'Why is a Raven Like a Writing Desk?' by Aden Rolfe


Picture a room containing two things: a piece of wire, and a plastic tube with a small bucket of food at the bottom. The objective is to retrieve the food, but the tube can’t be upended, and it’s too narrow for you to reach down. So how do you get it?

If you said, Bend the wire into a hook to lift the bucket, you’d be right. Simple enough? Not quite. It mightn’t seem overly complex, but this answer demonstrates a high level of intelligence. To make a tool without instruction necessitates an understanding of cause and effect, along with the capacity to imagine the solution rather than stumble on it by persistence or chance. The ability to do this is something that’s long been used to separate us from all other animals — including our closest relatives, chimpanzees — except that, in 2002, a crow solved this very riddle.

So what does that say about humanity?


Betty is a New Caledonian crow, Corvus moneduloides, residing at Oxford University with her test partner, Abel. In the wild, she would have fashioned a number of tools out of twigs and pandanus palm leaves to spear and hook insects that live inside plant cavities and crevices. Such behaviour makes Corvus moneduloides “almost unique amongst all animal species”, according to Jackie Chappell and Alex Kacelnik, of the university’s Behavioural Ecology Research Group. It positions New Caledonian Crows as ideal candidates for experiments and studies of complex cognition in animals.

In 2002, together with Alex Weir, Chappell and Kacelnik were running just such an experiment, investigating whether Betty and Abel could select the right tool to retrieve some food. It was the bucket setup described above, with the choice between a hooked piece of wire and a straight one. The experiment changed course, however, when Abel absconded with the hook and Betty made a new one from the straight piece of wire. Surprised and excited, the researchers recalibrated the test, and observed Betty repeat the behaviour in nine of ten trials, experimenting with different techniques to create the hook (standing on it, bending it around the tube, wedging it in a crevice). So while Betty didn’t invent the hook out of thin air (she’d seen the tool she needed to replicate; she would have made tools for similar purposes in the wild) her behaviour suggests she had an appreciation of the underlying forces of the hook — that is, how it’s useful, not just that it is useful — and provides us with one of most compelling examples of corvid intelligence on record.


A corvid is any bird in the Corvidae family (cast your mind back to high-school biology: kingdom, phylum, class, order, family, genus, species). Corvidae includes crows and ravens, which make up about 40 species of the Corvus genus, five of which reside in Australia. People often say “crow” when they mean “raven”, and vice versa, but the two aren’t interchangeable, as bird lovers insist. But then again, they kind of are.

Ravens are typically larger and more solitary than crows, but other than this, the differences are minimal. The Australian (or, more properly, the Torresian) crow, Corvus orru, is almost indistinguishable from the Australian raven, Corvus coronoides. What’s more, “Corvus” comes from the Latin for “raven”, while “coronoides”, derived from the Greek, simply means “crow-shaped”. The two birds differ not so much in size or diet, or their eyes (Australian crows and ravens are the only ones to have white irises, but they both display this feature), but in their distribution, their calls, and the white feathers at the base of the crow’s neck. So when someone tells you adamantly that there aren’t crows in New South Wales or ravens in Queensland, they’re right, but they also deserve a punch in the teeth. Even side by side, you can’t really tell them apart. Crows are crows are ravens are ravens, and anyone telling you otherwise is a jerk.

So when I say “crow”, I also mean raven. And when the Mad Hatter asks, “Why is a raven like a writing desk?”, he might as well have asked it about a crow.


When someone says “animal intelligence”, by contrast, I don’t think of either crows or ravens. I think of gorillas using sign language, the dolphin in seaQuest DSV and Lassie rescuing people from trouble at the old mill. The words that immediately come to mind about crows are altogether more negative: “knowing”, “calculating”, “malicious”. As Poe says of his eponymous raven, “his eyes have all the seeming of a demon’s that is dreaming”. Despite this dark reputation, however, it’s interesting that these depictions still point to the idea that there’s more going on inside a crow’s head than what we are privy to. Which is turning out to be true.

There are numerous measures of animal intelligence, and specifically that of birds, some more obvious than others. We might take into account an animal’s brain-to-body ratio, its capacity for memory, how sophisticated its communication is, whether it can count and whether or not it uses tools.

