the discipline; they are the operational form of the discipline as much as the commitments are. This section names the prohibitions, explains their grounds, and applies them reflexively to the present document. A meta-document about evidence discipline could itself fall into the failure modes the discipline forbids. The reflexive application of the prohibitions to the document about the prohibitions is what allows the document to argue for the discipline while observing it.

No grandiosity

The program does not describe itself or its methods as groundbreaking, revolutionary, unique, unprecedented, or first-of-its-kind. The prohibition extends to softer versions: pioneering, transformative, paradigm-shifting, definitive. These descriptions are marketing language. They describe what a program would say if its goal were to attract attention rather than to engage readers with the merits of its claims.

The grounds for the prohibition: grandiosity is rhetorically expensive. A program that describes itself in grandiose terms is asking the reader to take the program's self-assessment as part of the evidence base, when the program's self-assessment is not evidence and should not be substituted for the claims the program is making about dogs. A reader who encounters grandiose language is being given a marketing signal rather than an evidentiary one. The grandiose language pulls attention away from the actual claims the document is making.

The application: editorial discipline at every level of the canon. Documents are reviewed specifically for grandiose language, with the words and phrases listed above flagged for revision. The discipline is operational: a writer who drafts a sentence describing the program in grandiose terms is signaled to revise the sentence regardless of how true the writer believes the description to be. The program's claims about dogs should earn the reader's attention; the program's self-description should not be doing that work.

The reflexive application to this document: a document on evidence discipline could itself be grandiose. A document arguing that evidence discipline is uncommon in the dog world and that the Just Behaving program practices it could easily slide into "we have developed a uniquely rigorous framework that the field has been waiting for." The discipline forbids this slide. The present document argues for the practice of evidence discipline; it does not celebrate the program for having adopted the practice. The voice is grounded. The program's contribution is described as one program's articulation of a discipline that other programs and practitioners could implement in their own forms. The contribution is offered, not promoted.

No industry-as-monolith framing

The program does not describe the dog training industry, the veterinary behavior community, the breeding community, or other adjacent fields as a monolithic opposition. The prohibition extends to softer versions: framings that imply the field uniformly fails in some way the program corrects, framings that group practitioners into "those who do it right" and "those who do it wrong," and framings that caricature individual practitioners as exemplars of broader field failure.

The grounds for the prohibition: many practitioners in the dog training industry, veterinary behavior, applied animal behavior, and the breeding community are skilled, ethical, and serious. The structural problems in the field are not the fault of individual practitioners; they are structural features of how the field is organized, how evidence circulates, how families encounter claims, and how the commercial environment shapes the production of advice. A critique that frames the field as monolithic opposition flattens the actual variation across practitioners and across communities. The flattening is both empirically inaccurate and rhetorically self-defeating: it converts what could be a serious structural argument into ad hominem critique.

The application: critique is structural and named at the level of frameworks, methods, or systemic patterns rather than at the level of individual practitioners. The program names the alpha framework as empirically retracted (Mech 1999) and as continuing to circulate in popular discourse despite the retraction. The program does not name individual trainers who continue to use the framework. The program critiques cross-species laundering as a pattern across the dog world's writing. The program does not critique specific writers as exemplars of the pattern. The structural critique engages with the structure; the structural critique does not personalize.

The reflexive application: this document discusses field-wide problems with evidence quality (Section 1), the methods that the welfare-cost literature documents costs from (Section 1, Section 3), and the persistence of retracted frameworks in popular discourse (Section 1, Section 3). At no point does the document name individual practitioners as exemplars of field failure. At no point does the document frame the field as a uniformly-failing monolith. The critique is structural. The structure can be engaged with on its merits. Individual practitioners across the field include both those whose practice the program would consider exemplary and those whose practice the program would consider less so; the critique is not at that individual level.

No "unique to JB" framing

The program does not claim that its specific methods are unique to itself or that other programs do not approach similar practices. The prohibition extends to softer versions: framings that imply the program has invented evidence discipline, that the Five Pillars are a Just Behaving creation rather than the program's naming of patterns the dogs themselves display, that the program's practice has no analogues elsewhere in the field.

The grounds for the prohibition: many serious breeding programs and individual practitioners approach evidence-discipline-similar practices in their own forms. The contribution of the Just Behaving program is in the specific architecture (the five primitive evidence tags, the wiki-facing composites, the SCR, the slippage audits, the dual-tagging discipline) and in the comprehensiveness of the practice across a documented canon, not in being the first or only program to think of approaching the work this way. Claiming uniqueness would misrepresent the broader field of careful practice and would do an injustice to the practitioners who have been doing comparable work in their own contexts.

The application: documents describe the program's methods without claiming uniqueness. The program's specific tools (the five-tag system, the SCR architecture, the rhetorical-ceiling enforcement) are presented as one particular implementation of evidence discipline, not as the only or the optimal implementation. The general principle that evidence discipline is achievable and valuable is presented as a principle that any program or practitioner could implement in their own form.

The reflexive application: this document is the program's articulation of its evidence discipline. The document does not claim that the Just Behaving program is the only program practicing evidence discipline. The document does not claim that the program's specific architecture is the only way to implement the discipline. The document presents the program's specific tools as one implementation, with the underlying principles (evidence is graded, cross-species inference should be marked, slippage requires ongoing audit, prohibitions are part of the discipline) as general principles available to anyone who wants to implement them in their own form. The reader is invited to evaluate the program's specific architecture on its merits and to consider what their own implementation might look like.

No "settled science" claims when the science is not settled

The program does not present heuristic claims at the rhetorical level of documented findings. The prohibition extends to softer versions: declarative language for claims that the SCR holds at heuristic confidence; "the research shows" for claims where the research is contested or insufficient; "we now know" for claims where the field has not yet resolved the question.

The grounds for the prohibition: this is slippage discipline applied to the specific case of unsettled questions. The slippage discipline (Section 5) addresses the general pattern of confidence drift across a document; the present prohibition addresses the specific case where the underlying question is genuinely unsettled in the field. Presenting unsettled questions at settled-science confidence misrepresents the state of the literature and produces false confidence in the reader's understanding.

