Sie sind auf Seite 1von 80

L(a

2006 3 18

ii

A Pragma-Cognitive Approach to Visual Poetry: A Case

Study of L(a By Ren Yuxin

Under the Supervision of Professor Chen Xinren

Submitted in Partial Fulfillment of the Requirements

For the Degree of Master of Arts

English Department School of Foreign Studies Nanjing University March, 2006

ii

I hereby declare that this submission is my own work and that, to the best of my knowledge and belief, it contains no materials previously published or written by another person or material which has to a substantial extent been accepted for the award of any other degree or diploma at any university or other institute of higher learning, except where due acknowledgement has been made in the text.

Signature: Name: Ren Yuxin Date: March, 2006

ACKNOWLEDGEMENTS

I would like to take this opportunity to express my gratitude to those who have offered me all kinds of help in finishing this thesis. First, my heartfelt gratitude is due to my supervisor, Professor Chen Xinren, who has offered me constant support, kind encouragement and expert advice through all the stages of development and revision of this thesis. He impressed me not only as a qualified teacher but also as a man of wisdom, devotion and patience. Without his patient guidance and kind encouragement, the study would not have taken the present shape. I am also deeply obliged to Professor Zhang Ren, who not only offered me a lot of insightful suggestions for this thesis but also took great pains to read the second draft of the thesis, and from whom I often ask questions about cognitive linguistics. I am also grateful to Professor Wen Qiufang, who acquainted me with research methods in applied linguistics, and Professor Ting Yenren, who familiarized me with the basic theories in applied linguistics and discourse analysis. My great thanks also go to my classmates, Nie Yuefang, Wang Xueyu, Wu Jue, Chen Hairong, Chen Ying, He Fang, Huang Niya, and Wang Yuanfei for their critical and valuable suggestions for this thesis during the Thesis Seminar. Furthermore, my special thanks also go to my good friends as well as my roommates, Jia Guangmao, with whom I often discussed the problems I encountered in applied linguistics and in writing this thesis, He Huiming, who always helped me solve the computer problems, and Feng Dong, with whom I discussed some statistical problems in the pilot studies of this thesis and to whom I often raise questions about literature. Finally, I also want to thank my good friend, Li Min, for his encouragement and help in finishing this thesis. My hearty appreciation also goes to those students who participated in this study.

THESIS: A Pragma-Cognitive Approach to Visual Poetry: A Case Study of L(a


i

SPECIALIZATION: English Language and Literature POSTGRADUATE: Ren Yuxin Professor Chen Xinren MENTOR:

Visual poetry has drawn peoples attention for a long time. However, their interests in it have been focusing on the interpretation rather than the exploration of the mechanisms of how this special kind of literary discourse works. Since poetic effects are the central value of a poem, probing into the poetic effects communicated by a visual poem will help people know the mechanisms on which visual poetry works as well as facilitate peoples understanding of the special discourse. This study aims to investigate how visual poetry achieves poetic effects, from a pragma-cognitive point of view, by exploring one of E. E. Cummings poems entitled L(a as a typical case. The study distinguishes itself from previous studies in two ways: first, L(a, which was written in the vertical style, is analyzed within the framework of the relevance theory to identify what elements in L(a determine its rich poetic effects and in what way L(a achieves its poetic effects. Second, a questionnaire is designed to validate the theoretical analysis and two groups of students with great differences in their background knowledge about English literature are chosen as subjects. Comparisons are conducted between the vertical style and the reproduced horizontal style of L(a (i.e. A leaf falls: / loneliness.) in theoretical terms and supported by an empirical test. The analysis of the data from the questionnaire yields the following findings: (1) The poetic effects of a poem are fundamentally determined by the number of ostensive points it contains. The vertical style, L(a, containing more ostensive points, communicates more poetic effects than its reproduced horizontal style does. (2) Readers cognitive environment, which refers to the set of all the facts that they can perceive or infer, plays a very important role in their interpretation of poetry. Due to different cognitive environments, different readers perceive different numbers of ostensive points and consequently derive different numbers and types of weak implicatures from the same poem.

ii

(3) In interpreting visual poetry, it is the specific assumptions about visual poetry but not the background knowledge of literature that determine whether readers can arrive at a successful reading. (4) The difficulty of visual poetry is recognized by most readers, who are nevertheless tolerant of such difficulties in their interpretation. Theoretically, this study probes into the nature of visual poetry by analyzing L(a. Compared with the traditional linear arrangement in the printed format of poetry, the vertical format of a visual poem may greatly enlarge the number of its ostensive points, which may activate readers to derive more weak implicatures, which in turn lead to richer poetic effects. In this light, the study will facilitate the readers understanding of visual poetry and provide people with a pragma-cognitive model to interpret such discourses as well. The empirical part of the study also serves as an indirect test of the relevance theory. Methodologically, the present study exemplifies how theoretical analysis can be combined with empirical research in literary pragmatics. In addition, it has some pedagogical implications for poetry teaching with respect to the readers interpretation process.

L(a 2003

iii

Cummings L(a L(a 1 L(a 2 3 4 L(a

iv

Table of Contents

Acknowledgements.............................................................................................................i Abstract in English.............................................................................................................ii Abstract in Chinese...........................................................................................................iv Tables of Contents.............................................................................................................vi List of Tables.....................................................................................................................ix Chapter One Introduction..........................................................................................1 1.1 Object of the Study...............................................................................................1 1.2 Significance of the Present Study.........................................................................2 1.3 Structure of the Thesis..........................................................................................4 Chapter Two Literature Review................................................................................5 2.1 Definition and Explanation of Terms...................................................................5 2.1.1 Visual Poetry................................................................................................5 2.1.2 Ostension.....................................................................................................6 2.1.3 Weak Implicatures and Poetic Effects.........................................................7 2.1.4 Ostensive Points........................................................................................10 2.1.5 Cognitive Environment..............................................................................11 2.2 Pragmatic Studies of Poetry................................................................................12 2.2.1 Non-relevance-theoretic Studies of Poetry...............................................12 2.2.2 RT Based Analysis of Poetry and Poetic Effects......................................14 2.3 Empirical Studies of Poetry................................................................................17 2.3.1 Empirical Studies of Poetry.......................................................................17 2.3.2 Studies on Visual Poetry............................................................................18 2.4 Studies on L(a.....................................................................................................18 2.5 Summary.............................................................................................................20 Chapter Three The Conceptual Framework.......................................................22 3.1 Relevance theory..................................................................................................22 3.1.1 Ostensive-inferential Communication...........................................................22

vi

3.1.2 Two Principles of Relevance......................................................................24 3.1.3 Effects and Effort........................................................................................26 3.1.4 Style and Weak Implicatures: Poetic Effects.............................................28 3.2 The Theoretical Analysis of L(a Within the RT Framework...............................29 3.2.1 L(a Viewed in the RT Framework..............................................................30 3.2.2 RT Based Analysis of L(a...........................................................................31 3.2.2.1 L(a as Ostensive-inferential Communication...............................32 3.2.2.2 Influence of Cognitive Environment on Deriving Weak Implicatures.........................................................................36 3.2.3 RT Based Comparison Between the Vertical Style and the Horizontal Style............................................................................37 3.3 Summary..............................................................................................................39 Chapter Four Methodology.....................................................................................41 4.1 Research Questions.............................................................................................41 4.2 Subjects...............................................................................................................42 4.3 Instrument...........................................................................................................42 4.4 Data Collection...................................................................................................43 4.5 Data Analysis......................................................................................................44 Chapter Five Results and Discussion.................................................................45 5.1 Ostensive Points and Poetic Effects...................................................................45 5.2 Influence of Readers Cognitive Environments on Their Perception of Ostensive Points..................................................................................................49 5.3 Readers Attitudes Towards the Two Poetic Styles............................................52 Chapter Six Conclusion............................................................................................55 6.1 Summary of the Study........................................................................................55 6.2 Major Findings....................................................................................................56 6.3 Implications of the Study....................................................................................57 6.4 Limitations of the Study and Suggestions for Future Research.........................58 References....................................................................................................................60

vii

Appendix.......................................................................................................................64 Questionnaire.............................................................................................................64

List of Tables

viii

Table 5-1: Comparisons between the numbers of the ostensive points perceived from the two poetic styles by the readers in the same group........................46 Table 5-2: Comparisons between the numbers of the ostensive points perceived from the same poetic style by the two different groups of readers...............49 Table 5-3: Readers assumptions of visual poetry...........................................................50 Table 5-4: Comparisons between the numbers of the ostensive points perceived by the readers who have read some visual poems and those who have not in both groups.........................................................................51 Table 5-5: Readers attitudes towards the two poetic styles............................................53

ix

Chapter One Introduction

This chapter is organized into three sections. The first section introduces the object of the present study. It is followed by presenting the significance of carrying out this study. The last section describes how the thesis is structured.

1.1 Object of the Study


The question to be addressed in the present study concerns how visual poetry achieves poetic effects. The probe will be conducted from a pragma-cognitive perspective. Visual poetry, as Gross (1997, p. 16) points out, is presented as spatial (rather than linear) arrangements in order to activate the visual composition on the page as an element of signification. Visual poetry has drawn many readers attention for a long time. However, only very few scholars have explored the mechanisms of how this special literary discourse works. The present study attempts to use one of E. E. Cummingss poems entitled L(a (presented on the next page) as a typical case to explore how visual poetry achieves poetic effects. L(a, written in 1958, is the first poem of a 95-poem collection composed by Cummings. It is considered by literary critics and scholars as one of the representatives of visual poetry.
1

The following is L(a: l(a le af fa ll s) one l iness Since this poem was published, many readers have been working at its interpretation. Of thousands of readers, Kidder (1979) is the most prominent one, who gave a detailed interpretation of the poem, attempting to figure out the symbolic meaning of the elements contained in the poem from the perspective of literary text analysis. Following Kidders interpretation, this poem has been interpreted from many other perspectives such as stylistics, aesthetics and cognitive linguistics (e.g. Duan, 2002; Hiraga, 2005; Ye, 2003; B. H. Zhang, 2001; Zhao, 1994) as will be reviewed in 2.4. However, almost all of these readers as well as scholars pay much attention to the interpretation rather than the exploration of the mechanisms of how L(a works. No one has probed into how L(a achieves its poetic effects, which are considered to be the central value of poetry. Therefore, the purpose of this study is to explore how L(a achieves its poetic effects within the framework of relevance theory (RT henceforth) from both theoretical and empirical standpoints. The empirical part of the study is used to validate the theoretical analysis.

1.2 Significance of the Study


The present study distinguishes itself in two significant ways. First, it employs a pragmatic theory to analyze a literary discourse, L(a. Second, it exploits an empirical
2

method to complement the theoretical analysis. Hanauer (1998, p. 566) argues that the empirical method, as a very important research tool, can be used in literary studies to deepen our understanding of literary issues. This tool has been used by many scholars, such as Richards (1929), Goodblatt (2001), Yaron (2002, 2003), Gibbs, Jr. (2002), Gibbs, Jr. and Nascimento (1996), Gross (1997), Barney (1996), Steen and Schram (1996) and Hanauer (1998) in experiments that examine literary discourses, especially poetry. On the other hand, some scholars have also exploited pragmatic theories to interpret literary discourses, thus giving rise to a new subfield of study, that is, literary pragmatics. Representative figures, who have conducted pioneer studies in this field, include Levin (1976), Verdonk (1991), Pilkington (1991, 2000), Sell (1991), van Peer (1991), Tu (2002) and Y. Zhang (2001). Since pragmatics deals with meaning in language use and reading literary works is one kind of communication from the communicative perspective, combining the exploration of literary study with pragmatic theories will provide people with a new perspective to understand literary works on the one hand, and enrich the resources for pragmatic research on the other. The present study is thus conducted in the scope of literary pragmatics. However, almost all previous studies focused either on using empirical methods to examine literary discourses or on using pragmatic theories to interpret literary discourses. The present study is unique in combining them in a single attempt. In addition, the present study focuses on visual poetry. The analysis of L(a from a pragma-cognitive point of view will probe into the nature of such discourses, thus helping readers interpret the poetic genre, know more about the mechanisms on which visual poetry works and finally facilitate readers understanding of visual poetry. Meanwhile, through theoretical analysis and empirical test, the present study will provide evidence for the argument that the existence of visual poetry is reasonable and valuable from a pragma-cognitive point of view. It has been almost twenty years since RT was formally proposed in 1986. The theory has been widely accepted and applied by many scholars. However, previous application of RT is impressionistic, and very few studies have involved experiments within the

framework of RT (Carston & Ucnida, 1998; Rouchota & Jucker, 1998; Yus, 1998; He & Ran, 2001). The present study attempts to carry out an experiment to test the theoretical analysis on the basis of RT. This will serve as an indirect test of RT with a view to promoting our understanding of the theory. Finally, the readers involved in the experiment are students. By examining the process of the readers interpretation, this study may offer some suggestions for literary teaching, especially poetry teaching.

