The Play You Keep

In a laboratory at Queen Mary University of London, forty-five bumble bees were given a clear, unobstructed path to a dish of sugar water. Along the sides of that path, out of the way, somebody had left some small wooden balls.

The bees did not need the balls. They could reach the sucrose without going near one. Nothing was hidden underneath them, nothing was gained by touching them, and no bee was ever rewarded for doing so.

Over the course of the study, forty-five bees rolled those balls 910 times.

Some bees went back again and again. They pushed the balls in different directions, over different distances, in patterns that were not the same twice. Then they went and got their sugar, which had been there the whole time.

The paper is called Do bumble bees play? — and the reason it takes thirteen pages to answer is that the answer had to be defended against every cheaper explanation first.

This is the part worth stopping on, and it is easy to read past. Before anyone could say whether a bee was playing, the field had to agree on what play is — formally, in criteria you can fail.

That work is Gordon Burghardt’s, in The Genesis of Animal Play (2005), and the criteria are deliberately strict, because play is exactly the thing a hopeful observer sees everywhere. Ball rolling was tested against all five.

The authors also excluded, by the design of the apparatus and by watching what the bees did, that the rolling was foraging, clearing clutter out of the way, or mating behaviour.

Read the first row again. The opening requirement for something to count as play is that it must not accomplish anything. Uselessness is not the absence of a category here. It is the entry condition, it is written down, and it can be checked.

The fourth criterion is the subtle one, and it is the one that carries this zine’s argument furthest.

Play has to be repeated — otherwise it is a one-off accident, an animal bumping into something. But it must not be stereotyped: not a fixed loop run identically each time. The amount of ball rolling varied within individual bees and between them. A bee came back to it, and did it differently.

Repetition was not the problem. Sameness would have been.

That distinction is worth holding, because we are so often handed the opposite one. A returning, absorbing, repeated interest gets read as a rut — something to be interrupted, redirected, or counted in a column. The formal literature on play makes precisely the reverse move: it asks whether the repetition varies, and treats variation as the evidence of something alive rather than something stuck.

A thing can be done over and over and still be free. The bees were not running a program. They were doing it again, which is a different word.

Criterion two says the behaviour has to be rewarding in and of itself. That is a strong claim about an insect, so the authors went and tested it separately.

Bees were given two chambers of different colours. In one, the balls could be moved. In the other, they could not. The bees learned the difference.

Then both chambers were emptied — no balls in either — and the bees were offered the choice again.

They chose the colour of the room where the rolling had happened.

There was nothing to roll. There was nothing to eat. What was left was the place, and a colour associated with having done something worthless there, and that was enough to bring them back.

This is what makes the finding hard to argue away. A bee that rolls a ball might be confused about something. A bee that returns to an empty room because of what it once did there is reporting on how the doing felt.

Ask what the literature on Autistic lives is about, and the honest answer for a long time was: what is hard. Deficit, burden, cost, risk, outcome. The vocabulary for the other half arrived late and arrived thin.

Elliot Wassell’s study of Autistic joy (Disability & Society, 2025) asked 86 of us a question the field had mostly skipped: what makes you happy? A majority reported experiencing joy often. Three themes came back — that the Autistic sensorium is itself a source of joy, that our interests matter enormously, and that joy needs circumstances being just right.

Notice what the third one costs, and hold it for two spreads.

Joy is not the receipt for having endured something. It does not settle a debt, and it is not evidence that the hard parts were worth it.

The bees are useful here for one narrow reason: nobody asked what the ball rolling was for before agreeing it counted. The paper did not have to show that play made the bees better foragers to publish the finding that they play. The activity was permitted to be its own justification, which is a courtesy the science extends to a bumble bee more readily than the world extends to us.

Bees rolled more when they were younger — most of all between three and seven days from eclosion, which is when a bee emerges as an adult. The pattern rises and then falls, which is the shape play takes in young mammals too.

It was not about experience. Bees that had met balls before and bees that had not showed the same age effect. Being young was the thing.

And there is a second fact sitting next to it. A bumble bee’s mushroom bodies — the parts of the insect brain that handle sensory information and memory — are at their most plastic in the first seven days after eclosion. The play and the plasticity occupy the same window.

What the age curve does not support is the reading we get handed constantly: that play is a stage, and growing up means coming out the other side of it. The older bees did not stop. They rolled less.

Nobody in this study called the older bee cured.

Male bees rolled individual balls for longer than female bees did. The authors offer an explanation, and it is the most useful sentence in the paper.

The females had constant access to pollen and nectar — which is to say, they had a job. The colony’s work was in front of them and available at all times. The males had no such call on their hours. In the authors’ own phrasing, the males had more “free time” in laboratory conditions: less motivation to stop.

Read plainly, that is a claim about environments rather than about bees. Play did not expand into the animals with more capacity for it. It expanded into the hours that function had not already claimed.

The variable was not who wanted to play. It was who was free to.

We know this one from the inside. The question “why don’t you do things you enjoy any more” is almost always answered by an audit of the week rather than an audit of the person — and the audit is rarely the thing anybody offers to change. Wassell’s third theme was circumstances being just right. Circumstances are somebody’s decision.

A finding this charming is exactly the kind this collection is supposed to be careful with. So here is what the authors say against their own result, in their own paper, without being asked.

They cannot exclude captivity. Object play in bumble bees may, they write, follow a lack of stimulation that the natural environment would otherwise provide. Captive animals are given play opportunities as enrichment for exactly that reason. The bees were in a box.

It has not been shown in the wild. Their own recommendation is that future work should examine whether play-like actions occur outside the laboratory. Nobody has watched a bumble bee do this in a meadow.

The motivation is unknown. Further research is needed, they say, to identify what actually drives the rolling.

And the conclusion hedges its own verb. What they report is a behavioural phenomenon resembling object play — not a proof of an inner life. They say it adds to a growing body of evidence for a form of sentience in insects, which is a careful sentence and should stay careful.

None of that is a retreat. It is what a finding looks like when the people who found it would rather be right than quoted.

We are keeping the whole of it, hedges included, because a fact that has to be trimmed to carry an argument was never carrying it. The criteria were met. The place preference held. The age curve is real. What is unsettled is unsettled on the page.

There is a habit of mind that hears “a bee played” and immediately asks what advantage it conferred. That reflex has a history, and it is not a neutral one.

Robert Chapman’s Empire of Normality (2023) gives the structural account: normality is manufactured and enforced rather than found — a standard produced under industrial capitalism, against which people are sorted by what they yield. Nick Walker names the same machinery as the neuronormative paradigm, in Neuroqueer Heresies (2021). Neither is a claim that difference is imaginary. Both are claims about who wrote the ruler.

This zine adds one narrow thing to that, and it comes from the method rather than from us:

A formal scientific standard exists for doing something worthless, it has five criteria, and not one of them asks what the behaviour was for.

Burghardt’s framework does not require play to pay. It requires the opposite — it disqualifies anything that pays. A whole research literature runs on a definition in which usefulness is disqualifying, and it has been running for decades without collapsing into incoherence.

So the demand that our interests justify themselves by output is not the rigorous position in the room. It is the parochial one. The rigorous people gave a bumble bee the benefit of the doubt and wrote down criteria that would let them be wrong.

A zine about play is unusually easy to turn into homework. So, plainly: