The Effects of Free Indirect Style in George Eliot's ‘Middlemarch’

15 PAUL SOPCAK, DON KUIKEN, AND DAVID S. MIALL The Effects of Free Indirect Style in George Eliot's Middlemarch: A Reader Response Study Introduction George Eliot's novel Middlemarch (1871) has been dismissingly characterized as the "classic realist text" for what is perceived as its presentation of the relationship between language and reality as unproblematically mimetic and for its neutralization of multi-perspectivism and polyvocality (Bakhtin 1981) through the presence of an omniscient narrator subordinating all competing subjectivities (voices and perspectives; MacCabe 1979, 15-18).1 Just as Bakhtin's concept of polyvocality is also epistemological, so is the criticism leveled at Middlemarch. In this line of argument, Eliot's work, as a prototype of the Victorian novel, not only gives its readers the illusion that language provides a direct, objective, and unambiguous "window on reality" (MacCabe 1978, 15), but it also represents a classical episteme. According to Foucault, the classical episteme is characterized by a naïve belief in the objectivity of knowledge, as yet undisturbed by the epistemological doubt, constructivism, subjectivism, multiperspectivism, and polyvocality of the modern episteme (1970, 58-61; 244-73). He uses spatial imagery to differentiate (a) the hierarchical organization of knowledge around one center in a unified space (classical episteme) from (b) fragmented organization around multiple centers and dispersed rays in a fractured space (modern episteme).2 Such an unambiguous attribution to Middlemarch (as the "classic realist novel") of these alleged shortcomings has met with resistance on several levels (e.g., Roberts 1982; Miller 1987). For the purpose of this paper, we will focus on claims regarding the narrator, specifically those rejecting the portrayal that an omniscient and 1 2 A more favorable view of Middlemarch is presented in Alan Palmer's (2005; 2010a; 2010b) discussions of the novel from the perspective of socially distributed cognition, or "intermental thought." Palmer argues that the collective inhabitants of the town of Middlemarch in the novel function like an extended social mind. Since Foucault's terminology was directly adopted in the questionnaires of the empirical studies presented below, we provide the relevant passage from The Order of Things in full length here: "It is thus apparent that the theory of sub-kingdoms does not simply add a supplementary taxonomic frame to the previous traditional classification; it is linked to the constitution of a new space of identities and differences. A space without essential continuity. A space that is posited from the very outset in the form of fragmentation. A space crossed by lines which sometimes diverge and sometimes intersect. In order to designate its general form, then, it is necessary to substitute for the image of the continuous scale, which had been traditional in the eighteenth century, […] that of a radiation, or rather of a group of centres from which there spreads outwards a multiplicity of beams" (1970, 272). Anglistik: International Journal of English Studies 31.1 (Spring 2020): 15-29. Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) PAUL SOPCAK, DON KUIKEN, AND DAVID S. MIALL omnipotent narrator is the sole organizing principle that subordinates all other subjectivities (voices and perspectives). The presence of such a subordinating, omniscient narrator, who has often been associated with the guard of the central watch tower of Bentham's Panopticon (e.g., in Bender 1987, 203; Miller 1988, 24; and Seltzer 1984, 54),3 would effectively place Middlemarch in the classical episteme described earlier. According to J. Hillis Miller, for instance, one feature of Eliot's writing contributing to its multi-perspectivism and polyvocality is her constant shifting between the implicit gender of the narrator: "[…] readers of Middlemarch or Eliot's work as a whole will know that a contrast between male and female imaginations is a major feature of her work" (1987, 68).4 Another observation along these lines, which applies not only to Eliot's novel but the Victorian novel in general, is that their narrator is no longer truly omniscient (Roberts 1982, 43). Rather than consistently functioning as an appropriating conduit for characters' subjectivities, the narrator's perspective and voice, although pervasive, stand alongside other, co-present, subjectivities in the narrative presentation. Miller puts this point as follows: The term 'omniscient narrator' has tended to obscure clear understanding