A fascinating study by researchers at Johns Hopkins University shows that the same principle also governs how we read and interpret language. The study finds evidence that the way we process information visually is influenced by our knowledge and experience.
The study, published in the Journal of Experimental Psychology, asked two groups of 25 university students to study the letters of the Arabic alphabet. The first group had no prior exposure to written Arabic, while the second group could read the language proficiently. The researchers flashed two letters on the screen and the study participants had to indicate whether the two Arabic letters were physically identical or not. The researchers measured how long it took participants to make the decision and how accurate their decisions were for 2,000 pairs of letters.
As the experience of bird lovers, car aficionados and fashionistas suggest, those who were literate in Arabic perceived different details in the letters than those who were not. The pairs of letters that looked similar to experts weren’t the same as those that looked most similar to novices.
But what is fascinating about the study is what influenced those decisions. The group that didn’t know Arabic often confused letters that were similar shapes. The letters that they had trouble differentiating both had curves that were open to the left, or dots underneath them. But the expert group was biased by other, non-visual information that they knew about the language, including the motor movements and strokes that people use when they write the letters, as well as how the letters sound.
Some of the letters that experts confused were what’s called “allographs” – different versions of the same letter that in Arabic are written slightly differently depending on their position in the word. For example, line A below shows different allographs of the same letter, "ba." These “knowledge-based predictors” accounted for about half of the total difference in the time it took speakers and non-speakers to discriminate between letters, and about 10 percent of the variance in accuracy, the study says.
Even when accounting for these knowledge-based factors, the experts performed differently from the novices in an interesting way. As the letters got more complex, with more lines, dots and loops, the non-expert group tended to do worse at distinguishing them from other letters. But the trend was opposite in experienced observers — they tended to do better, suggesting that complex letters might be easier to recognize to those familiar with them.
The chart below shows how non-Arabic speakers and Arabic speakers differed in terms of reaction time. The height at which the cluster branches indicates how similar participants thought the letters were — those that branch lower down are perceived to be more similar. If the letters are placed close together on the chart, it means that people took longer to differentiate them. But if participants could quickly and easily say that they were different, the letters would be much farther apart.
The chart is color-coded to show how the choices of non-Arabic speakers differed from Arabic speakers. For example, look at the red cluster on the left. The top chart shows that non-Arabic speakers typically confused these shapes. But the second chart shows that Arabic speakers more easily differentiated among them.
The study reveals an aspect of how we learn: that the sound and meaning of a letter automatically becomes part of our understanding of it, without us even realizing it. But it also has implications for more than language.
Robert Wiley, the lead author of the study, says that letters are just one example that is easy to study, and he points to past studies that have looked at how people differentiate between cars, birds, faces and imaginary little people called “Greebles.” “Any class of object like this, where you become an expert and you’re very exposed to them, the visual system is going to be performing the same kind of operations,” he says.
You might also like: