Access
This article is part of Nature's premium content.
Published online 31 March 2008 | Nature | doi:10.1038/news.2008.713
News
Babies have an eye for statistics
Infants intuitively understand probability.
Babies have many talents — such as the ability to charm a roomful of adults. But statistical reasoning? It's not the first skill that springs to mind regarding a gurgling 8-month-old.
To read this story in full you will need to login or make a payment (see right).
Comments
Reader comments are usually moderated after posting. If you find something offensive or inappropriate, you can speed this process by clicking 'Report this comment' (or, if that doesn't work for you, email webadmin@nature.com). For more controversial topics, we reserve the right to moderate before comments are published.
i don't think that it is surprising.Because they are just some things which human possess . One wouldn't lose these possessions just because of the change of age.
Alternativey, the reason humans 'start so well' and 'end so poorly' can be attributed to the fact that we probably become socially more aware (take fewer risks, maybe even without knowing) as we grow up, which most certainly helps with reproduction I would imagine. In some sense, there could be biases which develop in the brain over time which make us "not want" to think in the rational (with bounds) way we are capable of, just for the sake of prolonged interaction or long term reward. This not wanting to think could also be a tool which came about so as to make us preserve energy for 'doing' as opposed to 'thinking' stuff in social scenarios. This can also be related to the fact that in an ultimatum game, chimps were supposedly found to be more rational than humans.
I don't think the seamingly intelligent behavior comes from reasoning. The neural structure of the brain is geared to expect these reactions. Natural selection has whipped of the world those animals that could not foresee that if two predators hide in a place and one leaves it, there still is another one ready to eat you. This intuitive math can appear in many ways and neural structures can mimic them.
It comes from the fact that the neural circuitry is there and the brain becomes more personal as we grow (as shown by the changes in white matter with age) and the reason it shouldn't be as personal initially could just be because while we are moving about (up or down) the ladder of rationality, we are still incapable of explaining many things, which results in the build up of inductive biases, so as not to expend all the energy in trying to explain (or else, either we as a species won't exist or there wouldn't be a need to do science or even reproduce or think because everything would be known and mundane). Having said that, knowing everything does not necessarily finish everything, because rationality survives (it seems).
correction: knowledge (which I called rationality) survives
The authors of this study say "The academic discipline of statistics formalizes these intuitive statistical inferences." Strictly this is incorrect. It is probability theory and inductive logic that attempts to formalize these intuitions and their operation. The operative term here being "attempt." There are two types of probability, "degree of confirmation" and "the likelyhood that one thing will happen over another" (Carnap). In both cases, as probability theory and its statistical predictions work, they require historical record. If it can be shown that inductive predictions of the kind explored in this study are present without prior record, that they are in effect skills genetically inherited, then this will suggest that these inductive mechanisms are nothing like the statistical inferences we are familiar with. And that, my friends, is a big deal.
Xu and Garcia (in their abstract) seem to over-interpret the results. In a footnote to the abstract, they suggest babies are implicitly making probability calculations, but clearly their experiment demonstrates no such thing. A generous interpretation is it provides evidence that babies seem to have a sense of when one outcome (drawing 4 white balls in a sample of 5) is unlikely to have arisen from certain conditions (such as an urn containing 70 red and 5 white balls). I believe current neurocognitive models adequately predict this, but neither the models nor this research imply the babies are doing probability calculations. I also believe that adults do not lose the skills tested in this experiment, as Luca Bonatti suggests at the end of Ms. Ledford's article. That would be an interesting hypothesis to test with a batch of students about to take their first statistics class!
There is no "neurocognitive model" of which I am aware that makes predictions like this as Huber suggests. Exactly what model was Huber thinking of? (Although, I clearly agree that there are no "probability calculations" as we understand them going on.)