Running head: THE SYMBOLIC INTERPRETATION OF POETRY



Peskin, J. and Wells-Jopling, R. (2012). Fostering symbolic interpretation during adolescence. Journal of Applied Developmental Psychology, 33(1),13-23.

Abstract

Although by 11 years children demonstrate impressive performance on various tasks that assess symbolic thinking in language development, research suggests that few young adolescents demonstrate evidence of symbolic processing when reading literature. This study investigated whether the difficulty might be due to a lack of adequate exposure to domain-specific knowledge. Students in the experimental groups in three age groups - preadolescence, middle adolescence and later adolescence - received concrete scaffolds designed to foster domain-specific knowledge of the symbolic process. A comparison of the experimental and control groups showed that students at all three ages who had experienced the scaffolds demonstrated significantly greater symbolic interpretation. Furthermore, despite concerns that the scaffolds might dampen the readers’ personal response, the experimental groups at all three ages provided significantly higher enjoyment ratings of the test poems.

Keywords: adolescence; symbolic interpretation; domain-specific knowledge; poetry; concrete scaffolds; computational skills

Fostering symbolic interpretation during adolescence

Although the language of poetic expression is highly symbolic (Gibbs, 1994), in a recent study on the interpretation of poetry during the school years, Peskin (2010) found that 8th graders did not demonstrate greater symbolic interpretation of poetic texts than the 4th graders in the study. Similarly, Harker (1994) described how 10th graders essentially gave prose translations of a poem’s literal meanings rather than exploring the imaginative symbolic possibilities; and Svensson (1987) who provided students with poems rich in symbolism, found that only 8% of responses of 11-year-olds, 18% of responses of 14-year-olds, and 42% of responses of 18-year-olds at vocational and academic schools provided even partly symbolic interpretations of poems. The major purpose of the present study is to investigate whether the difficulty might be a lack of domain-specific knowledge: Will an intervention that targets the knowledge required for symbolic interpretation improve adolescent performance at three age-points, pre-adolescence, middle adolescence and later adolescence?

Conceptual metaphors are so pervasive in human language that we are virtually unaware of their metaphorical character or even their existence (Johnson, 1991). For instance, in the conceptual metaphor, “People are plants,” people are understood metaphorically in terms of the life cycle of plants, as “in full bloom” or “withering away” (Crisp, 2003; Lakoff & Turner, 1989; Turner, 1991; Winner, McCarthy & Gardner, 1980). Even children easily understand many conceptual metaphors, for instance, “Good is up” and “Bad is down,” resulting in everyday expressions such as “I am over the moon” or “I’m really down in the dumps.” Johnson & Lakoff (2002) explain this ease in terms of “embodied realism” which holds that as humans have the same kinds of bodies and environments they have shared image schemas. Unlike the processing of a conceptual metaphor which appears to be largely unconscious, the processing of literary symbolism appears to be an intentionally selected, conscious and effortful literary strategy (Gibbs, 1996; Steen, 1989). However, as creative literary metaphors are often novel variants of common conceptual metaphors it is somewhat of a mystery why adolescents have such difficulty with symbolic thinking when reading literature, and in particular, poetry.

Development of an Understanding of Non-literal Language

Symbolic thinking emerges in linguistic expression at 18 to 24 months with the representation of objects, events or activities by means of words (Fischer & Bidell, 1998). In the preschool years there is a further development with the understanding and production of visual image metaphors, which involve the appreciation of similarities between very different domains of knowledge, as in a comment such as, “the chimney is a house-hat” (Billow, 1981; Harris, 1982; Winner, McCarthy & Gardner, 1980).

After about four years of age children begin to develop a metacognitive awareness of non-literal language as they become able to distinguish “what is said” and “what is meant,” (Lee, Torrance, & Olson, 2001; Robinson et al, 1983; Torrance & Olson, 1985). For instance, they appreciate that a listener will find the request for a “blue flower” ambiguous if there is both a small and a large, blue flower (Peskin & Olson, 2004; Ruffman et al, 1990). This distinction is conceptually congruent with children’s new found ability to represent other people’s mental states, or what is called theory of mind: Theory of mind tasks (which involve thinking about someone’s ignorance of something that the child knows to be true) as well as say-mean tasks (which involve thinking about what someone both said and meant) are dependent on children’s recently established ability to co-ordinate two different representations or perspectives.

