HPS203 Lecture Notes - Lecture 3: Bigram, Distributed Knowledge, Tachistoscope
HPS203 Week 3
Describe the effects of familiarity and recency on word and letter recognition and the
effect of well-formedness on word recognition
• Tachistoscope: a device specifically designed to present stimuli for precisely controlled
amounts of time
• Post-stimulus mask: often just a random jumble of letters
• Can people recognise these briefly visible stimuli? à Depends on
many factors, including how familiar a stimulus is
o More familiar - more likely to be recognised
• Recency of view- first exposure primes the participant for the second exposure
o Priming: a process through which one input or cue prepares a person for an
upcoming input or cue
• Word-superiority effect: the data pattern in which research participants are more accurate
and more efficient in recognising words (and word-like letter strings) than they are in
recognising individual letters
o Words are easier to perceive, as compared to isolated letters
o This effect is usually demonstrated with a ‘two alternative, forced choice’ procedure
• Effect of well-formedness on word recognition:
o Easier to recognise a letter if it’s in the right context
o Context such as “FIKE” or “LAFE” vs. context such as “SBNE” or “HZYE”
o Wordlike strings are recognised more than letter strings like “JPSRW”
o Pronounceability – easily pronounceable strings provide a context benefit, generally
easier to recognise compared to unpronounceable strings
o Probabilities- how often particular letter combinations occur which are more
probable in English spelling?
o Englishness- the more English-like the string, the more easy it is to recognise that
string, and the greater the context benefit the string will produce
Describe a simple feature net and how it accounts for word recognition
• Feature net: a system for recognising patterns that involves a network of detectors, with
detectors for features as the initial layer in the system
• This flow of information is bottom-up
• Each detector in the network has a particular activation level: a measure of the current status
for a node or detector
o Increased if the node or detector receives the appropriate input from its associated
nodes/detectors
o Activation level will be high if input has been received frequently or recently
Explain how feature nets can account for the effects of recency, frequency, and well-
formedness on word recognition, and the word superiority effect
• The activation level will eventually reach the detector’s response threshold: the quantity of
information, or quantity of activation, needed in order to trigger a response
• Activation level is dependent on principles of recency and frequency
• Another layer of the net is detectors for letter combinations
• Another layer is of bigram detectors- detectors of letter pairs
• Detectors for groups of familiar letter combinations have high
find more resources at oneclass.com
find more resources at oneclass.com
