Cawsal Reasoning

A few years ago, the psychologists Saxe, Tzelnic, and Carey looked at how children as young as 7-months old would react when a bean bag is thrown from behind a screen, with the screen then being raised to show either a human hand or puppet, or an inert object like a toy train¹. What they found was that, even in these young children, there was evidence of causal reasoning in the form of the children showing signs of surprise after the screen was removed to reveal an inert object. That is, the children were utilising an abstract understanding of how causal agents (the hand or puppet) can affect its environment, which remains true even if the causal agent cannot be seen.

At this point you might be asking: "What's with the horrible pun in the title?". The answer is found when we compare the novel aspect of Saxe's work with the novel aspect of the recent work of Taylor, Miller and Gray. The former is interesting for extending evidence of causal reasoning to very young children, and the latter is interesting for extending evidence of causal reasoning to crows.

In their latest paper, "New Caledonian crows reason about hidden causal agents"², Taylor and colleagues set up an analogue situation to that used by Saxe that was obviously adapted for crows (or perhaps they just could not find any puppets and toys trains at short notice). Their design is best characterised in the figure below:

For the first condition (on the left), a crow would observe two people enter the aviary and one person would go behind the "hide" (a screen preventing the crow from seeing the person) whilst the other remains motionless within the room. This condition was termed the "Hidden Causal Agent" (HCA) condition as the hidden person would move a stick in the baited hole where the crow would forage for food, and for those capable of causal reasoning, they should be able to infer that the movement of the stick was being caused by the human in the hide. After moving the stick in and out fifteen times, the person would leave the hide and the aviary completely (all being seen by the crow). The second condition (on the right) was the "Unknown Causal Agent" (UCA) condition, where only one person would enter the aviary but they would remain motionless in plain view of the crow and the stick would move in and out fifteen times with no apparent cause (the experimenters were manipulating it with a hidden string).

The logic behind the experiment is that the crow should hesitate when attempting to retrieve their food from the baited tube, as movement of the stick could result in them being poked in the side of the head (importantly, as the authors take effort to note, this was in actuality impossible as they only moved the stick when setting up the condition, before the crow began to forage for food). In the HCA condition, if the crow accurately infers the association between the movement of the stick and the human, then there should be little-to-no hesitation when foraging for food because the agent causing the movement of the stick leaves before the crow enters the baited tube. In other words, there is no reason to fear being poked in the side of the head. With the UCA condition, however, the crow has no clues about what could be causing the movement of the stick so it has no information on whether the stick will move again or not (thus presenting a risk). The results are shown below:

The pretty lines and dots tell us that in the HCA condition, the crows spent significantly less time investigating the stick and surrounding before foraging for food. Since the only difference in the two conditions was the observation of a second experimenter entering and then leaving the hide, our explanation must utilise this variable. The authors argue (and successfully, in my opinion) that the best explanation for this behavior is one that is consistent with the Saxe research - that the crows are demonstrating causal reasoning as they have less reason to fear being poked in the head due to the fact that the agent believed to be causing the movement of the stick is no longer present.

It is pretty amazing research as even though some observational reports had alluded to these possible abilities in animals (for example, Darwin's example of dogs barking at a parasol being blown across a garden by the wind), this is the first example of it being demonstrated in experimental conditions. The authors rightly speculate that the methodology they used could be applied to other animals to assess the causal reasoning abilities of various species, and this could give us some information about possible selective pressures producing this ability.

For anyone interested, the lead author Alex Taylor did an "AMA" (Ask Me Anything) on Reddit yesterday and answered a number of questions put to him on the study. You can find it here: Caws and Effect – IAM Alex Taylor, Evolutionary Psychologist and lead researcher on the recent paper, "New Caledonian crows reason about hidden causal agents". AMA.

REFERENCES:

<sup>1. Saxe, R., Tzelnic, T., Carey, S. (2007) Knowing who dunnit: Infants identify the causalagent in an unseen causal interaction. Developmental Psychology, 43:149–158.↩

2. Taylor, A.H., Miller, R., Gray, R.D. (2012) New Caledonian crows reason about hidden causal agents. Proceedings of the Academy of Sciences, Published Online First: 17 September.↩</sup>