Triangle vs dual route cascaded reading

Dual route cascaded model and triangle model.

Dual route cascaded model was developed to test the dual model theory using computer programming. For the programme to run the theory needs to be complete, and the computer can then run simulations of reading using the model. The closer the DRC programme matched the reading behaviour of a person, the more accurate the model was assumed to be. The DRC model splits reading into three routes: lexical semantic, lexical nonsemantic, and grapheme-phoneme.

The DRC is a model to explain the process of reading aloud for a skilled reader. Dual refers to either recognising a word and retrieving it from the mental dictionary (lexicon), or breaking the graphemes down into phonemes that can be said without comprehension. The lexical routes are faster for more frequently used words, but they cannot process unknown or pseudowords. For these words the grapheme-phoneme route is quicker. Correct pronunciation of irregular words that are not said the way they are written (eg. Worcester) is only possible when they are stored in the lexicon.

It is more difficult to use the grapheme-phoneme route in written languages like English which do not conform to a set of rules for how each grapheme is pronounced (eg “o” does not sound the same in “son” and “on”).

Irregular words take longer to process than regular words which suggests that both routes are activated while reading and then the person has to decide whether to say the lexical or grapheme-phoneme version because they conflict.

Reading is fastest for frequently used regular words where the faster lexical route can be used and the grapheme-phoneme route does not give a contradictory answer.

Lexical processing may require less attention because the word only needs to be retrieved, whereas in the grapheme-phoneme route each grapheme must be processed.

Surface dyslexia may be present in patients with dementia. Patients are able to read words that are regular (eg Manchester) and pseudowords, but not irregular words (eg Worcester). Patients may be able to pronounce irregular words that they encounter regularly. The dual route model is able to account for this by suggesting the patients have damage to the lexical route but are still able to read many words through the grapheme-phoneme route.

Phonological dyslexia patients are unable to read unfamiliar or pseudowords and only appear to be able to read words stored in memory. The dual route model explains phonological dyslexia as the lexical route being intact but the grapheme-phoneme route impaired.

Data from non-impaired patients shows that regular-frequent words are read fastest and irregular-infrequent words slowest, these findings are accounted for by the dual route model because the lexical route quickly processes familiar words and there is no conflict from the grapheme-phoneme route when the word is regular.

The triangle model says that three areas work in parallel to read text and produce speech: orthography, semantics, and phonology. Letter patterns are stored in orthography, sound patterns in phonology, and word meanings in semantics. Each of the three areas contains many processes and reading a word involves extensive activity back and forth between the three areas. Reading is a result of a combination of inputs from phonology, orthography and semantics.

Connections between the areas are strengthened by repeated use in frequent words.

All words are processed in the same way, whether they are regular or not.

Initially readers rely on orthography and phonology, but learn to utilise semantics over time.

The existing connections support reading of new words because the existing patterns are followed to make an educated guess.

A computational programme was also built to test the triangle model. It simulated lesions in the brain by deleting one of the three domains and then comparing the computer’s performance to a patient’s. The triangle model was supported by these computer studies. Damage to orthography impaired reading of irregular words and damage to phonology impaired reading of pseudowords.

The triangle model explains performance in healthy people as due to the strength of connections between the areas accounting for ease of reading infrequent and irregular words.

The triangle model accounts for the role of learning, wheras the dual route cascade model is limited by not explaining how learning changes reading.

Data in healthy people and patients with types of dyslexia can be explained by either model.

Neither model explains the neurological processes underlying reading.

The cortical network of reading has been studied using brain imaging and the different areas mentioned in both models appear to be processed in different parts of the brain.