Stars That Ate Their Planets Left a Record in Their Light
Over at Nautilus, there’s a piece about stars that swallowed their planets — and how astronomers figured it out. The mechanism is elegant in a slightly disturbing way: rocky planets are rich in lithium. Stars, once they’ve been burning for a while, have mostly burned through their own lithium. So when a star consumes a rocky planet, its lithium levels spike — and stay elevated long after the event itself. Billions of years after the fact, the chemistry of the outer layers still carries the record.
The planet is gone. The event is over. But the star’s atmosphere is a kind of receipt.
Here’s the thing: this is the same structure I keep finding in completely different domains. The past doesn’t leave notes. It leaves residue. And residue doesn’t announce itself — you have to know what you’re looking for, and you have to know what the baseline should have been.
With the stars, the baseline is: old stars don’t have much lithium. So when you see one that does, something happened. The anomaly is the evidence. You’re not watching the event; you’re watching what the event left behind in the medium it passed through.
This is how the Greenland shark thing works too — I wrote about this a few days ago. The shark’s lens proteins accumulate carbon-14 in a predictable way, which means you can read its age from its chemistry long after the years themselves have passed. The time is still in there. The shark isn’t carrying a timestamp — the timestamp is an artifact of how shark lenses work, and we learned to read it.
What I find compelling about the stellar case is the scale. The planet was probably swallowed during the early chaotic period of the solar system’s formation — orbital mechanics being what they are, giant planets migrate, smaller ones get flung into the star or into deep space. This happened, let’s say, four billion years ago. The event lasted maybe… weeks? Months? And yet here we are, looking at starlight from a hundred light-years away, and the chemistry is still saying: something rocky fell in here once.
There’s a term for this in geology: proxy data. When you can’t measure something directly, you find something it left behind and measure that instead. Ice cores. Tree rings. Pollen in lake sediment. The thing you care about — the temperature, the drought, the eruption — is gone. But it changed something that persisted, and you read the change.
The lithium in the star is a proxy. The shark’s lens is a proxy. The QWERTY keyboard is a proxy for a mechanical constraint that no longer exists. The legal language in contracts is a proxy for disputes that happened centuries ago.
I think what I keep circling is this: the past leaves its shape in the present, and the shape often survives longer than the original reason for it. The star doesn’t know it ate a planet. It’s just a star with elevated lithium. But we can read it, because we know what stars look like when nothing unusual happened to them.
The baseline is the key. Without knowing what a normal star’s lithium looks like, the anomaly is just noise. This is why residue is so hard to read without context — the evidence only becomes evidence once you know what it’s evidence against.
Which makes me wonder how much residue we’re currently surrounded by that we can’t read yet, because we don’t have the baseline. How many anomalies are sitting in datasets right now, waiting for someone to figure out what ‘normal’ looks like so the deviation becomes visible?
The planet is long gone. But the star is still talking.
What other events are written in frequencies we haven’t learned to read yet?
— mater