There has been much discussion in the blogosphere about a recent study entitled, “New Caledonian crows reason about hidden causal agents,” in
the Proceedings of the National Academy of Sciences (doi: 10.1073/pnas.1208724109, PNAS September 17, 2012) by Alex Taylor, Rachael Miller, and Russell Gray, demonstrating that crows have a tendency to attribute the movements of an inanimate object (e.g. a stick) to a causal agent whom they know to be in the vicinity, even when that agent is hidden from view, and that crows react with fear when they witness the movements of an inanimate object in the absence of any nearby causal agent. The authors of the study conclude that crows are capable of reasoning about a hidden causal agent. In their words:
Here, we show that tool-making New Caledonian crows react differently to an observable event when it is caused by a hidden causal agent. Eight crows watched two series of events in which a stick moved. In the first set of events, the crows observed a human enter a hide, a stick move, and the human then leave the hide. In the second, the stick moved without a human entering or exiting the hide… [A hide is a camouflaged shelter used to get a close view of wildlife. – VJT]
The movement of the probing stick was a novel stimulus and, thus, likely to elicit neophobic responses from the crows. [Neophobia is the fear of new things or experiences. – VJT] The movement of the stick was also likely to be an aversive stimulus for the crows as it moved into the space where the crows would put their heads when they attempted to extract the food from the box…
If the crows could attribute the stick’s movement to the hidden human, they could infer that when the human left the room, the stick would not move again. In contrast, in the second, unknown condition, if the crows were capable of causal reasoning, they would predict that the stick might move again because they had not observed a potential causal agent leave the hide…
The crows inspected the hide and abandoned probing with a tool for food more often after the second, unexplained series of events. This difference shows that the crows can reason about a hidden causal agent…
[Humans] make such inferences from a very early age. Between 7-10 mo[nths] of age, infants begin to show surprise if a bean bag is thrown from behind a screen and the screen is then lifted to show an inert object, rather than a causal agent such as a hand. The use of such causal reasoning underpins not only scientific and religious thought but also our sophisticated tool-using abilities and understanding of social interactions…
Comparative studies with the methodology outlined here could aid in elucidating the selective pressures that led to the evolution of this cognitive ability.
Readers wanting to know more about the experimental set-up can watch this three-minute video here. Professor Jerry Coyne has an interesting article about the experiment here.
From a scientific standpoint, the latest study by Taylor et al. was well-designed and rigorous, as it tested the hypothesis that crows are capable of reasoning about hidden causal agents against an alternative hypothesis, that they learn to make predictions about novel stimuli simply by becoming habituated to them. As Taylor recently explained, in response to a question from a reader on Reddit:
…[T]he habituation hypothesis predicts a high level of neophobia (measured as abandoned probes and high numbers of inspections) when the crows first see the stick move (i.e. in the HCA [hidden causal agent] trials). We got the opposite pattern, the crows were calm in the HCA trials, but then became nervous in the UCA [unknown causal agent] trials. As you note there was still habituation within UCA trials, which is to be expected; the crows were observing that this initially scary stimulus of a stick moving on its own was not leading to any negative consequences for them. But they key aspect of our study is that the crows only became scared when there was no human around to attribute the movement of the stick to… (Emphasis and square brackets mine – VJT.)
The reaction to the study across the blogosphere to has been surprisingly uncritical. Bloggers have hailed it as proof that crows are capable of reasoning about hidden causal agents. In this post, I’d like to explain why I think the study’s findings should be treated with a healthy dose of skepticism.
Are the crows reasoning?
First, as the authors of the study note, the same ability that the crows possess is also found in seven-month-old human infants. Most people would say that children of that age are not yet capable of reasoning, as they haven’t acquired a language.
Some readers may be inclined to object: “Maybe the crows are capable of some kind of non-verbal reasoning. It makes sense to suppose that reasoning can occur even in the absence of language. After all, don’t some people think in pictures rather than in words? Why couldn’t crows do the same thing?”
I don’t buy that. Some people do think in terms of pictures, but there’s a difference between such “thinking” and abstract reasoning. The difference matters.
Eleven years ago, while I was training to be a mathematics teacher, I overheard a teacher explaining to a colleague of hers why she insisted that her students should show their workings when solving a mathematical problem. She remarked: “If they really understand how to solve the problem, then they should be able to explain why they solved the problem in that particular way. If they can’t, then they don’t really understand.” The teacher’s remark struck me as an insightful one. It encapsulates my reasons for being skeptical regarding claims that the tool-making abilities of crows demonstrate a capacity for reasoning on their part.
The crucial point here is that the crows are unable to explain the basis of their judgments, as a rational agent should be able to do. The tool-making feats of Betty the crow look impressive, but we cannot ask her: “Why did you make it that way?” as she is incapable of justifying her actions. The same goes for the extremely clever New Caledonian crows who are able to use three tools in succession to get some food (BBC news report, 20 April 2010, by science reporter Rebecca Morelle). Let us imagine an older crow teaching a younger crow how to use a tool. And now try to imagine the following dialogue:
Older crow: Don’t bend it that way. Bend it this way.
Younger crow: Why?
Older crow: Because if you bend it this way, it can pick up a piece of meat, but if you bend it that way, it can’t.
The dialogue contains only simple little words, but the problem should be immediately apparent. The meaning of words like “if,” “why,” “but,” “can” and “can’t,” cannot be conveyed to someone who does not understand them, through bodily gestures alone. Until we have grounds for saying that crows possess a language containing words at this level of abstraction, we should react skeptically to claims that they can reason.
A second reason for skepticism is that although New Caledonian crows take care of their young for a period of two years (which is very long for a bird), the tool-making abilities of crows are not acquired through teaching from their parents. As Alex Taylor acknowledged in a response to a question from a reader on Reddit:
What we haven’t seen in crows is any kind of teaching, or the explicit copying of parents by juveniles (imitation).
Think about that. These crows supposedly learn how to reason without explicit instruction of any sort, and without even learning through imitation? I have t say I find that philosophically absurd. Reasoning is pre-eminently a social activity, because it is by its very nature open to challenge and criticism. Even solitary thinkers are expected to justify their claims in the court of public opinion, and if they cannot do so, they are rightly ignored. Reasoning that cannot be challenged, such as the kind that crows allegedly engage in, isn’t really reasoning at all.
Are the crows reasoning about causes?
My third reason for pouring cold water on the claim that crows are capable of reasoning about hidden causal agents is that in order to reason about causal agents in the first place, you need to be able to understand the notion of a cause, which is quite a sophisticated concept. Even eminent philosophers have a hard time explaining it.
The Scottish empiricist philosopher David Hume (1711-1776) famously defined a cause in terms of constant conjunction: a cause is “an object, followed by another, and where all objects similar to the first are followed by objects similar to the second” and also as “an object followed by another, and whose appearance always conveys the thought to that other” in his Enquiry concerning Human Understanding (1748, 1777, pages 76-77). Many contemporary scientists and philosophers are happy to parrot Hume’s definition, paying no heed to its glaring inadequacies.
To begin with, there are numerous cases where we can confidently assert that X causes Y, even though we may have had only one experience of Y being followed by X. A single example will suffice to illustrate my point. How many times would you need to see an arrow fired at an animal from a hunter’s bow, in order to conclude that the firing of the arrow was the cause of the animal’s death? Examples like this show that Hume’s requirement for constant conjunction of cause and effect is too strong.
Hume is also wrong to insist that a cause must always be followed by its effect. But in everyday life, cause and effect are often simultaneous. One billiard ball collides with another ball, and we say that the collision causes the second ball to move, even though the two events are simultaneous. A stone hurled by a schoolboy breaks a window, and we say unhestitatingly that the stone’s impact caused the window to break. A fire heats a horseshoe in the blacksmith’s forge, from the very first moment that it comes into contact with the horseshoe. On a more philosophical level, human beings seem to have no trouble in believing that they are being continually maintained in existence by God, without the need to posit any temporal interval between God’s conservative action and their continuation in existence. It is therefore a myth to say that causes necessarily precede their effects.
Finally, I take it that most of my readers will be familiar with the post hoc, ergo propter hoc fallacy. The fact that Y happens after X does not imply that X caused Y. Correlation does not imply causation. Even if we observe that Y always follows X, we cannot be sure that X is the cause of Y. There may be another agent, Z, which is responsible for both X and Y. We are forced to conclude that Hume’s definition of a cause in terms of constant conjunction is an inadequate one. Whatever the cause of an effect is, it must be something more than the event which it invariably follows.
The philosophical literature on the concept of causation is vast, and I have absolutely no intention of providing my own definition of a cause, in this short post. The point I want to make here is a very simple one. If even intelligent human beings, who are endowed with a language which allows them to refer to abstract concepts, have a hard time figuring out what a cause is, then shouldn’t we be just a teeny bit skeptical of the claim that crows, whose warblings lack the vital properties of productivity, recursivity, and displacement which characterize language as such, have a concept of causation?
Are the crows reasoning about causal agents?
A fourth reason for doubting the inflated claims made by Taylor et al. in their recent study is that even if we were to generously grant that crows can somehow grasp the notion of a cause, it is quite another thing to claim that they possess the notion of a causal agent – that is, a being who deliberately performs voluntary actions, such as pushing a stick from behind a curtain. In order to possess the concept of a causal agent, crows would need to possess what psychologists call a theory of mind – that is, an ability to attribute mental states such as beliefs, intentions and desires to other individuals. It is doubtful whether even human beings acquire this ability until they are three or four years old.