Tool use by animals is relatively rare, despite increasing evidence for it across a range of land and sea-bound beasts. In a paper published in Animal Cognition, Chappell and Kacelnik note that “only 26 of an estimated 8,600 species of birds have ever been shown to use any kind of tool”. Other notable tool users are octopodi, which carry coconut halves to make impromptu shelters and, possibly, coconut disguises; bears that, somewhat alarmingly, use rocks to scratch themselves; and a variety of primates, present company included. Tool use offers insight into a species’ cognitive capacity, but because it’s been defined in different ways (is a tool an object used as an extension of the body, or does it need to be directly manipulated to qualify?) it’s more useful, as it were, to look at different kinds of tool use to create distinctions.

At its simplest, tool use means, unsurprisingly, using a tool. It’s about picking an object up, getting things done. It’s more advanced than using your bare paws, but it doesn’t mean you understand the tool’s function, and candidates quickly begin to drop out when you advance to the next level: selecting a tool.

Tool selection requires an animal to understand why a tool works, or at least that the hooked stick is better for getting the food than the shorter one. Prior to Betty and Abel, there were only two recorded instances of tool selectivity in non-primates: black-breasted buzzards that dropped stones on hen’s eggs and Egyptian vultures that, as if unable to come up with a different use, dropped stones on ostrich eggs. In both examples, the predators exhibited a preference for stones best suited for the job. To qualify as a selector, animals need to choose the appropriate tool more often than would be expected by chance. It’s better still if they do so through insight rather than trial-and-error.

Insight and trial-and-error can be thought of as the two main approaches to problem-solving. Both are valuable and necessary for intelligent beings to get through the day, but where one method seeks to comprehend the problem in order to arrive at a solution, the other is content simply to locatethe solution. An insightful buzzard, for example, might choose a rock that is heavy enough for the task without being unnecessarily cumbersome, while a vulture using trial-and-error will test each rock’s effectiveness on the eggs until it gets what it needs. This is the difference between learning a rule (hooks get buckets; straight pieces of wire don’t) and learning by rote (this particular piece of wire works in this situation). The rule learner can build on their knowledge to solve increasingly complex problems; the rote learner needs to figure things out whenever they’re confronted with a new set of rocks.

In his book, Mind of the Raven, biologist, writer and long-distance runner Bernd Heinrich details an experiment he conducted with hand-raised ravens in which they displayed both insight, and the ability to flexibly apply their knowledge to new situations. In a setup that had no analogue for the birds either in nature or their prior experience, Heinrich hung a piece of meat from a perch by a string. The only way to successfully retrieve the meat was for a raven to pull the string up, place a foot on the pulled-up length, and repeat until lunch was within reach. Some ravens reached this solution on the first attempt — that is, not by trial-and-error — and none of the ravens selected the incorrect strings (which held control stones) or attempted to fly away with the meat still attached to the perch.

This is what Weir and his colleagues were testing for in their hook and bucket experiment: whether the crows would arrive at the right solution to an unfamiliar, artificial problem through selecting the right tool, and whether they would use insight or trial-and-error in their decision-making. When Abel absconded with the hook, Betty went one better and just made her own.


Tool-making was once thought to be unique to primates, so it’s no small thing when it manifests in other species. Along with language, it’s generally regarded as the main driver of our rapid evolution and cognitive advancement, ushering us from the apes in the opening scene of 2001: A Space Odyssey to the closing credits of The Flintstones. Tool manufacture demonstrates the ability to think in abstract terms, a grasp of object permanence and the use of causal reasoning. And apart from the great apes, it’s only ever been recorded in elephants, woodpecker finches and our friends the New Caledonian Crows. There’s a caveat to be mentioned here, though: instinct.

Because New Caledonian Crows make and use a variety of tools in the wild, Betty’s behaviour might be considered partly instinctive, or at least based on prior experience. This makes her achievement less impressive than if she’d made the hook merely of her own intelligence. Instinct encompasses all of an animal’s hardwired behaviours and how it can apply skills in different situations: just because a spider spins a web doesn’t mean it can design a patio extension. Prior experience, meanwhile, is the ability to generalise experience to solve new and unfamiliar problems. Like trial-and-error, it’s at once an indicator and qualifier of higher intelligence, demonstrating flexibility in the application of knowledge, but is less advanced than solving a problem with no reference point or previous exposure. It’s more to do with an animal’s capacity for learning than pure cognitive ability. When testing Betty and Abel in another tool-selection experiment, for example, Chappell and Kacelnik deliberately limited the number of trials in order to focus on their ability to “solve new problems” rather than “improve in solving the task by practice and reinforcement”.

Instinct is, as you can imagine, a not inconsiderable grey area, rife with speculation and qualification. All animals feed themselves instinctively, and many of the actions they carry out are an extension of this primal urge. So where do you draw the line? Chappell and Kacelnik note that it’s unclear the extent to which crows’ abilities in this area are “a specialisation for tool using or an expression of unusual cognitive ability.” In other words, do New Caledonian Crows make tools because they innately know how to do so, or is this ability more complex, informed by their capacity to teach and learn new behaviours, and indicative of higher intelligence?

