The Loop That Domestication Built

In April 2015, Takefumi Nagasawa and colleagues published a study in Science that is the strongest evidence in this course for attachment running between species on shared biological machinery.

The design is elegant. Dogs and their owners were left together and the interaction recorded; urinary oxytocin was measured in both before and after. Pairs that exchanged long mutual gazes showed increases in oxytocin, in the dog and in the human. Pairs with shorter gazing did not.

Then the causal half. Oxytocin was administered to dogs intranasally, and the dogs gazed at their owners more. The owners' oxytocin then rose in response. That closes a loop: looking raises the hormone, the hormone increases looking, and the effect crosses between the two animals.

This same loop is known to operate between a human mother and her infant, where it is part of how the bond is established and maintained.

The control is what makes the study matter. Hand-raised wolves, socialised to humans, rarely hold human eye contact and did not produce the effect. Since wolves and dogs share an ancestor, the absence of the loop in wolves suggests it emerged during domestication rather than being inherited, which is a claim about the last few tens of thousands of years rather than about canids in general.

So what has been established, stated precisely?

That a physiological mechanism which evolved to support attachment between human caregivers and infants also operates between humans and dogs, bidirectionally, and that it appears to be a product of domestication.

What has not been established is that your dog loves you in the sense you mean when you use the word. A hormone assay cannot reach what an animal experiences, and the researchers do not claim it does.

There is a deflationary reading available, and it should be taken seriously rather than waved away. On this account, dogs evolved to exploit a response humans already had. The loop is a parasitic adaptation, a hijack of existing machinery, and calling it love is sentimentality about a successful evolutionary strategy.

The trouble with that reading is that it works equally well one step over.

Infant faces exploit a caregiver response. Infant crying is shaped by selection to be difficult for an adult to ignore. The entire apparatus of human parental bonding can be described in exactly the same deflationary vocabulary, and nobody concludes from this that parental love is fake or that a mother is being manipulated.

So consistency demands an account of why the same mechanism, producing the same hormonal signature, counts as love in one case and as exploitation in the other. The only available answer that does not simply assert human exceptionalism is that we mean something by love that is not captured by the mechanism at all.

Which is a question about the word rather than about the biology, and it is where nearly every argument in this course eventually arrives.