1.3 Structure of the Thesis


This thesis consists of six chapters. Chapter One presents the object, aims and significance of the present study. Chapter Two reviews, explains and defines some terms related to the present study. It also gives a brief look at the related studies previously conducted from different perspectives about poetry, visual poetry and L(a. Chapter Three first presents the conceptual framework of the study, and then analyzes how L(a achieves its poetic effects within the framework. Chapter Four reports the research questions, subjects, instrument, procedures of data collection and data analysis. Chapter Five discusses the results yielded from the empirical study to answer the research questions, which were designed to test the theoretical analysis and to examine the readers attitudes towards different forms of poetry as well. Chapter Six summarizes the major findings, proposes the implications, points out the limitations of the present study, and finally offers some suggestions for future research.

Chapter Two

Literature Review

This chapter is organized into five sections. In the first section, some terms will be defined and explained. In the next section, pragmatic studies of poetry will be briefly reviewed. In the third section, some empirical studies on poetry and some studies on visual poetry are to be introduced. In the fourth section, the previous studies on L(a will be reported. Finally, a very brief summary will be made to point out the direction the present study takes.

2.1 Definition and Explanation of Terms


In this section, some terms closely related to the present study will be defined and explained such as visual poetry, weak implicatures, poetic effects, ostension, ostensive points and cognitive environment. 2.1.1 Visual Poetry Various terms such as pattern poems used by Gross (1997), picture poetry and concrete poetry used by Tsur (2000), all refer to visual poetry. This study refers to visual poetry as those poems that are shaped like a picture by linguistic objects such as

letters, words and sentences in the printed format. That is, visual poetry contains not only linguistic stimuli but also visual stimuli. According to Verdonk (1991), a poem can be regarded as a verbal composition which represents a discourse between the author and the reader. By poetic discourse, he means that the poets text becomes a meaningful discourse only at the time when it is being read (1991, p. 96). In other words, from a communicative perspective, poetry can be treated as one mode of discourse, and a poem can also be called a poetic discourse. Luo (2003) also argues that poetry is one kind of literary discourse and pragmatic purposes is inherent in it. 2.1.2 Ostension Sperber and Wilson (1995, p. 49) define ostensive behavior or simply ostension as behavior which makes manifest an intention to make something manifest. To put it simply, showing someone something is a case of ostension. So an utterance or, in certain context, a gesture, even a body movement such as leaning back or sniffing in Sperber and Wilsons view can be an ostension, through which the communicator intends to make something manifest to the addressee. Consider the following dialogue which is often used in the literature: (1) a. Peter: Do you want some coffee? b. Mary: Coffee would keep me awake. In (1a), Peter as a communicator intends to make some assumptions manifest to the addressee Mary, that is, he thinks Mary may need some coffee, and wants to offer some coffee to Mary. Here, Peter makes his intention manifest to Mary through an ostension, that is, the utterance Do you want some coffee? On the other hand, instead of answering Peters question directly, in (1b), Mary makes an ostension by way of an utterance, Coffee would keep me awake, which will draw Peters attention to infer what has been communicated by Mary through this ostension. In (1), the utterances made by both Peter and Mary are ostensions in terms of RT, which are realized solely by linguistic forms. Imagine Mary says Coffee would keep me

awake with a yawn in (1b). Under the new circumstances, both the utterance and the movement of yawning are considered ways of ostension, which implicitly communicates that Mary does not need any coffee or Mary refuses Peters offer. However, this ostension consists of two kinds of stimuli, that is, the linguistic stimulus as well as the physical stimulus. 2.1.3 Weak Implicatures and Poetic Effects Pilkington (1991, p. 60) points out that the value of a poem lies in its poetic effects which the poet tries to communicate and the reader is expected to take his responsibility to obtain. Our discussion of poetry cannot avoid mentioning poetic effects. Many scholars have discussed poetic effects (e.g. Jacobson, 1960/2000; Sperber & Wilson, 1986/1995; Blakemore, 1992, 1993; Pilkington, 1991, 2000). In Jacobsons (1960/2000) view, poetic function, which corresponds to poetic effects, is one of the six basic functions of verbal communication (i.e. emotive, conative, referential, poetic, phatic and metalingual functions) and is the dominant one in poetry. According to Jacobson, the set (Einstellung) towards the message (i.e. the physical structure of a text) as such, focus on the message for its own sake, is the poetic function of language. This says, as Waugh (as cited in Zwaan, 1993, pp. 7-8) points out, the poetic function of a given message is an intrinsic quality of that message itself; thus the focus of the message is an inherent quality of a poetic discourse. For example, people often say, as Jacobson argues, Joan and Margery, yet never Margery and Joan. The reason is that people take the former but not the latter as a well-ordered shape of the message to achieve certain effect (e.g. rhythm). It is the poetic function of language that leads to poetic effects of utterances. In RT, poetic effects are associated essentially with weak implicatures. The following part of this section deals with weak implicatures and poetic effects proposed by RT. Before we discuss these two terms, that is, weak implicatures and poetic effects, we have to get an idea of another indispensable term, assumption. In RT, an assumption is a structured set of concepts, and the assumptions are different from knowledge in that they need not be true. Communicators communicate certain assumptions to addressees and if

any assumption is implicitly communicated, then it is an implicature. More technically, an implicature is a contextual assumption or implication which a speaker, intending her utterance to be manifestly relevant, manifestly intended to make manifest to the hearer (Sperber & Wilson, 1995, p. 194). The strength of implicatures of an utterance may vary in degree. For instance (Sperber & Wilson, 1995, pp. 194-199):
(2) a. Peter: Would you drive a Mercedes?

b. Mary: I wouldnt drive ANY expensive car. The propositional form of (2b) does not directly answer the question in (2a). However, it gives Peter immediate access to his encyclopedic information about expensive cars, which includes, for example, the information in (3): (3) A Mercedes is an expensive car. If processed in a context containing assumption (3), (2b) would yield the contextual implication (4): (4) Mary wouldnt drive a Mercedes. In addition, on the basis of (2b) it would be reasonable for Peter to add premises (5) and (6) to the context, and derive conclusions (7) and (8): (5) (6) (7) (8) A Rolls Royce is an expensive car. A Cadillac is an expensive car. Mary wouldnt drive a Rolls Royce. Mary wouldnt drive a Cadillac.

Furthermore, imagine that Peter believes (9) and finds it worth his effort to derive (10) from (9) and (2b):
(9) People who would not drive an expensive car would not go on a cruise either. (10) Mary would not go on a cruise.

In these examples, the strongest possible implicatures are those fully determinate premises or conclusions, such as (3) and (4), which must actually be supplied, and for which the speaker takes full responsibility. Strong implicatures are those premises and conclusions, such as (5)-(8), which the hearer is strongly encouraged but not actually

forced to supply. The weaker the encouragement, and the wider the range of possibilities among which the hearer can choose, the weaker the implicatures, such as (9) and (10). In other words, the weak implicatures such as (9) and (10) in the example are marginal assumptions, which the speaker does not encourage the hearer to derive from (2b). However, the marginal assumptions might be activated to be salient in interpreting (2b). Therefore, weak implicatures and strong implicatures are a matter of degree. Based on the notion of weak implicature, in RT, poetic effect is defined as the peculiar effect of an utterance which achieves most of its relevance through a wide array of weak implicatures (Sperber & Wilson, 1995, p. 222). In other words, an utterance achieves its poetic effects mainly through the associations activated by some aspects of its propositional form such as the syntactic, phonological or semantic properties. These properties may activate the hearer to derive a wide range of weak implicatures. Consider the following examples (Sperber & Wilson, 1995, p. 219): (11) My childhood days are gone, gone. (12) My childhood days are gone. First, what is worth discussing is the repetition exploited in (11). Compared with the utterance (12), the repetition in (11) can be considered as one special syntactic property. To justify the repetition of gone, the hearer must think of all the implicatures that the speaker might have expected him to derive from (12) and further assume that there are still some premises and conclusions which the speaker intends to convey. Thus, the hearer must expand his context (i.e. supply more possible assumptions stored in memory) in which he interprets the utterance. Consequently, the hearer can obtain more from this repetition; for example, it is a great pity that the speakers childhood days are gone, which might remind the hearer of his own childhood days. Second, the semantic property of childhood may activate the hearer to explore his encyclopedic entry in relation to the concept of childhood. The hearer may access such assumptions as the following (Pilkington, 2000, p. 126): (13) a. Childhood days are innocent. b. Childhood days are ingenuous. c. Childhood days are carefree.
9

Based on these contextual assumptions, the hearer may derive the following weak implicatures: (14) a. The speakers days of innocence are gone. b. The speakers days of ingenuousness are gone. c. The speakers carefree days are gone. In sum, poetic effects are achieved through the derivation of a wide range of weak implicatures in pursuit of relevance. Pilkington (2000, p. 169) points out: Poetic effects result from a special kind of mental processing a wide-ranging activation and accessing of contextual assumptions, triggered by the search for an interpretation consistent with the principle of relevance. In addition, the derivation of weak implicatures is dependent on ostension. In other words, the number and the type of yielded weak implicatures are determined by ostension. Some ostensions contain more ostensive points, which may result in more weak implicatures. Different ostensions, even including those which share the same propositional form, may yield different numbers and types of weak implicatures, which in turn affect their poetic effects. Thus, poetic effects are a matter of degree. The wider the range of potential implicatures and the hearers responsibility for deriving them, the more poetic effects there will be. 2.1.4 Ostensive Points For the purpose of our quantitative study, the present study defines those points contained in one ostension as ostensive points, from which the addressee can derive implicatures. In other words, ostensive points are those points embedded in an ostension that prompt the addressee to explore encyclopedic entries thoroughly and exploit acceptable assumptions to derive a wide range of possible implicatures. For example, in (11), the utterance My childhood days are gone, gone as a whole is an ostension; the propositional form of the utterance (i.e. My childhood days are gone) and the repetition of gone are two specific ostensive points, which may activate the hearer to explore more encyclopedic entries with a greater effort and supply more contextual assumptions to derive a wide range of weak implicatures, with the salient ones listed in (14). Take L(a as another example. L(a as a whole is an ostension which contain many
10

ostensive points such as a, l, af / fa, one, ll and iness. These ostensive points may activate the reader to derive a wide range of possible implicatures through the pictures they shape and the propositional forms or the concepts they encode, some weak and some strong. However, for L(a, most implicatures derived from its ostensive points are weak implicatures because they may not be strongly communicated. The poet may or may not have intended to convey some information through them, but the reader can take his responsibility to derive a wide range of acceptable implicatures from them in interpretation. For instance, ll has two ones as Kidder (1979) interpreted and it can be associated with one of the salient themes loneliness. This may be intentionally communicated by the poet. In addition, ll in Kidders eyes resembles an upended equal mark, through which the reader can imagine that ones life has half passed if he associates the five stanzas contained in L(a with ones life. This may be weakly communicated or may not be intentionally communicated by the poet. In spite of the poets unintentional communication, the reader can take his responsibility to derive the above acceptable implicature. ll, being one of the ostensive points contained in L(a, activates the reader to derive possible implicatures and thus makes contribution to the poetic effects of L(a. Simply put, ostensive points are the components of an ostension. They include not only the propositional form of an utterance as the above discussion on (11) shows, but also some concepts conveyed by linguistic stimulus or visual stimulus which the communicator may not very consciously exploit to convey her intention but from which the addressee can take his responsibility to derive possible implicatures, as the discussion on L(a in this section indicates. 2.1.5 Cognitive Environment Sperber and Wilson (1995, p. 39) define cognitive environment as follows: A cognitive environment of an individual is a set of facts that are manifest to him. According to Sperber and Wilson, to be manifest means to be perceptible or inferable. An individuals total cognitive environment is the set of all the facts that he can perceive or infer.

11

They (1995, p. 41) further argue that since individuals physical environments are never identical and their cognitive abilities are influenced by previously memorized information, they will never share their total cognitive environments, which differ in many respects from one person to another. In addition, an individuals cognitive environment is a set of assumptions available to him and which assumptions in his cognitive environment he is likely to construct and process is consistent with the principle of relevance. The readers cognitive environments play an important role in literary communication. The individual readers cognitive environments will decide whether they can obtain a successful interpretation of the literary discourses they are working on.

