Article — "A Jupiter-mass 'exomoon' orbiting a brown dwarf forces astronomers to rethink definitions": Article — "A Jupiter-mass 'exomoon' orbiting a brown dwarf forces astronomers to rethink definitions": Read the arXiv preprint's disclaimer once and the whole story reorganizes itself: the detection rests on "a satellite candidate with a minimum mass of 0.743 Jupiter masses and an orbital period of 169 days," possibly joined by a second body of 0.277 Jupiter masses near 87 days, the two sitting close to a 2:1 resonance — the same pattern that locks Jupiter's Galilean moons.
The authors add, plainly, that in peer review "the results of the accepted work, specifically which of the presented satellite models is favored and the exact parameters of those models, have meaningfully changed."
So the object is real enough to argue about, and unfinished enough that its precise mass is still in motion. Good. Because the argument it triggers doesn't depend on the third decimal place.
Here is the trap, and it is a genuine logical closure, not a mystery. Take the IAU's 2018 exoplanet definition and apply it to the companion in CD-35 2722. Run the reading twice.
First reading, by the explicit rule: a planetary-mass object orbiting a brown dwarf is an exoplanet. The rule doesn't require the brown dwarf to be the largest thing in the neighborhood. So a body around 0.7 to 0.9 Jupiter masses orbiting a 37-Jupiter-mass brown dwarf is, unambiguously, a planet.
David Kipping of Columbia states it flatly: "If you accept that this is a brown dwarf, that kind of forces the companion to be a planet." And then, closing the door: "We can argue about if the IAU definition is wrong or right, but by the current IAU definition it's a planet."
Second reading, by the principle the same document rests on. The IAU authors wrote that planethood is less about size or composition than about a body moving around another "in a hierarchical configuration." Motion has defined planets since the Greeks named the "wandering stars."
But motion is exactly what makes this thing look like a moon. The brown dwarf reportedly orbits the star CD-35 2722 on a roughly 5,000-year loop; the companion orbits the brown dwarf every 169 days. It is, in lead author Kevin Hoy's words, a "moonish thingy" — something orbiting a bigger something orbiting a bigger something.
That nested arrangement is the definition of a satellite. A moon in our solar system is precisely a body that orbits a planet that orbits a star.
So one careful pass through one document yields two irreconcilable verdicts. The explicit clause says planet. The hierarchical principle underneath it says moon. Same object, same rulebook, opposite answers.
One careful pass through one document yields two irreconcilable verdicts. The explicit clause says planet. The hierarchical principle underneath it says moon.
This is not sloppy drafting. It's a structural fault, and it becomes obvious once you ask where the categories came from.
Our words — star, planet, moon — were reverse-engineered from a single solar system in which the three tiers never overlap in mass. The Sun dwarfs the planets; the planets dwarf their moons. That tidy stratification let us pretend that "orbits a star" and "orbits a planet" were natural kinds rather than accidents of one arrangement.
CD-35 2722 breaks the convenient separation. The tiers now overlap. A Jupiter-mass object plays the role of satellite; a 37-Jupiter brown dwarf plays the role of primary.
Hoy has a thought experiment that exposes how thin the labels always were. The largest moons, Titan and Ganymede, are roughly Mercury's mass and size. "We'd be comfortable calling Mercury a moon if it orbited Jupiter and calling Ganymede a planet if it orbited a star," he says.
Read that slowly. It means the label was never about the object. It was about the object's address. Move the same ball of rock to a different orbit and its category flips — a strange property for a definition of a thing to have.
That's why coining "exosatellite," as Hoy's team did, doesn't resolve anything. It just gives the paradox a place to sit. Mary Anne Limbach of the University of Michigan endorses the word — "the field needs language for companions that do not neatly fit into the traditional categories" — but she is careful about what the word is supposed to buy us.
For Limbach, definitions aren't bookkeeping. They're claims about how something formed. "An object can reasonably be called a moon if it formed through a process analogous to the formation of the moons around the Solar System's planets," she says — that is, by core accretion, clumping together out of a debris disc around its parent.
Under that criterion, "moon" and "planet" stop being about who orbits whom and start being about origin story. And that reframes the whole fight. The real question isn't what to call the CD-35 2722 companion. It's whether it accreted from a disc or collapsed from a cloud.
Hoy would bet on collapse. "If I had to put money on it, I would doubt that the CD-35 2722b satellite formed via core accretion," he says, citing the implausibly large disc you'd need to build something Jupiter-scale, the system's youth at only 150 million years, and the companion's eccentric orbit — moons born in place tend to settle into circular ones.
His best guess is that the brown dwarf and its companion collapsed directly out of a gas cloud, "in much the same way that stars do." If that's right, this "satellite" formed by a stellar process, not a lunar one — which by Limbach's criterion makes calling it a moon actively wrong, whatever its orbit looks like.
And there is precedent for pairs that condense together rather than one growing around the other. According to Limbach, JWST reportedly found roughly 40 pairs of rogue binary giant planets — Jupiter Mass Binary Objects, or JuMBOs — drifting free in the Orion Nebula's Trapezium cluster. Limbach sees CD-35 2722 as close to a binary giant planet, and notes we can't yet rule out that the companion is a second brown dwarf, since we only know its minimum mass.
So collect the possibilities on the table for one object: exoplanet, exomoon, exosatellite, half of a binary giant planet, member of a binary brown dwarf. Five names, one body, and the choice among them is really a choice among formation histories we haven't measured yet.
This is where the honest move stops being lexical. The temptation is to keep minting buckets until each oddball has one. But the CD-35 2722 case shows the buckets were never carved at nature's joints — they were carved at our solar system's joints, and those turn out to be local features.
Kipping puts it without ceremony: "Nature doesn't give a damn about how humans want to neatly bucket things."
That's the resolution, and it's an uncomfortable one, because it asks us to give up the premise that a good enough word exists. There is no word that makes a Jupiter-mass body orbiting a brown dwarf orbiting a star into one clean thing, because it isn't one clean thing — it's a point in a continuous space of masses, orbits, and origins that our vocabulary chops into three false categories.
The near future guarantees more of these. Hoy's team is already chasing another candidate; a separate group reportedly reported a tentative half-Jupiter satellite around a brown dwarf in a different system. ESO expects its 39-metre Extremely Large Telescope to pull smaller companions out of the noise, and each will "further reconsider how we label planetary objects," in the observatory's own words.
The flood of hierarchical oddballs won't be a naming emergency to be managed by committee. It's the moment classification changes jobs.
For four hundred years, to classify a body was to name it by where it wandered. The next generation of instruments will make that impossible to sustain, because "where it wanders" no longer tracks anything real about what a thing is or how it came to be.
What survives is the harder, better question the labels were always standing in for: not what do we call it, but how did it form. CD-35 2722 doesn't need a new word. It needs an origin story — and the admission that the origin story was the point all along.