Cognitio 2011

Abstract: Visual patterns are a human universal - all human societies have a pattern culture, manifesting itself in weavings, jewellery, tattoos, tilings etc. No animal species is known to intentionally produce patterns. Gombrich (1984) suggests that the human mind naturally strives to impose order on visual input it receives and also to create an aesthetically pleasing structured output. Szilagy and Baird (1977) showed that humans spontaneously and without instruction produce highly symmetrical patterns in one, two and three dimensions. While Gestalt psychologists have attempted to derive laws of visual perception and rules for the 'goodness' of a form (e.g. Wertheimer 1923), they were never quantified or tested empirically. Hence, it is still unclear where the human drive to make patterns comes from, and why we like them so much.  One distinct possibility is that visual patterns are a by-product of the cognitive capacities that evolved to parse and produce language, another uniquely human universal.



In order to address the question to what extent visual patterns are unique to our species, we are taking a doubly comparative approach to study pattern perception. We are comparing different human groups and their performance at pattern perception and production tasks as well as adapting these tasks for animal species, in this case the pigeon (Columba livia). Thus, we are studying how patterns are processed in different kinds and stages of the human mind, and how they may be processed in the minds of animals. This research is funded by ERC grant SOMACCA (Syntax of the Mind: A Computational Comparative Approach).



The human groups we have studied include undergraduates, children, adults with Autism Spectrum Disorder (ASD) and a control group of age and IQ-matched adults. Participants were shown patterns consisting of a single motif repeated on a square grid (like the arrangement of bathroom tiles) with two production rules: translational or rotational symmetry. Some patterns contained a structural violation, in that one of the motifs had an additional 90° rotation. The participants had to indicate by button box whether or not the image shown contained a violation. No feedback was given.



Pigeons were trained to respond to images on a touch screen. In a Go/No-Go paradigm they had to peck on images of one category (e.g. flawed) and refrain from pecking on images of the other (e.g. unflawed).  The birds received auditory feedback and a food reward for correct responses.



Despite intensive training, pigeons consistently failed to detect a structural violation in both rotational and translational patterns. It has been shown that pigeons are extremely good at recognising complex figures under any rotation (Hollard and Delius 1982), and their failure to detect a structural violation may actually be due to their ability to abstract an object away from its orientation. The birds however did successfully discriminate between patterns that contained a colour violation and those that did not.



All human groups found the task easy and performed at a very high level. Results show that visual pattern processing capabilities in humans are already present at a young age. Furthermore, individuals with ASD perform at the same level as matched controls, and have faster reaction times for certain patterns than other adult groups, which is consistent with previous findings (O'Riordan 2004). Thus, like language, pattern perception develops reliably and robustly in humans with no explicit instruction.



References:

Gombrich, E. H. (1984). The Sense of Order. A Study in the Psychology of Decorative Art. London: Phaidon Press.

Hollard, V. & Delius, J. (1982): Science, 218, 804.

O'Riordan, M. (2004). Superior visual search in adults with autism. Autism, 8(3), 229-248. Szilagyi, P. & Baird, J. (1977): Perception and Psychophysics, 22(3), 287.

Wertheimer, M. (1923). Psychological Research, 4(1), 301.