Biological motion¶
A human or animal action pattern in which body parts change their relative positions over time, distinct from the body's travel as a whole.
Core Idea¶
In the bounded articulated-body sense used here, biological motion is the changing pattern of relative movement among the parts of a human or other animal body in action, or a source-linked depiction of such an action. The pattern unfolds through time: a walk changes the relation among legs, arms and trunk; a bird's flight changes the relation among wings and body. It can be described separately from the route the body takes through a room or through the air. A recorded or photographed sequence can carry this pattern after the actor has left; a viewer's success in recognizing it is a different event.[1][2][3]
This is an explicit narrow use of a broader title. Ordinary “biological motion” can mean many kinds of movement by organisms. Johansson's original abstract named characteristic animal and human locomotor patterns and showed human actions through moving main-joint spots, without giving a universal definition of every biological movement. Troje later distinguished intrinsic deformation of an acting body from its common travel through space and reserved “biological motion” for that intrinsic sense within his chapter. The entry adopts that stated research distinction; it does not silently recast cell migration, all organism movement or every use of the phrase as articulated-body motion.[1][2]
Structural Signature¶
- Action correspondence. The pattern is an action of a human or animal articulated body, or a depiction whose trajectories are traceably based on such an action. Mather and West built animal joint-point sequences from photographed locomotion; Hagen and colleagues built bird-flight displays from video. An unrelated jointed machine does not become biological motion simply because its moving parts look lively. The source relation defines the biological qualification here, while the actor need not be present when a display is viewed.[4][3]
- Temporal part configuration. Several body parts have positions or orientations that change through an ordered sequence. One photograph can show an articulated posture but cannot, alone, instantiate its movement. Johansson's human main-joint displays, Mather and West's successive animal frames, and Hagen's 75-frame bird series all preserve ordered part configurations.[1][4][3]
- Coordinated intrinsic change. Some parts move relative to other parts, with a temporal organization of the action. This is more specific than translating an unchanged animal-shaped object as a single rigid piece. Hagen's bird display tracked one beak point, two body points and four wing points across frames; Baier, Gatesy and Dial separately measured wing-joint timing in chukars. Those are distinct studies and measurement levels, but both make the wing/body relation concrete.[2][3][5]
The intrinsic-versus-common-motion test follows from these roles rather than adding a fourth causal stage. Describe how parts change in body- or part-centered coordinates and compare that with the movement of the whole body. A centered display can retain wing or limb articulation even when much common travel has been removed.[2][3]
What It Is Not¶
Biological motion in this sense is not the observer's perception of biological motion. Recognizing a walker, judging an action, showing an expertise effect, or recruiting a neural pathway is an outcome or method of studying a motion signal. Hagen's bird-expertise experiment measured observers, but the bird-flight point trajectories were its stimulus. The experiment did not find the predicted bird-expertise advantage, and no such advantage is part of this motion identity.[3]
It is not a required point-light format. Dots can expose part trajectories while stripping away much surface form, but photographs, video and source-linked animation can also present the underlying pattern. Nor does mere apparent animacy suffice: a rigid triangle or an animal figure carried intact may move through space without the intrinsic part-relative deformation that distinguishes the present sense.[2][4]
Scope of Application¶
The literal scope is articulated human and other animal actions, together with depictions that preserve a warranted relation to their part motion. Human walking and running are established examples; bird wing action during flight supplies a non-gait example. The entry concerns the motion pattern, not a universal biomechanical cause, a particular species, an experimental display recipe or an automatic inference about intention.[1][4][3][5]
For a synthetic display, source linkage matters more than a claim that it “looks biological.” Mather and West digitized animal photographs and adjusted playback timing to make the sequence appear natural, so source linkage does not demand frame-exact preservation of every original timing parameter. The cited sources do not supply a universal numeric fidelity threshold for arbitrary robot or animation trajectories. Without documented correspondence to human or animal action, resemblance alone is insufficient for this bounded identity.[4][2]
Clarity¶
The phrase “motion of a bird” can name two different things: the bird's flight path and the changing relation among its wings, body and beak. The intrinsic pattern is the second question. Hagen's bird-flight image remained centered while seven tagged locations changed over 75 frames; centering made common travel less prominent without removing the articulated motion. Baier and colleagues separately reported wing-joint timing and whole-body trajectory, illustrating that those measurements answer different questions.[3][5]
Likewise, seeing a human walker in a sparse display does not create the gait. Johansson's abstract described moving main-joint combinations and an observer response to them. Treating the action's structure and the viewer's recovery of it as separate lets one ask whether a display contains the relevant temporal relations before asking whether a particular viewer identifies the action.[1][2]
