Ethogram¶
An operationally described repertoire of animal behaviors used to recognize and compare acts within a declared species or study scope.
Core Idea¶
An ethogram is an organized repertoire of animal behaviors, with named units and descriptions by which an observer can recognize them. It can aim at a broad species repertoire or a study-specific subset. Stanford's laboratory-mouse resource, for example, lists roughly one hundred behaviors and explicitly says individual protocols may combine, omit or emphasize entries; an original chimpanzee cooperation study uses a much narrower task ethogram of competitive acts and responses. Both are ethograms because the animal actions are made into an inspectable observation vocabulary, not because they code every instant of life.[1][2]
An ethogram is a representation and potential coding instrument, not the activity stream itself. A particular project may additionally choose scan sampling, all-occurrences recording, mutually exclusive state codes, actor/recipient fields, rates or transition statistics. Those design choices can improve a particular analysis, but none is a universal requirement for the identity. The original chimpanzee paper says some responses in its ethogram can co-occur, directly contradicting an “exactly one behavior per moment” rule.[2]
Structural Signature¶
Sig role-phrases: declared animal scope — named behavioral units — operational discrimination — purpose-relative grouping — interface to a separate observation protocol.
- Animal and scope. A species-level inventory and a narrow experiment-specific inventory cover different ranges of acts. Scope tells the reader whether a behavior absent from the list is unobserved, deliberately out of scope, or thought not to occur.[1][2]
- Unit inventory. The ethogram names recurring postures, movements, interactions or states that can be sought again: mouse grooming and nesting; chimpanzee freeloading attempts and displacement; macaque hair pulling and grooming.[1][2][3]
- Operational discrimination. Descriptions identify observable differences and, when needed, thresholds. In the macaque study, pulling hair by a forceful jerking movement was distinguished from smoother grooming. A name such as “aggression” alone would not do that work.[3]
- Organization. Units can be nested or cross-referenced under groups suited to the question. Stanford arranges mouse behaviors by active/inactive/unknown and function-related groups; the chimpanzee task separates competition types from responses. Hierarchy and functional grouping are possible, not mandatory in one fixed form.[1][2]
- Protocol interface. A later study chooses how to observe, which subset to code, whether overlapping actions receive multiple labels, and what counts or durations to calculate. The inventory enables such choices but is not the dataset or a transition model.[1][2]
What It Is Not¶
An ethogram is not a raw video, event log or actogram. The video contains activity, the event log records observations using selected units, and an actogram arranges time-binned activity across repeated cycles. The ethogram itself supplies the behavior descriptions and repertoire against which those records are interpreted.[1][2]
It is not necessarily a complete, mutually exclusive partition of all animal activity. A species resource may aspire to breadth yet acknowledge unknown categories and study-specific subsets. The chimpanzee table explicitly allows some response codes to overlap within one event. Nor is every ethogram free of functional vocabulary: Stanford's mouse inventory groups acts by adaptive meaning. Good practice requires clarity about when a label describes visible form and when it embeds a functional hypothesis, not a universal ban on the latter.[1][2]
It is also not one species' behavior itself or a proof of the cause of that behavior. An ethogram can distinguish macaque hair pulling from grooming by observable motion, while the motivation or welfare consequence requires separate evidence.[3]
Scope of Application¶
Ethograms are literal tools in ethology, behavioral ecology, comparative medicine and animal-welfare observation wherever a researcher needs an explicit animal-act repertoire. A broad resource can support many later questions: Stanford's mouse ethogram includes exploratory, social, maternal, maintenance and abnormal behaviors, with protocols choosing appropriate subsets. The grouping is not a claim that each animal performs every entry at every age or in every setting.[1]
In narrowly defined experiments, an ethogram can instead cover just the acts relevant to a mechanism or question. The chimpanzee cooperation study lists freeloading, displacement and reactions during a group task. A macaque study lists social hair pulling, grooming, aggression and subordination and records actor and recipient for those interactions. In each case the repertoire is bounded by observational aims and by what the units can reliably distinguish.[2][3]
Use beyond animal behavior would be an analogy unless the animal-act carrier and operational ethological unit remain intact. A customer-interaction codebook may share the broader representational skeleton, but it is not automatically an ethogram under this entry's domain identity.
Clarity¶
The crucial question is “What could another observer see that warrants this label?” In the macaque ethogram, forceful jerking that removes hair is differentiated from the smoother handling of grooming. That distinction allowed the authors to investigate whether hair pulling really belonged inside “overgrooming” or aggression rather than deciding its identity from a loaded label.[3]
Some labels contain interpretation. Stanford's mouse ethogram uses functional categories such as exploratory and affiliative behavior. This does not invalidate it; it means the observable unit description and the higher-level interpretation should be legible at different levels. When an observation could satisfy two units, the protocol should state whether co-coding is allowed or whether a priority rule resolves the overlap. The original chimpanzee ethogram explicitly chooses the former for some responses.[1][2]
Manages Complexity¶
An animal's ongoing behavior is far richer than a study can record verbatim. The ethogram compresses that stream into a finite, reviewable repertoire of units. Stanford's broad mouse resource can be reused across questions because later protocols select subsets. The chimpanzee study narrows to competition and response acts, making a 94-hour video corpus scoreable while maintaining definitions that another rater can apply.[1][2]
Compression is lossy. A narrowly targeted ethogram may omit behaviors that later prove relevant; a very broad one can burden training and introduce ambiguous neighboring categories. The right scope follows the question, with explicit unknown/out-of-view or uncodable handling when completeness is claimed. The ethogram organizes observation; it does not itself solve sampling bias or guarantee that a frequency estimate is reliable.[1][2]
Abstract Reasoning¶
An ethogram lets a researcher move from an unstructured impression—“these animals are aggressive”—to discriminable observations: which act occurred, in what context, and according to which definition? The macaque study's operational distinction between hair pulling and grooming opened the possibility of testing different social and welfare associations. The behavior list made the question investigable; it did not predetermine the answer.[3]
It also separates reference vocabulary from inference. The chimpanzee ethogram defined responses, then observers scored video and the study calculated occurrences and interrater agreement. If two responses can overlap, treating their counts as a single exhaustive partition would be a coding error. If a study changes its unit definitions, apparent differences across groups may arise from the changed representation rather than animal behavior itself.[2]
Knowledge Transfer¶
The same ethogram structure transfers literally between mouse repertoire documentation, chimpanzee cooperation coding and macaque social-interaction research: state the animal/context, name acts, explain how observers distinguish them and indicate how a separate protocol will record them. The particular units and grouping cannot simply be copied between species or questions. Stanford's broad species resource and the chimpanzee study's narrow table are different realizations of the same type.[1][2][3]
The portable skeleton—target phenomena mapped to named, interpretable units—belongs to live prime Representation. Live Classification may describe the later action of assigning an observed bout to an ethogram unit, but the existence of the repertoire is not identical to that sorting action. A transfer to non-animal event codebooks is therefore transfer of the parent skeleton, not proof that the term Ethogram has escaped its ethological frame.
Examples¶
Laboratory-mouse species repertoire. Stanford Comparative Medicine's original mouse ethogram lists roughly one hundred behaviors, organized into active, inactive and unknown groups and more detailed exploratory, social, maternal, maintenance and abnormal classes. The project explains that experimental protocols may combine, exclude or emphasize behaviors from this larger resource.[1] Mapped back: animal and scope = laboratory mouse species repertoire; unit inventory = named exploratory, grooming, nesting and interaction acts; operational discrimination = behavior pages and category descriptions; organization = nested active/inactive/unknown and functional groups; protocol interface = narrower experimental subsets chosen later.
Chimpanzee cooperative-task ethogram. Suchak and colleagues' original study enumerates freeloading attempts, displacement, withdrawal, protest and other social responses for video coding in an open-group cooperation task. It specifies which behaviors count and states that some response labels can co-occur in one event; a separate observer checked a sample of coded video.[2] Mapped back: animal and scope = chimpanzees during the cooperative task; unit inventory = competition types and responses; operational discrimination = observable inclusion criteria in the SI table; organization = competition versus response groups with qualified overlap; protocol interface = video scoring and later frequencies/reliability, not an every-moment exclusive partition.
Structural Tensions¶
Broad repertoire versus focused coding. A species-wide inventory aids unexpected-behavior recognition and comparison across questions but takes more effort to define and train; a small study ethogram is easier to score yet can omit context that later matters. Neither maximal coverage nor minimal coding burden wins without reference to the inquiry.[1][2] Diagnostic: Is this intended as a reference repertoire or a protocol-specific subset, and which omitted acts would change the conclusion?
Observable discrimination versus functional interpretation. Fine descriptions can improve agreement on what happened, while function-level groupings help answer what it meant. Collapsing several movements into “aggression” gains interpretive economy but can obscure distinct acts; a purely microscopic motion list can lose the context the study cares about. Both layers can be retained if distinguished and made revisable.[1][3] Diagnostic: Can another observer recognize the unit from observable criteria, while its functional grouping remains separately testable rather than built in as an unquestionable fact?
Structural–Framed Character¶
Ethogram is structural as a representation type but strongly framed by animal-behavior practice. Its unit-inventory relation recurs in unlike species and studies; its named identity still requires an animal repertoire rather than just any codebook.