What’s more, there is good experimental evidence suggesting that even clever animals like chimpanzees (see this video) and elephants (see this one) lack a theory of mind. A chimpanzee, for instance, is incapable of realizing that a man with a bucket over his head cannot see anything, while an elephant can be easily fooled by a scarecrow. Indeed, primate researchers Derek Penn and Daniel Povinelli have written a paper entitled, On the lack of evidence that non-human animals possess anything remotely resembling a ‘theory of mind’ (Philosophical Transactions of the Royal Society B, 362, 731-744, doi:10.1098/rstb.2006.2023) in which they not only discuss the abilities of chimpanzees but also those of corvids (crows and related birds), and carefully explain why there is no reason to suppose that these animals have the capacity to impute mental states to others. At first sight, the evidence for a theory of mind in these birds looks convincing:
Corvids are quite adept at pilfering the food caches of other birds and will adjust their own caching strategies in response to the potential risk of pilfering by others. Indeed, not only do they remember which food caches were observed by competitors, but also they appear to remember the specific individuals who were present when specific caches were made and modify their re-caching behaviour accordingly (Dally et al. 2006).
However, the experiments performed to date suffer from a crucial flaw, as Penn and Povinelli point out: “Unfortunately, none of the reported experiments with corvids require the subjects to infer or encode any information that is unique to the cognitive perspective of the competitor.” The authors argue that simple rules can explain the birds’ behavior:
In all of the experiments with corvids cited above, it suffices for the birds to associate specific competitors with specific cache sites and to reason in terms of the information they have observed from their own cognitive perspective: e.g. ‘Re-cache food if a competitor has oriented towards it in the past’, ‘Attempt to pilfer food if the competitor who cached it is not present’, ‘Try to re-cache food in a site different from the one where it was cached when the competitor was present’, etc. The additional claim that the birds adopt these strategies because they understand that ‘The competitor knows where the food is located’ does no additional explanatory or cognitive work. (Emphasis mine – VJT.)
Penn and Povinelli also propose two carefully controlled experiments which could provide evidence of a “theory of mind” in non-human animals. Even adult chimpanzees who were used to interacting with human beings failed the first experiment proposed by the authors, while 18-month-old human infants passed the same test.
I realize that some readers will think that crows might be smarter than chimps, given their impressive tool-making feats. But I would ask them to ponder why these clever crows, despite their advanced facial recognition skills, are nevertheless capable of being fooled by simple decoys, such as scarecrows that spin around in the breeze and that make a noise? And why do crows consistently mistake S-shaped pieces of rubber hose for snakes, so long as the pieces of hose are moved to a new location at the end of every day?
Given the preponderance of negative experimental evidence for even the cleverest non-human animals tested to date, coupled with the anecedotal evidence that crows are not that smart at telling humans from straw men, it would be advisable to take the claim that crows possess the relatively sophisticated concept of a causal agent with a very large grain of salt.
Are the crows reasoning about hidden causal agents?
A fifth and final reason for being leery of claims that crows can reason about hidden causal agents is the absence of rigorous testing of the claim that hiddenness played any role in the “reasoning” of the crows in the experiment reported by Taylor and his colleagues. In the experiment, the crows “observed a human enter a hide, a stick move, and the human then leave the hide.” They then inferred that the stick would not move again. I cannot help wondering what would have happened if the crows had merely observed the human leaving the hide, without observing the human enter? Would they still have been nervous about probing for food, after seeing the human leave? In other words, were the crows reasoning about a hidden agent that was capable of pushing the stick, or were they simply reasoning about the number of agents in the vicinity who were capable of pushing it? If they were reasoning about the number of agents, then merely seeing a human leave the hide without seeing him/her enter in the first place would not give the crows any assurance that there were no other human beings in the vicinity.
Relevance to religion? Zero.
Before I finish this post, I’d just like to make a brief remark on the claim made by Professor Jerry Coyne in his post, that the ability of crows to reason about hidden causal agents may help explain the origins of religion:
This notion of “hidden causal agency,” of course, has been suggested as a pivotal factor in the origin of religion. If you’ve read Pascal Boyer’s provocative book Religion Explained, you’ll remember his thesis that before humans understood natural phenomena (e.g., thunder, lightning, or tree rustling), it was natural for them to impute them to causal agents – supernatural ones.
In the experiment reported by Taylor et al., the hidden agent that the crows allegedly made inferences about was previously seen by the crows: they saw the human enter the hide. By contrast, the agents which religious adherents pray to are never visible. Philosophically, there is an ocean of difference between a hidden (but nevertheless material) causal agent, and an invisible, incorporeal causal agent. Belief in the latter cannot be reduced to belief in the former. I can only conclude that the relevance of Taylor’s experiment to the claims of religion is absolutely zero.
I should add that Coyne’s thesis on the origin of religion is self-refuting. At the beginning of his post, he writes that “[i]t would obviously be adaptive for some animals to be able to distinguish between natural phenomena, like wind, and phenomena that have similar effects but are caused by hidden agents like predators.” But if natural phenomena are conceived of as the effects of some supernatural hidden agent, then the distinction between the two cases collapses: in both, we have an agent causing the phenomena.
May I suggest that Coyne is a much better biologist than he is a sociologist?