Manages Complexity¶
Many anatomical details, camera choices and display formats reduce to three checks: is there a warranted human or animal action source, is there a temporal configuration of parts, and do those parts change relative to one another? The checks allow a human biped and a flying bird to be compared without asserting that legs and wings have identical anatomy or that both are recognized with equal ease. They also prevent a single frame, a rigid translation and an ungrounded animation from becoming positives merely because each involves an animal-like shape.[1][3][2]
That compression has limits. A sparse display can retain local point movement while losing the coherent global figure. Hagen's spatial scrambling preserved the local motion of each dot but disrupted the global object percept. Thus “scrambled” is not a universal negative for every local biological kinematic cue. The target pattern and the level at which coherence is claimed must be stated.[3]
Abstract Reasoning¶
To test an instance, first identify the action source or the documented path by which its movement was depicted. Then follow at least two parts through time. If their relative geometry changes in an action-organized way, ask whether that change remains after common body travel is conceptually removed. This makes a human step and a bird wingbeat comparable at the level of intrinsic articulation while leaving their different anatomy and measurement methods intact.[2][4][3]
Counterfactuals sharpen the test. Freeze the sequence to one posture and the temporal role disappears. Carry a rigid animal figure across a room and the intrinsic-change role disappears. Replace a recorded or documented animal-action trajectory with an arbitrary jointed machine and the action-correspondence role is no longer warranted. None of those counterfactuals says that the resulting image cannot evoke animacy in an observer; that is a different classification.[2][4]
Knowledge Transfer¶
Within biological-motion research, the same three checks apply to different articulated actions and displays. Human joint-point sequences, digitized animal locomotion and bird-flight video-derived points can be compared for their temporal part relations even though the studies use different animals, frame counts and observer tasks. What transfers is the motion-pattern test, not a prediction that the same viewer will detect all stimuli equally well.[1][4][3]
Outside that field, “coordinated relative motion of parts” is a possible broader formal skeleton. A robot can share that skeleton, but the biological-action correspondence of this named entry is then a separate question. Applying the phrase to any jointed machine without evidence of such correspondence would turn a useful analogy into an unsupported literal instance.[2]
Examples¶
Human walking in moving-joint displays. Johansson's publisher abstract reports human walking, running and dancing evoked by adequate combinations of 10–12 moving main-joint spots. Mather and West later made joint-point sequences from photographed human and nonhuman locomotion. Mapped back: the human action or photographed sequence supplies action correspondence; successive spots provide temporal part configurations; changing limb-to-body relations provide intrinsic articulation. Recognition by observers was measured, but is not one of those three roles. Johansson's full paper was not inspected for this entry, so no further method claim is drawn from it.[1][4]
Bird flight in a video-derived display. Hagen and colleagues tagged one beak, two body and four wing locations in each of 75 frames from a bird-flight video. The centered display retained changing relations among those points. Baier and colleagues independently measured chukar wing-joint timing during flight, confirming an anatomical route for avian relative motion without measuring Hagen's particular birds. Mapped back: the recorded bird action supplies the source; the tracked frames supply time order; wing/body changes supply intrinsic articulation. The two papers are complementary evidence, not a single experiment.[3][5]
Structural Tensions¶
No opposed pressure is constitutive to the motion pattern itself in the inspected sources. A researcher may trade display realism against isolation of kinematic cues, but that is a methodological choice about observing the pattern. Troje discusses stylized stimuli and their limits.[2] Neither maximal realism nor maximal simplification is a necessary condition for an animal's articulated action. The practical diagnostic is which information a chosen display removes before interpreting a viewer result.
Structural–Framed Character¶
The part-relative versus common-travel distinction is structural: Troje explicitly describes intrinsic deformation in body-centered coordinates and contrasts it with common whole-body motion.[2] The distinction can be tested through trajectories even when the actor's species or the display changes. The named category is also strongly domain-framed because “biological” requires warranted correspondence to human or animal action, and because the narrow term comes from a particular research tradition. The motion is not intrinsically good or bad; any evaluative weight belongs to later uses or interpretations. Human practice affects how researchers record and stylize actions: Mather and West animated photographs, while Hagen and colleagues tracked points in bird-flight videos.[4][3] Those practices do not create the underlying part changes in a walking person or flying bird. The category's institutional origin is a vision-science naming and research practice, not a rule that brings those animal movements into existence. The vocabulary may travel to robotics or animation, yet the biological qualification is imported literally only with warranted action correspondence; otherwise one recognizes an analogy to its broader moving-parts skeleton. Its character: structurally testable relative motion within a strongly biological-action-framed research category.