of the narrating voice in Victorian fiction. The theological overtones of the word "omniscient" suggest that such a narrator is like a God, standing outside the time and space of the action, looking down on the characters with the detachment of a sovereign spectator who sees all, knows all, judges all, from a distance. The narrators of Victorian novels rarely have this sort of omniscience. The perfect knowledge is rather that of pervasive presence than that of transcendent vision. (1968, 63-64) Powered by TCPDF (www.tcpdf.org) Winter Journals for personal use only / no unauthorized distribution 16 Like Miller, but for different reasons, David Lodge (1992) rejects the notion that the Victorian novel's omniscient narrator represents the classical episteme with its related naïve "window on reality" theory of language. Lodge argues directly against MacCabe's notion that the omniscient narrator's subjectivity in Middlemarch functions as the supreme, univocal ordering principle to which all other "discourses" are subordinated. He suggests instead that "free indirect speech" complicates such univocal categorization: If we are looking for a single formal feature which characterises the realist novel of the nineteenth century, it is surely not the domination of the characters' discourses by the narrator's discourse […] but the extensive use of free indirect speech, which obscures and complicates the distinction between the two types of discourse. (Lodge 1992, 52) Whether or not readers perceive this co-presence of subjectivities that "free indirect speech" allegedly introduces, and if they do, whether it has a dispersion effect resulting in the lack of any privileged conscious subjectivity, as Lodge would have it, is an open empirical question. In this paper, we will present two empirical studies aimed at contributing to the preceding discussion by examining whether free indirect 3 4 We thank Jan Alber for pointing this out to us. On this aspect of Eliot's writing, see also Schabert (1992). Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) THE EFFECTS OF FREE INDIRECT STYLE IN GEORGE ELIOT'S MIDDLEMARCH 17 style (henceforth FIS5) introduces actual readers to co-present subjectivities (multiperspectivism and/or polyvocality), and if so, whether it introduces the "gnawing epistemological doubt" (Lodge 1992, 47) through a dispersion of subjectivities that would warrant seeing the Victorian novel as the transitional genre Roberts holds it to be (1982, 43). We will first provide a definition and example of FIS, followed by a brief discussion of terminology and an overview of the scholarly debate regarding its effect. Free Indirect Style and Its Near Neighbors Theories of free indirect style (FIS) have provided and continue to provide diverging accounts of its characteristics and of its effects on the reader. This is at least in part due to terminological disagreement and imprecision and, at times, conflation of free indirect speech with free indirect thought and discourse. Following Eric Rundquist (2017) and Violeta Sotirova (2006; 2011; 2013), we take FIS to encompass not only these different forms of indirect speech and thought presentation, but rather represented subjectivity or consciousness in general. We adopt Rundquist's definition of FIS as: […] the unsubordinated expression of a character's subjectivity alongside narratorial deictics for tense and person. It represents the consciousness of a third-person subject in a language that is not necessarily their own. Often, the 'speaker' or locutionary agent of the language, whether narrator or author, is obfuscated and all but effaced by a character's subjectivity, so that one has the impression of gaining direct, unmediated access to the character's mind through the mimetic, representational function of the discourse. (2017, 45) The following example from the opening pages of Eliot's Middlemarch, which formed part of the passages that participants in the studies reported here responded to, will serve to illustrate FIS. The description of the different FIS criteria follows Brinton (1980), as presented in Sotirova (2006). The passage presents the protagonist Dorothea's reflections on the prospects of marriage for herself and her sister Celia: That he should be regarded as a suitor to herself would have seemed to her a ridiculous irrelevance. Dorothea, with all her eagerness to know the truths of life, retained very childlike ideas about marriage. She felt sure that she would have accepted the judicious Hooker, if she had been born in time to save him from that wretched mistake he made in matrimony; or John Milton