2.2 Pragmatic Studies of Poetry


2.2.1 Non-relevance-theoretic Studies of Poetry Since a poem is composed of linguistic symbols, two questions arise. First, what kind of linguistic act (i.e. speech act) is performed in the production of a poem? Second, what kind of linguistic objects are employed in a poem? Let us deal with the two questions one by one. Speech act theorists may answer what kind of linguistic act is performed in the production of a poem. Inspired by the theory of speech acts proposed by Austin, Levin (1976) points out that poetry performs a special speech act. He makes his investigation on the basis of Austins speech act theory, holding that in poems, the implicit higher sentence is I imagine myself in and invite you to conceive a world in which, which expresses the kind of illocutionary force that the poem is to be taken as having. That is to say, every poetic discourse, even every literary discourse, communicates such illocutionary force implicitly to readers when it is read. For I imagine myself in (a world), Levin argues that the illocutionary force of this utterance lies in that the poet projects himself into a world that he imagines, in which he is free to make as different from the actual world as he pleases. For invite you to conceive a world, Levin argues that the invitation to conceive a world has indeed been extended and the reader should contemplate a world different from

12

the actual one, that is, a world that the poet imagines and that can be discovered only from the poets account of it. This may be the first time for a pragmatic theory to be used to facilitate the understanding of poetry. However, the speech act theory itself has some faults in interpreting poetry. As Cook (1999, pp. 44-45) points out, speech act theory encounters some problems when applied to literature. First, when a verb related to an illocutionary act is not used in its expression, the inference of the illocution depends upon the writers correct assessment of the readers knowledge. Yet literary discourses often have, along with other written texts for unspecified readers, a degree of uncertainty about the knowledge of the readers. Second, since literary discourses, including poetry, lack an overt purpose, it seems likely that that there is often no perlocutionary force at all in literary communication. In addition, Levins exploration does not provide a procedural account of how the poet makes his imagining manifest to the reader. With regard to the second question, many scholars, especially stylists, who deal with linguistic objects in poetry, notably Fowler (1975) and Leech (1969), have done a lot of exploration. Despite their great contributions, our review will only focus on those studies conducted in the scope of pragmatics. Verdonk (1991, p. 94) introduced three stages of the writing of a poem as suggested by Philip Larkin: The first is when a man becomes obsessed with an emotional concept to such a degree that he is compelled to do something about it; what he does is the second stage, namely, construct a verbal device that will reproduce this emotional concept in anyone who cares to read it, anywhere, any time; the third stage is the recurrent situation of people in different times and places setting off the device and re-creating in themselves what the poet felt when he wrote it. A verbal device constructed by the poet in the second stage to reproduce his emotion is a special kind of linguistic object. Verdonk argues that such kind of linguistic objects, including textual devices like deviation and parallelism, which are often applied by the poet to write and the reader to interpret in literary communication, should receive attention. According to Verdonk (1991, p. 98), deviation refers to a writer conscious or unconscious violation of some linguistic rule of phonology, morphology, syntax, or
13

semantics, or an infringement of a particular linguistic usage or of some literary genre or convention. Parallelism is based on extra regularities, not irregularities, in poetry and it is usually achieved by repetitive patterns on any of the already indicated levels of language organization. Verdonk further argues that deviation and parallelism in linguistic poetics are usually subsumed under the concept of foregrounding, which is a dynamic concept, and which seeks to provide a link between specific textual features on the one hand and aspects of text production as well as of text reception on the other. Foregrounding devices in literary discourses is therefore pragmatic by nature. Though Verdonk suggests, from a pragmatic point of view, that some aspects such as deviation, parallelism and foregrounding are worthy of attention in interpreting poetic discourses, he does not explain how these aspects interact in the process of interpretation. 2.2.2 RT Based Analysis of Poetry and Poetic Effects Pilkington (1991, 2000) explores poetic effects from a RT perspective. Claiming the need for a theoretical literary pragmatics within literary studies, to describe and explain the communication of poetic effects, Pilkington discusses formalism, semiotics and structuralism, and conventionalism, and then points out the faults of these approaches; for example, they inevitably fail to provide realistic psychological accounts of the reading process (1991, p. 46). As a result, he adopts RT to explore poetic effects because he thinks it can describe and explain communication more adequately than previous semiotic models. Within RT, Pilkington argues that such aspects as context, poetic metaphors and indeterminateness of poetry should invite attention. He (1991, pp. 59-60) goes on to argue that it is impossible for a poetic discourse to offer a set of determinate alternative meanings to choose from. Images, symbols and metaphors within the poetic discourse interact to yield a wide range of weak implicatures. Pilkington (2000) develops the basic idea of his argument in 1991 and has continued to hold the firm belief that it is necessary to provide a psychological and cognitive account of literary discourses. In his view, literariness is a universal form of aesthetic experience, and aesthetic response involves mental representations that occur when a literary discourse is read, or on a smaller scale, when a rhetorical device is used to create

14

poetic effects. Commenting on Pilkingtons arguments, Mackenzie (2002, p. 199) points out that literariness refers to a very special kind of mental activity. When the reader approaches a literary works (e.g. a poetic discourse), he is activated to identify the writers intention in line with the principle of relevance. The extensive exploration of encyclopedic entries (i.e. information stored in our brains about concepts, and objects, events or properties which instantiate the concepts) will be involved in this process. Due to the high indeterminateness of the poetic discourse, a lot of marginal assumptions stored in the readers memory would become salient when the reader takes his responsibility to interpret the poetic discourse. Pilkington (2000) explores how metaphors in utterances contribute to their poetic effects from a pragma-cognitive point of view. Consider the following example from Pilkington (2000, p. 101), which has also been used by Sperber and Wilson (1995) in their discussion on metaphor. The remark is made by Flaubert of the poet Leconte de Lisle: (15) Son encre est pale His ink is pale. In his analysis, the contextual assumptions are supplied by exploring the encyclopedic entries attached to the concepts ink and pale. This utterance includes an instance of metonymy, as well as one of metaphor, with ink standing for writing, in the sense of work rather than handwriting. So the reader or hearer may derive the assumption: (16) Leconte de Lisles writing is pale. After searching through the encyclopedic entry attached to the concept pale, the reader or hearer may derive implicatures for the whole utterance such as: (17) Leconte de Lisles writing is weak. By exploring encyclopedic entries more thoroughly and expanding context in this way, more possible implicatures can be derived, as Pilkington lists: (18) a. Leconte de Lisles writing lacks contrast. b. Leconte de Lisles writing may fade.
15

c. Leconte de Lisles writing is sickly. d. Leconte de Lisles writing will not last. e. Leconte de Lisles does not put his whole heart into his work. That this utterance achieves a wide range of weak implicatures mainly depends on a thorough exploration of the encyclopedic entries attached to the concepts ink and pale, and metaphors in this utterance thus achieve their poetic effects. In addition, Pilkington (2000, p. 100) points out that the metaphors that are relatively conventional and the metaphors that are relatively creative vary in terms of the range and strength of the assumptions that they make more salient. If these assumptions are considered as implicatures, the richer and more creative the metaphor is, the wider the range of weak implicatures will be. Pilkington (1991, 2000) re-interpreted traditional literary theories which were divided as to where the meaning of a literary discourse lies (e.g. meaning is only the authors intended interpretation, or the objective meaning of the discourse itself, or the readers personal interpretation of it), thus facilitating the present exploration of poetic effects from the RT perspective. However, Pilkington (2000) focuses on small-scale poetic effects. His study is limited because it only concentrates on the effects achieved by individual rhetorical devices rather than by a complete poetic discourse. After all, for a complete poetic discourse, to achieve its poetic effects involves dynamic interaction among its ostensive points. For example, if the reader associates ll in L(a with an equal mark, which symbolizes that ones life has half passed, the reader needs to imagine the five stanzas of the poetic discourse as five stages of ones life. These weak implicatures are derived through dynamic interaction among different ostensive points as well as the propositional form possessed by the ostension, L(a. For the above reason, the present study attempts to provide a pragma-cognitive account of how L(a as a complete poetic discourse achieves its poetic effects as well as to probe into the nature of visual poetry.

2.3 Empirical Studies of Poetry


2.3.1 Empirical Studies of Poetry
16

Goodblatt (2001) applies the rigorous experimental approach of cognitive psychology to the study of poetic metaphor and to the process of its comprehension by the real reader. The study suggests different types of reading strategies, which were found to be adopted by various readers, including the integration of alternative readings in order to fill semantic gaps, the simplification of structure through the use of local comparisons, and the elaboration of the structure through the addition of a semantic field. The study therefore contributes to the understanding of the reading process. Yaron (2002) reports two experiments that examine the strategies readers adopt when confronted with the task of processing difficult but valued discourses, such as obscure poems. Two Cummings poems, that is, what a proud and what if a much, which are viewed as obscure poems, are used as materials. The results of the study show that even though the poems are extremely difficult, they cannot be totally misunderstood by readers; and alternative processing mechanisms are activated, a salient one being selection. Yaron (2003) conducts another study to examine the mechanisms of combination in the processing of obscure poems. She tries to deal with the question: how is an obscure poem, which violates communicational norms, processed by the reader? In the study, another pair of Cummings poems, that is, what a proud and pity this busy monster are chosen as materials. The two empirical studies are conducted to examine two hypotheses: (1) there is a progression from the assimilated elements to items which are stereotypically connected to those elements previously stored; and (2) there is a progression from stereotypical elements to those which are semantically associated with them. The results indicate that readers combine different elements belonging to the same semantic field to the detriment of linearity. The above studies, by providing some evidence for the readers interpretation process in the actual reading of poetry, can help us better our experiment design. In spite of their contributions, there needs to be a further step to take, that is, to probe into how such difficult or obscure poetic discourses like L(a achieve their poetic effects in actual reading.

17

2.3.2 Studies on Visual Poetry Despite a great amount of literature concerning the studies of poetry, there are very few studies on visual poetry available. The following two are but the only ones found in the literature. Tsur (2000) adopts a cognitive-semiotic approach in the exploration of picture poetry. In Tsurs view, language is a hierarchy of signs, that is, the graphemic string signifies a phonological string, which signifies units of meaning, which in turn signify referents in extra-linguistic reality. Poetic language draws attention to itself, that is, to the hierarchy of signifiers. Tsur also points out the reason why visual patterning is not admitted in manneristic styles, that is, speech sounds are special in our cognitive economy and visual patterning cannot achieve the naturalness of their patterning (2000, p. 751). Based on the fact that in reading, physiological activity and mental activity are correlated, Gross (1997) discusses three visual poems, which she calls pattern poems. In her view, visual poems violate the fundamental law of written language and upset our customary economy of reading, that is, the text generates meaning on both the symbolic and the iconic level, and the reader is required to read it on both levels. The study thus suggests a model that readers exploit to process visual poetry. The study conducted by Tsur (2000) from cognitive-semeiotics perspective probes into the visual poetry theoretically while the study conducted by Gross (1997) examines the actual process of the readers interpretation. Both of the studies have made great contributions to the exploration of visual poetry. However, a lot of questions about visual poetry remain unanswered. One of them is how visual poetry achieves poetic effects, which will be addressed by the present study.

2.4 Studies on L(a


With regard to the studies on L(a, we cannot avoid mentioning Kidders interpretation. Kidder (1979) provides careful readings of a collection of Cummings poems, including L(a. The following is the central part of Kidders (1979, p. 200) interpretation of L(a.

18

Since we base our discussion in Chapter Three partially on this interpretation, we have purposefully used this long citation. The brevity, the evocation of an experience visually apprehended, the intervening colon functioning as an equals sign, and the implied analogy between natural image and human feeling would qualify it for inclusion in any imagist anthology. But Cummings has done more. Focusing on oneness, the poet structures his letters to highlight the numerous appearances of the number one. The first line, l(a, uses both the number and the indefinite article; the second line uses the French definite article le; line 5 (ll) has two ones; line 7 (one) spells out the word; line 8 (l) isolates the digit; and line 9 (iness) suggests both Cummingss customary first person pronoun(i) and lower-case Roman numeral one, attached to a suffix making them into nouns of condition or quality. Even the number (1) which precedes the poem is significant. Visually, Cummings structures the poem so that two different aspects of its design the overall shape and the inner movements among the lines draw attention to the two sides of its equation, oneness and the leaf. The overall shape is that of a tall figure 1 resting on a flat base. And the patterns within the poem suggest a leaf drifting downward from a tree. The eye, like a leaf, moves slowly earthward in a gently rocking and somewhat repetitious motion: l(a shifts to le; af flips over into fa the quick straight drop in the middle of the poem (ll, resembling an upended equals sign) is slowed by the extended line one; and the eye finally lands on the longest line of the poem, resembling a little heap of leaves already fallen. Cummings central perception here is that loneliness contains oneness or rather one-one-one-iness and that the single leaf falling is metaphor for both physical and spiritual isolation. Furthermore, by splitting open loneliness and inserting the brief sentence about the leaf, he has made the word for the human condition contain the natural image. Aside from Kidders (1979) interpretation, this poem has been interpreted from other perspectives by many readers as well as scholars. Zhao (1994), from the perspective of stylistics, discusses the visual effects produced by the poem. She points out that the visual effects of the poem was influenced by three factors: first, the omission of punctuation; there is no capitalized letter as the start signal of the poem, no stop as ending signal; second, the arrangement of the letters in the poem; and third, the shape of the poem. Duan (2002), from the perspective of aesthetics, argues that the value of Cummings poems, including L(a lies in their high indeterminateness, which encourages the reader to take the full responsibility to interpret the poems. By contrasting the interpretation of L(a with

19

many similar themes expressed by Chinese poetry, B. H. Zhang (2001) argues that people have similar cognition of the world in spite of their differences in cultural background, ideology and aesthetics. From a cognitive perspective, Hiraga (2005, p. 109) suggests a metaphorical interpretation of L(a. In his view, there are two input mental spaces, nature and mind as epitomized in the image of a falling leaf and a feeling of loneliness. These spaces are blended in the juxtaposition in a way that a falling leaf is mapped onto loneliness. It is a leaf which is falling, a metaphor of decline, which is also visually presented by the vertical layout of letters and symbols. L(a as a visual poem has drawn a lot of readers and scholars attention. However, almost all of them put their emphasis on its interpretation rather than the exploration of the mechanisms through which L(a works as a visual poem and of how L(a achieves its poetic effects.

