Credit: Jesse Bering

Last week I found myself in a web video dialogue on This is a sort of video blogging website sponsored by the New York Times where the two participants are meant to rally big ideas back and forth, or at least get a decent dialogue going about the topic at hand. I spoke with (soon-to-be) Yale philosopher Joshua Knobe on on the natural foundations of religion. Our discussion was the first of a John F. Templeton Foundation series called Percontations, a term defined as “questionings or inquiries, especially those requiring more than a yes or no answer.” In the present context, this inquiry presumably concerns the achingly dull question over whether or not God exists.

The recording format for Bloggingheads is quite odd, as the two discussants are conversing over the phone (usually wearing an embarrassing headphone) while recording their own video images using the QuickTime Pro software program. The editors then splice the two videos together so it creates the impression, at least, that the two individuals can actually see each other throughout the conversation. But in reality it’s more like staring into a mirror while chatting over the phone. Anyway, when it finally went up online and I sat down to watch the first few minutes, it was a stranger experience than I’d anticipated. For one, I think I can safely say that this was the first time I’d ever watched myself watching myself having a conversation. The fact that I continually averted my own gaze throughout the video is revealing; I often fail to make eye contact with other people—a habit a Belgian colleague of mine once told me was quite rude—but I didn’t realize I’d even treat myself so disrespectfully. I also didn’t realize that I had such a gargantuan forehead and that I used the words “essentially” and “basically” in nearly every sentence, so I’ll have to work on that.

The whole experience reminded me of a series of experiments conducted some years ago by University of Louisiana at Lafayette cognitive scientist Daniel Povinelli. Povinelli and his colleagues were interested in pinpointing the exact age at which children develop an autobiographical sense of self. Although children first begin to “recognize” themselves in mirrors between the ages of 18-24 months of age, Povinelli points out that this fact does not mean that children at this age possess a rich, subjective understanding of the self.

Given infants’ extensive familiarity with mirrors prior to 18 months, it is likely that they are already familiar with the constellation of features that comprise their facial appearance. Thus, the ability to correctly label the image may only reflect the fact that the infant knows that the particular constellation of features he or she sees in mirrors are referred to by the infant’s proper name, or the personal pronoun “me.”

Even once children do understand that the image represents the self as a subjective entity (allowing him or her to attribute first-order, private psychological experiences to the person reflected in the mirror), this doesn’t necessarily mean that they have an autobiographical sense of self. That is to say, a self that endures through time or, in the psychologist William James’s words, an understanding that “I am the same self I was yesterday.” Rather, it could be the case that young children’s minds—lucky them—are locked into a sort of here-and-now existence, one where antecedent events fail to clearly connect with the circumstances of the present. To get at this idea, Povinelli and his colleagues developed a rather clever series of experiments where children are videotaped playing a game in the laboratory and then shown the video footage a short time later. But, and here’s the clever part, during the course of the videotaping, one of the experimenters pats the child’s head in praise and in so doing surreptitiously places a large, brightly colored sticker on their head. (This was preceded by a sticker-less “sham pat” where children were simply accustomed to being patted on the head.) Although there were a number of hypotheses tested, the critical research question was whether upon seeing the previously recorded footage, children of different ages would be more or less inclined to reach up and touch their head. If they did so, this would indicate their general understanding that their past is causally bound to their present.

In the first study to use this general paradigm, published in the journal Child Development in 1996, Povinelli and his colleagues Keli Landau and Helen Perilloux reported that none of the two-year-olds and only 25 percent of three-year-olds reached up to touch their heads when the videotape was played back to them after a three-minute delay. In contrast, 75 percent of the study’s four-year-olds did so. These same age-related patterns were replicated in a similar study where, rather than watching themselves on delayed videotape playback, children were shown Polaroid images of the experimenter marking them with the sticker. Yet although they largely failed to connect the past with the present by reaching up to touch the sticker, nearly all of the three-year-olds in the study accurately identified themselves when asked who was shown in the image. Interestingly, however, the authors found that the three-year-olds were significantly more likely to refer to themselves in the third person (using their first names rather and saying that the sticker is on “his” or “her” head) than were the four-year-olds, who used first-person pronouns (“me” and “my head”) almost exclusively.

In a follow-up study published in a 1998 issue of Developmental Psychology, Povinelli and his research colleague Bridget Simon used nearly the identical procedure, but this time the study included five-year-olds and also included a seven-day comparison condition. That meant that, for half of the children, the time duration between the covert sticker-marking event and the video playback was a full week. In this study, 88 three- to five-year-old children were randomly assigned to either the brief delay (3 min) or the extended delay (7 days) condition. Similar to the results from the earlier study, less than half of the three-year-olds responded by reaching up to their heads regardless of the length of time that separated the two events. In contrast, nearly all of the four- and five-year-olds in the brief condition did so, and furthermore their same-age peers in the extreme delay condition did not. “That is,” the authors write regarding the delayed condition findings, “as age increased, the number of children who reached up tended to decrease…. Four- and especially five-year-olds displayed a clear understanding that although briefly delayed visual feedback is causally relevant to transient aspects of the present self, extremely delayed feedback is not.”

Povinelli has pointed out the relevancy of these findings to the phenomenon of “infantile amnesia,” which tidily sums up the curious case of most people being unable to recall events from their first three years of life. (I spent my first three years in New Jersey, but for all I know I could have spontaneously appeared as a four-year-old in my parent’s bedroom in Virginia, which is where I have my first memory.) Although the precise neurocognitive mechanisms underlying infantile amnesia are still not very well-understood, escaping such a state of the perpetual present would indeed seemingly require a sense of the temporally enduring, autobiographical self.

In this column presented by Scientific American Mind magazine, research psychologist Jesse Bering of Queen's University Belfast ponders some of the more obscure aspects of everyday human behavior. Ever wonder why yawning is contagious, why we point with our index fingers instead of our thumbs or whether being breastfed as an infant influences your sexual preferences as an adult? Get a closer look at the latest data as “Bering in Mind” tackles these and other quirky questions about human nature. Sign up for the RSS feed or friend Dr. Bering onFacebook and never miss an installment again.