Figurative language comprehension and links to autistic traits

Figurative language comprehension and links to autistic traits

Anamarija Veic

2018

Figurative language is used quite frequently in both speech and writing, as to express our creative and abstract thoughts. Traditionally, it was thought that metaphors are ornamental in nature, as well as they are used rarely compared to literal language. However, today’s research suggests that people use metaphors in everyday communication. Moreover, people seem to pay more attention to sentences which are emotionally evocative, rather than neutral ones. In addition, it has been extensively reported that socio-communicative skills might be related to the successful comprehension. Special populations, such as autistic individuals, often struggle with both figurative language comprehension and acknowledging properly other people’s emotions. However, no prior research has explored both different types of sentences and their content (emotional or neutral). Sixty-two participants took an online questionnaire measuring their comprehension abilities and the Autism-Spectrum Quotient (AQ) test, in order to measure their socio-communicative skills. Significant results were found for both the type of sentences, and the content. No significant effect of socio-communicative skills affecting comprehension was found. The results are discussed in terms of their theoretical and clinical importance.

figurative language
comprehension
emotions
autism

Participants
Sixty-two typically developed participants (M=31, F=31) between the ages of 18 and 62 (M= 24, SD=9.32) were got involved in the study. The majority of sample were students at Lancaster University (N=51). Participants were recruited in Lancaster (United Kingdom) via SONA or email. Twenty-nine participants were paid £5 (five British pounds) for taking part in the study. The remaining participants were not re-imbursed for their time. Only the adults (minimum age of 18) who were British English native speakers could have taken part in this study. Participants were not aware of a true aim of the study. Participants were simply told that the project is about figurative language comprehension, as to avoid any possible bias. At the end of their participation, they were informed about the details and the aims of the study. The study has been approved by the ethics committee.
Apparatus and materials
The participants were asked to complete an online questionnaire developed with the Qualtrics survey software. Upon recruitment, participants were sent a Qualtrics link to the survey. All participants were exposed to the same stimuli but each of them got a different randomised order. Approximately ten minutes were sufficient for participants to take part in the study. Participants could start answering the questionnaire and then finish it at another point of time if needed, as their answers were automatically saved for seven days after they opened the questionnaire on their browser. No more than 10 sentences were shown per page, as to avoid fatigue.
Both literal sentences and novel metaphors used as stimuli in this project were originally structured by Cardillo, Schmidt, Kranjec, and Chatterjee (2012). Their aim was to construct a design of matched metaphoric and literal sentences as to test the role of novelty and different metaphor types involved in metaphor comprehension. The authors managed to control the next ten dimensions: dimensions: length, frequency, concreteness, familiarity, naturalness, imageability, figurativeness, interpretability, valence, and valence judgment reaction time. What makes these sentences even more different than previous work is the fact that the same word was used in both literal and novel metaphors. As such, literal sentences and novel metaphors were further analysed and selected in a laboratory by Francesca (my supervisor) Citron’s students. The students selected the stimuli based on existing value of valence and imageability, so that sentences from different condition would differ in emotional valence, but not in the imageability. Conventional metaphors were structured by the same students, as well. Students created simple sentences which contained similar structure as the existing ones. Yet, it was not possible to use the same word as from literal sentences and novel metaphors, so conventional metaphors were a bit more diverged. The content of sentences was controlled in a way that half of the sentences were positive, and another half of them was neutral, so that their level of imageability would have been similar to novel metaphors and literal sentences.
Finally, for the current research, the conventional metaphors were edited as to make them shorter to be more alike to both literal sentences and novel metaphors. The length was calculated and analysed statistically, for both the content and the types of sentences. There was no significant difference neither between the number of words nor the number of letters, both regarding the content and the types of sentences, p>.05. It is important to note that the current study did not replicate what Cardillo, Schmidt, Kranjec, and Chatterjee (2012) already explored since their main interest was to investigate neural processes underlying metaphor meaning.
The questionnaire consisted of 120 short questions such as ‘To which extent do you understand this sentence?’ containing one type of a metaphor expression (e.g. ‘The woman dove into the pool.’). Participants were required to rate the ease of the comprehension on a scale from one (‘It does not make any sense at all.’) to five (‘It makes perfect sense.’). The questionnaire included 20 sentences of each of the following groups, which are presented in the Table 1.
The Autism-Spectrum Quotient (AQ)
The AQ test was used at the end of the questionnaire. It is a self-report measure of autistic traits and presents a valuable instrument for rapid quantifying where any given individual is situated on the continuum from autism to normality (Ruzich et al., 2015). The test was constructed by Baron-Cohen, Wheelwright, Skinner, Martin, and Clubley (2001) since no prior instrument at that time could have measured such factor. It can be administered to adults of at least average intelligence with autism or to nonclinical controls but can also be administered to clinical control groups (e.g., individuals with depression) (Ruzich et al., 2015). The AQ consists of 50 questions assessing five different areas: social skill, attention switching, attention to detail, communication, and imagination. Thus, participants’ scores could range between 0 and 50. Approximately half the items were worded to produce a “disagree” response, and half an “agree” response. This was to avoid a response bias either way. Following this, items were randomized with respect to both the expected response from a high-scorer, and with respect to their domain (Baron-Cohen, Wheelwright, Skinner, Martin, & Clubley, 2001).
Design and procedure
The dependent variable was the ease of figurative language understanding. The within-participants independent variables were type of a sentence (conventional, novel, and literal) and content (positive or neutral). Conventional metaphors represent expressions commonly used in everyday setting, whereas novel metaphors were made up for this occasion. The between-participants independent variable was the degree of autistic-like traits (either high or low). To obtain this latter variable, participants were divided into two groups based on their AQ scores. The median score was used to split them. Participants were instructed to rate their understanding of metaphors in 120 sentences. There were 20 sentences of each type × content (e.g., conventional positive) (see Appendix A). Thus, six different mean scores were calculated for each participant (conventional positive, conventional neutral, literal positive, literal neutral, novel positive, novel neutral).The Likert scale consisted of five points (1-‘It doesn’t make any sense at all’, 2-‘It doesn’t make much sense’, 3- ‘It makes some sense’, 4- ‘It makes sense’, 5-‘It makes perfect sense’). The following coding rules were applied to calculate the AQ score: “definitely agree” or “slightly agree” responses scored 1 point on items number 1, 2, 4, 5, 6, 7, 9, 12, 13, 16, 18, 19, 20, 21, 22, 23, 26, 33, 35, 39, 41, 42, 43, 45, 46. “Definitely disagree” or “slightly disagree” responses scored 1 point on items number 3, 8, 10, 11, 14, 15, 17, 24, 25, 27, 28, 29, 30, 31, 32, 34, 36, 37, 38, 40, 44, 47, 48, 49, 50 (see Appendix B). Subsequently, the AQ scores were divided in two groups based on the median score (Med = 19.5). Any results above the median threshold were categorised as high, and those below were categorised as low. Half of the sample (N = 31) scored high, while the other half (N= 31) achieved a low score. Results were analysed using a 3x2x2 mixed analysis of variance (ANOVA).

Lancaster University

Data/SPSS.sav

Veic2018

Ellie Ball

Open

None

English

Data

LA1 4YF

Francesca Citron

MSc

Clinical Psychology
Cognitive Psychology
Psycholinguistics

62 participants (31 males and 31 females)

Mixed ANOVA