when his blindness had come on; or any of the other great men whose odd habits it would have been glorious piety to endure; but an amiable handsome baronet, who said "Exactly" to her remarks even when she expressed uncertainty, – how could he affect her as a lover? The really delightful marriage must be that where your husband was a sort of father, and could teach you even Hebrew, if you wished it. (Eliot 1997, 10) In this passage there is a gradual shift from the narrator's discourse (i.e., diegesis in Plato's sense) to the character's discourse (i.e., mimesis in Plato's sense). In the first 5 More on this terminological choice and how it relates to free indirect speech and discourse below. Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) 18 PAUL SOPCAK, DON KUIKEN, AND DAVID S. MIALL sentences, as well as those preceding this passage, Dorothea's subjectivity is clearly presented through the internal focalization of a third-person narrator. Gradually, this narratorial mediation and subordination is replaced with an increasingly mimetic and direct presentation of her subjectivity. Arguably, this introduces multi-perspectivism and polyvocality in which readers may perceive the co-presence of two autonomous subjectivities, the narrator's and Dorothea's. Drawing on Brinton and Sotirova, the following paragraphs will present a few of the linguistic markers of FIS responsible for this shift. In the first sentence of the presented passage, use of the third-person reflexive pronoun "herself" (instead of the personal pronoun "her") begins slightly to undermine subordination of the expression of Dorothea's subjectivity to that of the narrator. The sentence immediately following reads like a textbook example of an omniscient third-person narrator by referring to the protagonist by name (rather than using a pronoun). That mode of reference is a clear expression of a somewhat patronizing narratorial subjectivity, telling the reader about Dorothea from an omniscient perspective. Besides the stylistic markers, the content of the second sentence clearly establishes the narrator's superordinate subjectivity, since it comments on a naïveté of Dorothea's that she herself is not aware of. However, in the sentence beginning with, "She felt sure that she would have accepted the judicious Hooker," Dorothea's subjectivity, as expressed through her voice, thoughts, feelings, or perceptions, begins to emerge as an unsubordinated expression. This becomes clear not only from what is presented, namely Dorothea's somewhat naïve and immature Puritan passion mingled with a penchant for pathos, but also how it is presented stylistically. The proper name is replaced by the thirdperson pronoun "she" and becomes the subject of consciousness, and it presents a veritable barrage of "lexical items that express the character's emotions, attitudes, judgements, evaluations and beliefs" (Sotirova 2006, 112). This mode of expression seemingly provides unmediated access to Dorothea's thoughts, feelings, and perceptions, as expressions of her subjectivity ("judicious Hooker," "wretched mistake," "great men whose odd habits it would have been glorious piety to endure; but an amiable handsome baronet; delightful; even;" our italics). Moreover, after "but an amiable handsome baronet," present time deictics ("even when she expressed uncertainty") are woven into the narrative past tense ("who said 'Exactly'") "to suggest simultaneity of the [depicted] moment of consciousness with an event in the narrative past" (Sotirova 2006, 111). Lastly, the FIS passage quoted above includes repetitions, clauses with initial conjunctions ("or"), a "non-embeddable, independent clause of direct quotation" (Sotirova 2006, 111) in the form of a direct (rhetorical) question ("how could he affect her as a lover?"), as well as the inclusive use of the second-person pronoun "you" ("The really delightful marriage must be that where your husband was a sort of father, and could teach you even Hebrew, if you wished it."). Arguably, these FIS features collectively give the impression that we are gaining access to the immediate, unsubordinated expression of Dorothea's subjectivity. Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) THE EFFECTS OF FREE INDIRECT STYLE IN GEORGE ELIOT'S MIDDLEMARCH 19 Whether FIS, such as present in the passage above, introduces two autonomously co-present subjectivities, and by extension metalinguistic polyvocality and perspectival indeterminacy through dispersion, as Lodge, Miller, and Roberts argue, has been a topic of sustained debate. Although the positions within this debate are normally referred to as the single (e.g., Fludernik 1993; 2009; Gunn 2004) vs. dual voice perspective (e.g., Ducrot et al. 1991; Pascal 1977; Sotirova 2006), Rundquist has recently suggested that it is crucial to move away from a model of FIS within which all consciousness and subjectivity takes linguistic form. In his view, there is much to be gained from expanding the concept to include free indirect perception, which potentially includes characters' non-linguistic experiences besides speech, discourse, and thought (2017, 1-63; see also Banfield 1982; 1991). He argues that "the narrator and character can co-exist simultaneously in FIS as two subjects. Instead of instantiating a 'dual voice' [polyvocality], these situations are more adequately described as dual subjectivity" (52; original emphasis). The perhaps most well-known scholar to reject the notion that FIS may present the reader with two autonomous and co-present subjectivities is Fludernik (1993). In her view, the potential ambiguity FIS creates in relation to whose voice or subjectivity is presented is not to be resolved by linguistic description. Rather, she claims, "the reader's inferencing activity" (Fludernik 1993, 452) will establish the passage as either the narrator's expression or a character's utterance. The few empirical studies investigating readers' "inferencing activity" in response to FIS (e.g., Bray 2007; Fletcher and Monterosso 2016; Hakemulder and Koopman 2010; Sotirova 2006) have not provided conclusive evidence in favor of either the single- or the dual-voice (subjectivity) hypothesis. In our view, the ambiguous results of these studies are partly due to the fact that they have not followed Banfield's (1982; 1991) and Rundquist's (2017) lead in shifting to the language of 'subjectivity' as a more adequate description (and assessment) of what is at stake in FIS. Thus, when reconceptualizing the debate in these terms, one possibility is that in FIS the narrator represents the subjectivity of the character; consequently, the reader perceives the character's thoughts, feelings, and perceptions as mediated and presented through the narrator's voice and perspective (single voice/subjectivity hypothesis). However, another possibility is that FIS presents the reader with two autonomous co-present subjectivities, the narrator's and the character's (dual voice/subjectivity hypothesis). Rather than one of these functioning as "overarching center" (Banfield 1991, 28), these subjectivities are involved in a subtle interplay and introduce multi-perspectivism and polyvocality (Bakhtin 1981). One version of the second proposal, following Banfield (1982; 1991) and aligning with Foucault's discussion of the episteme, suggests that this interplay disperses the autonomous subjectivities, equally destabilizing the subjective "presence" of the narrator and the subjective "presence" of the character, which results in a lack of any privileged conscious subjectivity: The alternatives, therefore, counterpose not a theory centered on a single unitary subject and a polyphonic theory but rather one in which a plurality of isolated and noncommunicating points of view or centers coexist in a narrative style in which there is no first-person, single omniscient voice, imposing a personal unity, and one in which Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) PAUL SOPCAK, DON KUIKEN, AND DAVID S. MIALL 20 polyphony consists in a hierarchy of voices, each conceived on the model of the other, yet one providing a single, overarching center. (Banfield 1991, 28) In a second version (e.g., Rundquist; Sotirova), the FIS-induced co-presence of subjectivities creates a tension that nonetheless retains coherence. In what follows, we present two studies that, we think, help to clarify whether readers of FIS perceive the characters' subjectivity as subordinated and mediated through the narrator's subjectivity or, rather, autonomous co-present subjectivities, and, if the latter, whether they are dispersed or coherent. Study 1 Method Participants. Sixty-four introductory students of psychology at the University of Alberta and eighteen high school students at the Victoria High School for the Visual and Performing Arts participated in this study; the former for partial course credit, the latter without compensation. Fifty-eight were women (Mage 18.82, range = 15-22), and twenty-four were men (Mage 19.57, range = 16-23). Materials and Procedure. Participants read the first five pages of George Eliot's novel Middlemarch, divided into six passages of roughly equal length (ca. 350 words). We evaluated the presence or absence of six formal