The say-mean distinction is the cognitive underpinning in the comprehension of non-literal language, including jokes, riddles, and sayings such as, “Don’t change horses midstream.” Although prior to age 7 children will process the language of idioms and metaphors literally, by 9 years of age research using various task factors suggests that children show an impressive understanding of figurative modes of thought (Berman & Ravid, 2010; Cacciari & Levorato, 1989; Gibbs, 1994; 1996; Vosniadou, 1987). At around 11 years of age pre-adolescents comprehend metaphors that are based on abstract relations (Billow, 1975; Gentner, 1988). However, the ability to use figurative language in a creative way is still fairly undeveloped at the age of 11 (Levorato & Cacciari, 2002), and like the symbolic processing of poetry continues to develop through adolescence.

Domain-Specific Knowledge in the Development of Symbolic Interpretation

An important question in applied developmental psychology, as it has huge import for pedagogy, is whether adolescent difficulties with symbolic thinking in literary reading might be a result of a lack of exposure to a field of knowledge or whether there is a fixed neural timetable such that teachers might need to wait (Byrnes, 2007) before expecting children to engage in symbolic interpretation when reading literature. Domains are recognized bodies of subject matter knowledge which are structured around concepts that are core to these fields (Alexander, 1997). Research in various domains has allowed for the identification of domain-specific strategies or heuristics (e.g., Peskin, 1998; Wineburg, 1991), and it has been argued that domain-specific knowledge acquisition is the greatest contributor to proficiency in any domain (Ericsson, 2006; Feigenbaum, 1989; Minsky & Papert, 1974). Deep structures of knowledge, or schemata, enable experts to see meaningful patterns. The corollary is that a lack of such knowledge is the primary handicap for novice literary readers (Hall, 2005). It has been suggested that adolescent difficulties with literary symbolism might be a result of deficits in literary education (Pirie, 2002; Smith & Wilhelm, 2004; 2006).

Most of the symbolism involved in poetry involves metaphor (Gibbs, 1994), and processing a creative literary metaphor is thought to involve a mapping of two different conceptual domains: the symbolic element, also called the “source” or “vehicle” (Richards, 1936) which is always expressed in the text, and the particular underlying topic, also called the target, which is what the symbol is about and may not be explicit in the text (Lakoff & Johnson, 1980). For instance, in the few lines of a poem translated from Swedish (Svensson, 1987, p. 503), “It is this persistent wind/ It just gets stronger// One after the other the old trees fall/ here in the garden// before the fruit has had time to/ ripen,” a young reader might see the poem as ultimately about trees falling. More experienced readers of poetry, however, assume that a seemingly banal poem will mean more than may be apparent in the mere lexical definitions of the words (Groeben & Schreier, 1998). These readers are likely to provide more creative, and often personal, symbolic interpretations, such as people dying before fulfilling their goals; or the power of new technologies; or perhaps life forces destroying the creative process. Expecting the poem to involve “polyvalence” or multiple meanings (Schmidt, 1989), such a reader examines which parts of the poem might be taken as symbolic. These underlying assumptions are part of the domain-specific knowledge required for literary symbolic interpretation.

Further, domain knowledge in symbolic interpretation involves the “grammar” of symbols that recur frequently in literature (e.g. Frye, 1994). Frye argued that “poems are made out of the same images, just as poems in English are all made out of the same language,” (Frye, 1994, p. 275). The fundamental structure of these symbols or archetypes involves the parallels between the cycles of human life and the cycles of the world. Once similarities between, for instance, the cycle of Spring to Winter and the human cycle of birth to death are perceived, symbolic interpretation has occurred.