2.5 Summary
In this chapter, we first defined and explained some terms, which the present study employs, and then reviewed the previous studies, which are closely related to the present study. To sum up, though some studies have explored poetry and poetic effects in the scope of pragmatics, thus facilitating our understanding of poetry, their discussions are by far inadequate. At the same time, the empirical studies of poetry and those studies on visual poetry have examined the process of interpretation from various perspectives. They have provided us with some hints regarding further study. However, a lot of questions remain unanswered. The following are the three of the prominent ones. Two of them are concerning poetic effects achieved by visual poetry: what elements in visual poetry determine its poetic effects? And in what ways does visual poetry achieve poetic effects? The third one is related to readers attitudes towards visual poetry, that is, how do readers react to visual poetry in actual reading? The present study therefore attempts to answer these questions from a pragma-

20

cognitive point of view by probing into the mechanisms on which visual poetry works so as to facilitate our understanding of such discourses. The study will be carried out in two steps: first, we will deal with these questions theoretically; second, we will validate the theoretical analysis with an empirical test. One of Cummings visual poems, L(a, will be singled out for our case study.

21

Chapter Three

The Conceptual Framework

This chapter is composed of two sections. In the first section, a conceptual framework will be constructed. In the second section, L(a will be analyzed within the framework to address how L(a achieves its poetic effects. Furthermore, L(a will also be compared with its reproduced horizontal version, i.e., A leaf falls: / loneliness to investigate whether L(a can achieve more poetic effects than its horizontal version.

3.1 Relevance Theory


3.1.1 Ostensive-inferential Communication Relevance Theory explores the process of human communication, especially the utterance comprehension in communication from a pragma-cognitive point of view. Rather different from coding-decoding model established by semiotics in verbal communication, the theory advocates an ostensive-inferential model of communication, defined as below: The communicator produces a stimulus which makes it mutually manifest to communicator and audience that the communicator intends, by means of this stimulus, to make manifest or more manifest to the audience a set of assumptions I. (Sperber & Wilson, 1995: 63) To obtain a good understanding of this claim, we may begin by taking a look at
22

informative intention and communicative intention involved in an ostensive-inferential communication. According to Sperber and Wilosn (1995, p. 50), an ostension provides two layers of information to be identified. One is the information which has been picked up; the other is the information that the first layer of information has been intentionally picked up. Corresponding to these two layers of information provided by an ostension are the informative intention and the communicative intention. In their view (1995, p. 58), a communicator produces a stimulus to intend to make manifest or more manifest to the audience a set of assumptions I. This is the communicators informative intention. Consider the following example (Sperber & Wilson, 1995, p. 58): (1) Passenger: When does the train arrive at Oxford? Ticket-collector: At 5: 25. In (1) the ticket-collectors informative intention is to make manifest to the passenger the single assumption that the train arrives at 5:25. With regard to communicative intention, Sperber and Wilson (1995, pp. 60-61) argue that: To communicate intentionally by ostension is to produce a certain stimulus with the aim of fulfilling an informative intention, and intending moreover thereby: (communicative intention) to make it mutually manifest to audience and communicator that the communicator has this informative intention. Consider the example (1) again. The ticket-collector makes her informative intention manifest to the passenger that the train arrives at 5:25. Furthermore, the ticket-collector makes it manifest to the passenger that she intends to make it manifest that the train arrives at 5:25. If this manifestation is realized by the passenger, then the ticket-collector communicative intention is fulfilled. Ostension and inference are two sides of communication. From the communicators (e.g. the speakers or the writers) viewpoint, communication is the process of ostension, that is, the communicator tries to convey her informative intention by exploiting, for example, utterances and body movements. However, for the addressee (the hearer or the

23

reader), communication is the process of inference, that is, the addressee combines the communicators ostension (e.g. an utterance) with his contextual assumptions to achieve contextual effects as well as to obtain the communicators communicative intention. In Sperber and Wilsons (1995, p. 54) view, inference and ostension are one and the same process, but seen from two different points of view, that is, the communicators and the addressees points of view. The communicator is involved in ostension while the addressee is involved in inference. 3.1.2 Two Principles of Relevance According to Wilson (1994, p. 44), relevance theory takes a few very simple assumptions as its foundation. First, every utterance has a variety of possible interpretations, all compatible with the information that is linguistically encoded. Second, not all these interpretations occur to the hearer simultaneously; some of them need more effort to think up. Third, hearers are equipped with a single, very general criterion for evaluating interpretations as they occur to them. Wilson (1994, p. 44) also argues that human cognition is relevance-oriented: we often pay attention to information that seems relevant to us. Based on these assumptions, Wilson (1994, p. 46) claims that the most basic assumption of RT is that every aspect of communication and cognition is governed by the search for relevance. The fundamental claim about cognition, which is now termed the Cognitive Principle of Relevance, is that Human cognition tends to be geared to the maximization of relevance (Sperber & Wilson, 1995, p. 260). This claim means that peoples cognitive resources, including both internal and external ones, tend to be allocated to the processing of the most relevant inputs available. On the other hand, all ostensive communication is governed by the Communicative Principle of Relevance, which states: Every act of ostensive communication communicates a presumption of its own optimal relevance. (Sperber & Wilson, 1995, p. 260) The presumption of optimal relevance in this claim is further expressed in two parts

24

as follows (Sperber & Wilson, 1995, p. 270): (a) The ostensive stimulus is relevant enough for it to be worth the addressees effort to process it. (b) The ostensive stimulus is the most relevant one compatible with the communicators abilities and preferences. This says that the communicator entitles the addressee to expect a level of relevance high enough to make his attending necessary to the stimulus, and moreover, the communicator is capable of achieving her given means and goals by making the stimulus the most relevant one. It is the case that the addressee may obtain a wide range of possible interpretations, which would yield enough effects to draw the addressees attention. However, the addressee should choose the most accessible interpretation, the one that is easiest for him to construct. Moreover, this interpretation consistent with the principle of relevance is the only one to yield adequate effects for the minimum justifiable effort (Wilson & Sperber, 1993, p. 6). Wilson (1994) strengthens this idea with the argument that the first acceptable interpretation is the only acceptable interpretation. Consider the example we mentioned above. (2) a. Peter: Would you like some coffee? b. Mary: Coffee would keep me awake. In interpreting Marys utterance, Peter would be expected to supply the contextual assumption in (3) and derive the contextual implication in (4): (3) Mary doesnt want to be kept awake. (4) Mary doesnt want any coffee. However, this is not the only possible interpretation. In certain circumstances, for example, if Peter knows that Mary is going to prepare her class for tomorrow, then he would supply the contextual assumption in (5) and derive the contextual implication in (6): (5) Mary wants to stay awake. (6) Mary wants some coffee. But how does Peter know which interpretation was intended by Mary? Following from clause (b) of the definition of optimal relevance, if in certain circumstances, the
25

contextual assumption in (3) is highly salient, and leads on to a satisfactory interpretation, then (4) is the only interpretation that the speaker is free to intend the hearer to choose. Similarly, if the contextual assumption in (5) is highly salient in certain circumstances, then (6) is the only interpretation that the speaker intends the hearer to choose. Yus (2002) supports this idea by claiming that there should be a first interpretation which provides the highest interest to the addressee in exchange for the least processing effort. However, in his view (2002, p. 619), in literary communication, the reader is often willing to put more cognitive effort in exchange for a wider array of aesthetic effects in his reading activity. For poetic language, Fabb (1997, p. 10) makes a similar claim that there is no single tightly constrained set of meanings intended by the poet, just some set of meanings that can be inferred from the utterance by the reader. He takes the line (7) as an example to illustrate the idea. (7) My love is a red red rose. The hearer may use (7) as evidence that the speaker intends to tell him that the loved person is beautiful, precious, and will not live for ever. The analogy to a flower means that various characteristics of the flower will be carried over to the loved person. As a result, the reader is encouraged to supply a great many contextual assumptions and derive a wide range of acceptable implicatures. It is clear that cognitive effort and contextual effects are interact to define relevance, and in literary discourses the reader is encouraged to derive more possible interpretations rather than the most salient one. But how do people balance the processing effort and contextual effects? And how does the reader derive more possible interpretations from literary discourses like poetry? These two questions will be addressed in 3.1.3 and 3.2 respectively. 3.1.3 Effects and Effort According to Sperber and Wilson (1986/1995), relevance is defined in terms of contextual effects and processing effort. Contextual effects can be achieved when newlypresented information interacts with a context of existing assumptions in one of the following three ways, that is, first, by strengthening an existing assumption; second, by
26

contradicting and eliminating an existing assumption; third, by combining with an existing assumption to yield a contextual implication. Moreover, newly-presented information is relevant in a context when and only when it achieves contextual effects in that context, and the greater the contextual effects, the greater the relevance. However, contextual effects do not come free but cost some mental effort to derive. The processing effort required to comprehend an utterance is decided by the effort put to construct a suitable context and the psychological complexity of the utterance itself. Wilson (1994, p. 46), in terms of relevance, summarizes the relationship between effort and effects as follows: Relevance, then, depends on contextual effects and processing effort. The greater the contextual effects, the greater the relevance; but the greater the processing effort needed to obtain these effects, the lower the relevance. Wilson further discusses the conditions in which the optimal relevance is achieved. She (1994, 47) points out: An utterance, on a given interpretation, is optimally relevant if and only if: (a) it achieves enough contextual effects to be worth the hearers attention; (b) it puts the hearer to no gratuitous effort in achieving those effects. On the effect side, what the hearer is entitled to look for is enough effects to make the utterance worth his effort to process. On the effort side, the hearer must invest some effort to achieve effects. Moreover, Wilson (1994, p. 53) claims that any element of indirectness in an utterance requires the addressee to put additional processing effort, and thus encourages the addressee to search for additional effects that a more direct formulation would not have achieved. This implies that extra effort would be offset by extra effects. Consider the following example from Pilkington (2000, p. 77). (8) A: Drink? B: Im driving. In (8) instead of saying No, B says Im driving, intending A to supply the assumption that when driving it is not permitted by law to drink and derive the conclusion that B does not want to drink. B, aiming at the optimal relevance, intends the indirect

27

answer to achieve these additional contextual effects, which are not achievable by the direct answer No. The extra contextual effects are achieved by the way that A puts more effort to process Bs indirect answer Im driving. Consider another example from Sperber and Wilson (1995, p. 219) (9) Theres a fox, a fox in the garden. Compare it with the following utterance: (10) There is a fox in the garden. The repetition in (9) requires more effort to process than (10). It is easy to imagine in both cases that one or two implicatures are strongly communicated, for example, The chickens are in danger. However, to justify the repetition of a fox, the hearer thinks of all the implicatures that the speaker could have expected him to derive from (9), for example, a fox is rare in neighborhood; or the speaker only shows her surprise to see a fox in garden, or express her excitement to see a fox. The hearer needs to put more effort to process the intention conveyed by the repetition. The extra effort is offset by extra contextual effects more acceptable implicatures being obtained. This idea is strengthened by Pilkington (2000, p. 79), the greater effort involved in accessing the intended assumption is repaid by an increase in contextual effects: a wider range of implicatures being communicated. 3.1.4 Style and Weak Implicatures: Poetic Effects Style is an ever-lasting topic in literary discourses. In RT, style is the relationship (Sperber & Wilson, 1995, p. 217). In their view, people can use the style involved in a communication to infer such things as what the communicator thinks of as the addressees cognitive capacities and level of attention, how much help she is prepared to give him in processing her utterance and their emotional closeness or distance. Poetic effects are often achieved by the style the communicator employs. Some depend on metaphors and some on verse devices and some on such devices as deviation and parallelism as we discussed in 2.2.1. Consider the following examples (Sperber & Wilson, 1995, p. 222):

28

(11) Mary lives in Oxford, Joan in York, and Lily in a skyscraper. (12) Mary came with Peter, Joan with Bob, and Lily with a sad smile on her face. In (11), the hearer is entitled to find some contextual assumptions in which the facts that Mary lives in Oxford, Joan in York, and Lily in a skyscraper have either identical or indirectly contrasting implications. Some basic assumptions about Oxford, York and skyscrapers suggest that Mary and Joan do not live in skyscrapers, and that Lily does not live in an old town. Moreover, if the overall interpretation is to be consistent with the principle of relevance, the speaker must have conveyed more than that: for example, she might have been trying to communicate a wide range of weak implicatures showing that the way Mary and Joan live is more influenced by the kind of town they live in than by the kind of building they live in, while Lilys way of life is influenced by the kind of building rather than the kind of town they live in. In the case of (12), the hearer is entitled to find some contextual assumptions in which the facts that Mary came with Peter, Joan with Bob, and Lily with a sad smile on her face have either identical or directly contrasting implications. What might be indicated is that Lily had no one to come with, that she was sad because she had no one to come with, that there was a story behind her sad smile in which Mary, Peter, Joan and Bob were involved, and that an imaginative hearer could seek for more along these lines. In this manner, the required parallelisms of context and contextual effects could be maintained and a wide range of weak implicatures would be derived. Sperber and Wilson (1995, p. 218) claim that the choice of style is something that no speaker or writer can avoid. Aiming at relevance, the speaker must make some assumptions about the hearers cognitive abilities and contextual resources, which will necessarily be reflected in the way she communicates, that is, the speaker must have a basic idea about what she chooses to make explicit and what she chooses to leave implicit in her utterance. Stylistic differences are just differences in the way relevance is achieved. Poetic style is a consequence of the speakers aim of producing an utterance consistent with the principle of relevance. Poetic effects result from the accessing of a large array of very weak implicatures characterizing a poetic style.

