Describe the 4 th experiment discussed in Guasti et al. (2005),…

JusticeFreedomMagpie33Describe the 4 th experiment discussed in Guasti et al. (2005),…Describe the 4th experiment discussed in Guasti et al. (2005), ‘Why children and adults sometimes (but not always) compute implicatures’ – the materials that were used, the procedures, participants, etc. What were they aiming to show with this experiment? Why do they think that the Truth-Value Judgement Task (TVJT) that they used is preferrable to the Statement Evaluation task (SET) that was used by Noveck (2001)? What were the results of this experiment? Do these show what Guasti et al. (2005) set out to prove? Why or why not?  EXPERIMENT 4: CONTEXT MATTERS Let us now consider the factors that contribute to the computation of implicatures, by asking, first, about the contribution of such factors when a person computes an implicature in natural conversation. One factor that is relevant for the computation of implicatures is the awareness that statements can differ in the quantity of information they convey—some statements are more informative than others (e.g., all-statements are more informative than some-statements). A second factor is the activation of a scale that includes the relevant terms (the use of ‘some’ activates a scale including ‘all’). A third factor is the assumption that a speaker does not utter a some-statement when she knows or has evidence that an all- statement holds. In turn, the listener assumes that if a some-statement is produced by a speaker, then the speaker does not know or has insufficient evidence to conclude that the corresponding all-statement is true. If the listener has her own evidence for an all-statement, then she can overtly reject the speaker’s some-statement. In the Noveck study, the third relevant factor for the computation of scalar implicatures was not controlled, in the sense that the evidence for the some-statements was not provided in the experimental workspace. In the absence of concrete evidence pertaining to the subjects’ judgements, however, we cannot be sure why adults (albeit less than children) sometimes assented to statements like some giraffes have long necks. Perhaps the adult subjects took the opportunity to conjure up a subset of giraffes, e.g., baby giraffes, which made the statement express a true proposition. In this case, the evidence against which the statement was evaluated would have included both baby and adult giraffes; using this domain as the evidential basis, the statement could have been accepted as a reasonable description of the facts. An experimental subject that disagreed with the same statement, by contrast, might have had just prototypical giraffes in mind, i.e., these subjects might have taken the statement to be a description of a typical property of giraffes. In this case the statement would have been an unreasonable description of the facts, since giraffes typically have long necks. Similarly, some bikes have a handle-bar might be accepted because the listener might evaluate the statement against the domain including bikes that are broken and lack handle-bars versus bikes that are in good condition. As this discussion shows, the evidence against which statements were evaluated by subjects has not always been controlled in previous experiments; the evidence in question was left up to the subjects to construct. No step was taken to ensure that all subjects evaluated the statements using the same domain, i.e., that of a prototypical individual or entity. Therefore, in the Noveck experiment and in our replication of it (Experiment 1), one possibility is that many children and adults produced a high percentage of acceptance of critical statements because of this feature of experimental design, one that makes the derivation of implicatures irrelevant. Moreover, on this scenario, it is not surprising that children accepted the critical statements more than adults. First, children may have experienced even more difficulty in figuring out the ‘right’ or the ‘experimenter intended evidence’ against which to evaluate the test statements. In addition, children may have been more biased than adults to mentally construct situations that make the experimenter’s statements true (for discussion, see Crain & Thornton 1998; Grimshaw & Rosen 1990). To eliminate this potential drawback to the experimental design, we chose to adopt a different methodology in the present experiment, namely one that permits the experimenter to control the evidence that is used by subjects in the evaluation of the test materials. In this case, we can be sure that a potential failure to compute implicatures is not due to the extraneous factors mentioned in the preceding paragraph. The change in method was to opt for the Truth Value Judgement Task (Crain & Thornton, 1998), which is a task that allows the experimenter to control the situation and, thereby, to establish the conditions that are prerequisite for computing scalar implicatures. Subjects. Fifteen Italian-speaking children (age range 7;0-7;5 years, mean age 7;2) from San Pellegrino Terme (Lombardia) participated in this study and 12 undergraduate students from the University of Milano- Bicocca. Materials. Subjects were asked to judge five statements including some which were true, but under-informative in the context of use (e.g., Some monkeys are eating a biscuit in a situation in which all monkeys were eating a biscuit). Since in previous experiments, quantified statements including some and all in felicitous contexts were not problematic for 7-year-olds, we did not include such statements in the present experiment. To ensure that children could reject false statements and accept true ones, and to be sure that they were paying attention, the targeted statements were interspersed with fillers, some of which were clearly true and others which were clearly false. Procedure. This experiment used a video-taped version of the Truth Value Judgement Task (TVJT) (Crain & Thornton, 1998). In our version of the TVJT, children watched a video featuring an experimenter acting out stories using props and toys and holding a puppet, Carolina, who was watching the stories alongside the child. At the end of each story, Carolina said what had happened in the story. The child was instructed to say whether Carolina’s statement was a good or a bad description of what happened, and to explain her answer whenever she judged Carolina to have ‘said the wrong thing.’ Children were previously familiarised with Carolina, and had been informed that Carolina was still a baby and, for this reason, she would sometimes be unable to correctly describe what happened in the stories. Children were tested individually in a quiet room where they watched the video together with an experimenter. They were invited to indicate their answer to the experimenter who filled a score sheet and took note of the explanation. Adults were also shown the video and were given a score sheet on which to write their answers. On a typical trial for the target sentence with some there were five characters performing some action. For example, one story featured five soldiers that had to go far away to collect a treasure and could either go by motorbike or ride a horse. Initially there was some discussion among the soldiers; some soldiers said that they would like to go by motorbike, since motorbikes are fast; other soldiers argued that gasoline is expensive and that it would be better to ride a horse. After this discussion, they all choose to ride horses. Then, Carolina was asked to say what was happening in the story. In the present case, Carolina’s description would be: Some soldiers are riding a horse. Then, the child was invited to say whether what Carolina had said was ‘right’ or ‘wrong.’ Results and discussion The main finding of this experiment is that children rejected the critical statements nearly as often as adults: the rejection rate was 75% for children and 83% for adults. A one way ANOVA was conducted, with age (Children, Adults) as a factor and logically correct responses as the dependent measure. The analysis revealed that, unlike in Experiment 1, there was no significant difference between children’s and adults’ responses: F(1, 25) 1/4 0.31, p 1/4 .58 (acceptance of critical statements by children M 1/4 .25, SD 1/4 .41; adults M 1/4 .166, SD 1/4 .389). On control items, children responded correctly 100% of the time and adults 97% of the time. No reliable difference between these means was found: F(1, 25) 1/4 2,77, p 1/4 .108. Finally, the distributions of the responses by children and adults were similar, as shown in Figure 3. Figure 3. Distribution of children (3a) and adults (3b) depending on the number of times they accepted critical statements in the Truth Value Judgement Task (TVJT). Image transcription textNumber of subjects Number of subjects 15 _|. O Ul 15 10′ (38) Distribution at children in the TVJT 7 2///Number of times critical statements were accepted (3b) Distribution oi adults in the TVJT W e i 2 e 4 5 Numberof times critical statements were accepted … Show moreMost of the subjects always rejected the target statements. The proportions of children and adults who accepted the critical statement three or more times did not differ (p 1/4 .585, n.s.). Thus, the use of different material and procedure had a dramatic effect; while in Experiment 1 a clear developmental effect was found between children and adults, this effect disappeared completely in this experiment, when a different method was used. Children rejected the critical statements and always explained their responses by invoking the stronger term of the scale; for example they explained that some soldiers are riding a horse was a bad description of what was happening because all the soldiers were riding a horse. Thus, we can conclude that 7-year-olds do compute implicatures when the evidence for evaluating under-informative statements is clearly in front of them, and when the task is clear about the nature of judgement that they were required to make. The same observation holds for adults. As in the Papafragou and Musolino (2003) study, our adult subjects rejected the test statements; that is, they computed the implicatures much more often than the corresponding adults had in Noveck’s experiment and in our replication of it. In fact, they almost always computed it. We return to this point in the general discussion.  Social SciencePsychology