Use this URL to cite or link to this record in EThOS:
Title: Familiarity effects in visual word recognition
Author: Gontijo, Possidonia de Freitas Drumond
Awarding Body: University of Edinburgh
Current Institution: University of Edinburgh
Date of Award: 1998
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
This thesis is an investigation of two different aspects of familiarity processes involved in visual word recognition. The first is how capitalisation influences visual word recognition. The second is the role played by onset, nucleus and coda in nonword recognition. A familiar aspect of proper names in English, is that they are printed with an initial capital letter. Two experiments investigated the effects of the capitalisation of the initial letter of nonwords. It was found that subjects generate fewer pronunciations for initially capitalised nonwords than for those which were not capitalised. I suggest that in English initial capitalisation acts as a cue strong enough to prompt readers to perceive unfamiliar strings of letters as belonging to the category of proper names. As a result, the phonological domain used to retrieve the pronunciation of initially capitalised strings becomes more restricted than that used for the non-capitalised unfamiliar strings. These results extend the applicability of Brennen's theory for proper names, which is based on the size of the set of plausible phonologies of a word. In a third experiment, pairs of nonwords had their familiar visual appearance manipulated in terms of first and last letter capitalisation, in a same-different matching task. Faster response times were obtained for those nonword pairs that kept a more familiar aspect (e.g., pairs in which the first letter was capitalised as opposed to others in which the last letter was capitalised). These results are explained in terms of Besner and Jonhston (1989) "orthographic familiarity route". I propose the transformation model as an explanation for the mechanisms by which this route operates. Nonwords are an important aspect of this thesis. A new algorithm was developed for the creation of monosyllabic nonwords in which the frequency of their onsets, nuclei and codas could be controlled carefully. This gave us the opportunity to study the influence of orthographic neighborhood in visual word recognition. The findings here are in agreement with previous studies which show the recognition of an item to be influenced by the presence of neighbours. It has been hypothesized that familiarity effects in visual word recognition can only be found in tasks where identification mechanisms are not implicated. Here, a new category of words, namely brand names, was used to test this hypothesis. There are many reasons why brand names are a more appropriate class of words than acronyms to be used in this type of investigation. The results obtained confirm the hypothesis above. Previously, acronyms had been the only class of words used to test this hypothesis. Finally, a computational assessment of the nature of the mappings from letterto- sound in British English was carried on. A program was developed to estimate the pronunciation of any string of English graphemes based on the probabilities of grapheme-phoneme correspondences. The algorithm was assessed by examining its behavior for nonwords. This was done by using a corpus of nonword transcriptions, collected in an experiment with trained phoneticians. The results confirm the fact that the statistical information about grapheme-phoneme correspondences alone is not sufficient to predict English pronunciation. Also, a method was developed that allows the quantification of the different orthographic depth for various languages.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available