Vocabulary travel: units, descriptions, scope and coding conventions transfer literally from mouse to chimpanzee or macaque ethograms, but calling a market event list an ethogram would normally import a metaphor. Evaluative weight: the inventory is a descriptive instrument; which behaviors matter for welfare, treatment or experimental success is a later evaluative choice, though selection and grouping make priorities visible. Human-practice dependence: animals act without observers, but an ethogram exists only when observers construct and maintain a description system; observer agreement and revised definitions matter. Institutional origin: Stanford, laboratories and journals supply examples and standards of practice, not an authority whose approval constitutes an ethogram. Import versus recognition: a new species list with operational acts is another ethogram even if independently devised; reusing mouse categories on another species without checking actual action boundaries is an import error.[1][2][3]
Its character: a reusable animal-behavior representation with a clear structural spine and a constitutive ethological carrier. The substrate-neutral encoding skeleton belongs to Representation, not to an unsupported prime promotion of Ethogram.
Structural Core vs. Domain Accent¶
The structural core is selected target acts → named and discriminable units → an interpretable catalogue for later recording. Live Representation already carries the general target-to-medium mapping and selective faithfulness claim. An ethogram is a narrower instance because the target is observable animal behavior and the medium is an operational behavior repertoire.[1][2]
The domain accent is not incidental. A species or study context, behavior-unit descriptions, animal action boundaries and observer recognition are what make the inventory an Ethogram. A generic rubric for human actions shares representational structure but lacks the animal-ethological carrier. The named node therefore does not clear the prime bar; claims about cross-domain applicability belong to its Representation parent, while overlap with Classification belongs to the separate act of scoring observations.
Instantiates / Related Primes¶
This entry is a kind of Representation.
DAG parent — Representation. The behavioral repertoire is a target; named acts, definitions and groups are a distinct medium; operational descriptions specify what features are preserved for later interpretation. This directly instantiates the live prime's target–medium–mapping signature.
Related, not strict parents. Classification describes assigning an observed event to a category according to rules. That may follow an ethogram, but the ethogram as a reference inventory exists before any particular observation is sorted. Live Schema is a generalized cognitive template with defaults and slots, not a published behavioral repertoire. Actogram displays activity in time, and Display behavior can name one kind of ethogram entry; neither subsumes the whole inventory.
Relationships to Other Abstractions¶
Current abstraction Ethogram Domain-specific
Parents (1) — more general patterns this builds on
-
Ethogram is a kind of Representation Prime
An ethogram maps selected animal acts into named, discriminable units under a declared observation convention.Live Representation requires a target, a distinct medium and an interpretation convention preserving selected features. An ethogram takes a species or study's animal-behavior repertoire as target and expresses selected acts as named, operationally described units or groups for observation. Its animal carrier and ethological discrimination rules make it a narrower form of representation; the coded observations and statistics are later products, not the inventory itself.
Hierarchy path (1) — routes to 1 parentless root
- Ethogram → Representation → Abstraction
Neighborhood in Abstraction Space¶
Ethogram sits in a sparse region of the domain-specific corpus (83rd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Cognitive & Behavioral Theories (16 abstractions)
Nearest neighbors
- Behaviorism — 0.84
- Cue Validity — 0.84
- Occupancy–Abundance Relationship — 0.82
- Ecological psychology — 0.81
- Psychological Model — 0.81
Computed from structural-signature embeddings · 2026-10-08
Not to Be Confused With¶
- Observation record: a dataset of scored bouts, counts or durations produced using a selected ethogram, not the behavior-unit definitions themselves.[2]
- Actogram: a repeated-cycle visualization of activity timing, not an inventory of behavior categories.
- Pure motive list: “fear,” “aggression” and “courtship” alone do not tell observers what to recognize; even functionally organized ethograms need interpretable acts and boundaries.[1][3]
- Universal one-code-per-instant rule: some protocols impose one mutually exclusive state at a time, but an original chimpanzee ethogram explicitly permits certain simultaneous response codes.[2]
- A transition model: transition probabilities can be calculated from a separately coded sequence; they are not required to make the underlying repertoire an ethogram.
References¶
[1] Stanford University Department of Comparative Medicine, Mouse Behavior Ethogram and Ethogram Index, original institutional species repertoire, “Introducing the Species Ethogram,” “About Ethograms,” and index. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r
[2] M. Suchak et al., “How chimpanzees cooperate in a competitive world”, PNAS (2016), Materials and Methods and SI Materials and Methods, “Behavioral Coding” ethogram table. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r ↩s ↩t
[3] A. Heagerty et al., “Social hair pulling in captive rhesus macaques (Macaca mulatta)”, American Journal of Primatology 79 (2017), “Behavioral observations,” Table 1 and accompanying distinction between hair pulling and grooming. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j