By providing a way to symbolize and communicate our thoughts, does language enable us to reason? Or are inference, deduction, and other forms of logical reasoning independent of our ability to put words to them? It’s hard to figure out whether babies can think, given that they can’t tell us, which makes separating language from reasoning even harder.

Ernő Téglás, at the Babylab in Budapest, researches “how infants acquire the conceptual sophistication necessary for abstract combinatorial thought involved in everyday reasoning.” His team has just published a paper describing the precursors of logical reasoning in pre-verbal infants. One group of infants was aged 12 months and the other was 19 months old; babies at these ages are just at the cusp of language learning and speech development, but they definitely precede the development of extensive language.

Wrong expectations

Like 20-something adults given the same tests, these babies expressed distress when their deductions did not hold true. Distress came in the form of staring at the inconsistent outcomes, which is how baby cognition is often measured.

The babies were shown movies to test whether they could frame what are called “disjunctive syllogistic hypotheses.” Whether you recognize the term or not, we do this all the time: if not A, therefore B. They saw two different toys that looked the same from the top—a dinosaur and a flower. Both toys were hidden in a box, and the babies could just see the very top of the toy over the edge. Then one of them was placed in a cup. But which toy is it?

At this point, the dinosaur was taken out of the box, precipitating the nail-biting moment of suspense that had the researchers on the edge of their seats: the Potential Deduction Phase. Would the babies grasp that the toy in the cup must then be the flower?

They did. The researchers know this because in some cases, they arranged for the toy in the cup to also be a dinosaur, a result inconsistent with the babies’ logical deduction. In these cases, recordings from the experiment show the babies stared for longer at this inconsistent outcome. Since the infants can’t yet verbalize “What the… ”, this is the best we’ve got.

Confusion

But this experiment can only measure the response after an inference has been made. To try to watch the infants make this inference in real time, the researchers altered the experiment by showing the babies which toy was in the cup. So they knew that the other toy must still be in the box; they didn’t need to figure it out. In this experiment, the infants still stared at the inconsistent result for longer.

But in the first experiment, when an inference was required, their pupils dilated more during the all-important Potential Deduction Phase. The researchers claim that this increased pupil dilation “suggests increased cognitive ability, possibly due to inference-making.” Why do the researchers think this? They used millennials as a control.

Millennials can, usually, tell us what they are thinking—that’s how we know they have reasoning skills. But when they are presented with the same movies of plastic dinosaurs and flowers in cups and boxes, their pupils likewise dilated only when they had to infer which toy was where. (They also stared longer at inconsistent outcomes.)

Téglás and co. conclude that precursors of logical reasoning are thus independent of, and even preliminary to, language. This work doesn’t come close to explaining the relationship between the parallel developments of reasoning and language. But because these visual responses remain stable throughout development until adulthood, this type of deductive reasoning may be a primitive and essential facet of the human mind.

Science, 2018. DOI: 10.1126/science.aao3539  (About DOIs).

https://arstechnica.com/?p=1274929