The application: at every mention of a contested claim, the tag is maintained. Examples of contested questions in the canine behavior context include certain heritability values for specific behavioral traits; certain developmental-window specifics; the cross-population variability of free-ranging dog behavior; the question of whether the species shows true imitative learning versus emulation; certain debates in the canine cognition literature about the relative roles of evolved capacity versus social learning. For each of these, the program holds the relevant claims at heuristic or ambiguous confidence rather than presenting them as settled.

The reflexive application: this document makes claims about the state of evidence discipline in the dog world. These claims are themselves contested. The claim that evidence discipline is uncommon in the dog space is a structural observation, not a controlled comparison of evidence quality across programs. The claim that Just Behaving's discipline serves dogs better than alternative approaches is a structural argument, not a documented finding of superior outcomes. Both claims are held at [Heuristic] confidence throughout this document. The discipline is not relaxed for the document's own arguments. The document does not claim more than the underlying observations support.

No cross-species laundering

The program does not present cross-species findings as canine findings without explicit acknowledgment. The prohibition extends to softer versions: presenting rat findings with the species qualifier dropped; describing primate attachment patterns as universal mammalian patterns without the cross-species inference being marked; using human-developmental findings without the human qualifier.

The grounds for the prohibition: cross-species inference is a legitimate epistemological move when handled with discipline; cross-species laundering (presenting the rat finding as a dog finding without marking the inference) is not. The discipline produces material that the reader can evaluate; the laundering produces material that misleads the reader about what species the original research was conducted in.

The application: dual-tagging at every cross-species translation, as developed in Section 3. The tagging is applied at every mention of a cross-species claim, not just at the first mention. The slippage discipline (Section 5) applies particularly to cross-species claims, where the natural drift of prose is toward dropping the species qualifier as the writing develops.

The reflexive application: this document discusses cross-species findings (the Baumrindian framework from human-developmental research, the Meaney and Champagne work in rats, the mammalian parenting convergence, the canine attachment literature derived from the Bowlby and Ainsworth tradition in human infants). At each cross-species discussion, the species qualifier is preserved. The reader is told what species the original research was conducted in. The canine application is held at heuristic confidence where the canine literature has not directly tested the framework.

No self-promotion

The program does not use its documents as marketing material for its breeding program. The prohibition extends to softer versions: calls to action embedded in documents; references to the program's available puppies, breeding schedule, or placement process embedded in intellectual material; rhetorical framings that orient the document toward conversion of readers into customers rather than toward engagement with the document's intellectual content.

The grounds for the prohibition: the documents are intellectual work. Their purpose is to make the program's framework and evidence base visible to outside readers, to enable external scrutiny and engagement, and to serve as references for families, professionals, and other interested readers. A document that includes marketing material is doing two things at once, and the two things compete: the marketing pulls attention away from the intellectual content, and the intellectual content is compromised by the marketing's commercial framing. The cleaner architecture is to keep the intellectual work in the documents and to keep the breeding-program operations elsewhere.

The application: the documents do not include calls to action, advertisements, or promotional language for the breeding program. A reader interested in the breeding program can find information through other channels. The documents are useful as intellectual references regardless of whether the reader has any interest in the Just Behaving breeding program specifically.

The reflexive application: this document is a public-facing meta-document on the program's epistemology. The document could have been written as a marketing piece, framing evidence discipline as a feature of the program that prospective families should value. The discipline forbids this framing. The document is intellectual work. The intellectual content is the substance. Readers who are not considering the breeding program should find the document useful as a reference on evidence-based practice in canine behavior writing. The document earns its place as a reference by being intellectual work rather than marketing material.

No epistemic permissiveness

The program does not accept all evidence as equally valid. The prohibition extends to softer versions: framings that treat anecdotal evidence as equivalent to peer-reviewed research; framings that elevate traditional practice to the level of documented finding; framings that treat all opinions in a field as worthy of equal consideration regardless of their evidentiary support.

The grounds for the prohibition: not all evidence is equally valid. The whole point of evidence discipline is that the gradations matter. A program that treats anecdotal evidence as equivalent to peer-reviewed research has, in effect, abandoned the discipline regardless of what it says about its commitment to evidence. The discipline requires distinguishing among evidence categories and holding claims at the confidence levels their actual support warrants.

The application: peer-reviewed research, clinical observation, breeder observation, anecdotal report, traditional practice, and marketing claim are treated as different categories with different evidence weights. A claim supported by peer-reviewed research is presented at higher confidence than a claim supported by anecdotal report. A claim supported by clinical observation by a credentialed practitioner is presented with the practitioner's clinical experience identified as the evidence base. The categories are distinguished operationally through the tagging system.

The reflexive application: this document makes claims about the state of evidence in the dog world. The document's claims about peer-reviewed research, clinical observation, breeder observation, anecdotal report, traditional practice, and marketing claim are all anchored to specific examples and specific evidence. The document does not treat all sources as equivalent. The discipline is what the document is arguing for; the discipline is also what the document observes in its own claims.

The discipline as a unified architecture

The seven prohibitions above are not a list of independent rules. They are facets of a single architectural commitment to evidence discipline. Grandiosity is a marketing posture; industry-monolith framing is ad hominem critique; "unique to JB" framing misrepresents the field; settled-science claims when the science is not settled are confidence overreach; cross-species laundering is inference-hiding; self-promotion compromises intellectual work with commercial framing; epistemic permissiveness abandons the discipline. Each prohibition addresses a specific failure mode the discipline forbids; together they form the operational shape of what the discipline forbids.

The discipline's commitments and prohibitions together define the practice. The commitments (the tagging system, the SCR, the slippage audit, the cross-species discipline, the methodology of maintenance) are what the discipline does. The prohibitions are what the discipline does not do. The combination is the architecture. The architecture is the practice. The practice is what produces the canon.

Section 8: The Methodology of Evidence Maintenance

Evidence discipline is not a one-time architectural commitment. It is a daily practice that has to be maintained over time. New findings enter the field. Some existing claims are challenged. The literature develops. Documents in the canon are revised. AI tools accelerate the work but introduce their own discipline questions. External reviewers surface issues the program's internal review missed. The maintenance methodology is what keeps the discipline operating across these ongoing conditions rather than degrading as the canon grows and the field changes.

The role of primary-source verification

Only primary-source verification upgrades confidence. This is the core principle of the maintenance methodology. When a claim is proposed for upgrade from [Heuristic] to [Documented], the supporting research must be retrieved, read, and evaluated. The species, methodology, sample size, statistical analysis, and finding's applicability to the proposed claim are all checked against the SCR's standards.