Structural Core vs. Domain Accent¶
The structural core is a temporal relation: several linked parts change position or orientation relative to one another while a whole body may also travel; Troje states the intrinsic/common distinction explicitly.[2] The domain accent is a human or animal articulated action, or a depiction traceably related to one. Mather and West's photograph-derived sequences and Hagen's video-derived bird points are source-linked displays of that kind.[4][3] Remove the accent and one has a broader kind of coordinated articulated movement, not automatically biological motion. That broader skeleton is a future Prime question, not a Prime established by this entry: it would require unlike cross-domain cases and an all-instance boundary test. The present evidence supports the bounded domain-specific identity.
Instantiates / Related Primes¶
No current strict parent passes the full all-instance test, so this node is an approved isolated root with zero edges. Biological Process requires a living bearer, ordered biological mechanism and biological consequence that a source-linked display does not itself carry. Perceptual Process and Structure from Motion concern what an observer does with a stimulus. Motion Geometry concerns rigid geometric transformation, whereas intrinsic articulated motion changes part relations. Pattern requires repeatable organization, which a single nonrepeating articulated action need not supply. Relation and Transformation add formal commitments that are not constitutive of every physical action. These current-catalog comparisons prevent topical similarity from becoming an unsupported DAG edge.
Neighborhood in Abstraction Space¶
Biological motion sits in a sparse region of the domain-specific corpus (97th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Unclustered & Miscellaneous (2721 abstractions)
Nearest neighbors
- Chronotope — 0.77
- Visual reasoning — 0.77
- Diffeomorphometry — 0.76
- Vehicle Axes Conventions — 0.76
- Descriptive Geometry — 0.75
Computed from structural-signature embeddings · 2026-10-17
Not to Be Confused With¶
The narrow motion-side entry should not absorb a separate biological-motion perception candidate. Point-light recognition, inversion effects, expertise, and neural processing describe observers and experiments. Hagen's scrambled displays are also not simply “no biological motion”: the local dot trajectories were preserved while global figure perception was disrupted.[3] Troje explicitly notes broader colloquial usage of the term.[2] This technical entry therefore does not declare other organism movements nonbiological in everyday usage.
References¶
[1] Gunnar Johansson, “Visual perception of biological motion and a model for its analysis,” Perception & Psychophysics 14 (1973), 201–211, doi:10.3758/BF03212378. Original publisher abstract inspected at https://link.springer.com/article/10.3758/BF03212378 . The full article was not inspected. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h
[2] Nikolaus Troje, “What Is Biological Motion? Definition, Stimuli and Paradigms,” author-hosted prepublication chapter manuscript. PDF viewer pp. 3–4 and 11 inspected at https://www.biomotionlab.ca/Text/TrojeChapter2013.pdf . Its terminology choice is explicitly scoped to the chapter; the final printed edition was not verified. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o
[3] Simen Hagen et al., “Bird expertise does not increase motion sensitivity to bird flight motion,” Journal of Vision 21(5):5 (2021), doi:10.1167/jov.21.5.5. Full original article inspected at https://pmc.ncbi.nlm.nih.gov/articles/PMC8107655/ , especially Introduction, Methods/Stimuli and Abstract. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p
[4] George Mather and Sophie West, “Recognition of animal locomotion from dynamic point-light displays,” Perception 22 (1993), 759–766. Full original author-hosted PDF inspected, especially pp. 759–762, at https://www.georgemather.com/PDF/Mather_west_1993.pdf . registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k
[5] David B. Baier, Stephen M. Gatesy and Kenneth P. Dial, “Three-Dimensional, High-Resolution Skeletal Kinematics of the Avian Wing and Shoulder during Ascending Flapping Flight and Uphill Flap-Running,” PLoS ONE 8(5):e63982 (2013), doi:10.1371/journal.pone.0063982. Full original article inspected at https://pmc.ncbi.nlm.nih.gov/articles/PMC3655074/ , especially Results, Wingbeat Timing and General Body Kinematics. registry ↩a ↩b ↩c ↩d