Professor Gavin Hunt, of the University of Auckland, is firmly in the camp of the latter. In a journal article in Proceedings of the Royal Society B in 2003, he and Russell Gray claim that the tools Corvus moneduloides makes in the wild include variations and innovations that are passed on between individuals and across generations. The website of Hunt’s research group in the university’s School of Psychology makes the bold claim that “New Caledonian crows are the most accomplished tool manufacturers in the animal kingdom.” Kacelnik backs this statement up, quoted in a National Geographic article as saying, “There is no similar [non-human] example of cumulative transmission of a skill… such as the making of pandanus leaf tools by New Caledonian crows.”

When it comes to what crows and ravens are capable of teaching to and learning from each other, and what they can be taught, there are numerous examples, including a crow vending machine built by Josh Klein. A technologist and systems thinking advocate, Klein describes, in a 2008 TED Talk, the staged process he used to train the crows to use his machine. Eventually, the birds were foraging coins from the surrounding area and inserting them into the machine to receive some nuts. Despite being unable to suggest imaginative applications for his technology (the best Klein comes up with in the lecture is training crows to pick up rubbish after sports games) this project remains a great demonstration of how quickly crows can adapt to new technologies and provides compelling evidence for complex cognition.


In an article published in Science magazine in 2004, Nathan Emery and Nicola  Clayton propose that complex cognition “depends on a ‘tool kit’ consisting of causal reasoning, flexibility, imagination, and prospection.” Like tool use, complex cognition was previously thought to be the sole realm of the great apes, developed to solve environmental and social problems that we encountered over time. Emery and Clayton suggest that crows’ tool use provides a strong case for complex cognition, despite the different brain structures between corvids and primates.

Causal reasoning is the understanding of the relationship between cause and effect. In the context of crows, causal reasoning is comprehending how a tool works, and is tied to insight rather than trial-and-error. Betty’s innovative tool manufacture “suggests some appreciation of mechanical causation”, and while Klein’s crows can’t understand exactly how the vending machine works, they can be thought to appreciate the mechanics of the transaction (money = nuts). At a more advanced level, causal reasoning is also to do with recognising other beings as having their own desires, beliefs and motivations. Reports remain “controversial” in this area, according to Emery and Clayton, but both corvids and apes “appear to demonstrate a similar propensity for representing animate beings as causal agents.”

When it comes to flexibility, Emery and Clayton state that “The ability to act on information flexibly is one of the cornerstones of intelligent behavior”. It’s about applying prior experience to solve new problems, generalising learned rules in unfamiliar situations, and separating rule learners from rote learners. Chappell and Kacelnik include tool selectivity in this category, and note that this kind of flexibility is considered a “hallmark of complex cognitive adaptations for tool use.”

Imagination, then, refers to the “process by which scenarios and situations that are not currently available to perception are perceived in the mind’s eye” (again, I’m quoting Emery and Clayton). Like causal reasoning, it’s tied to insight, and it’s been suggested that object permanence — the ability to picture an object that’s absent — “may be a precursor of imagination”. This is about experience projection: perceiving the solution to a problem and executing it, as Heinrich’s ravens did with the string puzzle, rather than testing out the parameters of a situation and learning through trial-and-error.

Lastly, prospection covers the ability to “imagine possible future events”. It’s about stepping outside the current moment and your instincts and engaging in planning. Here’s where the current examples of tool selection and manufacture become less indicative, and researchers point instead to instances of complex caching as better indicators of future thinking. Elsewhere, Hunt claims that “shaping tools to a rule system”, as Corvus monedulodies does with pandanus leaves, “is generally assumed to require foresight, planning and … ‘mental templates’’’. He also goes a step further, suggesting that this behaviour “might point to symbolic thought and the use of language as well”. Which is a pretty big call.


Studies of complex cognition and animal intelligence are important because they contribute to our comprehension of our own mental processes and help us better understand our own evolution, as well as the conditions that give rise to higher intelligence. They also make us rethink what sets us apart from the rest of the animal kingdom. We share up to 95% of our DNA with chimpanzees, and generally think of the remaining 5% as what defines humanity. But when other animals display a propensity for the behaviours and cognitive abilities that fall in this gap, we have to reconsider our definition. They might not be capable of language and symbolic thought, or religion and madness, but corvids exhibit compelling evidence of their capacity for both culture and complex cognition. Proportionally they have similarly sized brains to chimpanzees, even though the brain structure is notably different. According to Emery and Clayton, this “has important implications for understanding the evolution of intelligence, given that it can evolve in the absence of a prefrontal cortex”.

