Rayner, K., Pollatsek, A., and Starr, M. S. (2003). Reading. In Healy, A. F., and Proctor, R. W. (Eds.), Handbook of psychology (Vol. 4, pp. 549-574). New York: Wiley.
This chapter from the ‘Experimental Psychology’ volume of this collection gives a clear overview of the main areas of investigation related to reading within the field of psychology. It focuses on five central types of experimentation:
- Word identification — where words are presented to participants and a response is illicited; response time is observe
- Sound coding in word identification
- Eye movements in reading (in which we observe first-fixation duration, gaze duration and the probability of skipping a word)
- Word identification in context
- Models of eye movement control
No definitive model has been accepted thus far for object recognition, let alone for word recognition. Models of pattern recognition generally fall into one of two classes. The first are template models, whereby incoming visual input is compared to existing memory-based object templates that enable object identification. There could potentially be a number of templates per category, depending on viewpoint or viewing condition. The second class involves feature models, where an object is identified based on its combination of visual features. Although we do not understand the process completely, we do know that words are identified very rapidly (in 150-200 ms) through an automatic process.
The extent to which sound coding is key in the reading process is controversial, but has been identified as highly significant, especially in research on dyslexia. According to this model, the written language would mainly be a means to accessing the spoken form of the language, and knowledge of the spoken language would facilitate the learning of the written form. Sound coding has been linked to how we access a visual word’s meaning(s), and in assisting short-term memory while reading.
Eye movements during reading have been studied at length. Through a series of alternating saccades (discrete jumps) and fixations (periods where the eyes remain fixed), our eyes focus on a small area of text (called the fovea) while also drawing limited information from the word immediately to the right of the fovea (which will be identified fully on the next fixation). Our eyes “typically move forward approximately eight characters at a time” (p.558) but not all eye movements are forward — “10-15% of eye movements shift backwards in text and are termed regressions” (p.558).These eye-tracking studies have also identified that short words are skipped completely most of the time, while longer words that are highly predictable (i.e. if they recur in the same passage) are sometimes skipped. Words seen more regularly have shorter fixations, while unusual words are fixated on for longer periods of time.
A key component in word and meaning identification is the context in which it occurs. Although the reading process is one that acts on an unconscious level, a word that is a predictable part of a sentence (dictated by context and meaning cues) will require a shorter fixation time that one that is less predictable. Syntactic ambiguity (as in the sentence While Mary was mending her sock fell off her lap) can also interfere with the automated reading process.
Also briefy mentioned is the increasing use of brain imaging (i.e. MRIs) to “study issues related to which parts of the brain are activcated when different types of words are processed” (p.550). I will be seeking more recent articles that focus on this aspect of reading research.
Although the research outlined in this chapter is helpful in understanding the cognitive processes applied in reading, it is also clear that many areas are not fully understood. For example, reading depends on much more than the mere recognition of words. The construction of meaning is in itself a huge area to be explored.