The absence of such knowledge in the adolescents’ literary reading repertoire could substantially impede progress in arriving at symbolic interpretations of poems. It is in this sense that an understanding of symbolism may be said to represent the core domain knowledge present in the reading of poetry. Various genres involve poetic effects such as symbolism, but poetry is the genre where the poetic function “changes from latent to patent and manifests itself most palpably and intensely” (Jakobson, 1960, p. 373). Adolescents’ poor performance in studies on symbolic processing may be because they do not yet have the genre-related expectation that exploring symbolic content enriches one’s reading nor the interpretive strategies to do so. Research on the teaching of English has shown the effectiveness of a proactive focus on the knowledge and strategies underlying high literacy (Langer, 2001a, 2001b). This fits well with the theories of developmental psychologists such as Karmiloff-Smith (1992) who claim that only when the representations of concepts and processes become explicit are they accessible to consciousness, verbal articulation, and cognitive control and flexibility.

Possible Cognitive Computational Constraints in the Development of Symbolic Interpretation.

For a symbolic construal of a poem, readers need to inhibit their tendency to only focus on what the poem is literally saying, that is, the surface and salient meaning (Gardner, 1975). However, recent neuropsychological work suggests that certain computational processes, such as “inhibiting the salient response” are associated with the prefrontal cortex, the area of the brain involved in executive functions (Rubia, Smith, Brammer, & Taylor, 2003) and that at puberty there is a proliferation of synapses in the prefrontal cortex causing less efficient cognitive functioning (Giedd, Blumenthal, & Jeffries, 1999). Indeed, there are concerns that at 11 to 12 years there is actually a decline in performance on some tasks which involve response inhibition (Blakemore & Choudhury, 2006). Furthermore research on understanding metaphor has recently shown that when participants read novel metaphors, the kind involved in poetry, there is involvement of regions in the prefrontal cortex that overlap with those regions that continue to undergo refinement through adolescence (Lee & Dapretto, 2006; Rapp et al., 2004; Stringaris et al., 2007). If the prefrontal cortex plays such an important role in metaphoric processing, perhaps the young adolescents’ conceptual skills are not yet sufficiently developed to process the complex relations involved in symbolic understanding when reading poetry. Neuro-developmental constraints during the early adolescent years might make symbolic interpretation very difficult to teach.

The first aim of the study will be to examine whether providing the requisite domain-specific knowledge will, indeed, foster symbolic interpretation during adolescence. This will be examined at three age points – preadolescence, middle adolescence and later adolescence. We hypothesized that domain-specific instruction would foster symbolic thinking in the two older age groups but possibly not during preadolescence. With this aim in mind, we planned to expose an experimental group at each of the three ages to domain-specific knowledge and compare their pre- and posttest literary symbolic interpretation skill with a control group at each age, statistically controlling for cognitive computational ability.

The Reader-Response/Explicit-Instruction Dichotomy

In any consideration of adolescents’ difficulty with symbolic processing in poetry one must examine a fundamental tension specific to poetry instruction. Because of the aesthetic, personal and emotional nature of poetry, many teachers fear that making explicit the structures and processes of literary symbolism will generate antipathy to literature. They believe that the knowledge needed for effective symbolic interpretation is best acquired tacitly through untutored perception, rather than by making the process explicit (Pirie, 1997; 2002).

Dressman and Faust, who recently surveyed more than 600 articles about teaching poetry written primarily by teachers, from 1912 to the present, describe two opposing traditions or approaches that emerged: A “Populist” emphasis on texts to be played with, in which the reader’s personal response is highlighted, but which is frequently seen to be at the expense of critical rigor; versus a “Formalist” emphasis on explicit teaching and critical rigor, which is frequently seen to be at the expense of personal enjoyment (Dressman and Faust, 2007; Faust & Dressman, 2009). In the last quarter century this dichotomy has been most evident in the polarization between what is called structural or textual analysis in instruction (van Schooten & de Glopper, 2003) and reader response theory, which emphasizes how the literary work of art “comes into being through the reader’s attention to what the text activates within him” (Rosenblatt, 1985. p. 38). However, this dichotomy in pedagogical approaches to poetry has deep historical roots, summed up by Glicksburg and Gordon’s claim (1939) that “the living poem is often destroyed in the process of salvaging its hidden meaning (p. 550).

