PSYC3331 Lecture Notes - Lecture 14: Saccade, Eye Tracking, Parafovea
Is it the case that when a word is presented in context, we process it in a qualitatively different wat to
when it is presented in isolation?
Meyer and Schvandeldt (1971) were the first to show that a target word (e.g., NURSE) is recognised
more quickly when preceded by a semantically related prime (e.g., DOCTOR), than by an unrelated
prime (e.g., NATION). However, it has also been shown (by Neely, 1976) that a neutral condition (e.g.,
NURDE preceded by XXXXXX) falls in between the other two conditions. That is, related words show
facilitation
(relative to the neutral baseline) while unrelated words show
inhibition
.
"nurse" is a predictable response to the word "doctor" in a semantic association task. Does one also find
facilitation to the target when the prime is related but not predictable? Yes, e.g., INJECTION primes
NURSE relative to the neutral (XXXXXX) baseline.
Neely explains the pattern of priming results in terms of there being two processes: 1) Automatic
spreading activation leads to facilitation of related words. 2) A conscious strategy where attention is
directed to the relevant part of semantic memory also leads to facilitation of related words, but
inhibition of unrelated words.
Is lexical processing qualitatively affected by its sentence context?
Schuberth & Elimas (1977), Fischler & Bloom (1979), and others:
Lexical decision responses to BONE faster when preceded by
"
The dog gnawed happily on the"
than "
xxx xxx xxxxx xxxxxxx xx xxx
"
(equivalent to DOCTOR NURSE vs. XXXXXX NURSE: Predictable vs Neutral)
Also inhibition for unrelated (anomalous) contexts (equivalent to NATION NURSE):
"
The girl ran hurriedly through
the BONE) > "
xxx xxx xxxxxx xxxxxx xx
xxx BONE"
However, unlike INJECTION NURSE, related, but not predictable sentence contexts produce no priming:
"
Inside the cave we found a large
BONE" = "
xxxxxx xxxx xxxxxx xx xxxx
BONE"
Lexical decision responses to a target word show facilitation when predictable from the sentence
context, inhibition for unrelated (anomalous) contexts, and no priming when simply plausible in the
context (but not predictable).
Forster (1981) explains this pattern of results as demonstrating that the main impact of context is to
inhibit words that don't fit the context
-
a
post
-
access
account of context. The facilitation observed with
highly predictable targets is seen as a special case of guessing, and is not the typical thing we do when
reading (because most words are not so predictable from their context).
Inhibition in normal reading is required when a word is misperceived
-
the misperceived word does not
fit the context, so the letter
-
string is re
-
analysed.
Inhibition is also important when a word is ambiguous (e.g.,
port
). If the wrong meaning is accessed, its
failure to fit the context alerts you to this, and you try another meaning. This implies that all meanings of
an ambiguous word are activated, and that context is used to suppress the inappropriate ones.
Research by Swinney (1979) supports this:
Spoken sentences like:
After spraying the room with insecticide, we were still able to find some bugs in the carpet.
Influence of context on lexical processing
Monday, 18 June 2018 12:34 PM
Lecture 14 Page 1
Document Summary
Meyer and schvandeldt (1971) were the first to show that a target word (e. g. , nurse) is recognised more quickly when preceded by a semantically related prime (e. g. , doctor), than by an unrelated prime (e. g. , nation). However, it has also been shown (by neely, 1976) that a neutral condition (e. g. , Nurde preceded by xxxxxx) falls in between the other two conditions. That is, related words show facilitation (relative to the neutral baseline) while unrelated words show inhibition. nurse is a predictable response to the word doctor in a semantic association task. Neely explains the pattern of priming results in terms of there being two processes: 1) automatic spreading activation leads to facilitation of related words. 2) a conscious strategy where attention is directed to the relevant part of semantic memory also leads to facilitation of related words, but inhibition of unrelated words. Schuberth & elimas (1977), fischler & bloom (1979), and others: Lexical decision responses to bone faster when preceded by.
