Assessing metaphor comprehension as a metasemantic ability in students from 9- to 14 years-old

Resumo

This article presents a test for assessing metaphor comprehension (MCT) in explicit linguistic form for subjects aged 9- to 14 years-old, i.e. in a transitional age where significant qualitative changes appear for this type of ability. Metaphor is viewed as a form of semantic conflict induced by the anomalous combination of the conventional meanings of its main constituents - tenor and vehicle - and metaphor comprehension is framed as a metasemantic ability based on the analysis of these meanings (Gombert 1990), that can have implications both for teaching and for learning strategies.The authors propose a functionalist piagetian frame, based on Piaget’s latest equilibration model (1975), for analysing how this semantic conflict can be faced and solved by children in the developmental span considered. The test is paper-and-pencil, composed of 12 items subdivided into 2 groups of metaphors: Psycho-physical (PP) and Conceptual (C), mainly drawn and adapted to Italian language from international literature on metaphor comprehension. The sample is composed of 874 Italian children from 4th to 8th grade, with gender balance, of average social background. By means of Principal Components Analysis, with oblimin rotation, a two-factor solution emerged, that espouses the C/PP metaphors distinction. All corrected item-total correlation coefficients >.30 were representative and acceptable. Correlation coefficients between MCT and Standard progressive Matrices (SPM38) and some validated metalinguistic subtests were all significant at p <0.01 level, showing good convergent validity. Cronbach’s Alpha coefficients for the two subscales were: .62 (C), .51 (PP) and .70 for MCT total score. Test-retest correlation coefficients for the two subscales were .79 (C), .69 (PP), and .83 for MCT total score. Cohen’s Kappa coefficients for interrater reliability are: .75 (4th-5th), .74 (6th), .67 (7th), .81 (8th). A twofactor ANOVA (gender and age) showed that the test is gender-sensitive in favour of females but that age is an even stronger factor, which underlines the developmental character of the test. Limitations and possible research developments are analysed in the Discussion.