In developmental studies of fiction reading, there is some empirical support for teachers’ fears. In a longitudinal study van Schooten and de Glopper (2003) provide some evidence for the negative effect of formal instruction on personal response in reading fiction. Using Miall and Kuiken’s (1995) “Literary Response Questionnaire,” a measure of personal response to literature, van Schooten and de Glopper (2003) showed that literary response decreases with age from 7th to 11th grade, when formal literary instruction is introduced and then intensifies. Furthermore, between grades 7 and 9 teaching methods which emphasize structural analysis have particularly negative effects on various personal responses, e.g., students’ reports of empathy with story characters. Similarly with regard to reading fiction, Greenleaf, Schoenback, Cziko and Mueller (2001) demonstrated that students frequently stop reading for pleasure in middle school and even begin fake “reading” during their silent-reading periods. However, in the reading of poetry, although the dichotomy between formalist and populist pedagogical approaches continues to be a rich area of debate and theoretical discussion, to our knowledge there is no scientific evidence to support the notion that explicit teaching of symbolic processing when reading a poem lowers students’ enjoyment ratings when reading poetry.

The second aim of the present study is to explore whether providing adolescents with the domain-specific knowledge of symbolic interpretation will inhibit students’ literary reading experience, in particular their enjoyment when reading poetry. Based on the evidence from studies on reading fiction we hypothesized that lessons designed to provide explicit knowledge of the symbolic process in poetry might dampen students’ personal responses to the poems.

Method

Participants

Participants were 137 students from six classes. Table 1 presents age and gender distribution by grade and lesson group.

--------------------------------------Insert Table 1 here-------------------------------------

An additional 17 children were excluded from the study: two ESL students and three students with learning disabilities, as advised by their teachers; and 12 students who were absent for one or more of the testing or lesson sessions. All the grade 9 and 12 participants attended the same high school, and the grade 6 participants attended an elementary school that is a feeder school for the high school. Both schools draw on children from middle-class and upper middle-class neighborhoods in a large Canadian city. The choice of grade 6 as the lowest grade was based on Piagetian and neo-Piagetian notions of the beginning of symbolic thinking at 11 years (Case, 1992; Inhelder & Piaget, 1958) as well as the neurophysiological research on synapse proliferation during puberty and possible performance decline on tasks involving response inhibition (Blakemore & Choudhury, 2006). One of the classes in each grade was randomly assigned to be the experimental group and the other class served as control group.

Materials

Test poem selection process. Many poems are narrative or dramatic and may not be rich in symbolism. The poems used in the present study involved invisible symbolism, that is, it was possible to take each poem either in a purely literal manner or to create various imaginative, symbolic construals. If these particular poems are taken literally, though, the meanings would be quite banal. Culler (1994) notes that trained readers expect that a poem should make a point or refer to something beyond itself, so when a poem seems to be platitudinous, the reader asks if this is the entire meaning (Svensson, 1987). Numerous poems, which we chose based on whether they involved invisible symbolism and language suitable for students in 6th to 12th grade, were read by our research group members plus an English instructor in the teacher education program, and rated according to “difficulty” (“1 = very easy to 10 = very difficult”), as well as “salience of the symbolic content” (“1 = not very salient to 10 = very salient”). Poems were eliminated if rated as difficult and/or if the symbolic content was not salient. The teacher from each participating 6th grade class then assessed the language of the remaining poems as suitable for their students (in a middle to upper-middle class neighborhood), or if a minimal change to a word could make the text suitable. Although in most disciplines it is unusual to find texts appropriate from Grade 6 to Grade 12, poetry is somewhat different in that it can be interpreted at various levels. As Bruner (1986) noted, literary texts “call upon the reader, in Barthes’ sense, to become a writer, a composer of virtual text in response to the actual” (p. 24) and a simply worded poem may be read literally by some, but also generate sophisticated interpretations when read by others.

