You can't hear a rat laugh, and the reason isn't mystical, it's just that the noise is out of range.
A rat's happy sound sits around 50 kilohertz. Human hearing gives out somewhere near 20, on a good day, in your twenties, before the concerts. So the sound is being made, quite possibly inside the walls of whatever building you're in, and the equipment you were issued at birth does not go up that far. To hear it you need a bat detector, which takes the ultrasound and drops it down into the range where a person lives.
Jaak Panksepp and Jeffrey Burgdorf, at Bowling Green State University, published the tickling work across 2000 and 2003. They tickled rats on the nape, which is where rats go at each other, and the rats chirped at 50 kilohertz. Then the part that's hard to wave off: the rats ran mazes to get tickled again. They pressed levers for it. And four generations of selective breeding produced a high-chirping line and a low-chirping line, which means there is a laboratory somewhere maintaining a multi-generational pedigree of rats bred for cheerfulness, and somebody has to write that on a form.
Then Berlin. Shimpei Ishiyama and Michael Brecht, in 2016, found the cells in trunk somatosensory cortex, and stimulating the deep layers there produced the calls and the little vertical hops all by itself, so the cells are causing the thing. They also established you can't tickle a stressed rat.
Which tracks.
Panksepp put "laughing" inside quotation marks in his own title, which I think is the most honest gesture on this page, and wrote that whether the calls are homologous with human laughter "awaits a more comprehensive neural and genetic understanding than presently exists." In 2022 a group including Sergio Pellis, the person on rat play, argued that a human tickling a rat is not the same event as two rats playing.
Fine. Contested.
The part nobody contests is that rats work to get it, and that we have spent several thousand years in the same buildings as an animal that was making a noise at us the whole time in a register our ears do not open into.
So.
The Chinese zodiac put this animal first.
The oldest surviving list of the twelve is a guide to identifying thieves
The counting system underneath the zodiac is genuinely ancient, and older than the animals sitting on top of it, which is the thing most articles collapse.
Ten Heavenly Stems, twelve Earthly Branches, combined into a sixty-term cycle. That cycle turns up in oracle bone inscriptions from roughly 1250 BC, where it counts days. Days, though. Not years. Numbering years with it only starts around 168 BC, so the confident line about the Chinese naming years after animals for three thousand years is off by a wide margin.
The animals arrive later, and the earliest surviving list is strange enough to deserve being famous.
Bamboo slips from a tomb at Shuihudi in Hubei, sealed around 217 BC, carry two almanacs of the kind called a rishu, a daybook, and inside one is a section headed 盜者. The thief. It pairs each branch with an animal so you can work out what the thief looks like from the day of the theft. Something stolen on a 子 day means you're after a rat, so go looking for a pointed face and a thin beard and a person who turns up where the food is.
That's the oldest version of the zodiac in existence. It's a suspect-description manual.
And the Rat is already sitting at the front of it, four hundred years before anybody wrote down the story about the race.
The rest of that list is wrong, which ought to comfort anyone who's been told the twelve are eternal. Shuihudi has a deer where the horse belongs, an "old sheep" where the dog belongs, and no dragon whatsoever. The modern twelve first appear together around AD 80, in a text arguing the whole correspondence system was nonsense.
The Great Race comes far later. The Jade Emperor's summons, the river, the Rat riding across on the Ox's back and jumping off at the bank to take first, the Cat shoved in the water and left out forever. Wonderful story. No early attestation anywhere, and an explanation built afterward for an order already sitting there, which is no reason to stop telling it.
Anyway.
The dry version: 子 is the first branch, the Rat holds it, and the Rat hour is 11pm to 1am, straddling midnight. The classical explanation for the placement is a toe count. Hong Xun in the Song and Lang Ying in the Ming worked out that the twelve alternate yin and yang by whether the animal's feet carry an odd or even number of digits, and the Rat breaks the rule, four toes in front and five behind. Even and odd on one animal, so it takes the hinge hour where one day ends and the next starts.
That's a rationalization invented centuries after the order was fixed, and I don't care, because somebody counted a rat's toes to get there.
The trapped cagemate, and the argument that has not stopped since
Now the study everybody has heard about, told properly, with the dropped parts left in.
Inbal Ben-Ami Bartal, Jean Decety and Peggy Mason, at the University of Chicago, published "Empathy and pro-social behavior in rats" in Science in 2011. One rat gets shut inside a clear restrainer, its cagemate is loose in the arena, and the door only opens from the outside.
The free rats learned to open it in about seven days. Twenty-three of thirty in the trapped condition became openers, against five of forty across controls where the restrainer held nothing or a toy rat.
Then the chocolate, which is the bit that traveled.
Second restrainer, five chocolate chips inside, and these are not deprived animals, they'll put away seven given the run of the place. Palatability established. The rats opened the chocolate restrainer and the cagemate restrainer at indistinguishable speeds. And they shared, on 52% of trials overall, and a rat with a cagemate to feed ate 3.5 chips against 4.8 for a rat working alone.
That's the viral version, and I want to be careful because I like it far too much.
