Is this it?
West and Stanovich report fluent readers’ performance being facilitated by two automated processes; sentence context (essentially semantic priming) and word recognition. According to chapter 3, fluent readers use phonological recoding if automated word recognition fails.
elsiep wrote:That's the one.
It is funny how two people can read the same words and come to completely different conclusions.
What Stanovich & West actually says is: (p20 Progress in Understanding Reading)
...the present study has led to some important conclusions regarding the operation of automatic recognition processes and interword redundancy in reading. First of all it appears that sentence contexts can automatically facilitate reading performance. However, the performance of more fluent readers seems to be dominated by rapid, probably automatic (in the sense of LaBerge and Samuels 1974), word recognition. These recognition processes occur so fast that effects due to slower acting contextual factors (my emphasis) are reduced.
I cannot draw any conclusion from this other than fluent readers primarily use rapid, automatic word recognition.
Have I disagreed with that?
maizie wrote:I can see that in the congruous sentence prime condition it might be argued that the fluent readers prioritised context over automatised word recognition because it was the less cognitively demanding process but that makes a mockery of the term 'automatic' (as applied to the automatised word recognition process) as it is not automatic if it has to be chosen in preference to another strategy. In fact, surely the act of 'choosing' would slow the process rather than speed it? The same would apply in the incongruous sentence condition; a decision would have to be made as to whether to use the 'automatic' process or the context for identifying the word, once again slowing the process.
Readers do not choose
between two automated processes. Both automated processes occur simultaneously. Because they are automated, the readers can't stop them happening. The question is; which process is the faster in any given reader in any given situation?
In skilled readers, in the congruous sentence condition, context speeds up word recognition due to priming. In the incongruous sentence condition context doesn't slow it down because word recognition is faster than the process of switching to the other context.
The section ends with a strange surmise:
Adults use of context is probably greater when they read more difficult material.
May be this early on in their research career they didn't have a very clear picture of what 'automatised word recognition' might consist; of although earlier in the chapter (p14) they suggest that fluent reades might 'automate letter and word identification...'
I think you've missed what's meant by 'automatised word recognition'. It means word
recognition - recognition at the whole word level. On p.38, there is a clear reference to 'three specific abilities' - recognizing words automatically, the ability to recognize words and sub-word units, and the ability to recode print items into phonological form.
There is no question that skilled readers use a number of strategies when reading, including using context, whole word recognition, recognition of sub-word units and phonological recoding. What's important to the reader is which gets there first.
- If the context is congruent, it will facilitate word recognition due to priming.
- If the context is incongruent, but word recognition is faster than activating another neural pathway, word recognition won't be slowed down by context.
- If word recognition fails - ie the reader doesn't automatically recognise an unfamiliar word - then other cues, such as context, sub-word units and phonological decoding will be used.
maizie wrote: I would propose that a truly 'difficult' text for a skilled reader would be one in which there were words unfamiliar in both appearance and meaning. In which case context would be no help whatsoever.
But they're not talking about a 'truly difficult text for a skilled reader'. They are talking about adult readers in general and pointing out that relatively simple words were used in West & Stanovich's study, thus increasing the likelihood of automated word recognition. They were speculating on what would happen if more difficult material were used. This could include words that readers understood but didn't read very often, in which case word recognition wouldn't be automated and there might be increased reliance on context.