Test poems. Five poems were selected for both the pretest and posttest booklets. One of the poems consisted of the first 14 lines of Stephen Spender’s poem “What I Expected” in an anthology for school students (Spender, 1970). In this poem Spender writes poignantly about his life expectations and later disillusionment. He describes how he expected “Thunder, fighting./Long struggles with men/And climbing…Then the rocks would shake.” What he had not anticipated was the “gradual day” weakening one’s will, “Leaking the brightness away,/The lack of good to touch,/The fading of body and soul.” In previous testing we have found that many students look at the surface words “climbing” and “struggles with men,” and interpret the poem as a narrative about climbing a mountain during a battle.

A second poem consisted of the first seven lines of “Washing” by Randall Jarrell (1969), with slight changes, such as replacing the word “abject” with “miserable” so that it was suitable for 6th graders. The remaining three poems adapted from Svensson’s (1987) translations have been provided in Appendix A. Slight changes were made mostly to improve Svensson’s translation, for instance, in “The Light and the Beetle,” “a shadow which he cannot run away from” was changed to “a shadow from which he cannot run.”

As reading poetry is so personal and idiosyncratic, to rule out a possible confound of content (Many, 1991; Purves, 1975) participants were given these five poems both in the pretest and the posttest. Interpreting poetry is somewhat different from other tasks in that there is not a correct or incorrect answer. Indeed, Schmidt’s (1989) “polyvalence” convention suggests that the same reader may provide different interpretations of the same poem at different times. As a check on the reliability of any result from the omnibus test of pretest and posttest skill in interpreting the five selected poems, we provided a sixth poem on the posttest that had not been seen by any of the students before. This poem, called “Poem (As the Cat)” by William Carlos Williams (1998), served as somewhat of a “litmus test” as it was rated by the members of our research group as more difficult to interpret, and having less salient symbolic content, than the other five poems. Like the other five poems it is written in simple language and can be interpreted literally, but also provides an opportunity for a variety of complex, creative and personal interpretations.

Measures

The symbolic interpretation task. On both the pretest and posttest the poems were presented in a booklet form, each on a separate page with the question below it, “What do you think this poem is about?” (Eva-Wood, 2004a, 2004b; Svensson, 1987). Space was provided below the poem for the students to respond.

Personal response ratings. On the page immediately after presentation of each poem students were asked to rate their personal response to it. The poem was presented again at the top of the page followed by three 1-10 rating scale items adapted from Levorato and Nemesio (2005) and similar to those used in the Poetry Reception Questionnaire (Hilscher & Cupchik, 2005; Nemesio, Levorato & Ronconi, 2006). In the first item, participants rated their enjoyment of the poem on a number line ranging from 0 to 10, with two anchors: “I did not enjoy reading this” as 0, and “I greatly enjoyed reading this” as 10. In the second item, they rated their emotions from 0, “I felt no emotion,” to 10, “I felt a lot of emotion.” The third item assessed the poem’s perceived difficulty, with “I found this very easy” as 0 and “I found this very difficult” as 10. In addition, on the pretest and posttest, students were required to rate how much they enjoyed reading poetry in general.

Computational skills measure.

In order to control for the central computational processes - cognitive decoupling, working memory, and executive functioning - participants completed the Verbal Reasoning subtest of the Canadian Test of Cognitive Skills (CTCS) (Canadian Test Centre, Inc., 1992). Cognitive decoupling and the computation of possible scenarios with the decoupled representations is the core ability measured by cognitive ability tests (Stanovich, 2009). Readers encountering symbolism must imaginatively consider the less salient, symbolic interpretation (i.e., cognitively decouple), holding both representational products in mind and integrating them, a high cognitive-load process involving executive functions and carried out in working memory. There are substantial correlations between cognitive ability tests and working memory (Kane, Hambrick & Conway, 2005) as well as between cognitive ability and executive functioning (Baddeley, 1992; Kane & Engle, 2002). The Verbal Reasoning subtest therefore was used as a proxy to both control for general computational skills as well as to estimate the likely amount of variance in posttest symbolic interpretation due to such computational skills. The disattenuated correlation between the Verbal Reasoning subtest and the full battery CTCS is .95 for the test for 6th graders, 1.00 for the test for 9th graders, and .98 for the 12th graders’ test.