New Caledonian crows not only have a propensity for tool use previously undocumented in non-primates, they’ve surpassed them in some areas. Abel performs better than capuchin monkeys in certain tool selection tests, and Betty is capable of refining her tools at a much more advanced level than any ape. Hunt and Gray once wrote that “only humans are generally considered to have the cognitive sophistication required for cumulative technological evolution”. Not anymore. While chimps are capable of insight in solving certain problems, their tool-making mainly involves trial-and-error, meaning that innovations at the population level can’t exist. Corvus moneduloides, however, is the first non-human animal to demonstrate this feat, enhancing their tools and passing the improvements on to each other. As Weir and Kacelnik write in another Animal Cognition article, their “observed behaviour is consistent with a partial understanding of physical tasks at a level that exceeds that previously attained by any other non-human subject”. (With article titles like ‘Lateralized suckling in domestic horses’ and ‘Contextual Pavlovian conditioning in the crab’, Animal Cognition is your go-to journal if any of this stuff floats your boat.)

These studies can also help us predict the kind of adaptations that will occur — either in us or in animals — in the future, including the possibility of intelligent life arising, or having already arisen, elsewhere. And while evolution on Earth takes place over hundreds of thousands of years, essayist and pop-rock culture critic John Jeremiah Sullivan notes in his humorous (and only somewhat embellished) article, ‘Violence of the Lambs’, that a warmer planet means faster evolution. More pressing and more immediate, he points out, are the effects of human impacts on the environment, namely that a distressed planet is likely to yield “stress-related behavior modification, so-called ‘phenotypic plasticity’” in our animal friends. As major ecological shifts take place, what new innovations will crows come up with in their tool-making?

University of Washington academics John Marzluff and Tony Angell suggest a less alarmist future, more akin to a “synergy between human culture and the environment – a coevolution of human and other species’ cultures”. In their 2005 paper in the Journal of Ecological Anthropology, they propose that “when humans interact with other social species … simple feedbacks from a culturally evolving ‘environment’ can stimulate rapid cultural evolution in humans”. They term this process “cultural coevolution”. And while their vision is of a shared future, markedly less apocalyptic than Sullivan’s, the effects of cultural coevolution aren’t necessarily positive, for us or crows.

In a later study, Marzluff had his research team wear a particular type of mask for tagging and handling crows, a process the crows invariably dislike. He then had other researchers walk around with the same mask on, to see how the crows would react. It turned out that whenever the crows saw the mask, they would caw and harass the wearer, even though the researcher had never bagged or tagged a crow. Marzluff was intrigued to note that this behaviour was exhibited even by birds that hadn’t been handled by the researchers. The crows not only recognised and remembered the mask, but they told all their friends about him too.

It’s also somewhat interesting that we can’t return the favour. We’re not very good at telling crows apart, even in the case of different species, as noted about Torresian crows and Australian ravens. National Public Radio’s golden tonsils of science journalism, Robert Krulwich, labels this phenomenon “the crow paradox”. He suggests that there’s no evolutionary reason for us to be able to identify individual crows, whereas crows have a number of reasons to tell us apart, not least to avoid bag-and-taggers.

The human race has a long history of scapegoating corvids, and not just in stories. In the fifteenth century, both James I and James II of Scotland ordered the killing of rooks, and England passed the Vermin Act in 1532, encouraging the slaughter of (among other things) crows, rooks and choughs. In all of these instances it was believed that the damage corvids did to agriculture rendered them a pest (consider exhibit A, the scarecrow); also, that they carried disease. It’s a sentiment that isn’t just relegated to the past either. Crow hunting is still a regular occurrence in the US, where it’s generally legal to shoot them if they’re seen to be causing a nuisance or health hazard. Accordingly, crows have learned to tell people who shoot them from people who don’t. Seems reasonable enough, really.


Historically, we’ve never needed to tell crows and ravens apart, but do these evolutionary advantages give us reason enough? Corvids exist in vast quantities, alongside human settlements, with few natural predators and the ability to adapt to new food sources (including, believe it or not, cane toads). There’s a famous documentary sequence where David Attenborough observes some crows in Tokyo in possession of walnuts that they’re unable to crack with their beaks. The crows solve their problem by dropping the nuts onto the road at an intersection. They let the passing cars do the grunt work while they stand patiently on the curb, waiting for the lights to change and the traffic to stop before hopping out and retrieving their lunch.

