Intrinsic motivation
Working at something because the work itself is the point — interest, curiosity, the pull of getting good at it — rather than for a reward bolted on from outside, such as a grade, a payment, or approval. The finding that carries weight is the interaction between the two: attaching an external reward to a task someone was already doing for its own sake often erodes that inner pull, and the loss can outlast the reward.
The orientation of working on a task because the work itself is the goal — interest, mastery, curiosity, the satisfaction of producing good work — rather than because of an external reward attached to completion. The contrast term is extrinsic motivation: working because of grades, scores, money, social approval, or threat of consequence. The empirical claim that matters for this topic is not that one is universally better than the other but that the two interact: introducing extrinsic rewards into a task that was being worked on intrinsically often reduces the intrinsic motivation that was already there, with the reduction persisting after the reward is removed.
Etymology§
The intrinsic / extrinsic distinction in this technical sense was developed across a cluster of work in the 1970s — Edward Deci's 1971 Journal of Personality and Social Psychology paper Effects of Externally Mediated Rewards on Intrinsic Motivation, Mark Lepper and David Greene's over-justification studies through the 1970s, and Deci and Richard Ryan's development of self-determination theory through the late 1970s and 1980s. The construct entered the assessment-and- feedback literature most prominently through Ruth Butler's 1988 paper on the effect of feedback type on divergent-thinking performance, and through Carol Dweck's task involvement versus ego involvement programme that ran in parallel.
The construct's central experimental result is the over-justification effect. A learner working at a task they find intrinsically interesting, given an extrinsic reward contingent on doing the work, comes to attribute their engagement to the reward rather than to the task; when the reward is removed, intrinsic engagement is lower than it was before the reward was introduced. The 1973 Lepper, Greene, and Nisbett study with preschool children and drawing materials is the demonstration most often cited: children who had been promised a good player award for drawing spent less time drawing in a later free-choice condition than children who had not been promised a reward.
For the skill-formation literature, the construct's main implication is what Butler 1988 demonstrated in the classroom. Students given grades along with formative comments shifted from a task-involving orientation (engaging with the work to improve it) to an ego-involving orientation (engaging with the work to find out how they compared to other students), and the shift was accompanied by a loss of the improvement effect that the comments produced when given alone. The mechanism is the attention-allocation account: when a score is available, the learner's processing capacity goes to interpreting the score rather than to acting on the comment, and the orientation follows the processing.
The construct sits in active conversation with the ungrading literature, which treats the removal of grades from formative assessment as a deliberate intervention to preserve the intrinsic engagement the grading apparatus erodes. The empirical case is not, in the literature, that grading is universally bad — terminal summative grades appear to have less effect on intrinsic motivation than the running attached-score feedback Butler tested. The case is more specific: that running formative feedback with attached scores is the configuration that collapses to the score and loses the formative function.
Research, writing, and working at a craft share a property that makes intrinsic motivation the relevant kind: the work carries no built-in external reward, so what has to sustain the long acquisition phase is interest in the work itself. The literature's implications — keep formative feedback separate from grading, give the learner visible criteria to read their own work against, allow time for revision rather than terminal judgement — are intended to protect that orientation rather than to augment it.