The past, present, and future of the written word
Friday, May 9, 2008, 3:30 – 5:30 pm Royal Palm 5
Organizers: Frederic Gosselin (Universit� de Montr�al) and Bosco S. Tjan (University of Southern California)
Presenters: Susana T.L. Chung (University of Houston), Dennis M. Levi (University of California, Berkeley), Denis G. Pelli (New York University), Gordon E. Legge (University of Minnesota), Mark A. Changizi (Rensselaer Polytechnic Institute), Marlene Behrmann (Carnegie Mellon University)
Gutenberg�s invention has democratized the written word: It is estimated that an average English reader will be exposed to over 100 million printed words before the age of 25. The scientific investigation of reading pioneered by Cattell in the 19th century was largely focused on single word recognition through the study of its cognitive, linguistic, and other high-level determinants (e.g., lexical frequency). Accordingly, in most of the influential theories of reading, the front-end visual processing remains unspecified, except with the assumption that it provides the abstract letter identities. This approach to reading greatly underestimates the complexity and the critical role of vision. Text legibility is strongly determined by the ease with which letters can be identified (Pelli et al., 2003), but it appears that standard fonts (e.g., Arial, Times) may be suboptimal as visual stimuli. For instance, the discriminability of a letter from the remainder of the alphabet, as indexed by identification accuracy with brief presentations, is inversely correlated with letter frequency, such that the letters most frequently encountered in texts are among the least discriminable. There is also a significant mismatch between the diagnostic spatial frequency spectra of letters and the human contrast sensitivity function, such that a large proportion of stimulus information is of poor use for the visual system (Chung et al., 2002; Majaj et al., 2002; Poder, 2003; Solomon & Pelli, 1994). Is there room for improvement? Previous attempts to improve reading speed in individuals with low-vision by bandpassing word images in the mid to high spatial frequency range led to equivocal results (Fine & Peli, 1995). However, we have recently witnessed significant advances in our understanding of foveal and peripheral vision and the mechanisms for letter identification and reading. Can this novel knowledge be applied to the development of fonts optimized for normal and impaired visual systems (e.g., developmental, letter-by-letter, or deep dyslexia, macular degeneration, cataract, diabetic retinopathy)? This is the challenge that the organizers of this symposium are submitting to the participants. We hope that this will be the first step toward vision science leading the way to a second Gutenberg-like revolution: Instant speed reading for all!
Enhancing letter recognition and word reading performance
Susana T.L. Chung
This talk will provide an overview of our efforts in enhancing letter recognition and word reading performance in the normal periphery and in patients with central vision loss.
Letter recognition, crowding and reading in amblyopia
Dennis M. Levi, Denis G. Pelli and Shuang Song
Crowding, not letter recognition acuity, limits reading in the amblyopic visual system.
Denis G. Pelli and Hannes F. Famira
“Legibility” means different things to visual scientists and type designers, and type design affects the different kinds of legibility in different ways.
The eyes have it: Sensory factors limit reading speed
Gordon E. Legge
Sensory constraints influence reading speed for normally sighted young adults, children, senior citizens, people with low vision and blind Braille readers.
The structures of letters and symbols throughout human history are selected to match those found in objects in natural scenes
Mark A. Changizi
New research supports the hypothesis that human visual signs look like nature, because that is what we have evolved over millions of years to be good at seeing.
Cognitive and neural mechanisms of face and word processing: Common principles
Marlene Behrmann and David Plaut
Through joint empirical studies (with normal and brain-damaged individuals) and computational investigations, we will argue that face and word recognition are mediated by a highly distributed and interactive cortical network whose organization is strongly shaped and modified by experience rather than by discrete modules, each dedicated to specific, narrowly-defined function.