The discipline is conservative in operation. A proposal to upgrade a claim's confidence level is not automatically approved on the basis of the writer's interpretation of the research. The reviewer reads the underlying research independently. If the research clearly supports the claim at the proposed higher confidence level, the upgrade is approved and recorded in the SCR's version history. If the research is consistent with the claim but does not directly demonstrate it at the higher confidence level, the upgrade is not approved; the writer is invited to either revise the proposed claim or to identify additional research that would directly support the upgrade.

The principle behind the conservative discipline: confidence is earned through evidence, not through writer conviction or rhetorical pressure. A writer who finds a claim compelling may be right about the claim and may be wrong. The discipline does not depend on resolving the writer's intuition. The discipline depends on whether the underlying evidence directly supports the proposed confidence level. The conservative threshold is what prevents personal-conviction overriding from entering the SCR.

The practice of primary-source verification has accelerated substantially with the development of AI-assisted literature retrieval. A writer or reviewer can now retrieve, read, and evaluate research faster than was possible a decade ago. The acceleration is genuine, and the program uses it. But the acceleration does not replace the underlying judgment. A retrieved study still has to be read by a human reviewer who evaluates its methodology and finding against the proposed claim. The AI tools surface the candidates; the human judgment evaluates them. The methodology is human review of AI-surfaced primary sources, not AI replacement of human review.

The role of AI assistants in the workflow

The Just Behaving workflow makes substantial use of AI assistants. AI tools are used to surface relevant literature, format citations, draft document sections, propose SCR entries, and run editorial checks across the canon. The acceleration is meaningful: the program's canon has expanded substantially with AI assistance in ways that would have been substantially slower without it.

The discipline that governs AI use: AI assistants do not independently upgrade confidence levels. The discipline is structural. An AI assistant might propose a [Documented - Dog] tag for a claim based on its training data and on the literature it has surfaced. The proposal does not become the SCR's position until a human reviewer has verified the underlying research, checked the species and methodology, and confirmed the tag's accuracy. The AI's proposal is treated as a starting point for human review, not as a finished evidence judgment.

The principle: evidence evaluation is judgment work. Judgment work requires human review of the underlying research in its original form, with attention to context, methodology, and applicability that AI tools cannot reliably provide. AI assistants can accelerate the work substantially, but the discipline cannot be automated. A program that treats AI-proposed evidence claims as automatically accepted into the canon has effectively delegated its evidence discipline to the AI tools, which is not the same thing as practicing the discipline.

The practical implementation: AI-drafted documents enter the canon only after human review against the SCR. AI-proposed SCR entries are reviewed against the primary sources before they are added. AI-flagged slippage detections are reviewed by human editors who confirm or override the flag. The AI tools are part of the workflow; the human judgment is what produces the canon.

The role of external review

The program seeks external review of its documents from readers with relevant expertise. The review can come from veterinary professionals, applied animal behaviorists, researchers in canine behavior, breeders with substantial experience, and skilled trainers from various traditions. The review is solicited specifically for the kind of audit that internal review may miss.

External reviewers are asked to flag specific concerns. Overclaiming: any claim presented at higher confidence than the underlying evidence supports. Slippage: any claim that shifts confidence levels across the document. Missed nuances: any findings or considerations the document has not adequately engaged with. Confidence-level errors: any tags that the reviewer believes do not accurately reflect the evidence base. Counter-evidence: any published research that complicates or contradicts the document's claims and is not acknowledged.

The reviewer's findings are integrated into revisions. The integration is not automatic; the program evaluates the reviewer's points and incorporates them where the evaluation supports the change. A reviewer's flag of overclaiming is taken seriously and addressed in revision; a reviewer's substantive disagreement with the program's interpretive position is engaged with on the merits, with the program either acknowledging the alternative position in the document, adjusting its own position, or maintaining its position with explicit engagement with the alternative.

The principle: internal review has known limitations. The writers and reviewers of a document share the document's frame of reference, share the program's interpretive commitments, and share the same blind spots. External reviewers operate from different frames of reference and surface issues that internal review reliably misses. The discipline of seeking and integrating external review is what keeps the program's documents engaging with the field rather than only with itself.

Handling new findings that complicate existing positions

New findings in the field sometimes complicate or contradict claims in the canon. The discipline: integration of new findings rather than dismissal.

The practice when a new finding emerges: the relevant SCR entries are reviewed. The new finding is read in its original form. The methodology is evaluated. The finding's implications for the program's claims are assessed. If the finding clearly complicates an existing claim, the affected SCR entries are revised, and the affected documents are flagged for revision. The change is logged in the version history.

The honest acknowledgment: this is uncomfortable. Writers can be reluctant to revise published claims, particularly claims that are part of the program's distinctive contribution to the field. The discipline is the willingness to do so when the evidence supports it. A program that does not revise claims when new evidence complicates them is not actually practicing evidence discipline regardless of what it says about its commitment.

A worked example. The Marshall-Pescini and colleagues 2017 Proceedings of the National Academy of Sciences paper documenting that wolves outperform dogs on conspecific cooperation tasks complicated simple "domestication made dogs better at everything social" framings. The Just Behaving program integrated this finding into JB_Free_Ranging_Dogs.md Section 4 rather than ignoring it. The integration required acknowledging that the comparative cognition picture is more nuanced than earlier framings had suggested. Both species share substantial cognitive infrastructure; they differ in specific domains tracking their socioecologies. The integration was a small revision to the framing of comparative cognition claims, recorded in the relevant SCR entries.

The Bonanni and colleagues 2017 Behavioral Ecology finding on age-graded tolerance-based hierarchies in free-ranging dogs complicates any flat "dominance does not exist in dogs" framing. The Just Behaving program integrated this finding consistently across the recent canonical documents that address the dominance question (JB_Indirect_Correction.md Section 1, JB_Structured_Leadership.md Section 3, JB_Free_Ranging_Dogs.md Sections 2 and 6). The integration: dominance exists in free-ranging dog social structure; the hierarchies are age-graded and tolerance-based rather than the captive-wolf alpha framework; the alpha framework that has been imported into dog training rests on a wolf-empirical claim that the field's primary scientific source publicly retracted (Mech 1999) and that does not match what the free-ranging dog literature documents. The integration preserves the dominance critique while honestly engaging with the empirical complexity the literature documents.

Retiring claims

Some claims do not survive scrutiny. They are retired. The retirement is recorded in the version history; the affected documents are revised to remove or rephrase the retired material; the program continues without the retired claim.

The discipline: retiring a claim is not a failure. It is the practice of intellectual honesty in operation. A program that has never retired a claim is either not engaging seriously with new evidence or is too young to have encountered the need. The willingness to retire claims is part of what makes the program's canon trustworthy: a canon that grows without ever retracting is producing growth without intellectual development.

The practice of retirement is rare but real. Most claims that survive the conservative initial entry into the SCR continue to be supported by ongoing evidence. The claims that are retired are typically claims that were entered at higher confidence than the evidence ultimately supports, claims that turn out to depend on research that has been criticized or failed to replicate, or claims that the program decides to drop in favor of a more careful formulation. The retirement is not common, but the practice of retirement is part of the discipline.

The version history's value: the version history is a record of the program's intellectual change. A program with a long-running canon and no version history is presenting a static intellectual position that the field's actual development does not support. The version history makes the change visible. The making-visible is part of what makes the canon credible.

The ongoing nature of the work

The maintenance methodology is not a state the program achieves. It is a practice that the program continues. The literature continues to develop. New findings continue to emerge. AI tools continue to evolve in their capacities and their failure modes. External reviewers continue to surface issues. The canon continues to grow. The discipline has to be maintained across all of this.

The honest position: the program does not claim to have arrived at a final or perfect form of the discipline. The program practices the discipline; the practice is ongoing. The discipline is refined through the work. The work continues because the practice is what produces the canon. The canon is what the program offers. The offering is what the program continues to do.

Section 9: Evidence Discipline as Service to Families

The case for evidence discipline can be made at the level of academic credibility, intellectual rigor, or field-wide responsibility. The case can also be made at the level that matters most for the Just Behaving program's daily work: families. A family bringing home a puppy is making consequential decisions under time pressure and emotional load. The discipline serves these families directly. This section makes the case at that level.

Families face consequential decisions under time pressure

A family bringing home a puppy has roughly two years of intensive developmental work ahead. The first three to four months are the most concentrated. Socialization windows open and close. The puppy's relational architecture is being built. The household's working rhythm with the dog is being established. The decisions the family makes during this period shape the dog the family lives with for the next decade.

The decisions are made without the family typically having professional training in canine behavior science. Most families have not read peer-reviewed papers in Animal Cognition or the Journal of Veterinary Behavior. Most families do not have a baseline for evaluating which sources of advice are anchored to rigorous research and which are anchored to anecdotal experience, traditional practice, or commercial marketing. The decisions are made under emotional load: the family wants to do well by the puppy, the family wants to feel confident about its choices, the family does not want to make mistakes that will affect the dog for the rest of its life.

The decisions are consequential at scale because dogs are the most common companion species. A family that makes well-informed decisions during the early developmental period typically ends up with a dog whose temperament, regulation, and capacity for partnership reflect the quality of the early decisions. A family that makes poorly-informed decisions, often through no fault of their own, often ends up with a dog whose behavioral challenges could have been substantially reduced by different early choices.

The implication: evidence quality matters at the family level, not just at the academic level. Bad evidence produces bad decisions. Good evidence produces decisions more likely to serve the dog. The family is not in a position to do the underlying research; the family is in a position to evaluate the sources that present research to them. The sources that earn the family's trust by being transparent about their evidence base are the sources that allow the family to make informed decisions. The sources that do not are the sources that ask the family to trust without the means to verify.

The accessibility of the discipline

Evidence discipline is sometimes treated as an academic concern that is inaccessible to lay readers. The Just Behaving program's practice of the discipline is intentionally accessible. Families do not need to read peer-reviewed papers to benefit from the discipline. Families need to be able to read evidence tags in plain text, understand what each tag means, and use the tags to evaluate what the program is and is not claiming.

A family encountering a Just Behaving document for the first time encounters tags like [Documented - Dog], [Heuristic - Dog], [Observed - JB], and [Documented - Human] [Heuristic - Dog] in the prose. On public wiki pages, the same discipline also appears through readable badges such as Documented-Cross-Species and Mixed Evidence when a page needs a composite display signal. The first time a family encounters these tags, the encounter is brief: a parenthetical or badge that the family may not initially understand in detail. After encountering the tags across a few paragraphs and noticing the brief glossary explanations that Just Behaving documents include for their tagging system, the family has the basic literacy needed to interpret what each tag or badge is signaling.

The literacy is not complicated. [Documented - Dog] means the claim is supported by peer-reviewed canine research. [Observed - JB] means the claim is grounded in direct Just Behaving program observation. [Heuristic - Dog] means the claim is the program's working position based on convergent observation and biological plausibility, without controlled-comparison evidence in dogs. [Documented - Human] [Heuristic - Dog] means the claim's empirical foundation is human research, with the canine application being the program's reasonable extension. Composite public badges compress these underlying distinctions for scanning, but the claim-level tags remain the primitive truth. The basic literacy can be acquired in minutes and used across the canon.

What the literacy gives the family: the ability to read the program's documents with a working understanding of what kind of evidence the program is offering at each claim. The family does not need to follow citations to evaluate every claim; the tag has done that work. The family does not need to do separate research to determine whether a confidently-stated sentence is backed by rigorous research; the tag has signaled the evidence level. The discipline has been designed to be transparent to outside readers without requiring outside readers to become experts.

The verification capacity

The discipline gives families the capacity to verify rather than trust. A family who wants to verify the relational-context firewall claim can read JB_Indirect_Correction.md Section 4, see the [Heuristic - Dog] tag, check the welfare-cost literature the document cites, evaluate the convergent evidence the document develops, and form their own judgment about whether the program's heuristic position is reasonable to them. The family does not have to accept the position on the program's authority. The family can evaluate the position on its merits.

A family who wants to verify the cross-species Baumrindian translation can read JB_Structured_Leadership.md Section 1, see the triple tagging, evaluate the human-developmental evidence, consider the cross-mammalian convergence, and form their own judgment about whether the canine extension at heuristic confidence is reasonable to them. The family is not asked to trust the framework without the means to evaluate it.

