W1.2 and 2.2- Computational Models of Reading and Acquired Dyslexia Flashcards
(18 cards)
What is acquired dyslexia (also referred to as alexia)?
Partial or complete loss of the ability to read after a brain lesion as opposed to developmental dyslexia
What two alternative ways did Forster and Chambers (1973) state coding could be assigned for us to understand words from written meaning?
Pronunciation either computed by an application of a set of grapheme-phoneme rules (based on letters and transcribing of sounds), OR by a direct dictionary look up which would only word for familiar words (recognising a word to understand its meaning)
What is the difference between a grapheme and a phoneme?
Grapheme- a letter or group of letters that represents a single phoneme (sound)
Phoneme- smallest unit of sound in a language
What are the two routes of reading and how can they be defined?
Nonlexical route- involves making use of rules relating segments of orthography (how it is written) to segments of phonology (how it sounds)
Lexical route- involves looking up words in LTM to retrieve knowledge about its meaning and pronunciation
What model did Coltheart et al (1993) propose in relation to lexical and non lexical routes?
The dual-route cascaded model (DRC) which states how we go from print to speech through either the lexcial or non lexcial route
Describe the processes of the LEXICAL route
Accessing representation in the orthographic output lexicon (set of memories you have for how words are spelled) -> can then access what this meansthrough the phonological output lexicon, in turn evoking phonemes at the phoneme level -> allows activation a word’s meaning from either its orthographic form or its retrieved sound form -> allows direct retrieval of pronunciation of known words, irrespective of whether the soundform is predictable or not (doesn’t work for novel words that are not part of the lexicon)
Describe the process of the NON-LEXCIAL route
Relies on a set of grapheme-phoneme correspondence (GPC) rules to derive pronunciation in an indirect fashion
Does not rely on lexicons
Print -> visual feature units -> letter units -> grapheme phoneme rule system -> phoneme system -> speech
How do the lexical and non lexical routes interact?
They do not race with eachother except when reading under time pressure e.g. regularising an exception by not recognising the string as a known word (pint, mint) and may also mistake a nonword for a word (pseudowords)
What is pseudoword reading and what does it allow for?
Letter combinations that resemble real words in a language but have no actual meaning, it allows for the assessment of knowledge about print to sound correspondences
How do error rates in pseudoword reading vary across languages?
English is not as clear cut about how letters or groups of letters should be read so children take more time in learning to read and master these correspondences (non-transparent). Whereas, more ‘transparent’ languages e..g. Dutch, Spanish and Italian have individual letters read by the same sound so there is no need to store exceptions in the mental lexicon (with the risk of not remembering what these are) so print to sound can be mastered more quickly
What six key effects have been found in both the DRC model and human readers?
- Words read faster than non words e.g. plane reads faster than brope
- High frequency words read faster than low frequency words e.g. table faster than maple
- Regular words read faster and more accurately than irregular words, especially for low frequency words
- The larger the orthographic neighbourhood of a nonword, the faster it is read out loud
- Nonwords that sound like words read faster than nonwords that do not sound like words (brane/grune)
- The more letters in a nonword there are, the slower it reads out loud, but the number of letters has little to no effect on reading aloud real worlds
Why do words read faster than non words (effect 1)?
Because words benefit from having both the lexical and non lexical route supporting their reading aloud but non words can only be sounded out
Why are high frequency words read faster than low frequency words (effect 2)?
This is because of a practice effect- high freq words have been practiced more so recognition and transcoding is faster and more accurate
Why are regular words read faster and more accurately than irregular words?
Because irregular words are exceptions and create a clash between responses given by the lexical and non lexical system but for regular words, the nonlexical route produces a correct response
Why does a larger orthographic neighbourhood result in a nonword being read faster out loud?
Because the neighbourhood refers to a pool of existing words similar to the target string- if a nonword is similar to many words than these evoked in the input and output lexicon which helps select the correct sounds in the phoneme system
Why do nonwords that sound like words e.g. brane read faster than nonwords that do not sound like words e.g. grune?
Because as soon as pronunciation of the nonword e.g. brane has been computed, recognition using inner speech will evoke existing sounds alike (brain) and in return, the existing entry in the lexicon will contribute to this share and confirm the sound form initially achieved using the non lexical route
Why does more letters in a nonword result in it being read out loud slower, but this is not the case for words?
Because non words require both segmentation of their individual graphemes i.e. the print segments corresponding to individual sounds and concentration of the sounds in order to obtain a plausible sound form within the non lexical route
What process is key to reading unknown words?
Grapheme-phoneme conversion