Measures of prior equivalence of lesson groups. To assess equivalence of the control and experimental groups prior to the intervention, in addition to the Symbolic interpretation task and Personal response ratings, students provided their age, gender and the English grade received on their previous report card. The self-report method for the English grades was selected because the school board did not allow the school to provide the information directly to the researchers. Although such self-reports may contain inaccurate information, there was no reason to believe that either lesson group would be more likely than the other to misreport their grade.

Design and Procedure

Students in both the experimental and control groups in each of grades 6, 9, and 12 participated in two sessions of testing (pretest and posttest) with a unit on poetry in between, which took approximately two-and-a-half weeks. A teacher from another school carried out the intervention and taught both the experimental and the control group lessons in all three grades. The teacher was chosen because he was considered to be an excellent teacher by his colleagues and had experience teaching English to a wide span of age groups.

The intervention. The aim of this study was not primarily to examine how to teach symbolic interpretation, but rather an attempt to investigate whether one can successfully teach symbolic interpretation explicitly to younger adolescents so as to enhance their ability to explore the imaginative, symbolic possibilities when reading poetry. The symbolism unit was developed by an interdisciplinary three-person team: A professor of developmental psychology, a doctoral student in developmental cognitive science who had previously taught English and French literature, and a highly regarded and experienced English teacher. The unit was based on empirical evidence from developmental science regarding the importance of concrete and visual representations in instruction (e.g., Gerlic & Jausovec, 1999; Marzano, Pickering & Pollock, 2005; Paivio, 1991; Schwartz & Fischer, 2004) and consisted of three concrete scaffolds designed to provide domain-specific knowledge of the process of symbolic interpretation.

Lesson plans were developed in detail for each lesson within the unit, one set for the control and one for the experimental groups in grades 9 and 12, and another two sets for the 6th graders. There was no reference to any of the test poems during the lessons for either the control or experimental groups. With regard to the two 6th grade classes, efforts were made to keep their lessons as similar as possible to those of the 9th and 12th graders, but a few of the poems were replaced with ones that had more simple vocabulary and more age-relevant content. One of the grade 6 teachers was consulted regarding the suitability of the material. We did not have permission to videotape or audiotape the classes and fidelity of implementation depended on the following: Explicit, detailed lesson plans that the teacher tried out on various groups of students prior to the intervention. An understanding that if any lesson plan were not completed during the intervention the class would be excluded from the study. When this happened with one grade 9 class, another grade 9 class was substituted, given the pretest, and the unit successfully implemented. Finally, as we were concerned about implementation in the Grade 6 classes (as the teacher had experience only from Grade 7 upwards), one researcher observed the teacher in both the Grade 6 experimental and control groups and confirmed that there was adherence to the lesson plans.

A detailed, descriptive report of the three scaffolds and intervention for implementation by teachers can be found in Peskin, Allen, & Wells-Jopling (2010) and a shorter description will be provided below. First, the teacher and each class jointly built up a visual image - called the Cycle-Wheel - of the most frequently recurring metaphors in literary works, what Frye (1994) called the “grammar” of symbols. It demonstrated the cyclical nature of e.g., the seasons as well as the polar phases and oppositions, such as summer and winter or creation and destruction (Friedman, 1972). The second concrete representation, a Venn diagram, was then introduced to make the similarities and differences between the source and target elements visually salient and concrete. For instance, in a poem that used a quilt to symbolize the patchwork nature of a family, students used the overlapping part of the Venn diagram to generate ideas about the commonalities between the source (quilt) and target (family), such as “warmth when cold”; “stitched together”; “may unravel”; “united but separated”; “common goal”; “parts may clash”; “parts may complement one another,” and so on. After working with a few poems focusing on the visual image of Frye’s “grammar” of symbols as well as the possible Venn diagrams that could aid symbolic interpretation, students were introduced to the third concrete representation: two-line image metaphors beginning with Ezra Pound’s (1998) famous couplet-poem “In a station of the metro” in which the first line consists of the target (“The apparition of these faces in a crowd”) and the second line involves symbolic elements to describe these faces emerging from the dark subway (“Petals on a wet black bough”), which is the source component. Other couplets were provided and discussed, e.g. An acrobat hanging from a tightrope/A spider dangling from his thread. Students then wrote a few of their own two-line image metaphors. Later, working in small groups, the students physically separated the two lines of their compositions, scrambled them, and their group’s pile of scrambled strips was given to another group to re-align into couplets. As each group worked on re-aligning their given set of scrambled couplets, discussions of the mapping of source and target domains ensued.