A family who wants to verify the modern-to-ancestral continuity can read JB_Free_Ranging_Dogs.md Section 7, see the explicit acknowledgment of the gap between modern free-ranging dogs and ancestral commensal proto-dogs, and form their own judgment about whether the continuity inference is reasonable to them.

The verification capacity matters because it shifts the family's relationship to the program from trust-based to evidence-based. Trust-based relationships require the family to evaluate the program as a whole rather than each claim on its merits. Evidence-based relationships allow the family to engage with each claim at the evidence level the program is offering it. The family can trust the program's discipline (the program does present claims at the confidence levels the evidence supports) without having to trust each specific claim independently of its evidence base.

What the discipline costs the program

The discipline has costs. Honesty requires naming them. The discipline produces less marketing-ready prose than the alternative. A document that says "the program produces well-mannered dogs through the application of mammalian developmental principles" is more rhetorically polished than a document that says "the program's interpretive synthesis, held at heuristic confidence, applies cross-species developmental principles to the canine context, with the canine extension grounded in convergent observational evidence and biological plausibility rather than in documented canine outcomes."

The discipline requires more editorial work. A document with carefully maintained tagging, slippage audits, cross-species discipline, and rhetorical-ceiling enforcement requires substantially more editorial attention than a document without these constraints. The editorial work is real and ongoing.

The discipline sometimes requires retiring claims the program would prefer to keep. A claim that the program has been making for years, that the program finds intellectually compelling, that the program has built parts of its canon around, may turn out to require revision or retirement when new evidence complicates it. The discipline requires the willingness to make these revisions.

The discipline costs the program rhetorical simplicity. Marketing wants simple, confident, easily-grasped claims. The discipline produces hedged, layered, evidence-grounded claims that are sometimes more complex than marketing prefers. The program accepts this cost.

These costs are real. They are also the price of the trust the discipline earns. A program that has paid these costs in its documents has earned a kind of trust that marketing-driven programs cannot earn through marketing alone. The trust is what the costs purchase.

What the discipline serves

The discipline serves several constituencies. It serves the dogs, who live with the consequences of the decisions families make on their behalf. Better evidence enables decisions that better serve the dog's developmental and welfare interests. The dogs are not in a position to advocate for evidence quality directly; the discipline is part of how the program advocates for them.

The discipline serves families, who make better decisions with better evidence. Families who can evaluate claims at their actual evidence levels are positioned to make informed choices rather than to choose between competing confident assertions without the means to evaluate them.

The discipline serves the program itself, by earning durable trust rather than marketing-driven loyalty. Marketing-driven loyalty is fragile because it depends on the marketing continuing to work. Evidence-based trust is more durable because it does not depend on the program maintaining marketing pressure; it depends on the program continuing to produce work that the discipline supports.

The discipline serves the broader field, by demonstrating that this kind of rigor is achievable for a working program operating in real-world conditions. The field benefits from any program demonstrating the achievability. The discipline is its own argument; the work is its own justification.

The discipline is not abstract. It is what the family encounters when they read a Just Behaving document. The encounter is the practice. The practice is the service. The service is what the program offers.

Section 10: The Wider Implications

The case for evidence discipline matters beyond the Just Behaving program specifically. The discipline is not a Just Behaving idiosyncrasy that other programs can take or leave on its own merits. The discipline is, in the program's view, what the broader dog world increasingly needs. This section considers the state of evidence discipline across adjacent fields, names why the dog world particularly needs greater discipline, considers what it would look like if the field moved toward greater rigor, discusses the role of model programs, and reflects on what publishing a document like this one contributes.

The state of evidence discipline in adjacent fields

Different communities within and adjacent to the dog world operate at different evidence standards. Naming the variation is part of the honest characterization of the field.

Veterinary behavior as a clinical specialty practices something resembling evidence discipline. The American College of Veterinary Behaviorists, which credentials specialists through residency training and board certification, maintains a discipline that integrates clinical experience with engagement with the published literature. The AVSAB (American Veterinary Society of Animal Behavior) position statements on punishment, on dominance theory, and on related topics articulate the field's clinical positions with citation to the supporting research. The Karen Overall Manual of Clinical Behavioral Medicine exemplifies the clinical literature's integration of evidence with practice. The veterinary behavior community is not perfect (the field has its own internal debates about evidence interpretation and clinical translation), but the field operates at a substantially higher evidence standard than the broader dog training world.

Applied animal behavior in academic settings practices peer-reviewed publication standards. Research conducted at the Family Dog Project in Budapest, the Wolf Science Center in Ernstbrunn, the Duke Canine Cognition Center, the Arizona Canine Cognition Center, the Indian Institute of Science Education and Research Kolkata, and many other academic research programs operates under the standard methodological and review processes of peer-reviewed science. The standards are imperfect (peer review has its known limitations), but they are substantially higher than the standards that govern most popular dog-behavior writing.

The dog training industry as a whole does not consistently practice evidence discipline. The field is heterogeneous. Some practitioners maintain high evidence standards in their own work and engage seriously with the research literature. Many do not. The certification landscape is fragmented: certifications from rigorous bodies (Karen Pryor Academy, the Pet Professional Guild's certifications, the IAABC certifications) coexist with certifications from less rigorous bodies, with the typical family unable to easily distinguish among them. Commercial pressure on claims is substantial: trainers competing for market share have incentives to present their methods as effective even when the evidence supporting the specific effectiveness claims is limited. The field is not uniformly low-discipline, but the field does not consistently produce evidence-disciplined writing.

The breeding community largely does not practice evidence discipline in the explicit, architectural sense. Breeders accumulate substantial practical knowledge across years of work; this knowledge is rarely documented in evidence-tagged form. Breed-club traditions and breeding-line practices carry knowledge across generations, but the knowledge is rarely audited against the published research literature. Some serious breeding programs (Just Behaving among them, and others in various breeds) approach evidence-discipline-similar practices in their own forms; these are exceptions to the general pattern rather than reflections of an established community standard.

The contrast between fields suggests that evidence discipline is achievable. It is not impossible. The veterinary behavior community demonstrates that evidence discipline can be embedded in a clinical practice tradition. Academic research demonstrates that evidence-disciplined work can be produced under peer-reviewed conditions. The discipline is a choice. Programs and practitioners that choose to practice it can do so; programs and practitioners that do not choose to practice it are not prevented by any external constraint from doing so.

Why the dog world particularly needs this

The dog world has several features that make evidence discipline particularly important.

Dogs are the most common companion species in the world. Estimates from the broader literature suggest a global dog population of seven hundred million to one billion individuals (Hughes and Macdonald 2013). The decisions families make about their dogs are consequential at scale because the population is large and the relationships are durable.

The dog world is highly commercialized. The pet industry generates hundreds of billions of dollars in annual revenue globally, with substantial portions going to food, equipment, training, veterinary care, and breeding services. The commercial pressure on claims is substantial. Companies and individual practitioners have economic incentives to present their methods, products, and services as effective. The pressure is not unique to the dog world (many fields face commercial pressure on claims), but the combination of large market size, fragmented standards, and family decision-making under emotional load makes the dog world particularly vulnerable to claims that exceed their evidence base.

The dog world is subject to popular evolutionary narratives that have outlasted their scientific support. The alpha framework is the most prominent example, retracted by its primary scientific source twenty-five years ago and still in popular circulation. The framework is not the only such narrative; various "domestication tamed everything" framings, various claims about wolf-pack social structure, and various claims about breed-specific behavioral inheritance circulate at confidence levels the underlying research does not support. The persistence of these narratives complicates families' encounters with the field.

AI-amplified content has flattened the landscape of evidence quality in ways that are difficult for families to detect. Large-language-model-generated material on dog topics has proliferated rapidly. Some of this material draws on rigorous sources; much of it draws on the broader mix of sources circulating in the field, with the AI layer producing confident prose that does not signal its underlying evidence quality. The proliferation is not bad in itself (AI tools have legitimate uses in producing accessible information), but the proliferation without evidence discipline produces material that confidently states claims at evidence levels that have not been verified.

The cumulative result: a field where evidence quality varies enormously and where families lack tools to evaluate the variation. The structural problem is upstream of any individual practitioner or program. The structural problem benefits from program-level responses.

What it would look like if the field moved toward greater discipline

A scenario in which the dog world moved toward greater evidence discipline at scale would have several recognizable features.

More programs would publish transparent evidence registers. The register form is not unique to Just Behaving; the GRADE system in clinical medicine, the Cochrane systematic review process, and the evidence-based veterinary medicine framework all use comparable approaches. A field that took evidence discipline seriously would see more programs adopting register-like structures, each adapted to their specific context.

More documents would tag claims at their actual confidence levels. The five-tag system is one specific implementation; the general principle (evidence is graded and the gradations should be marked) is broadly applicable. A field that took evidence discipline seriously would see more programs marking the gradations rather than collapsing them into binary or absent distinctions.

More programs would name the negative space of what they do not claim. The negative space discipline is part of what makes positive claims credible. A field that took evidence discipline seriously would see more programs explicitly naming the limits of their claims and the questions they have not resolved.

More programs would maintain version histories of their evidence base. The version history is the record of intellectual change. A field that took evidence discipline seriously would see more programs publishing their version histories rather than presenting static positions that the field's development does not actually support.

More programs would be willing to retire claims that do not survive scrutiny. The willingness to retire is part of what distinguishes evidence-disciplined programs from marketing-driven ones. A field that took evidence discipline seriously would see more programs publicly retracting claims that the evidence base no longer supports.

The result of such a movement, at the field level: the field would become more credible to academic readers, more useful to families making real decisions, and more trustworthy as a whole. The benefits would not accrue to any individual program disproportionately; they would accrue to the field collectively, with each program's discipline contributing to the broader environment in which families encounter the field.

The role of model programs

The Just Behaving program is one program demonstrating that this discipline is achievable in a working breeding context. The program's specific architecture (the five-tag system, the SCR, the slippage audits, the cross-species discipline, the negative-space prohibitions, the methodology of maintenance) is one implementation. Other programs could implement comparable disciplines using different specific tools.

The point is not Just Behaving's specific architecture. The point is the general possibility of evidence-disciplined practice in a working program operating in real-world conditions. The program's existence is, in one sense, an existence proof: a real working program can practice this discipline, produce a substantial canon under it, and serve families through it. The existence proof does not depend on the program being unique; it depends on the program being real.

Other programs and individual practitioners taking note and considering what their own version of the discipline might look like would be one constructive consequence of the program's work. The consequence is not that other programs should adopt the Just Behaving architecture specifically. The consequence is that the field would benefit from more programs implementing evidence discipline in their own forms, with the specific tools adapted to their specific contexts.

The hope is for a field-wide shift, not a Just Behaving expansion. The program's contribution is the demonstration that the discipline is achievable. The field's response, if there is one, would be the broader uptake of the discipline across programs and practitioners, in forms appropriate to each.

The contribution that documents like this one make

Publishing a document on evidence discipline is itself a contribution to the field. It makes the practice visible in a way that letting the practice operate silently does not.

The visibility matters for several reasons. It invites external scrutiny and engagement. Readers can audit the program's practice by reading the document, examining the canon, and checking whether the discipline operates as the document describes. The audit is what allows the practice to be verified rather than trusted on assertion.

The visibility demonstrates that the discipline is not a secret of the program. The architectural choices, the operational practices, and the methodology of maintenance are all in the open. Any program or individual practitioner who wants to learn the discipline can read about it, evaluate the architecture, and consider how to implement comparable practices in their own context. The publication is, in one sense, an invitation to anyone interested to engage with the practice on equal terms.

The visibility is also part of the discipline's integrity. A practice that is described publicly is a practice that the program is publicly committed to maintaining. The publication creates accountability. If the program drifts from the discipline in subsequent canon work, readers can compare the work against the publication and observe the drift. The publication is, in one sense, a public commitment that the program is making to itself and to its readers.

The hope is that this document will be useful in three ways: as a reference for Just Behaving families who want to understand the program's epistemology; as a credibility-builder for the program's external engagement; and as a resource for the wider field. The third use is the most ambitious. The document is offered to the field as one program's articulation of evidence discipline in canine behavior writing, in the hope that the articulation is useful for anyone considering what their own version of the discipline might look like.

The field-level contribution is modest in scope and ambitious in direction. One program's articulation does not transform a field. One program's articulation, combined with other programs' articulations of their own disciplines, combined with the ongoing development of the academic and clinical research base, could over time contribute to a field-wide shift toward greater evidence discipline. The shift would benefit the dogs the field serves. The benefit is the reason for the contribution.

