Flip - Reading processes when reading mirrored scriptDoctoral thesis
Date of Examination:2022-06-29
Date of issue:2023-03-03
Referee:Prof. Dr. Sascha Schroeder
Referee:Prof. Dr. Uwe Mattler
Files in this item
EnglishThere is a common consensus that orthographic processing is deeply rooted in the structure of the visual system. One central aspect that is particularly important in visual processing is mirror-generalization, which is the tendency to treat visual objects and their mirror-image reversal equivalently. Most lower-case letters of the modern Latin alphabet were designed to be read from left to right and from the top the bottom of a page. Some of these letters, in particular those which have a mirror-image counterpart (e.g., "b", "d", "p", and "q") receive different interpretations, although they have the same shape and differ mainly in their orientation. Young children between the age of 5 to 6 tend to produce spontaneous mirror-confusions of letters, and in particular of those letters which are reversible. This behaviour is thought to be rooted in general visual principles of mirror-generalization and it commonly disappears by the age of 8. Mirror-generalization is thought to occur mainly across the vertical (left-right) axis and it is thought to be unlearned, suppressed or inhibited with reading acquisition. However, recent research using a priming procedure - which taps into early, automatic stages of visual word recognition - has shown that even skilled adult readers unconsciously mirror letters in reading. In this context, it has been shown that words which comprise only non-reversible letters (e.g., e, r, c, s) produce priming effects (i.e. they boost word recognition) whereas words which comprise vertically reversible letters (e.g. b, d, p, q) do not produce priming effects, indicating the presence of some inhibitory effect of reversible letters on word recognition. This raises several questions: First, may the propensity of letters and words to be mirror-confused (their mirror-confusability) be a visual property which moderates the ease with which words can be recognized? Second, are involuntary confusions of letters in reading confined to vertical confusions (e.g. b vs d) or do they also occur across the horizontal (up-down) axis (b vs p)? Third, when mirror-confusions are purposely induced through mirroring the letters within text, can a word be recognized before all its constituent letters are identified? Or are words only recognized once all letters have been identified? Three studies were designed to tackle these questions. The first was a masked priming study placing special emphasis on the locus and nature of early, automatic mirror reversals of letters during word recognition in adults. This was done by examining whether both vertically and horizontally mirrored letters produce priming effects on word recognition in two different tasks: one tapping into lexical (lexical decision task) and one tapping into pre-lexical processes (same-different match task). Results show that mirror-priming effects occur across both mirror axes and that they generalize to non-words (which do not have a lexical representation). Furthermore, mirror-priming effects are reduced for targets which comprise confusable letters (e.g., d, b, p, q, f, t, u, n). This indicates that general visual principles of mirror- generalization operate both vertically and horizontally and that mirror-priming effects are pre-lexical by nature. To examine whether a word’s mirror-confusability moderates word reading times in adults, study 3 involved three experiments. First, the mirror-confusability of the Latin alphabet was quantified in a letter-based score (Exp 1). Second, the mirror-confusability of target words was quantified and manipulated based on the score and words were categorized as high and low confusability words in a lexical decision task (Exp 2). Third, the eye-movements of participants were recorded as they silently read these high and low confusability words when they were embedded in sentences (Exp 3). The results of study 3 imply that letters vary considerably in their mirror-confusability and that a word’s average mirror-confusability moderates word reading times in adults. Study 2 addressed the question of whether interference effects induced by mirrored letters are confined to early, visual-orthographic processing of letters or whether they also affect lexical stages of the reading process. Results show that mirroring letters disrupts also later, language related processes on the word level, before individual letters are identified, suggesting that processing is cascaded across levels. Taken together, the results from the three studies provide a comprehensive overview on the nature and time-course of mirror-confusions in functional reading adults. On the ground of these findings, I present how cascaded models of visual word recognition can account for these mirror-effects in adults across different stages and units of processing. In particular, my findings indicate that during an early, automatic and purely visual stage of processing, letter features are generalized to their vertical and horizontal mirror-image counterpart and that this produces priming effects visual word recognition. During a later stage of visual and orthographic processing, mirroring produces interference effects which permeate to the word level. These interference effects are more pronounced for words with a high mirror-confusability. The dissertation provides comprehensive empirical evidence and a theoretical frame-work that advances our understanding of if, how and when mirroring letters affects the reading process in adults.
Keywords: Eye movements; Reading; Visual word recognition; Masked priming; Mirrored letters