The control group lessons were based on the lessons that this English teacher usually uses to teach his classes but an attempt was made to employ the same instructional strategies as in the lessons for the experimental group (i.e., emotional response, large and small group discussion, writing of short poems, short poem developed into longer poem): In both the experimental and control groups, there was an initial appeal to students’ emotions and an exploration of personal reactions. In the experimental group this was tied to the introduction of the first scaffold, the Cycle Wheel (i.e., emotions were described metaphorically in terms of feeling depressed "downward direction” or happy “upward”) and the notion of loneliness in nature was explored in the poem, “leaf loneliness” by e. e. cummings. This poem was then used to introduce the second scaffold, the Venn diagrams. In the control group the introduction to the unit involved discussion of poetry as expressing emotional responses to an interior or exterior landscape. Beginning with an exterior landscape, students’ personal responses to nature and emotive responses of wonder were explored in Raymond Souster’s (1965) poem, "On Georgian Bay" with discussion of poetic techniques such as word sounds and onomatopoeia; repetition of words/lines; progressions in time/distance/event/ intensity; allusions; arrangement of ideas; word choice; and rhyme scheme.

In both the experimental and control groups the homework assignment was couplet writing. In the experimental group these were three image metaphor couplets (i.e. for the third scaffold). In the control group the students wrote a paragraph and then rewrote the concluding sentence as a couplet. In both the experimental and control groups there were then small group discussions about the homework couplets written by the group members. In both groups a few examples were chosen and they were then switched with another group. Later both groups developed the couplets into a longer poem.

In the rest of the unit, in both groups five to six more poems were discussed: In the experimental group these poems were selected and explored within the framework of the three concrete scaffolds for symbolic thinking, i.e. the teacher made frequent reference to the Cycle-Wheel”; used Venn diagrams to aid symbolic interpretations; and made reference to the alignments that had been made explicit by the scrambled couplets.

In the control group, the poems were selected and explored within the framework of how the poet uses poetic techniques to appeal to our emotions and imagination. For instance, in teaching Shakespeare’s famous sonnet, “When in Disgrace with Fortune and Men’s eyes,” to the 9th and 12th grade control groups, the lesson plan stated: “A. Intro: 1. Begin with statement that we are shifting from an external landscape to an internal one. 2. Ask for feedback from students on what causes them to feel depressed, and what brings them out of that feeling - connect to the poem… B. Method: 1. Define words that need explaining… 2. Find repetition – e.g., ‘like to him’, ‘like him’ - purpose of this repetition? (emphasize isolation, sense of inferiority); 3. Progressions - from depressed to joyful…4. Allusions…; 5. Arrangement of ideas….; 6. Word choices - elevate emotional feeling: ‘disgrace’.. ‘deaf heaven,’ ‘bootless cries,’ ‘sullen earth,’...; 7. Rhyme Scheme and Metre… iambic pentameter rhythm matches the rhythmic beating of our own hearts - e.g. ‘I all alone be weep my out cast state’ - / - / - / - / - / .” Although symbolic interpretation was not explicitly addressed in the control group, it must be noted that, as symbolic language is the basis of poetic expression (Frye, 1978), symbolic interpretation is frequently implicit in discussions about poetry. As can be seen in the above example, discussions of many of the word choices clearly involved references to metaphoric language, e.g. ‘deaf heaven,’ ‘bootless cries,’ or ‘sullen earth.’ However, the control groups were not provided with the concrete scaffolds, e.g., a Venn diagram or image couplet to explicitly compare a person’s joy after depression to a ‘lark at break of day arising/From sullen earth,’ and the teacher emphasized the emotion generated from such comparisons and word choices, such as “sullen earth,” but did not explicitly analyze the metaphors as metaphors.

