Induced errors
المؤلف:
Paul Warren
المصدر:
Introducing Psycholinguistics
الجزء والصفحة:
P73
2025-11-02
46
Induced errors
Using the SIIP (Spoonerisms of Laboratory-Induced Predisposition) technique, a series of studies has induced sound errors, specifically spooner isms (Baars & Motley, 1974; Motley, 1985; Motley & Baars, 1976; Nooteboom, 2005a). The methodology involves the rapid and brief presentation on a computer screen of pairs of words that have the same repeating pattern of initial consonant sounds. Typically, the consonant sounds used in this task are phonetically quite similar, as errors are more likely if this is the case, just as they are in spontaneous speech. The three pairs in 5.3 all have the /ʃ / /h / pattern. The word pairs are presented one at a time for silent reading, except that occasionally an additional prompt asks participants to say aloud the pair they last saw. The key condition is when such a pair has a different order of the initial consonants from the sequence that participants have got used to. An example of such a target pair following the sequence in 5.3 is given in 5.4. The prediction is that if a spoonerism is induced by the sequence in 5.3, then the participant will say shot hurt’ for 5.4.
The issues being addressed in this task are varied, but include the extent and nature of the monitoring that speakers make of their own output. To measure this, the experimenters used different target pairs in the key condition, including examples like those in 5.4 above and 5.5 below, again following a sequence such as that in 5.3.
What the experimenters found was that participants, under otherwise identical conditions, would produce a spoonerism 20 of the time when the key condition was like that in 5.4, but only 6 of the time when it was as in (5.5) (Baars et al, 1975). The crucial difference between these conditions is that a spoonerism in the first case would produce a sequence of real words (shot hurt), but in the second it would not (shide hame). This indicates that speakers monitor their output and filter out nonsense words, even in this rather unnatural task. If this extends to spontaneous language production then it explains the tendency for speech errors to produce real words rather than nonsense, as we saw in Chapter 4. That is, the real word bias may be due at least in part to the speaker filtering out the nonsense.
When confronted by a stimulus like that in 5.6, however, participants were less likely to produce a spoonerism than in either of the other conditions, producing them on only 4 of occasions Motley, Camden Baars, 1982.
Although the spoonerism would produce real words, this would be the sequence shit head. It seems that the monitoring system excludes inappropriate language, as well as nonsense words.
In a more complex version of this task, properties of the experiment setting were exploited Motley, 1980. The basic set up again involved sequences of stimulus pairs with the same pattern of initial consonants leading to a possible spoonerism on a target stimulus pair. Within the experiment there were two types of potential spoonerism – one related to the notion of electric shocks e.g. (bad shock as spoonerism of shad back, following a sequence of b_ sh_ pairs), and the other to the notion of glamorous or attractive women (good legs as a spoonerism of lood gegs), following a sequence of g_ l_ pairs). All participants received the same sets of stimuli, but they were placed in one of two conditions. In one condition, participants were connected to fake electrodes and told that they might receive mild electric shocks during the experiment. The other participants did not receive this instruction, but were instead met by an attractive and provocatively dressed female experimenter. The first group turned out to be more likely to produce errors that related to the electric shock situation. The second group was more likely to produce errors related to the notion of glamorous women. This indicates that the setting in which the participants found themselves had an effect on their ability to filter out the errors – if the error was in some sense primed’ by the setting, then it was less easy to block.
We will return below to further discussion of the nature of the monitoring for and filtering of errors, but first we will take a closer look at the kinds of repairs that speakers carry out when they detect an error in their output.
الاكثر قراءة في Linguistics fields
اخر الاخبار
اخبار العتبة العباسية المقدسة
