Literaldoing

Literaldoing is an emerging term in linguistics and human–computer interaction used to describe the practice of carrying out actions strictly according to the literal content of spoken or written instructions, without accounting for pragmatics, context, or inferred intent. The concept highlights situations where an agent follows the exact wording of a command rather than interpreting broader goals or situational cues.

The term is relatively new and is not yet standardized across disciplines. It is often discussed in analyses of instruction-following behavior, command semantics, and the design of interactive systems. In studies of artificial intelligence and conversational agents, literaldoing helps distinguish between operations that are literally commanded and those that rely on user intention, context, or common sense.

In practice, literaldoing treats the surface form of a directive as the operational directive. This can lead to predictable outcomes when instructions are clear, but it can also produce rigidity or errors when the user’s intent differs from the literal wording. For example, a digital assistant asked to “start a new document” may create a blank file, or one instructed to “set a reminder for 9 a.m.” may schedule the reminder exactly at that time, regardless of other tasks the user has in mind.

Critics note that overreliance on literaldoing can ignore necessary context and safety considerations, while proponents argue that it provides a transparent baseline for evaluating how systems interpret natural language. Related concepts include pragmatics, instruction following, and literalism in semantics.