Skip to main content
business medicine-and-health sociology

Boundary object

Description

A boundary object is an artifact that inhabits several social worlds at once. It is plastic enough that each world can use it for its own purposes, yet robust enough to retain a recognizable common identity across those uses. The same specimen can be a collectible, an inventory item, and scientific evidence; the same medical record can organize different professional accounts of one patient’s trajectory. The load-bearing claim is not simply that stakeholders see an object differently. Their different readings remain anchored to one shared artifact, and that artifact enables cooperative work without first forcing agreement on a single interpretation. Boundary objects therefore solve a coordination problem by preserving partial commonality rather than eliminating difference. The structure is shared artifact + distinct social worlds + local uses + common identity. Remove the local variation and the object becomes an ordinary standard. Remove the common identity and the worlds no longer have a shared reference. Remove the joint work and the case collapses into mere polysemy.

Aliases

Susan Leigh Star and James R. Griesemer introduced the term in their 1989 study of Berkeley’s Museum of Vertebrate Zoology. The hyphenated form emphasizes that the object does not merely sit near a boundary: its identity is constituted through use on both sides of it.

Triggers

User-initiated: A user describes a document, model, prototype, specimen, chart, or map that several disciplines or stakeholder groups use differently, and asks why it succeeds despite unresolved disagreements. Vocabulary cues: “shared artifact,” “common reference,” “different meanings to different teams,” “coordination without consensus.” Agent-initiated: The agent notices repeated attempts to make every participant adopt one vocabulary when a stable shared artifact already supports their local work. Candidate inference: “Could this artifact be strengthened as a boundary object—keeping a common identity while making each group’s legitimate local use explicit?” Situation-shape signals: Cross-functional programs centered on a prototype or plan. Multi-professional records. Scientific collections assembled by heterogeneous contributors. Collaboration that works operationally even though participants disagree about goals or interpretation.

Exclusions

  • Uniform-use standards — a standard interpreted and applied identically by all participants lacks the plasticity across worlds that makes a boundary object distinctive.
  • Adapters and translators — adapters cross a seam by converting representations. Boundary objects instead let the parties continue using one object through different local readings.
  • Mere polysemy — multiple senses of a word do not by themselves coordinate work. The object must be enrolled in practices shared across social worlds.
  • Single-community artifacts — internal documents can support many tasks, but without a boundary between communities they do not instantiate this structure.

Structure

Internal structure of boundary-object: a table of its component slots and the concepts that fill them. The artifact’s common identity is deliberately thinner than any participant’s full interpretation. That under-specification is a coordination asset: each world can attach the detail its work requires without breaking the shared reference. The balance is delicate. Too rigid, and the object serves only one world; too plastic, and it stops being recognizable as the same object.

Relationships

Relationship neighborhood of boundary-object: a graph of the concepts it connects to and the concepts it is a part of.
  • seam — names the boundary where incompatible practices meet; boundary-object names one way cooperation can cross it.
  • polysemy — shares the one-form/multiple-meaning structure, but boundary-object adds communities, practices, and joint work.
  • adapter — performs translation into separate native forms; a boundary object remains common rather than translated away.
  • consensus — is not required. Agreement may emerge, but the concept’s distinctive power is coordination while interpretations remain different.
  • multi-channel-ingest — has a useful directional inversion: many representations become one normalized store there; one artifact supports many situated representations here.

Examples

Berg, Marc, and Geoffrey Bowker (1997), "The Multiple Bodies of the Medical Record: Towards a Sociology of an Artifact," https://doi.org/10.1111/j.1533-8525.1997.tb00490.x · medicine-and-health

Berg and Bowker analyze the medical record as an artifact that constructs several versions of the patient’s body for different occupational practices. Physicians, nurses, and administrative or audit functions work through the same continuing record, but foreground different facts and organize them for different purposes. The record’s identity as one patient’s trajectory ties those professional accounts together without making them identical.Inference: A shared record need not erase professional differences to coordinate care; it must preserve a stable referent while supporting the distinct views the work requires.

Star, Susan Leigh, and James R. Griesemer (1989), "Institutional Ecology, Translations and Boundary Objects: Amateurs and Professionals in Berkeley's Museum of Vertebrate Zoology, 1907-39," https://doi.org/10.1177/030631289019003001 · sociology

Berkeley’s Museum of Vertebrate Zoology coordinated professional scientists, museum workers, patrons, and amateur collectors through shared specimens, standardized forms, maps, and repositories. The participants did not need identical motives or interpretations: collectors could contribute field material while scientists incorporated standardized specimens and observations into comparative research. The objects remained recognizable across those social worlds while being useful within each one’s practice.Inference: When contributors cannot share a full theory or vocabulary, strengthen the common artifact and its identity-preserving conventions rather than demanding premature consensus.
Carlile studied boundary objects in new-product development, where drawings, models, and prototypes are shared across functions with different specialized knowledge. A product representation can let design, manufacturing, and commercial participants locate dependencies and negotiate changes while each reads it through local requirements. Its value is not only transmitting facts: it can make consequences across functions visible enough for knowledge to be transformed.Inference: When cross-functional discussion stays abstract, make the developing product concrete in a manipulable shared representation; disagreements can then attach to one object rather than pass each other in separate vocabularies.