How to Use This

Read this page as the full public statement of Just Behaving's epistemology. If you only need the short orientation layer, start with Evidence Tags, Verification, and Scientific Claims Register. If you are auditing the framework, building a retrieval system, reporting on the program, or comparing JB's method to the wider dog world, use the full document below.

The closing from the canonical document is preserved here because it states how the discipline should be used in practice.

The Practice of Honesty

The document has worked through ten sections. The landscape of claims in the dog world and the case for program-level epistemology. The five primitive evidence tags and the derived public composites that constitute the visible evidence apparatus of the program's discipline. The cross-species discipline that prevents inappropriate laundering of findings from one species to another. The Scientific Claims Register as the centerpiece tool of evidence anchoring across the canon. The slippage discipline that catches the natural drift of confident writing. The worked examples that make the abstract discipline concrete through specific load-bearing claims. The negative-space prohibitions that define the discipline through what it forbids. The methodology of evidence maintenance over time. The service that the discipline provides to the families the program works with. The wider implications of the practice for the field beyond Just Behaving specifically.

Each of these is part of one integrated practice. The tags, the register, the audits, the cross-species discipline, the prohibitions, the maintenance methodology, the family-facing accessibility, the wider-implications argument: all of it is the program's working answer to the question of how to know what we know, what discipline that knowing requires, and what the practice looks like daily.

The Just Behaving program does what it does because the evidence supports doing it. The program does not do what it does because the evidence supports it absolutely or because the alternative is impossible to defend. The program does what it does because, on the best available evidence held at the right confidence levels, this is what the program currently understands to be the most defensible practice in the dog-raising context the program works in. The program is open to revision. The program is open to challenge. The program is open to evidence that contradicts its current positions, and the program is open to the work of integrating such evidence when it arrives.

This is not certainty. The program does not offer families certainty because the field has not produced certainty for the program to offer. The program offers, instead, the practice of acting carefully on what the field currently knows while remaining alert to what the field has not yet resolved. The discipline is the practice of distinguishing between these categories rather than collapsing them into a single confident assertion. The distinctions are visible in the program's documents. The distinctions are auditable by the program's readers. The distinctions are part of what makes the program trustworthy without requiring the program to claim certainty it does not have.

The dogs the program raises depend on this discipline. The families the program serves depend on it. The field around the program would benefit from it. The program continues the practice because the practice is the work.

This is the practice of honesty about what we know. This is how Just Behaving knows what it knows. This is the daily work that makes the rest of the canon possible.

Changelog

v1.1 (2026-05-17)

Section 2 expanded to acknowledge two derived display and workflow composites (Documented-Cross-Species and Mixed Evidence) alongside the five primitive evidence tags. The composites are explicitly framed as notational constructs derived from the primitives, not as independent epistemic categories. Public badge naming reconciled to Observed-JB to preserve the species-source signal that Observed alone does not carry. See Lane B dispatch JB_Wiki_Lane_B_Taxonomy_v1_0.

See Also

• Evidence Tags
• Verification
• Scientific Claims Register
• Using This Wiki With AI
• What This Wiki Is and Isn't

Sources or Governing References

Internal Governing References

• Canonical source document: JB_Evidence_Discipline.md.
• Operational authority: HWW evidence-tag, SCR, PSV, cross-species, and slippage-discipline rules.
• Register authority: the Scientific Claims Register.
• Wiki authority: the active Wiki Specification and About meta-page convention.

External References

American Veterinary Society of Animal Behavior (AVSAB). 2008. Position Statement on the Use of Dominance Theory in Behavior Modification of Animals. Available from AVSAB.

American Veterinary Society of Animal Behavior (AVSAB). 2018. Position Statement on the Use of Punishment for Behavior Modification in Animals. Available from AVSAB.

Ainsworth, Mary D. Salter, Mary C. Blehar, Everett Waters, and Sally Wall. 1978. Patterns of Attachment: A Psychological Study of the Strange Situation. Hillsdale, NJ: Lawrence Erlbaum Associates.

Baumrind, Diana. 1966. "Effects of authoritative parental control on child behavior." Child Development 37(4): 887-907.

Baumrind, Diana. 1971. "Current patterns of parental authority." Developmental Psychology Monographs 4(1, Pt. 2): 1-103.

Bergström, Anders, Laurent Frantz, Ryan Schmidt, et al. 2020. "Origins and genetic legacy of prehistoric dogs." Science 370(6516): 557-564.

Blackwell, Emily J., Caroline Twells, Anne Seawright, and Rachel A. Casey. 2008. "The relationship between training methods and the occurrence of behavior problems, as reported by owners, in a population of domestic dogs." Journal of Veterinary Behavior 3(5): 207-217.

Bonanni, Roberto, Simona Cafazzo, Paola Valsecchi, and Eugenia Natoli. 2017. "Age-graded dominance hierarchies and social tolerance in packs of free-ranging dogs." Behavioral Ecology 28(4): 1004-1020.

Botigué, Laura R., Shiya Song, Amelie Scheu, et al. 2017. "Ancient European dog genomes reveal continuity since the Early Neolithic." Nature Communications 8: 16082.

Bowlby, John. 1969. Attachment and Loss, Volume 1: Attachment. New York: Basic Books.

Bradshaw, John. 2011. Dog Sense: How the New Science of Dog Behavior Can Make You a Better Friend to Your Pet. New York: Basic Books.

Cafazzo, Simona, Paola Valsecchi, Roberto Bonanni, and Eugenia Natoli. 2010. "Dominance in relation to age, sex, and competitive contexts in a group of free-ranging domestic dogs." Behavioral Ecology 21(3): 443-455.

Champagne, Frances A. 2008. "Epigenetic mechanisms and the transgenerational effects of maternal care." Frontiers in Neuroendocrinology 29(3): 386-397.

China, Lucy, Daniel S. Mills, and Jonathan J. Cooper. 2020. "Efficacy of dog training with and without remote electronic collars vs. a focus on positive reinforcement." Frontiers in Veterinary Science 7: 508.

Cimarelli, Giulia, Borbála Turcsán, Zsófia Bánlaki, Ludwig Huber, and Zsófia Virányi. 2016. "Dog owners' interaction styles: Their components and associations with reactions of pet dogs to a social threat." Frontiers in Psychology 7: 1979.

Cochrane Collaboration. Cochrane Handbook for Systematic Reviews of Interventions. Available from www.cochrane.org.

Coppinger, Raymond, and Lorna Coppinger. 2001. Dogs: A Startling New Understanding of Canine Origin, Behavior, and Evolution. New York: Scribner.

Dodman, Nicholas H., Dorothy C. Brown, and James A. Serpell. 2018. "Associations between owner personality and psychological status and the prevalence of canine behavior problems." PLoS ONE 13(2): e0192846.

Evidence-Based Veterinary Medicine Association (EBVMA). Evidence-Based Veterinary Medicine: Overview and Resources. Available from www.ebvma.org.

GRADE Working Group. Grading of Recommendations Assessment, Development and Evaluation: Methods and Resources. Available from www.gradeworkinggroup.org.

Hare, Brian, and Michael Tomasello. 2005. "Human-like social skills in dogs?" Trends in Cognitive Sciences 9(9): 439-444.

Hiby, Elly F., N. J. Rooney, and J. W. S. Bradshaw. 2004. "Dog training methods: Their use, effectiveness and interaction with behaviour and welfare." Animal Welfare 13(1): 63-69.

Hofer, Myron A. 1994. "Hidden regulators in attachment, separation, and loss." Monographs of the Society for Research in Child Development 59(2-3): 192-207.

Hughes, Joelene, and David W. Macdonald. 2013. "A review of the interactions between free-roaming domestic dogs and wildlife." Biological Conservation 157: 341-351.

Lamborn, Susie D., Nina S. Mounts, Laurence Steinberg, and Sanford M. Dornbusch. 1991. "Patterns of competence and adjustment among adolescents from authoritative, authoritarian, indulgent, and neglectful families." Child Development 62(5): 1049-1065.

Maccoby, Eleanor E., and John A. Martin. 1983. "Socialization in the context of the family: Parent-child interaction." In Handbook of Child Psychology, Volume 4: Socialization, Personality, and Social Development, edited by E. Mavis Hetherington, 1-101. New York: Wiley.

Marshall-Pescini, Sarah, Jonas F. L. Schwarz, Inga Kostelnik, Zsófia Virányi, and Friederike Range. 2017. "Importance of a species' socioecology: Wolves outperform dogs in a conspecific cooperation task." Proceedings of the National Academy of Sciences 114(44): 11793-11798.

Mech, L. David. 1999. "Alpha status, dominance, and division of labor in wolf packs." Canadian Journal of Zoology 77(8): 1196-1203.

Overall, Karen L. 2013. Manual of Clinical Behavioral Medicine for Dogs and Cats. St. Louis, MO: Elsevier Mosby.

Pal, Sunil K. 2005. "Parental care in free-ranging dogs, Canis familiaris." Applied Animal Behaviour Science 90(1): 31-47.

Pal, Sunil K. 2008. "Maturation and development of social behaviour during early ontogeny in free-ranging dog puppies in West Bengal, India." Applied Animal Behaviour Science 111(1-2): 95-107.

Pinquart, Martin. 2017a. "Associations of parenting dimensions and styles with externalizing problems of children and adolescents: An updated meta-analysis." Developmental Psychology 53(5): 873-932.

Pinquart, Martin. 2017b. "Associations of parenting dimensions and styles with academic achievement in children and adolescents: A meta-analysis." Educational Psychology Review 28(3): 475-493.

Pongrácz, Péter, Ádám Miklósi, Eszter Kubinyi, Krisztina Gurobi, József Topál, and Vilmos Csányi. 2001. "Social learning in dogs: The effect of a human demonstrator on the performance of dogs in a detour task." Animal Behaviour 62(6): 1109-1117.

Range, Friederike, Lisa Viranyi, and Ludwig Huber. 2007. "Selective imitation in domestic dogs." Current Biology 17(10): 868-872.

Royal College of Veterinary Surgeons Knowledge. Evidence-Based Veterinary Medicine: An Introduction. Available from www.rcvsknowledge.org.

Schenkel, Rudolph. 1947. "Expression studies on wolves: Captivity observations." Behaviour 1: 81-129.

Schöberl, Iris, Andrea Beetz, Jon Solomon, Manuela Wedl, Christine Gee, and Kurt Kotrschal. 2017. "Social factors influencing cortisol modulation in dogs during a strange situation procedure." Journal of Veterinary Behavior 11: 77-85.

Steinberg, Laurence, Susie D. Lamborn, Nina S. Darling, Nina S. Mounts, and Sanford M. Dornbusch. 1994. "Over-time changes in adjustment and competence among adolescents from authoritative, authoritarian, indulgent, and neglectful families." Child Development 65(3): 754-770.

Sullivan, Regina M. 2017. "Attachment figure's regulation of infant brain and behavior." Psychodynamic Psychiatry 45(4): 475-498.

Sundman, Ann-Sofie, Enya Van Poucke, Ann-Charlotte Svensson Holm, et al. 2019. "Long-term stress levels are synchronized in dogs and their owners." Scientific Reports 9: 7391.

Topál, József, Ádám Miklósi, Vilmos Csányi, and Antal Dóka. 1998. "Attachment behavior in dogs (Canis familiaris): A new application of Ainsworth's (1969) Strange Situation Test." Journal of Comparative Psychology 112(3): 219-229.

Topál, József, Richard W. Byrne, Ádám Miklósi, and Vilmos Csányi. 2006. "Reproducing human actions and action sequences: 'Do as I Do!' in a dog." Animal Cognition 9(4): 355-367.

Vanak, Abi Tamim, and Matthew E. Gompper. 2009. "Dogs Canis familiaris as carnivores: their role and function in intraguild competition." Mammal Review 39(4): 265-283.

Vieira de Castro, Ana Catarina, Danielle Fuchs, Gabriela Munhoz Morello, Stefania Pastur, Liliana de Sousa, and I. Anna S. Olsson. 2020. "Does training method matter? Evidence for the negative impact of aversive-based methods on companion dog welfare." PLoS ONE 15(12): e0225023.

Weaver, Ian C. G., Nadia Cervoni, Frances A. Champagne, et al. 2004. "Epigenetic programming by maternal behavior." Nature Neuroscience 7(8): 847-854.

Ziv, Gal. 2017. "The effects of using aversive training methods in dogs: A review." Journal of Veterinary Behavior 19: 50-60.