29

3.2 The Theoretical Analysis of L(a Within the RT Framework


Within the framework constructed in 3.1, we intend in this section to analyze how L(a achieves its poetic effects by comparing it with a reproduced style. We first analyze how L(a achieves its poetic effects and then compare it with a different style to probe into the mechanisms of how L(a works as a visual poetic discourse. We would like to use A to stand for L(a, the vertical style, and B to stand for the horizontal style, which was reproduced by Kidder (1979) in his interpretation of L(a. The following are the two styles: A l(a le af fa ll s) one l iness 3.2.1 L(a Viewed in the RT Framework In 2.1.1, we argued that reading is one mode of verbal communication. To be specific, in terms of RT, reading and interpreting L(a is an ostensive-inferential communication, in which L(a is the ostension, the poet, Cummings, being the communicator and the reader being the addressee. Cummings, as the communicator, employs L(a as an ostension to convey his informative intention to the reader. The reader, as the addressee, puts the ostension, L(a, into his cognitive environment and combines L(a with his contextual assumptions to obtain the poets communicative intention. The following will present how L(a works as an ostension consistent with the principle of relevance. B A leaf falls: loneliness.

30

The poet, aiming at achieving optimal relevance, assumes that the employed vertical style is the most relevant stimulus that contains as many ostensive stimuli as possible for the reader to obtain both informative intension and communicative intention, and which is compatible with his abilities and preferences. Since poetry is a kind of saying (Brooks & Warren, 2004, p. 1), from the poets perspective, the vertical style is the most appropriate way to say what Cummings intends to say to his reader. The vertical style then arises in pursuit of relevance. Moreover, when Cummings employs this vertical style as an ostension to convey his informative intention to the reader as well as expects that his reader can recognize his communicative intention through inference, he has made some assumptions about his readers cognitive and contextual resources, which are available for use by the reader in interpretation. On the side of the reader, the vertical style is the most relevant stimulus worth his effort to process. The reader is entitled to expect a level of relevance high enough to warrant his attending to the stimulus of L(a, which encourages him to greatly expand his context and thoroughly explore his encyclopedic entries related to the ostension to derive a wide range of weak implicatures. In the process of interpretation, the readers role is the most important one, just as Chrzanowska-Kluczewska (2001, p. 193) argues that it is the reader who decides what the text is about and the text speaks to the reader whenever it can be related in some way to his existential situation. This indicates that the readers cognitive environment will greatly influence the interpretation. This also suggests that in literary communication, the reader can take his responsibility to obtain possible interpretations from a literary works in line with the principle of relevance. To be specific, if a lot of contextual assumptions can be supplied in the process of interpreting L(a, it will be possible for the reader to exploit these contextual assumptions to derive a wide range of implicatures from L(a. 3.2.2 RT Based Analysis of L(a In this section, we intend to take a careful look at how L(a achieves its poetic effects, that is, what elements in L(a contribute to its poetic effects and in what ways they make their contributions.

31

Unlike other ostensions that are constructed by linguistic stimulus alone as in the horizontal style, L(a, the vertical style, contains not only linguistic stimulus but also visual stimulus. It is the visual stimulus of L(a that makes it different from conventional linear poems. However, the linguistic stimulus and the visual stimulus cannot be separated since they are encoded and shaped in one ostension. The linguistic stimulus alone or the visual stimulus alone cannot make L(a so attractive. It is an active interaction between these two kinds of stimuli that contributes to its poetic effects and the poet must have expected the reader to use these two kinds of stimuli in an interactive way in interpretation. In the following subsections, we will explain how L(a achieves its poetic effects in ostensive-inferential communication from two aspects: first, how L(a communicates itself to the reader through its overall propositional form (i.e. A leaf falls: loneliness) and the ostensive points contained in both linguistic stimulus and visual stimulus; second, how individual readers cognitive environments influence their derivation of weak implicatures. 3.2.2.1 L(a as Ostensive-inferential Communication For ease of discussion, we will first take a look at the linguistic stimulus of L(a. Its propositional form of the linguistic stimulus is A leaf falls: loneliness, which is embedded in the ostension and which is covert for the reader. However, the poet expects the reader to obtain this propositional form by careful perception. First, Cummings exploits a leaf falls to create a poetic image just as Kidder (1979, pp. 199-200) says had Cummings been merely an imagist, he might have written simply A leaf falls: / loneliness. The poet uses A leaf falls but not Two leaves fall or Leaves fall. It is this A leaf falls that creates loneliness. The poet links the propositional form of a leaf falls with the abstract concept loneliness together by creating an image of a leaf falls. To obtain the communicative intention, which the poet exploits a leaf falls: / loneliness to convey, the reader must imagine an image of a leaf falls, which the poet intends to make manifest. In other words, when the reader approaches the propositional form, the first thing for him to do is to search for relevance and try to establish a close link between the propositional form a leaf falls, which is

32

composed of several concepts and the concept of loneliness by exploring his encyclopedic entries attached to them. In so doing, the reader may supply some contextual assumptions and correspondingly derive some possible implicatures, such as in (13) and (14): (13) A falling leaf symbolizes falls coming, just as a Chinese saying tells that (That one leaf falls suggests autumn is coming). (14) A leaf falls can be associated with the end of a life, for example, a person died. Only a leaf falls may encourage the reader to obtain such implicature as a person died lonely. Besides, fall as a verb has the same spelling and pronunciation with the season fall in American English. This may strengthen the image that a leaf falls in the fall, which is often considered by poets as a sad and lonely season. On the other hand, fall can also be considered as a harvest season. If this contextual assumption is supplied by the reader in interpreting process, then the following possible implicatures may be derived in (15): (15) a. It is a great pity that a person lost his life in his harvest season. b. When a leaf falls in harvest season, it is mature. In addition to its overall propositional form, L(a contains a lot of ostensive points such as a, ll, one, l and i as well as af/fa, s, ( ) and ll/iness. These ostensive points as discussed in 2.1.4 can activate the reader to exploit contextual assumptions to derive a large number of possible implicatures. Stimulated by these ostensive points, the reader would supply a lot of contextual assumptions as listed in (16) and derive a wide range of implicatures as provided in (17): (16) a. a means one. b. a is an indefinite article. c. ll is like two Arabic number of 1. d. l is the Arabic number of 1. e. ll/iness can be combined to be illness. (17) a. a can be associated with loneliness. b. a can be associated with uncertainty. c. ll can be associated with loneliness.
33

d. l can be imagined to be a unique thing or person. e. ll/iness can be associated with the case that someone is ill. For L(a, most of these implicatures are weakly communicated. In other words, most of them might not be intended by the poet to communicate. However, the reader can take his responsibility to derive these implicatures consistent with the principle of relevance. Furthermore, the visual stimulus also contributes a lot to poetic effects of L(a. In the vertical style, the poet makes a dynamic picture of a leaf is falling down by arranging the letters and words carefully. Based on this kind of visual stimulus, the reader would supply the contextual assumptions and correspondently derive a lot of possible implicatures in (18): (18) a. a, like the shape of a leaf, stimulates the readers visual sensory. b. Two f, being fricatives, like wind, blows a leaf from left to right while it is falling down. c. s, also a fricative, like wind, keeps blowing to make the leaf rest on one for a while. d. ( ) can be associated with durance. e. af/fa may indicate that a leaf is blown from left to right or the world is in disorder. f. iness is the longest line, which may be associated with many leaves on the ground. g. the shape of the vertical arrangement is rather slim like a straight tree. We can see that all kinds of elements in L(a contribute to its poetic effects. As Verdonk (1991, p. 98) points out, each in its own way, the physical aspects of textual organization convey some secondary meaning and guide the readers experience and perspective. Based on the interaction among this dynamic picture, the overall propositional form and the ostensive points, the reader may expand his context search and explore his encyclopedic entries more thoroughly. Then a wider range of weak implicatures can be derived from the contextual assumptions this picture yields such as in (19): (19) a symbolizes a person, and five stanzas of the poem can symbolize the five
34

periods of a persons life: childhood, youth, middle-agedness, agedness and the end of life. Life is just like a falling leaf. a is the beginning of life, the birth of a person; in the second stage of life, that is, youth, in which people are greatly influenced by their environments, just as a falling leaf is blown from left to right by wind. In the third stanza, ll, like an equal mark, tells that half of the life has passed. The fourth stanza symbolizes the aging stage, in which l stresses the only way out for people, that is, the destination the last stanza death. The dynamic picture creates a wider context, which supplies certain assumptions within the encyclopedic entries that are explored to become more highly activated and, hence, makes them easier to process and use in the construction of further assumptions. In this sense, it is the visual stimulus that strongly encourages the reader to thoroughly explore his encyclopedic entries to derive as wide a range of weak implicatures as possible consistent with the principle of relevance. When the reader accesses a lot of contextual assumptions and derives a wide range of weak implicatures in line with the principle of relevance, L(a achieves its poetic effects. However, the dynamic picture, various types of one, loneliness, or other devices involved in this poetic discourse cannot make L(a rich enough for the reader to supply a great many contextual assumptions to derive as wide a range of weak implicatures as possible. Only the active interaction among these various aspects can make this poem rich enough in its poetic effects. Just as Pilkington (1991, pp. 59-60) points out, Images, symbols, metaphors within the poem interact to make manifest a vast range of weak implicatures. Through the active interaction between the overall propositional form of L(a and the ostensive points created by both linguistic stimulus and visual stimulus of the ostension, a wider range of weak implicatures might be derived in (20), (21) and (22) as follows: (20) L(a might encourage the reader to activate another picture in his mind that a person is wandering alone far from his home, without relatives around, without friends around, and even without a person to talk with. (21) L(a might also encourage the reader to imagine that no matter how far a person

35

is away from his home or motherland, his home or motherland would be his destination just as a Chinese proverb says (Falling leaves settle on their roots). (22) The reader might imagine that the poet is appreciating the beautiful natural scenery a leaf is falling down. The poet is so obsessed with such an instantaneous beauty that he pays great attention to each eye-catching moment. His mind is in absolute stillness. Of course, the derivation of a wide range of weak implicatures demands a great amount of processing effort to explore a lot of encyclopedic entries and supply adequate contextual assumptions. However, just as we discussed in 3.1.3, the greater effort involved in accessing the intended contextual assumptions is repaid by an increase in contextual effects: a wider range of weak implicatures being communicated, which results in richer poetic effects. 3.2.2.2 Influence of Cognitive Environment on Deriving Weak Implicatures In 3.2.2.1, we gave a careful explanation of how L(a achieves its poetic effects from a pragma-cognitive point of view. However, in our discussion, we have assumed an ideal reader, whom the poet expects to identify his informative intention and communicative intention by interpreting his ostension L(a. One question thus emerges. Can every individual reader supply the similar contextual assumptions and derive the similar amount of weak implicatures in actual reading? Our answer is in the negative. Just as we discussed in 2.1.5, since individual readers cognitive abilities and cognitive environments that are affected by previously memorized information may differ from person to person, they may supply different contextual assumptions and derive different numbers and types of weak implicatures. Actually, the derivation of weak implicatures in actual reading can be influenced by two factors. One is the number of the perceived ostensive points from an ostension (e.g. L(a) as will be discussed in 3.2.3, and the other is the individual readers cognitive environment. It is possible for different individual readers to perceive the same number of ostensive points from L(a, but it is impossible for them to construct the same contextual assumptions to interpret L(a because their cognitive environments will never be identical.

36

In this case, individual readers cognitive environments will be the crucial factor to determine the number and type of weak implicatures derived from L(a. The richer the individual readers cognitive environment is, the larger number and more types of weak implicatures he will obtain. In other words, if the individual readers cognitive environment is rich enough to supply adequate contextual assumptions to interpret L(a, he may derive more possible implicatures from it. Therefore, since individual readers can never share the same cognitive environments, they will never obtain the same number and type of weak implicatures from L(a. It may also be the case that readers cannot supply enough contextual assumptions to process the ostensive points perceived from L(a in actual reading. If there are no enough cognitive resources available for use in interpreting L(a, individual readers may not arrive at a successful reading of the poem. Pilkington (2000, p. 63) makes a similar claim that in the actual reading, real readers may have problems either when they do not have the assumptions they need for interpretation stored in memory (i.e. in the encyclopedic entries attached to contextual addresses that are triggered by words or other devices in the text), or when they are not able to construct the assumption easily. So another question comes up. How will readers react to visual poetry like L(a in actual reading? As reviewed in 2.3.1, Yaron (2002, 2003) argues that despite of the difficulty of obscure poems, readers will not reject them in actual reading. Following Yarons arguments, we make a similar hypothesis that the difficulties and obscurities of visual poetry like L(a will be recognized by readers, who are nevertheless tolerant of such difficulties and obscurities in their interpretation. This question will be dealt with by the empirical part of the study. 3.2.3 RT Based Comparison Between the Vertical Style and the Horizontal Style The most apparent difference between the vertical style and the horizontal style lies in their arrangement. The vertical style is carefully shaped to be vertical while the horizontal style is arranged as the conventional linear poems are. However, from a RT perspective, the most significant difference lies in their poetic effects. We attempt to deal with the difference in this section.