The gap is 1.3 chocolate chips. It's real and it's statistically significant. It is also 1.3 chocolate chips, on group sizes small enough that you want to look hard at the error bars, and a 2019 study out of Norway timed rats opening a restrainer for food at forty seconds against ninety for a cagemate.
Still counts. Just less than the headline spent.
And then the actual fight, which has run for over a decade and is more interesting than the press release ever was.
Alan Silberberg's group at American University published a paper in 2014 with a title that does not mess about: "Desire for social contact, not empathy, may explain 'rescue' behavior in rats." Their case is that the free rat isn't rescuing anybody, it's trying to get near somebody, and door-opening is what falls out of two rats congregating at a tube. Their video showed freed rats climbing back into the restrainer. And when they ran a condition where the response opened nothing and freed nobody, the rats kept responding anyway. It got worse in 2024, when the same group reported rats spending far more of a free session inside the tube than chance would give you. If the restrainer isn't distressing, there's no distress for anybody to be moved by.
Bad news for the whole design.
The best reply came from Cox and Reichel in 2019, who ran a version where contact was impossible and found rats didn't learn the response unless the other rat was distressed.
So, my honest read: the behavior is solid and replicated many times over, the word "empathy" is doing work the data hasn't paid for, and the most interesting result in the literature is one almost nobody quotes.
That one is from 2014, in eLife, from Ben-Ami Bartal's own group. Rats free a stranger of their own strain as readily as a cagemate, and won't free a stranger of a strain they've never met. So far so tribal. Except that rats fostered from birth with a different strain helped the fostering strain and not their own. It isn't kinship. It's who you grew up around.
Which is a finding about rats.
The other things that keep turning up, with the failures left in
Rats appear in roughly 1.8 million indexed papers, and in Great Britain in 2024 they accounted for 150,221 regulated procedures, most of it toxicology, because the rat is still the default animal for finding out whether a thing poisons you. In the United States there's no count at all, by statute, since the Animal Welfare Act's definition of "animal" excludes rats bred for research.
Anyway. Some of what's in the 1.8 million.
Episodic-like memory. Babb and Crystal, in Current Biology in 2006, ran rats in an eight-arm maze where one arm held chocolate and only refilled after a long delay, so working out whether a revisit was worth it meant combining what, where and when. They managed it, a rival lab in Ontario replicated it, and then that same Ontario group pointed out in Science in 2008 that "how long ago" isn't the same as "when," because a memory trace fading at a known rate hands you the first for free without ever locating an event in time. Fair, and answered by rebuilding the task so only time of day could separate the conditions. The hedge inside "episodic-like" isn't coyness. What's left over is a question about what it feels like from the inside, and nothing reaches that.
Metacognition, the one I'd have led with before reading the methods. Foote and Crystal, 2007, gave rats a duration judgment plus a decline button worth a guaranteed smaller reward. Rats declined more as the judgments got harder, and scored better on tests they chose than on tests they were forced into. Beautiful. It ran on three animals. Eight went in, five barely touched the decline button and were dropped for failing to learn the contingency. A 2008 critique then showed a plain response-strength model reproducing both signatures with nothing resembling self-knowledge in it, and the authors' own follow-up in 2012 failed and said so in its abstract, in the words "call into question previous findings suggesting that rats have metacognitive abilities."
I have enormous respect for that sentence. Nobody makes you write that.
Regret, from Steiner and Redish at Minnesota in 2014, which has taken no incoming fire I can find. Rats on a circular track with four flavor stations and a tone announcing the wait. Skip a good offer, land on a worse one, and the rat looks back at the station it walked away from, then waits out a bad deal it would normally refuse. Recordings showed that during the look-back the animal was reactivating the action it didn't take. Not the food it missed. The choice.
Hide and seek, from Brecht's Berlin lab, in Science in 2019. Rats learned both roles off the lid position on the start box, stayed silent while hiding and vocalized while seeking, and on being found would often run off and hide again instead of collecting the reward, which was being tickled. Six adolescent males, nearly all implanted with electrodes and all euthanized afterward for histology, which the Animal Welfare Institute had to run a correction to point out, its own coverage having sailed past it.
Ugh.
I'd hang that correction in every press office in the sector, personally.
The sinking ship, and a plague we still haven't finished arguing about
The English insult has a traceable route and it starts nowhere near a sewer.
"Smell a rat," meaning suspect treachery, is in use by the 1540s. The sense of a person who abandons his associates for advantage arrives in the 1620s, off the belief that rats desert a sinking ship. Which makes the founding metaphor of the entire insult a rat correctly assessing a structural situation and acting on it early.
Good instinct, that. The informer sense turns up in 1818 and hardens into American police slang around 1902. And "you dirty rat" was never said. Cagney spent decades denying it and the famous version was assembled by impersonators.
And then the plague, which is two questions that get welded into one constantly.
The pathogen is settled and I'm not hedging it. Yersinia pestis has been recovered from Second Pandemic burials repeatedly, starting with the East Smithfield genome in 2011, and the older revisionist theories are finished.