Coding of Symbolic Interpretations

For each poem, the response to the question: “What do you think this poem is about?” was coded into one of five categories adapted from Svensson (1987): “literal descriptive,” “literal interpretive,” “thematic” (which included mixed literal-thematic), “mixed literal-symbolic,” and “symbolic.” As readers construct meaning through the lens of their personal experience and imagination (Langer, 1993; Rosenblatt, 1978), interpretations coded as symbolic differed widely from each other. Descriptions of each coding category and examples of interpretive responses for each of the five categories have been provided in Appendix B. For ease of comprehension the examples have been taken from two of the texts, a short poem, “It is this persistent wind,” (See Appendix A) and the longest poem, Stephen Spender’s poem “What I expected” (both of which have been described earlier).

Coders were blind as to whether a subject was in the control or experimental group. After the coding scheme had been established, one researcher coded a third of the total number of poems. A second researcher scored half of these coded poems. Discrepancies between the two coders were resolved through mutual discussion. The first researcher then coded the remaining poems accordingly, and the second researcher coded a third of these. At this stage, agreement was more than 86%. All remaining discrepancies were again resolved through discussion.

Scoring of Symbolic Interpretations

As the focus was on whether the younger students in particular would be able to provide some form of symbolic interpretation, even if there was vacillation between a literal and symbolic approach, we divided the categories so as to create a dichotomous score: The “mixed literal-symbolic” category as well as the fully “symbolic” category were combined and coded as “symbolic,” and the four remaining categories, in which there was no demonstration of symbolic thinking, were coded as “non-symbolic.” Students therefore received either 0 (“non-symbolic” response) or 1 (“symbolic response”) for each of the five poems in the pretest and each of the six poems in the posttest. The dependent variable when testing for the effect of the intervention on symbolic processing was the mean number of poems read symbolically out of the total of six poems on the posttest (range 0 – 6). The mean of the five poems in the pretest (range 0 – 6) was a used as a covariate and also used in preliminary analyses.

Results

The results are presented in two sections. First, preliminary analyses, that is those which involve the pretest data: 1. Grade-level differences in symbolic interpretation of poetry prior to the intervention and 2. Equivalence of the two poetry-lesson groups on known or suspected covariates of symbolic processing. The second section presents the results from the actual intervention, i.e. the effect of the intervention on 1. Symbolic processing and 2. Personal response. As five omnibus analyses were planned, the acceptable familywise α for statistical testing was reduced to .02 per test, using a simple Bonferroni adjustment (i.e., .10/5) as suggested by Tabachnick & Fidell (2007).

Preliminary Analyses

1. Pretest Grade-level differences in Symbolic Interpretation of Poetry through Adolescence

Table 2 provides the percentage of responses in each category of Svensson’s coding scheme at each age level on the pretest. Symbolic (that is, mixed literal-symbolic or symbolic) responses were in evidence in 19% of 6th grade responses, 46% of 9th grade responses, and 71% of 12th grade responses.

--------------------------------------Insert Table 2 here-------------------------------------

The three grades were compared in terms of mean number of poems read symbolically out of the total of five poems on the pretest. A one-way analysis of covariance was conducted in which the Pretest Mean Number of Poems Read Symbolically was the dependent variable. The independent variable was grade with three levels: 6, 9 and 12. The Verbal Reasoning subtest of the CTCS was included as a covariate in order to (1) control for pre-existing differences in computational capacities, and (2) to statistically test the magnitude of the unique contribution of these capacities to skill in the symbolic interpretation of poetry. A Method 1 analysis (Overall & Spiegel, 1969) was employed, as it both allows testing of the unique contribution of the covariate to the dependent variable, after all other factors and interactions have been accounted for (Tabachnik & Fidell, 2007, p. 212), and is recommended in designs in which sample sizes are not equal. An initial analysis evaluating the homogeneity of slopes assumption indicated that the relationship between the covariate and pretest symbolic interpretation did not differ significantly as a function of grade. The ANCOVA was significant, F (2, 133) = 34.31, MSE = 1.70, p ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download