37

The horizontal style as an ostension only contains linguistic stimulus. However, the vertical style, as we discussed in 3.2.2, contains both linguistic stimulus and visual stimulus. The vertical style contains a great many ostensive points whereas the horizontal style contains a few ostensive points such as a leaf, fall, : and two short lines. The horizontal style achieves its poetic effects mainly through its propositional form and these ostensive points. In 3.2.2, we discussed how L(a achieves its poetic effects. We assumed that the great many ostensive points from both linguistic stimulus and visual stimulus activate the reader to supply rich contextual assumptions and correspondingly derive a wide range of weak implicatures consistent with the principle of relevance. The similar claim has been made by Gross from another perspective, In pattern poems, the text generates meaning through both symbolic and iconic signification, and the reader is asked to read it on both levels (1997, p. 17). In this process, many encyclopedic entries are thoroughly explored and a great many contextual assumptions are supplied to derive a wide range of weak implicatures. In other words, more ostensive points may activate the reader to supply more contextual assumptions to process. As a result, more weak implcatures may be derived, which in turn will lead to more poetic effects. Both the vertical style and the horizontal style possess the same propositional form, but differ in the total of ostensive points. It is obvious that the number of ostensive points in the vertical style is far more than that of ostensive points in the horizontal style, and it is the number of ostensive points that makes the two styles greatly different in the poetic effects achieved. Thus, theoretically, the vertical style achieves much more poetic effects than the horizontal style. Cureton (2002, p. 37) claims that what a poem conveys is always significant and must be considered carefully, and the task of poetic intention and poetic accomplishment usually falls on the side of form. This claim is verified by the analysis of L(a by comparing it with the horizontal style. So far we have made a relatively thorough exploration of L(a, leading to the conclusion that no matter what style a poetic discourse uses, its poetic effects are achieved through deriving as wide a range of weak implicatures as possible, which

38

depends on a thorough exploration of the readers encyclopedic entries. The most significant difference between different poetic discourses lies in the number of their ostensive points, which would result in a great difference in the number and type of weak implicatures derived, which in turn results in difference in the poetic effects they may communicate. However, it should be pointed out that the poetic effects potentially conveyed by a poetic discourse will not necessarily match those achieved by readers in actual reading. In other words, readers may not obtain the same number and type of poetic effects from a poetic discourse in actual interpretation as the poetic discourse intends to communicate.

3.3 Summary
In this chapter, we constructed a conceptual framework, within which how the visual poem L(a achieves its poetic effects was carefully analyzed. By comparing L(a with the horizontal style, we claimed that L(a can achieve more poetic effects than the horizontal style. The reason why this is so does not reside in the picture that the vertical style creates but lie in the number of the ostensive points the vertical style contains as well as the active interaction among these ostensive points. The picture shaped by the visual poem creates more ostensive points than the conventional linear form. It is the larger number of ostensive points contained in L(a, which activate the reader to explore more encyclopedic entries and supply more contextual assumptions to derive a wider range of weak implicatures, which plays a crucial role in achieving poetic effects. Based on the theoretical analysis, we now arrive at two hypotheses: (1) The vertical style as an ostension contains more ostensive points than the horizontal style does. More ostensive points will encourage the reader to supply more contextual assumptions and access more encyclopedic entries. As a result, more weak implicatures will be derived from the ostension, which in turn, will lead to more poetic effects. (2) The richer the readers cognitive environment is, the more weak implicatures he will obtain. Since the individual readers cognitive environments are different, it is

39

impossible for different readers to access the same number and type of encyclopedic entries, to supply the same number and type of contextual assumptions and to derive the same number and type of weak implicatures. In the next two chapters, we will attempt to put these two hypotheses to empirical test.

40

Chapter Four

Methodology

This chapter consists of six sections. First, some research questions are raised to test the theoretical hypotheses put forward in Chapter Three as well as the readers attitudes towards visual poetry as mentioned in 2.5 and discussed in 3.2.2.2. Second, the background information of the subjects is provided. Third, the experimental instrument employed in the study will be described carefully. Finally, a detailed explanation will be given regarding how the data were collected and analyzed.

4.1 Research Questions


To test the hypotheses from the theoretical analysis in Chapter Three as well as to examine the readers attitudes towards visual poetry as we briefly discussed in 3.2.2.2, the present study raised the following research questions: 1. Does L(a, the vertical style, communicate more poetic effects than the horizontal style in actual reading? a. How many ostensive points do readers perceive in the vertical style and the horizontal style respectively? b. Is there any difference between the numbers of ostensive points perceived from
41

two different poetic styles by the readers in the same group? If any, why? 2. Do individual readers cognitive environments influence their interpretation? a. Is there any difference between the numbers of the ostensive points perceived from the same poetic style by the readers in the two groups? b. 3. If yes, what leads to such difference? If no, why?

How do readers react to the two different poetic styles?

4.2 Subjects
The subjects involved in the study, all from Nanjing University, consisted of two groups. Group One were 62 first-year undergraduate students, whose major was accounting while Group Two were 56 first-year graduate students, whose major was English linguistics or English literature. They were all new students in the second week of their university study. In other words, the students in Group One had just finished their high school study and they had no experience of taking English literature course, whereas the students in Group Two had finished their undergraduate study and they had taken at least one-year English literature course during their undergraduate study. The reason why we chose these two different groups of readers as subjects was that we assumed that general knowledge about English literature stored in readers memory, which formed readers cognitive environments, would play an important role in interpretation. The readers in Group Two had taken at least one-year English literature during their undergraduate study. They had got the basic knowledge about English literature, so they were expected to perform better than those readers in Group One in finishing the task.

4.3 Instrument
Based on the two pilot studies, a questionnaire (see the Appendix) was finally designed to fulfill the purpose of the study. L(a, the vertical style marked A, and its reproduced form, the horizontal style marked B in the questionnaire were presented together to readers. One salient theme of the poem (i.e., loneliness) shared by the two poetic styles was made known to readers, who were then asked to perceive the two styles
42

and identify the details from the two styles respectively, the ostensive points in our terms, that may contribute to the theme. The tasks were presented on the same paper. Questions 1 and 2 in the questionnaire were designed to examine the subjects basic assumptions about the two poetic styles. Question 3 in the questionnaire consisted of two parts, requiring the subjects to identify the details that contribute to the given theme in the vertical style and the horizontal style respectively. Question 4 in the questionnaire first required the subjects to choose the style, which they thought could better convey the theme, and then to write down the reason why they did so.

4.4 Data Collection


With the permission of the two teachers who taught two courses to the two groups of subjects respectively, the researcher conducted the experiment in class. The purpose of the questionnaire and tasks were explained to the subjects. The subjects were allowed to ask questions concerning with understanding of the questionnaire. The subjects were required to finish the questionnaire in 20 minutes. Data were collected in terms of the answer to the first sub-question of Question 1 (i.e. Have you read or seen A?). That is, if the subjects ticked Yes to answer the first sub-question of Question 1 in the questionnaire, that meant the subjects had read or seen the vertical style. These data then were considered to be invalid. If the subjects ticked No to answer the first sub-question of Question 1 in the questionnaire, that meant the subjects had not read or seen the vertical style. Such data then satisfied the needs. As a result, 35 response papers from graduate students were qualified. Correspondingly, 35 response papers from the undergraduate students were randomly chosen as our data. In addition, the reasons presented by the subjects in Question 4 in the questionnaire were also collected as the qualitative data, from which the researcher randomly chose some from each group to support the theoretical analysis as well as the quantitative results. For ease of the discussion in Chapter Five, these data were numbered like G1-S1, G1-S2, and G1-S8, which stand for the data from Subject One in Group One, the data

43

from Subject Two in Group One,, and the data from Subject Eight in Group One; and G2-S1, G2-S2, and G2-S8, which stand for the data from Subject One in Group Two, the data from Subject Two in Group Two,, the data from Subject Eight in Group Two.

4.5 Data Analysis


First, the ostensive points perceived by readers, that is, the details that contribute to the given theme of the poetic discourse were identified for each form and then were scored one point for each ostensive point. For instance, provided that the reader read the vertical style and wrote down something like the words are separated as one or two letters, and this separation symbolizes the theme of loneliness, it was considered as one ostensive point and was scored one point. If the reader continued to write down something like l and one are associated with loneliness, it was treated as two ostensive points and was scored two points. Given that the reader finally identified only these three details, then he got three points in all. The same went for the horizontal style. It should be pointed out that some ostensive points demand readers to put more effort in perceiving them. These ostensive points should be scored higher. However, it is very difficult to identify how much effort was paid when different ostensive points were perceived by readers, so the researcher scored each ostensive point one point despite the fact that some ostensive points need more effort to be perceived. SPSS 11.5 (Statistical Package for the Social Sciences) was employed to analyze the data collected. The mean (Mean) and standard deviation (SD) for each group and each style were reported. The ostensive points perceived by the readers in each group as well as in each style were compared through t-test. Based on Question 2 and the first part of Question 4 in the questionnaire, the readers assumptions of visual poetry and the readers attitudes towards the two poetic styles were also examined by statistical comparison between two groups. Finally, the reasons presented in the response papers were qualitatively analyzed to complement the quantitative analysis.

44

Chapter Five

Results and Discussion

This chapter reports and discusses the results of the empirical study on the basis of both the quantitative and the qualitative data. The first section presents the number of ostensive points perceived from the two different poetic styles by the readers in the same group to test the first hypothesis in Chapter Three. The second section reports the number of ostensive points perceived from the same poetic style by the two groups of readers. It also deals with Question 2 in the questionnaire quantitatively to find out the role of individuals cognitive environments in actual interpretation to validate the second hypothesis. The third section discusses Question 4 in the questionnaire quantitatively and qualitatively to examine the readers attitudes towards the two poetic styles. In addition, qualitative data collected for Question 4 will be presented to examine the actual reasons given by the readers to supplement the quantitative analysis in each section.

5.1 Ostensive Points and Poetic Effects


This section addresses the ostensive points perceived from the two different poetic styles by the readers in the same group. Two comparisons are conducted to examine
45

whether there are differences in the numbers of the ostensive points perceived by the two groups of readers from the vertical style and the horizontal style respectively. Meanwhile, the individual differences in perceiving ostensive points will be explained on the basis of the qualitative data.
Table 5-1: Comparisons between the numbers of the ostensive points perceived from the two poetic styles by the readers in the same group Ostenp Mean SD T Sig. Group One A 5.51 2.20 9.85 .00 B 2.06 .80 Group Two A 5.49 2.59 7.86 .00 B 1.83 .92 (Sig. level is .05) Notes: Ostenp = ostensive points; A = the vertical style; B = the horizontal style; Group One = the subjects who are the first-year undergraduate students; Group Two = the subjects who are first-year graduate students.

Table 5-1 reports the results of the two comparisons between the numbers of the ostensive points perceived from the two poetic styles by the readers in the same group. For the readers in Group One, there is a significant difference in the number of the ostensive points perceived from the vertical style and from the horizontal style (T = 9.85, p = .00). For the readers in Group Two, there is also a significant difference in the number of the ostensive points perceived from the vertical style and from the horizontal style (T = 7.86, p = .00). The results suggest that the readers in both groups can perceive more ostensive points from the vertical style than from the horizontal style. The vertical style, which contains far more ostensive points than the horizontal style, provides the readers with a lot more opportunities to derive a large number of possible implicatures. Therefore, when the readers perceive more ostensive points from the vertical style and associate them with the given theme of loneliness, they must explore more encyclopedic entries and supply more contextual assumptions, and thus they can derive more weak implicatures from the vertical style. As a result, the vertical style communicates more poetic effects to both groups of readers. The results from Table 5-1 suggest that L(a, the vertical style can achieve more poetic effects than the horizontal style (A leaf falls: / loneliness.) in actual reading. This is

46

in agreement with one of the hypotheses from the theoretical analysis in Chapter Three, that is, more ostensive points in the vertical style result in more weak implicatures, which in turn lead to more poetic effects. The hypothesis is also supported by the qualitative data. Compared with the horizontal style, the vertical style is very novel in form and creates a special kind of effect by cutting the complete words and rearranging the separated letters. In other words, the vertical style contains more ostensive points than the horizontal style by splitting the complete words and rearranging the separated letters. The following excerpts from G1-S1, G1-S2 and G2-S1 are the typical accounts of this kind from both groups of readers to support this claim. G1-S1: A cuts the words into several short lines, which create strong visual effects. At the first glance, these short lines create an atmosphere of loneliness. In A, the arrangement of l-one-l strongly emphasizes the role of one person. Compared with B, A, in both form and content, makes the theme loneliness more salient. Whats more, the arrangement of falls in the poem strengthens the visual stimulus of a leaf falling down. A conveys the theme of loneliness not only through the content but also through the form. The vertical arrangement of letters, which are separated from the complete words, is like the process of a leaf falling down.