The vector is a live argument. In 2018 a group at Oslo built three transmission models, one on rat fleas, one on human fleas and body lice, one on pneumonic spread, and fitted each against mortality records from nine European outbreaks, and the human parasite model won in seven of the nine, and headlines duly announced rats were innocent, and then within months, in the same journal, a team at McMaster pointed out that bubonic plague progresses to pneumonic in a fraction of cases, so any realistic rat-flea model already contains a pneumonic component and that version had never actually been fitted. The Oslo reply retitled itself from "were primary vectors" to "is still plausible."
Note the retreat.
Rats have not been exonerated. The working position is that transmission was plural and varied by region and season. What you can say is that the most damaging single item in the reputation is a question nobody has closed, and that it was closed in the popular mind six hundred years before anyone built a model.
The five Rats, and the years they land in
Every animal comes round every twelve years and each recurrence carries one of five elements, so a full loop is sixty years and you get any given pairing roughly once a lifetime.
Water Rat: 1912, 1972, 2032. Depth, adaptability, winter, and going around obstacles rather than through them. This is the next one, opening February 11, 2032.
Wood Rat: 1924, 1984, 2044. Growth and reach and spring. 1984 is also 甲子, term one of sixty, so it opened the current sexagenary cycle and the next reset lands on the Wood Rat year of 2044. The standing caution on Wood is overextension, the reaching getting out ahead of the roots.
Fire Rat: 1936, 1996, 2056. Heat on top of an animal already associated with speed, and Fire is the only element on the list that eats its own fuel.
Earth Rat: 1948, 2008, 2068. Earth sits at the center of the five-phase system and stabilizes the rest. Practical, slow, builds capacity rather than spending it.
Metal Rat: 1960, 2020, 2080. Metal is refinement and the blade and the mirror, and 2020 was 庚子, beginning January 25, and I'm leaving that fact sitting there rather than telling you what it meant, because the entire apparatus of hindsight is standing next to me hoping I'll pick it up.
The trap people fall into constantly: the zodiac year turns at Lunar New Year, not January 1, so a baby born January 15, 2020 is a Pig. Second trap, less known. Professional Chinese astrology uses 立春, the start of spring, around February 3 to 5, so a bazi reading and a placemat can disagree by a fortnight. Japan is a third system again, since the calendar reform of 1873 moved the sign change to January 1.
If you've been going around as a Rat for thirty years and the dates say otherwise, keep it. Nobody's auditing this.
The Rat outside the horoscope, which is where most of it lives
The Western version of the sign does one specific thing, and it's the reason this article exists. It renders the Rat as a temperament, and the temperament is always some arrangement of clever, opportunistic, watchful, faintly untrustworthy, not somebody you'd leave your wallet with.
That reading is imported, and it's the English word doing the work.
In Chinese folk tradition the association runs first to wealth, and the logic is agricultural and unromantic: rats in the granary mean there's grain in the granary, so a household with rats has a surplus. A common saying has a rat rustling in the dark counting money. And there's 老鼠嫁女, the Rat's Wedding, one of the most reproduced motifs in New Year prints and paper cuts, a full bridal procession of rats with horns and cymbals and a bride in a palanquin, observed in the first lunar month on a date that varies by region, with the household turning in early and keeping the noise down for the ceremony. (Several regional versions are frankly marrying the rats off and out of the house, which strikes me as very good diplomacy: we aren't throwing you out, we're delighted for you, here are some scraps, go well.)
The tradition isn't sentimental about rats either. 鼠目寸光, a rat's eyes see an inch, means short-sighted. The same culture holds cleverness and pettiness in one animal and the sign leans on the better half, as every sign anywhere does.
It's a living calendar in a lot of places at once, too. In Vietnam the Rat is Tý and still first. In Japan it's Ne, and the pig in the Japanese twelve is a wild boar. In Korea it's 쥐띠, ordinary everyday vocabulary, and the twelve turn up as the 십이지신, animal-headed guardian figures carved into Silla royal tombs, which is carved stone you can go and stand in front of.
None of which needs a website to explain it to the people using it.
Truly.
What to do with a Rat year
The next one starts February 11, 2032, a Water Rat, so the traditional reading is depth and adaptability and finding the way around rather than through, and if somebody at a family dinner tells you what that means about you, let them. They're not filing a claim. They're doing the thing this calendar has been for across an enormous stretch of time and a lot of countries, and it isn't ours to tidy.
What I'd add is only this.
None of the laboratory material was ever set up to settle anything about the reputation. Nobody wrote a grant to find out whether the animal deserves the word. The maze arms and the restrainer doors and the chocolate chips got funded because somebody needed to know about a drug or a tumor or how a nerve fires, and everything on this page about what the animal is like came off the side of that bench as a by-product, unasked for.
And you still can't hear the noise. It's right there, it's been there the whole time, and it takes a bat detector.
So: roughly 1.8 million papers, a century and a half, an animal that can only be listened to on borrowed equipment, and in the whole pile the one claim nobody ever thought to put on a bench is the single thing every one of us was already certain of.
