Tools, rules, and division of labor jointly shape behavior

What it is

In an activity system, tools carry established methods, rules define acceptable behavior, and the division of labor assigns responsibility and visibility. Together the three determine what a user can see, what they can do, and what they must hand off to whom; an interface is only part of the tool set. This card follows directly from "the unit of analysis is a mediated activity, not a single interaction": that card established that activity theory looks at the whole activity system rather than a single click; this one takes the system apart and shows exactly which forces inside it jointly determine what a person does in front of an interface — the answer is never a single variable like "is the interface well designed," but the compounded effect of tools, rules, and division of labor together.

Why it happens

A tool shapes behavior because it fixes accumulated experience and method into a physical object — a default value, a template, a checklist is itself the material embodiment of "people before decided this is how it should be done." Facing a tool, a user's attention and action get guided by that embodied experience almost involuntarily, often without even realizing they are following a method the tool has already predetermined for them. Rules mark the boundary between "allowed" and "not allowed," a boundary made up of both formal regulation written into policy and tacit convention that everyone follows without ever writing it down; what a rule really decides is whether, faced with an exception, a person can use their own judgment to break from the routine, or must follow a fixed procedure regardless. Division of labor decides how information and responsibility are carved up across roles: who sees the whole picture, who sees only their own slice, who is accountable for the final outcome, who is responsible only for executing one intermediate step. These three are mutually entangled and cannot be adjusted in isolation: giving a role a new tool without adjusting their permissions leaves the tool's capability unusable, and users often find a way around the formal process to get things done anyway; changing the division of labor without updating how information flows through the interface can suddenly cut off someone who should have seen a piece of information, even though neither the tool nor the rules themselves changed.

Where it stops holding

Tools, rules, and division of labor cannot be treated as three mutually independent variables to freely manipulate in isolation when making design decisions — a real organization also carries substantial informal norms and power relationships that appear in no process document, yet genuinely determine whether a rule actually gets enforced or a division of labor is actually followed. A tool can also be repurposed by its users in ways the designer never anticipated — a table originally meant to track progress can quietly get repurposed as a coordination channel for dividing work, and this kind of repurposing is often exactly because the formal rules or division of labor have a gap the users are filling with their own actions. Analysis must distinguish "formal rules" from "the rules actually being followed in practice"; reading a system purely from its formal documentation makes it easy to take the status quo for granted and miss that users have already found workarounds to problems the formal design never solved.

Applying it

• For every new feature, list the tools, related rules, and role assignments it will affect, and consider all three together rather than making design decisions purely from interface interaction.
• Design permission settings, approval flows, and notification mechanisms against the reality of the actual division of labor, not the nominal division shown on an org chart.
• Document existing workarounds and orally transmitted informal rules as design signals that need addressing, rather than dismissing them as "users not following procedure correctly."
• How to check: before a formal change ships, run an impact walkthrough asking explicitly "who gains new information because of this change, who loses information they used to see, which role now needs to pick up a new responsibility" — list all three kinds of impact and check them, rather than focusing only on whether the change itself is easy to use.

Related

• Same group: B4.06.1 The unit of analysis is mediated activity, not a single interaction · B4.06.3 It suits understanding longitudinal and organizational context
• Nearby: R1 Design Systems · V1 Collaboration and Social Interaction
• Search terms: mediating artifact · rules and roles · division of labor · informal norms