G1-S2:

G2-S1:

These reports implicitly show that the readers can perceive more ostensive points from the vertical style than the horizontal style. They understand that compared with the horizontal style, the vertical style contains more ostensive points which are conveyed to them through both linguistic stimulus and visual stimulus. The larger number of ostensive points contained in the vertical style fundamentally determines the more poetic effects it communicates. Table 5-1 also shows that for the readers in the same group, SD varies greatly when they were presented to read the vertical style and the horizontal style (SD = 2.20, and SD = .80 for the readers in Group One; SD = 2.59, and SD = .92 for the readers in Group Two). This suggests that individual differences among the readers in the same group in perceiving the ostensive points from the vertical style are much greater than from the
47

horizontal style. There are at least two reasons for this. First, most readers are entrenched to hold that poems are usually arranged in linear manner. The horizontal style is thus more acceptable for them. Second, the arrangement of the vertical style results in much difficulty for readers to interpret while the short lines of the horizontal style are easily understood. These two reasons are also supported by the qualitative data, according to which the vertical style puts too much emphasis on its form, and consequently it is too obscure and too complicated as well. The excerpts from G1-S3, G2-S2 and G2-S3 are good examples. G1-S3: G2-S2: It seems that A is too complicated. As form influences my understanding and makes me feel uncomfortable. Compared with B, A puts too much emphasis on its form. For those who are not familiar with such forms, A is very difficult to understand. A is very different from traditional poetry. Although it is novel, it destroys the completeness of language. It is very difficult for some readers to understand A. Without referring to B, I dont know what A is, nor how to read A. In A, the poet plays language game by cutting the complete words and rearranging the separated letters.

G2-S3:

Compared with the vertical style, the horizontal style is very short, very simple and easy to understand. The excerpts form G1-S4 and G2-S4 lends support to this view. G1-S4: G2-S4: Compared with A, the purpose of B is to express the content. For readers, it is very easy to catch what B is going to express. B is a traditional form, but it can create an image. Besides, it is easier to understand and B conveys the theme more directly.

The obscurity of the vertical style and the simplicity of the horizontal style can explain why the individual readers differences among the two groups of readers is very low in statistics when they interpret the horizontal style and why the individual readers differences among the two groups of readers seems very high in statistics when they interpret the vertical style. In addition, we notice that the readers in both groups obtain a relatively lower number of perceived ostensive points (M = 5.51 for the readers in Group One, M = 5.49 for the readers in Group Two) in actual interpretation of L(a compared with a great many ostensive points it contains as we analyzed in 3.2.2.1. The possible reason is that the
48

readers are required to associate what they have perceived with the theme of loneliness, which is one of the salient themes of L(a. Therefore, it is likely that the readers have perceived some ostensive points but they might have not written them down since the readers could not associate them with the given theme, loneliness.

5.2 Influence of Readers Cognitive Environments on Their Perception of Ostensive Points


This section addresses the ostensive points perceived from the same poetic style by the readers in the two groups. Two comparisons are conducted to examine whether there are differences between the numbers of the ostensive points perceived from the same poetic style by the two groups of readers. Meanwhile, the individual differences in perceiving ostensive points will be explained carefully.
Table 5-2: Comparisons between the numbers of the ostensive points perceived from the same poetic style by the two different groups of readers Group Mean SD T Sig. OstenpA One 5.51 2.20 .05 .96 Two 5.49 2.59 OstenpB One 2.06 .80 1.11 .27 Two 1.83 .92 (Sig. level is .05) Notes: OstenpA = ostensive points perceived by the readers from A, the vertical style; OstenpB = ostensive points perceived by the readers from B, the horizontal style.

Table 5-2 reports the two comparisons between the numbers of the ostensive points perceived from the same poetic style by the two different groups of readers. For the vertical style, the two groups of readers perceived the similar number of the ostensive points (M = 5.51 for the readers in Group One and M = 5.49 for the readers in Group Two). There is no significant difference between the two groups of readers in perceiving ostensive points from the vertical style (T = .05, p = .96). For the horizontal style, the two groups of readers also perceived the similar number of the ostensive points (M = 2.06 for the readers in Group One and M = 1.83 for the readers in Group Two). There is no significant difference between the two groups of readers in perceiving ostensive points from the horizontal style, either (T = 1.11, p = .27).
49

The results indicate that readers assumptions about English literature might not play a role in interpreting L(a, the vertical style, and the horizontal style. As addressed in 4.2, there exists an obvious difference between the readers in Group One, that is, the first-year undergraduate students who lacked assumptions about English literature and the readers in Group Two, that is, the first-year graduate students who have stored some assumptions about English literature. However, this difference did not result in difference between the numbers of the ostensive points perceived from the same poetic style by the two different groups of readers. The general assumptions about English literature constructed by the readers in Group Two contribute very little to their perception of ostensive points. This is in conflict with our expectation that the readers in Group Two, who have got some basic assumptions about English literature, might perform better than those readers in Group One in finishing the task. From Table 5-2, we can also see that individual differences between the two groups of readers are very close in the number of the ostensive points perceived from the vertical style (SD = 2.20 for the readers in Group One, and SD = 2.59 for the readers in Group Two). Individual differences between the two groups of readers are much closer in the number of ostensive points perceived from the horizontal style (SD = .80 for the readers in Group One, and SD = .92 for the readers in Group Two). Despite of the statistical closeness, these individual differences are still observable. Why is there a disagreement between the results from Table 5-2 and our expectation that the readers in Group Two might perform better than those in Group One in finishing the task? And what leads to the observable individual differences between the two groups of readers in the number of the ostensive points perceived from the two poetic styles. These two questions will be carefully addressed one by one. The disagreement can be explained if we take the readers assumptions about visual poetry into account as shown in Table 5-3.
Table 5-3: Readers assumptions of visual poetry Q&A Group Group One (n = 35) Group Two (n = 35) Question 2 Yes 17 48.57% 18 51.43% No 18 51.43% 17 48.57% Notes: Q & A = question and answer; Question 2 = Question 2 in the questionnaire.

50

Question 2 in the questionnaire was designed to identify whether the readers had seen or read the poems similar to L(a in form. In this manner, we may examine whether the subjects assumptions about the poems similar to L(a in form will influence their perception of ostensive points and Table 5-3 reports the results. Table 5-3 shows that 17 (48.57%) readers in Group One have read some poems in Chinese similar to L(a in form and 18 (51.43%) readers in Group Two have read some poems in English and Chinese similar to L(a in form. The reason why the two groups of readers are very close in the number of the perceived ostensive points associated with the theme of loneliness is that the number of the readers in Group One who have read some visual poems in Chinese is very close to the number of the readers in Group Two who have read some visual poems in English and Chinese. Moreover, in both groups, the readers who have read some visual poems perceive more ostensive points from the vertical style than those who have not read some poems similar to L(a in form. Table 5-4 presents the results.
Table 5-4: Comparisons between the numbers of the ostensive points perceived by the readers who have read some visual poems and those who have not in both groups Group Answer to No. of Mean SD Question 2 Subjects Group One Yes 17 6.00 2.42 No 18 5.02 1.92 Group Two Yes 18 6.23 2.86 4.74 2.15 No 17 Notes: Question 2 = Question 2 in the questionnaire.

Although it is not significant to conduct a t-test between the numbers of the ostensive points perceived by the readers who have read some visual poems and those who have not in each group owing to our small sample, it is clearly observable that the former as a whole is larger than the latter in each group (M = 6.00 for the former and M = 5.02 for the latter in Group One; M = 6.23 for the former and M = 4.74 for the latter in Group Two). It seems likely that readers specific assumptions about L(a stored in their memory will greatly influence their interpretations of the poem. Therefore, we claim that whether readers can interpret visual poetry successfully is not determined by their background knowledge of English literature but by the fact

51

whether they can supply the contextual assumptions about visual poetry. In addition, the results also indicate that individual readers cognitive environments play a very important role in their interpretation. With respect to the individual differences between the two groups of readers in the number of the ostensive points perceived from the two poetic styles respectively, they are also caused by different individuals cognitive environments. The more ostensive points are perceived and used to associate with the theme, the more contextual assumptions need to be supplied from an individual readers cognitive environment. Furthermore, the more ostensive points are combined with more contextual assumptions, the more encyclopedic entries the reader is activated to explore to derive more weak implicatures. Therefore, different individual readers may derive different numbers and types of implicatures due to the differences of their cognitive environments. This verifies our theoretical analysis in 3.2.2.2 that individual readers may supply different numbers and types of contextual assumptions and derive different numbers and types of weak implicatures as well as supports Pilkington claim that individual readers will not access all the same implicatures or the same number of implicatures (1991, pp. 59-60). In addition, the fact that the horizontal style is simple in form and easy to understand as the qualitative data (e.g. the excerpts from G1-S4 and G2-S4) show in 5.1 and the qualitative data in 5.3 (e.g. the excerpts from G1-S7 and G2-S7) can explain why there is very little individual difference (SD = .80 for the readers in Group One, SD = .92 for the readers in Group Two) among the readers in the number of the perceived ostensive points they depend on to interpret the horizontal style.

5.3 Readers Attitudes Towards the Two Poetic Styles


In this section, we intend to examine the readers attitudes towards the vertical style and the horizontal style by using both the quantitative and the qualitative data to address the third research question as well as to reinforce our explanation in 5.2 and 5.3, in which we have made a great effort to deal with the other two research questions raised in terms of two hypotheses from the theoretical analysis in Chapter Three.

52

Table 5-5: Readers attitudes towards the two poetic styles Q&A Group Group One (n = 35) Group Two (n = 35) Question 4 A 28 80.00% 22 62.86% B 7 20.00% 13 37.14% Notes: A = the vertical style; B = the horizontal style; Q & A = question and answer; Question 4 = Question 4 in the questionnaire.

Table 5-4 describes how the readers answered the first part of Question 4 in the questionnaire. The first part of Question 4 was designed to investigate which form readers like more and which form, in readers opinion, can convey the given theme better so as to complement their responses to Question 3 in the questionnaire. Table 5-4 shows that in Group One, 28 (80%) readers thought the vertical style could convey the theme better and only 7 (20%) readers thought the horizontal style could convey the theme better while in Group Two, 22 (62.86%) readers thought the vertical style could convey the theme better and 13 (37.14%) readers thought the horizontal style could express the theme better. For those readers in both groups who chose the vertical style (A), their reasons were mainly summarized as follows: First, the vertical style is more creative in form, which is very novel and special. The excerpts from G1-S5 and G2-S5 are the typical accounts of the summary. G1-S5: G2-S5: A is novel in form and more importantly, A creates a kind of loneliness with creative arrangement of the letters and words. The form of A is novel. The complete words are cut and rearranged. At the first glance, this kind of separation gives readers a deep impression. A places loneliness at the beginning and at the end of this poem, which creates a large span. In addition, a leaf falls is bracketed into the loneliness. This indicates that the process of a leaf falls is lonely. And this also expands readers imagination.

Second, the vertical style contains visual stimulus, which creates a dynamic picture of a leaf falling down. The excerpts from G1-S6 and G2-S6 clearly give this point. G1-S6: The shape of A is like the trace of a leaf falling down. Compared with B, A is more implicit in expressing the content. Besides, A puts a leaf falls into brackets, which strongly emphasizes loneliness. A conveys loneliness not only from the content but also from the form. The vertical arrangement of letters, which were separated from the complete words, is like the process of a leaf falling down.
53

G2-S6:

Although the vertical style is very difficult, the fact is that most readers in both groups thought the vertical style could convey the theme better than the horizontal style. This verifies Yarons argument that despite the extreme difficulty of the poem, readers do not reject it (2002, p. 165). For those readers in both groups who chose the horizontal style (B), there are also two reasons. First, the horizontal style is very simple and is very easy to understand. It is very direct in expressing the theme. The following two excerpts from G1-S7 and G2-S7 are the typical account of this kind. G1-S7: G2-S7: Two short lines in B lead readers to very rich imagination. When I read B, I can imagine the situation in which a leaf is falling down. Based on the image, I can experience loneliness directly. However, when I read A, I notice the separated letters but cannot find out the link between the letters if I dont refer to B.

Second, compared with the horizontal style, the vertical style is too complicated and it is very difficult to understand. The following two excerpts from G1-S8 and G2-S8 are the typical reflection of this idea. G1-S8: G2-S8: A plays language game, and what A is meant to express need readers to guess. A is too obscure and its form is too abstract.

To sum up, for both groups of readers, the reason why they chose the vertical style is that the vertical style is very novel and it creates special kind of effects by splitting the complete words and rearranging the separated letters. For both groups of readers, the reason why they chose the horizontal style is that the horizontal style is very simple and very easy to understand. Moreover, most readers in both groups like the vertical style and are tolerant of the difficulty and obscurity of it in their interpretation.

54

Chapter Six

Conclusion

This chapter, which serves to wind up the whole research, consists of four sections. The first section briefly summarizes the present study. The second section concerns itself with the major findings of the present study. The third section presents the implications of the study. The last section points out the limitations of the study and offers some suggestions for future research.

