At Harvard, men paid to assemble Lego Bionicle models built an average of 10.6 when their finished work stayed on the desk — but only 7.2 when the experimenter dismantled it as they worked, handing back the same plastic pieces to be built again.

At Harvard, a small experiment gave men a choice about when to stop building Lego Bionicle models for money. Those whose completed figures remained on the desk built an average of 10.6. Those who watched the experimenter dismantle their work while they continued assembling built 7.2.

The result appeared in Dan Ariely, Emir Kamenica and Dražen Prelec’s 2008 paper, Man’s Search for Meaning: The Case of Legos. Forty male undergraduates were randomly assigned to two groups. The payment schedule was identical: $2 for the first model, with subsequent payments declining by 11 cents until reaching a two-cent floor.

A finished object could stay finished for a while

In the condition labelled Meaningful, completed models accumulated on the desk. In the Sisyphus condition, the experimenter alternated between two sets of pieces, taking one model apart while the participant built the other. The physical assembly task was otherwise the same.

In his later TED talk about motivation, Ariely emphasised how modest the distinction was. The participants were assembling toys, and even the preserved models would eventually be dismantled. They were not creating a permanent monument or doing something with an obvious social mission. Yet the difference between temporary preservation and immediate destruction was associated with a different willingness to continue.

The models did not have to last forever.

That makes the setup useful to think with. Finishing something and seeing it remain finished can be an experience in its own right, even when its eventual destination is unremarkable. A completed object gives a task a visible ending. Reusing its parts immediately makes that ending harder to hold on to.

The result concerns a decision to keep going

There is an important distinction between producing more units before stopping and being able to produce each unit faster. The headline describes the first. Reading it as a universal productivity increase would change the question the comparison answers.

It is also tempting to translate any striking average into a forecast for a team. That would require much more evidence. A short, individual laboratory task leaves out salaries, promotion prospects, colleagues, competing deadlines and the consequences of refusing an assignment. Those are central features of employment, not incidental details.

Nor does a difference between group averages describe every individual. Some people in one condition may work longer than some in the other. An average is a summary of the observations, not a rule that can be applied to each person who walks into an office.

The useful question is narrower: could the treatment of completed work influence whether doing the next piece feels worthwhile? That is a question about the working situation as well as the person doing the work.

Workplace diaries offer a different perspective

A separate research programme offers context from actual workplaces. In Harvard Business School’s account of The Progress Principle, Teresa Amabile describes work with Steven Kramer involving roughly 12,000 diary entries from 238 employees at seven companies.

The diaries linked progress in meaningful work with better emotions, perceptions and motivation. Setbacks featured prominently on bad days. Small advances could matter, rather than only the rare moment when an entire project succeeded. Amabile also highlighted practical conditions such as clear goals, sufficient resources, autonomy and support.

This is related evidence, not a replication of the Bionicle experiment. Diaries describe experiences unfolding in organisations; a controlled experiment changes a specific feature of a task. The methods answer different questions and have different limitations. Their connection is the attention they direct towards what happens to a person’s effort during the working day.

Revision can still be useful work

A literal reading would be a poor guide to building products. Prototypes are dismantled, code is rewritten and drafts are discarded for good reasons. Preserving everything would create its own problems. The practical distinction is whether a revision has a purpose that the people doing it can understand.

Consider a hypothetical engineering team asked to rebuild a component after a failed test. If the test exposed a weakness and the next design addresses it, the discarded version has contributed knowledge. The object may disappear while the project still moves forward.

Now imagine the same team repeatedly rebuilding the component because an unexplained instruction keeps reversing. The physical work might look similar, but the account of what has been achieved is different. These are illustrative scenarios, not conditions tested in the Harvard study.

Repetition alone does not establish pointlessness.

A small experiment, a concrete question

The Bionicle result does not establish that meaning can replace adequate pay, or that an inspiring explanation will make an unsustainable workload acceptable. Those claims would extend far beyond a choice about assembling another toy. The experiment is most revealing when its modest scale stays visible.

It invites attention to an easily overlooked part of a task: what happens after someone finishes it. Is there a usable result, a lesson carried forward, or an explanation for the next revision? Those possibilities give completion a place in a continuing project.

A desk full of plastic figures is a very small form of accomplishment. In this experiment, that small distinction was enough to make the next model a different proposition.