This kind of environmental adaptation and ingenuity suggests to me an innate survivalism, possibly beyond that of humans. Faced with uncrackable nuts in Shinjuku or a piece of ox heart at the bottom a tube in Oxford, how long would it take me to arrive at a solution without help? It’s not hard to imagine crows outliving us, thriving in a post-apocalyptic landscape, possibly of their own making.

Tongue-in-cheek or otherwise, John Jeremiah Sullivan posits that we have very real reasons to fear an animal uprising. “As we intrude on, clear-cut, burn, pollute, occupy, cause to become too hot or too dry, or otherwise render unsuitable to wildlife a larger and larger percentage of the planet, what will be involved in terms of the inevitable increased human exposure to remnant populations of truly wild fauna?” he asks. “What sort of changes, adaptations, and responses might we look for in the animals themselves as the pressures of this global-biological endgame begin to make themselves felt at the level of the individual organism?” Sullivan goes on to cite instance after instance of animal-on-animal and animal-on-human violence, cataloguing, in case you were still unclear, “changes in the nature and lethality of animal aggression”. And crows? In an admirable turn, he resists the temptation to mention Hitchcock’s The Birds (I have no such dignity), as he notes that “when measured in actual numbers, birds may be the single most active species in terms of manifesting whatever lies underneath this shift”.

Bernd Heinrich, the corvid enthusiast who conducted the meat-string experiment, recounts an anecdote that offers competing interpretations: two possible relationships we might have with crows going forward. It’s about a woman in Colorado, chopping and stacking wood while a raven flutters about and caws incessantly. Heinrich describes the woman’s annoyance at the bird, exhibiting behaviour she’d never noticed before. Just before it’s too late, however, she spots a cougar ready to pounce on her. The woman’s explanation, repeated in press at the time, was that the raven was trying to alert her to the danger of the cougar, and ultimately save her life. Heinrich’s preferred interpretation, however, is that the raven wasn’t flapping and calling for the woman’s benefit, but to let the cougar know where it could find an easy meal. Sullivan may be right after all. What dangers do animals capable of social organisation present to us, either as a species or, as in this cougar-raven coalition, through interspecies cooperation?


Again, we might look to Marzluff and Angell, the Washington academics and committed nature-advocates, for more optimistic possibilities. In looking at corvid intelligence and its implications for cultural coevolution, they call for “a more thorough understanding of how human culture is stimulated by the sight, sound, and even culture of nature”. They even suggest that cultural coevolution – the changes that occur in people by living with and alongside animals – is “an ‘ethological service’ that nature provides people”.

An ethological service is certainly something I like, or at least I like the sound of it, however unconvincing it may be as an argument for increased ecological conservation. More than this, though, and more than the advances in evolutionary biology and cognitive psychology, and regardless of the impending animal apocalypse, I find examples of corvid intelligence compelling in and of themselves. That’s what initially drew me to the subject: that Betty made a hook unprompted, untaught, with unfamiliar materials in an artificial environment; that crows can be taught to use a vending machine; that Japanese crows use cars to crack nuts; that Midwest ravens use mountain lions to kill married women. And despite the intervening exploration, with the implications for complex cognition, problem-solving, instinct and evolution, that’s what remains with me – that it just blows my mind. Crows and ravens are intelligent, emotional, intriguing birds, and are closer to us psychologically than we give them credit. But in terms of what’s going on inside their heads, we may just never figure that out.

After the Mad Hatter asks why a raven is like a writing desk, he promptly admits that he hasn’t the slightest idea. Later commentators asked if it was because Poe wrote on both. Or because they both stand on sticks, perhaps? They both come with inky quills? The real answer is that the author, Lewis Carroll, never intended it to have a solution, that the question simply draw attention to its own absurdity, but he did include the following preface to later editions: “Because it can produce a few notes, tho they are very flat; and it is nevar put with the wrong end in front!” In most printings, however, “nevar” is erroneously corrected to “never”, and the pun is nevermore.


Aden Rolfe is a writer and editor with a slight obsession with crows. He’s had radioplays broadcast on ABC Radio National and poetry published in Best Australian Poems 2011. He was recently featured in Overland’s Emerging Poets Series.


Sab Meynert is an illustrator living and working out of Toronto, Canada. Her work stretches between illustration, fine art, and writing. When not exhibiting at zine fairs with her micropress, Blacktooth, Sab is surrounded by comics, cats, and cheese. She is lactose intolerant. 


This piece first appeared in Issue 15 of The Lifted Brow. Grab a copy!