6.1 Summary of the Study


The present study aimed to investigate how visual poetry achieves poetic effects from a pragma-cognitive perspective. One of Cummings poems entitled L(a was used to conduct the study. The study was carried out in two steps: one is theoretical and the other is empirical. That is, the study first analyzed how L(a achieves its poetic effects from a pragma-cognitive perspective theoretically and then an experiment was designed to validate the theoretical analysis. Based on the theoretical analysis, we arrived at two hypotheses: (1) The vertical style as an ostension contains more ostensive points than the horizontal style does. More ostensive points will encourage the reader to supply more contextual assumptions and access more encyclopedic entries. As a result, more weak implicatures will be derived from the ostension, which in turn, will lead to more poetic effects. (2) The richer the readers cognitive environment is, the more weak implicatures he will obtain. Since the individual readers cognitive environments are different, it is
55

impossible for different readers to access the same number and type of encyclopedic entries, to supply the same number and type of contextual assumptions and to derive the same number and type of weak implicatures. Aiming to test the above hypotheses and to examine readers attitudes towards visual poetry, three research questions were raised and a questionnaire was designed to investigate how L(a achieves its poetic effects and how the readers cognitive environments influence their interpretation of visual poetry in actual reading as well as how readers react to visual poetry. Two groups of readers with major differences in English literature background from Nanjing University were chosen as subjects to finish the questionnaire. Their answers to some questions were scored by the researcher on the basis of the number of the ostensive points they perceive from the two different poetic styles, which share the same propositional form. Then the scores were put into SPSS 11.5 for statistical analysis. The readers responses to some other questions in the questionnaire were collected as the qualitative data.

6.2 Major Findings


The research yielded the following results: (1) The poetic effects of a poem are fundamentally determined by the number of ostensive points it contains. The vertical style, containing more ostensive points, communicates more poetic effects than the horizontal style does. In other words, it is the large number of the ostensive points contained in L(a that makes it significantly different from the horizontal style in achieving poetic effects. (2) Readers cognitive environment plays a very important role in their interpretation of a poetic discourse. Due to different cognitive environments, different readers perceive different numbers of ostensive points and consequently derive different numbers and types of weak implicatures from the same poetic discourse. (3) In interpreting visual poetry, it is the specific assumptions about the visual poetry but not the general knowledge of literature that determines whether readers can arrive at a

56

successful reading. (4) The difficulty of visual poetry is recognized by most readers, who are nevertheless tolerant of such difficulties in their interpretation.

6.3 Implications of the Study


The present study sheds some light on the exploration of visual poetry, the understanding of RT, and literature teaching as well. It yields the following theoretical and pedagogical implications: First, the present study concentrates on the exploration of visual poetry by examining L(a as a case. The analysis of L(a from a pragma-cognitive point of view probes into the nature of visual poetry, thus facilitating readers understanding of such discourses. In addition, the model which was used to analyze L(a in the present study can also be applied to the analysis of other visual poems. Second, the present study provides evidence for the argument that the existence of visual poetry is reasonable and valuable from a pragma-cognitive point of view. The reason why visual poems can achieve more poetic effects than their linear counterparts, which can often be reproduced, lies in a great many ostensive points created by both the linguistic stimulus and the visual stimulus conveyed by them. However, this does not mean linear (or horizontal) poems have their vertical counterparts. Third, the present study carries out an empirical test to validate the theoretical analysis on the basis of RT. This will serve as an indirect test of RT with a view of promoting our understanding of the theory. RT as a pragmatic theory can be applied to the exploration of literary discourses, including visual poetry. Therefore, the scope of the application of pragmatic theory can be extended and the understanding of literary discourses can be facilitated. Finally, this study may offer some suggestions to poetry teaching by examining the process of the readers interpretation. In teaching a visual poem, teachers can first rewrite it in the linear form, and then ask students to compare two different forms to figure out the similarities and differences between them, which can facilitate the students

57

understanding of the visual poem. Meanwhile, teachers should encourage students to imagine on the basis of the ostension shaped by the visual poem to derive possible implicatures and obtain possible interpretations.

6.4 Limitations of the Study and Suggestions for Future Research


The study is a preliminary attempt to explore, from theoretical and empirical perspectives, what elements in visual poems determine their poetic effects and in what way visual poems achieve their poetic effects. Though the researcher designed it as carefully as he could, there are still some drawbacks, which are to be overcome in future research. First, in the empirical study, the subjects in Group Two were first-year graduate students majoring in English linguistics or English literature. Although those subjects whose major is English linguistics have taken English literature course for one or two years during their undergraduate study, their assumptions about English literature have more or less faded away from their memory as time goes by. This might influence the finding that individual differences between the two groups of readers are very close in the number of the ostensive points perceived from the vertical style and the horizontal style respectively. Therefore, it will be much better to use subjects for Group Two whose major is English literature instead in the future study. Second, it may be the case that the readers put more effort in interpreting the vertical style because of its difficulty and obscurity. Meanwhile, the readers must have made much more effort to supply more contextual assumptions and explore more encyclopedic entries to derive more weak implcatures from the vertical style than from the horizontal style. One question thus comes up, that is, how much effort is needed in perceiving an ostensive point by an individual reader? Our empirical evidence did not provide answers to this question. This question can be dealt with in the future study by examining the time consumed by the reader to finish the task.

58

Third, we mentioned in 5.1 that the readers in both groups perceived the relatively lower amount of ostensive points from the vertical style in actual interpretation compared with a great many ostensive points contained in it. The possible reason is that the readers might have perceived some ostensive points but they did not write them down, since they are required to associate what they have perceived with the theme of loneliness, which is one of the salient themes of L(a. As a result, the readers might write down only those ostensive points, which they thought can be associated with the given theme. However, since L(a as a poetic discourse is highly indeterminate, there may be several possible interpretations of it. Therefore, natural interpretation (i.e. the readers interpretation happens naturally and should not be limited to one possible theme of the poem) should be taken into consideration in future research and it is possible to employ the way of thinkaloud to investigate the readers natural interpretation of L(a. Fourth, the ways the reader takes in interpreting two poetic styles may be different. The two poetic styles may have been stored with different types of representations in our brain, so readers may pick up the two poetic styles (i.e. the vertical style and the horizontal style) from their memory differently. The difference is worth examining by means of a think-aloud method. Last but not least, although L(a is one of the typical visual poems, it still seems that using only one visual poem in the empirical test is a little weak to validate the theoretical analysis. To make the empirical study more valid, more visual poems need be involved in the future study.

59

References

Barney, T. (1996). Phonetics and the empirical study of poetry. In R. J. Kreuz & M. S. Macnealy (Eds.), Empirical approaches to literature and aesthetics (pp. 309-328). New Jersey: Ablex Publishing Corporation. Blakemore, D. (1992). Understanding utterances. Oxford: Blackwell Publishers. Blakemore, D. (1993). The relevance of reformulations. Language and Literature, 2 (2), 101-120. Brooks, C. & Warren, R. P. (Eds.) (1997). Understanding poetry. Beijing: Foreign Language Teaching and Research Press. Carston, R. & Uchida, S. (Eds.) (1997). Relevance theory: Applications and implications. Amsterdam & Philadelphia: John Benjamins Publishing Co. Chrzanowska-Kluczewska, E. (2001). Semantic and pragmatic language-games in poetry. Journal of Literary Semantics, 30, 181-197. Cook, G. (1999). Discourse and literature. Shanghai: Shanghai Foreign Language Education Press. Cummings, E. E. (1985). L(a. In G. McMichael (Ed.), Anthology of American literature (pp. 1223). New York: Macmillan Publishing Company. Cureton, R. D. (2002). Temporality and poetic form. Journal of Literary Semantics, 31, 37-59. Duan, J. W. []2002. (3)96-102
60

Fabb, N. (1997). Linguistic and literature: Language in the verbal arts of the world. Oxford: Blackwell Publishers Ltd. Fowler, R. (1981). Literature as social discourse. London: Batsford Academic and Educational Ltd. Gibbs, R. W. Jr. & Nascimento, S. B. (1996). How we talk when we talk about love: Metaphorical concepts and understanding love poetry. In R. J. Kreuz & M. S. Macnealy (Eds.), Empirical approaches to literature and aesthetics (pp. 291-307). New Jersey: Ablex Publishing Corporation. Gibbs, R. W. Jr. (2002). Identifying and appreciating poetic metaphor. Journal of Literary Semantics, 31, 101-112. Goodblatt, C. (2001). Adding an empirical dimension to the study of poetic metaphor. Journal of Literary Semantics, 30, 167-180. Gross, S. (1997). The word turned image: Reading pattern poems. Poetics Today, 18, 1532. Hanauer, D. (1998). Reading poetry: An empirical investigation of Formalist, Stylistic, and Conventionalist claims. Poetics Today, 19, 565-580. He, Z. R. & Ran, Y. P. (Eds.). [, ]2001 . Hiraga, M. K. (2005). Metaphor and iconicity: A cognitive approach to analyzing texts. Chippenham and Eastbourne: Antony Rowe Ltd. Jacobson, R. (1960/2000). Linguistics and poetics. In D. Lodge & N. Wood (2nd Eds. rev.), Modern criticism and theory: A reader (pp. 31-56). Singapore: Pearson Education Asia Ltd. Kidder, R. M. (1979). E. E. Cummings: An introduction to the poetry. New York: Columbia University Press. Levin, S. R. (1976). Concerning what kind of speech act a poem is. In T. A. van Dijk (Ed.), Pragmatics of language and Literature (pp. 141-159). New York: NorthHolland Publishing Company. Luo, Y. M. []2003 . (5)345351
61

Mackenzie, I. (2002). [Review of Poetic effects: A relevance theory perspective]. Journal of Literary Semantics, 31, 199-209. Pilkington, A. (1991). Poetic effects: A relevance theory perspective. In R. D. Sell (Ed.), Literary pragmatics (pp. 44-61). London & New York: Routledge. Pilkington, A. (2000). Poetic effects: A relevance theory perspective. Amsterdam & Philadelphia: J Benjamins Publishing Co. Richards, I. A. (1929). Practical criticism: A study of literary judgment. New York: Harcourt Brace Jovanovich. Rouchota, V. & Jucker, A. H. (Eds.) (1998). Current issues in relevance theory. Amsterdam & Philadelphia: J Benjamins Publishing Co. Sperber, D. & Wilson, D. (1986). Relevance theory: Communication and cognition. Oxford: Blackwell Publishers Ltd. Sperber, D. & Wilson, D. (1995). Relevance theory: Communication and cognition (2nd ed. rev.). Oxford: Blackwell Publishers Ltd. Steen, G. & Schram, D. (1996). Literary aspects of reading and thinking out aloud. In R. J. Kreuz & M. S. Macnealy (Eds.), Empirical approaches to literature and aesthetics (pp. 329-338). New Jersey: Ablex Publishing Corporation. Steen, G. (1991). Understanding metaphor in literature: Towards an empirical study. In R. D. Sell (Ed.), Literary pragmatics (pp. 94-109). London & New York: Routledge. Tsur, R. (2000). Picture poetry, mannerism, and sign relationships. Poetics Today, 21, 751-781. Tsur, R. (2002). Aspects of cognitive poetics. In E. Semino & J. Culpeper (Eds.), Cognitive stylistics: Language and cognition in text analysis. Amsterdam & Philadelphia: J Benjamins Publishing Co. Tu, J. []2002 . (6)8184 van Dijk, T. A. (1976). Pragmatics and poetics. In T. A. van Dijk (Ed.), Pragmatics of language and Literature (pp. 23-57). New York: North-Holland Publishing Company. van Peer, W. (1991). But what is literature? Towards a descriptive definition of literature.
62

In R. D. Sell (Ed.), Literary pragmatics (pp. 127-141). London & New York: Routledge. Verdonk, P. (1991). Poems as text and discourse: The poetics of Philip Larkin. In R. D. Sell (Ed.), Literary pragmatics (pp. 94-109). London & New York: Routledge. Wilson, D. & Sperber, D. (1993). Linguistic form and relevance. Lingua, 90, 1-25. Wilson, D. (1994). Relevance and understanding. In G. Brown et al. (Ed.), Language and understanding (pp. 37-58). Oxford: Oxford University Press. Yaron, I. (2002). Processing of obscure poetic texts: Mechanisms of selection. Journal of Literary Semantics, 31, 133-170. Yaron, I. (2003). Mechanisms of combination in the processing of obscure poems. Journal of Literary Semantics, 32, 151-166. Ye, H. []2003. (3)138-141 Yus, F. (1998). A decade of relevance theory. Journal of Pragmatics, 30, 305-345. Yus, F. (2002). [Review of Poetic effects: A relevance theory perspective]. Journal of Pragmatics, 34, 619-628. Zhang, B. H. []2001 . (1)48-49 Zhang, Y. []2001 . (5)41-44 Zhao, C. L. []1994 L(a . (4)64-66 Zwaan, R. A. (1993). Aspects of literary comprehension: A cognitive approach. Amsterdam & Philadelphia: John Benjamins Publishing Company.

63

Appendix
Questionnaire
__________ __________ __________

A B A l(a le af fa ll s) one l iness A leaf falls: loneliness. B

1 1 A 2 B 2 A

3 1A A

64

________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ 2B B ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ 4 ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ ________________________________________________________________________________ A B

65

66

Das könnte Ihnen auch gefallen