criteria of FIS, as identified by Brinton (1980). Although the source of these criteria is dated, its use in our study is supported by the fact that they still are regularly referred to in current scholarship as sound criteria for capturing the features of FIS (see, e.g., Sotirova 2006; Rundquist 2017). The six FIS criteria are presented in full length in Appendix A. Three raters, in addition to the first author of this paper, evaluated the presence or absence of each of these criteria for each passage, and differences were discussed until resolved. Subjectival Co-presence – Open-ended Response. In responding to these passages, participants described which subjectivities they perceived in each passage in an openended form, by answering the question: "Whose point(s) of view is/are presented in the passage?" A single page-width line was provided for this response. This measure aims at providing empirical data with which to address the single vs. dual voice (subjectivity) question, that is, whether FIS is perceived as presenting the subordinated character's subjectivity through that of the narrator, or alternatively, whether it establishes the co-presence of autonomous subjectivities. Subjectival Dispersion Rating Scale. A three-item, bipolar, mini-scale measures the extent to which subjectivities are perceived as dispersed vs. unified. This miniscale, which is anchored in the Foucault passage mentioned above, aims at providing empirical data for the discussion of whether FIS introduces a dispersion effect that destabilizes the subjective "presence" of both the narrator and the character (Banfield 1982; Lodge 1992). Foucault uses spatial imagery to differentiate (a) the hierarchical organization of knowledge around one center in a unified space from (b) fragmented organization around multiple centers and dispersed rays in a fractured space. The Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) THE EFFECTS OF FREE INDIRECT STYLE IN GEORGE ELIOT'S MIDDLEMARCH 21 mini-scale asked participants to respond to the question "Which image better describes the passage?" by rating three items on a seven-point scale between the following poles: "unified space vs. fractured space," "closely focused rays vs. multiple dispersed rays," and "one centre vs. multiple centres." The average internal consistency of this three-item mini-scale was acceptable (α = .692). Results Of the six passages, the fifth is the one shaped by FIS. The other passages are written in more straightforward third-person narration, with only very occasional presence of the character's voice. To provide an appropriate contrast, we calculated the mean for all of our measures for passages 1 through 4 and 6, and compared these means with the responses to passage 5. Subjectival Co-presence. We coded the open-ended questions by marking references to the narrator with the number 1, to Dorothea with number 2, and references to multiple subjectivities with a number 3. In response to the low-FIS passages (1-4, and 6), on average 71% of participants perceived the narrator's subjectivity, 6% perceived Dorothea's subjectivity, and 23% perceived multiple subjectivities. In contrast, in the high-FIS passage (5), 44% of participants perceived the narrator's subjectivity, 28% perceived Dorothea's subjectivity, and 28% perceived multiple subjectivities. These findings clearly point toward an increased perception of Dorothea's subjectivity in the high FIS-passage, and a slightly higher polyvocality or co-presence of multiple subjectivities. Subjectival Dispersion. The FIS-established co-presence in passage 5 did not lead to the perception of dispersed subjectivities. Rather, passage 5 was perceived as the least dispersed and most unified of all six passages, as the responses to the Subjectival Dispersion Scale presented in Figure 1 below indicate. Discussion of Results Although the narrator's subjectivity is pervasive in the low FIS and the high FIS passages, the fact that, on average, the narrator's subjectivity is perceived 27% less and Dorothea's 22% more, and that 28% of participants perceive multiple subjectivities in the high FIS passage as opposed to the low FIS passage, supports the notion that FIS establishes the co-presence of two subjectivities, at least at the textual level. But to what extent are these findings conclusive regarding the single vs. dual subjectivity debate? On the one hand, they clearly show that for 22% of the participants in this study, FIS undermined the presence of an omniscient narrator who subordinates all competing subjectivities. On the other hand, however, the 5% increase in the perception of multiple subjectivities does not conclusively indicate that readers perceive an FIS-induced co-presence, as opposed to establishing the passage as either an expression of the narrator's or an expression of Dorothea's subjectivity, as Fludernik (1993, 452) suggests. Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) 22 PAUL SOPCAK, DON KUIKEN, AND DAVID S. MIALL Figure 1: Mean Subjectivity Dispersion. Repeated measures analysis of variance indicated differences in the mean ratings of the six passages presented here, F(1, 77) = 4.77, p < .003; and post hoc comparisons indicated that, in passage 5 (high FIS), subjectivity dispersal ratings were lower than in all of the other passages. In other words, in the FIS passage, ratings of dispersed subjectivity were lower than in other passages. To replicate the co-presence findings, while also pursuing the latter distinction empirically, we conducted a second study with a slightly modified subjectival copresence measure. This second study served also as an attempt to replicate the seemingly counterintuitive findings that the FIS-induced subjectival co-presence did not lead to subjectival dispersion, as Lodge (1992) and Banfield (1982) suggest, but rather to increased coherence. Study 2 Method Participants. Eighty-three undergraduate psychology students at the University of Alberta participated in Study 2 for partial course credit. Fifty-five were women (Mage 22.85, range = 19-47), and twenty-seven were men (Mage 23.41, range = 18-30). Materials and Procedure. As in Study 1, participants read the opening five pages of Eliot's Middlemarch, divided into six passages and completed a number of rating scales after each passage. Subjectival Co-presence Ratings. In an effort to replicate the subjectival copresence findings, while also being able to determine to what extent each reader perceived co-present subjectivities, or rather chose between the narrator's and Dorothea's, as Fludernik (1993, 452) suggests, we optimized our measure of subjectival co-presence. Instead of replying in an open-ended form to perceived subjectivities, participants after each passage were asked to separately rate the extent Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) THE EFFECTS OF FREE INDIRECT STYLE IN GEORGE ELIOT'S MIDDLEMARCH 23 to which they perceived the narrator's subjectivity ("This passage seemed to reflect the narrator's point of view") as well as Dorothea's ("This passage seemed to reflect Dorothea's point of view"). These items were rated on a seven-point scale, ranging from 0 ("I don't agree at all") to 6 ("I totally agree"). We assessed the relative presence of the narrator's and the protagonist's subjectivities by calculating the difference between (1) readers' ratings of the extent to which the segment reflected Dorothea's (D) subjectivity and (2) their ratings of the extent to which the segment reflected the narrator's (N) subjectivity. Calculating this difference score allowed us to capture not only whether and when the subjectivities were perceived as equally present, but also the degree to which one outweighed the other, where no co-presence was perceived. This method provided us with an indication of how balanced the perceived co-presence was, while also eliminating the effect of individual differences in ratings, such as consistently rating perceived subjectivities low, for instance. Subjectival Dispersion Rating Scale. In Study 2, we employed exactly the same method and procedure to measure subjectival dispersion as we did in Study 1. Segment Level Analysis of Passage 5. To look at what happens not only at the passage level, but also at the sentence level within a passage of FIS, we asked participants to reread passage 5 (high FIS), subdivided into twelve roughly sentence length-segments (see Appendix B). After each of these twelve segments, they again completed the Subjectival Co-presence Ratings. Results Subjectival Co-presence. Repeated measures ANOVA indicate that the difference between the perceived subjectivity of Dorothea as opposed to that of the narrator was least in response to passage 5 (F(1, 81) = 12.95, p < .001) which means that just as in Study 1, FIS evoked the co-presence of autonomous subjectivities. Subjectival Dispersion. Another result replicated from Study 1 was the lower subjectivity dispersal ratings in the high FIS passage, as opposed to those low in FIS. Repeated measures ANOVA indicated differences in the mean ratings of the six passages on the dispersed subjectivity ratings (F(1, 81) = 7.52, p < .001) and post hoc comparisons indicated that these were lower in passage 5 (high FIS passage) than in all of the other passages (except passage 4). In other words, FIS contributed to the perception of coherence, rather than dispersion. Segment-level Co-presence. New in Study 2 were results for the subjectival copresence ratings for each of the twelve sentence-length segments that passage 5 (high FIS) was divided into. Repeated measures ANOVA indicated that the segments containing five or more of Brinton's (1980) FIS markers (segments 6, 9, 10, 11, and 12) had higher average co-presence scores than the segments that contained three or fewer FIS markers (segments 1, 2, 3, 4, and 5), F(1, 80) = 33.86, p < .001 (MlowFIS = –1.02; MhiFIS = 0.67). This confirms at the sentence level what we found on the passage level, namely that FIS introduces the co-presence of subjectivities. However, this overall contrast between high FIS and low FIS segments obscures a temporal structure that becomes evident when the alternating pattern of low FIS Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) PAUL SOPCAK, DON KUIKEN, AND DAVID S. MIALL 24 (segments 1-5), high FIS (segment 6), low FIS (segments 7-8), and high FIS (segments 9-12) is considered. 0.6 0.4 Co-Presence of Subjectivities 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 -1.2 -1.4 low FID high FID low FID high FID Figure 2: A visual representation of two cycles within passage 5, within each of which there is a movement from low- to high-FIS. The closer to zero, the more Dorothea's subjectivity was perceived as equally co-present as the narrator's; the more negative the values, the more the narrator's subjectivity was perceived, and vice versa. Finally, repeated measures ANOVA indicated that participants' ratings on the subjectival co-presence scores parallel the oscillating pattern of these two cycles of low- to high-FIS segments. Discussion of Results Just as in Study 1, we again found the passage level co-presence and dispersed subjectivity effects. This replication of our findings adds further support to the notion that in FIS, rather than an omniscient narrator subordinating all competing subjectivities, autonomous, co-present subjectivities are established. Our new measure of subjectival co-presence not only allowed us to replicate this effect, it further provided a measure of the balance and degree of perceived co-presence, and the findings undermine Fludernik's notion that "the reader's inferencing activity" (Fludernik 1993, 452) establishes an FIS passage as either the narrator's or the character's subjectivity. Our findings clearly showed that some readers perceive a copresence of both. The replication in Study 2 of our subjectival dispersion findings from Study 1 provided further evidence that the FIS-established co-presence provides coherence, rather than dispersion, as Lodge (1992) and Banfield (1982) would have it. And, lastly, the segment-level analysis of the effects of FIS provided insight into the temporal unfolding of the perceived co-presence of subjectivities, in which the Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) THE EFFECTS OF FREE INDIRECT STYLE IN GEORGE ELIOT'S MIDDLEMARCH 25 perceived co-presence developed in parallel to the degree of FIS in the segments of passage 5. General Discussion The results of the two studies presented here and the pattern of successful replication within them pose a direct challenge to the single-subjectivity (voice) hypothesis of FIS (e.g., Fludernik 1993; 2009; Gunn 2004), namely to the notion that in FIS the reader perceives the character's thoughts, feelings, and perceptions as mediated and presented through the narrator's subjectivity, or that readers will infer a single subjectivity where the text affords a co-presence. Instead, this series of replicated results supports the notion that FIS presents the reader with two autonomous copresent subjectivities. But does this FIS-induced polyvocality and multi-perspectivism undermine the coherence of the perceived (discursive) space, and introduce "a plurality of isolated and noncommunicating points of view or centers" that Banfield (1991, 28), or the "gnawing epistemological doubt" that Lodge (1992, 47) suggest? Our findings indicate that the opposite is the case. Already at the passage level, readers in our studies when asked, "Which image better describes the passage?" repeatedly chose the unified over the fractured space, the closely focused rays over the multiple dispersed rays, and one center over multiple centers, when responding to the passage high in FIS (passage 5). That is, FIS established the co-presence of two autonomous subjectivities and coherence, rather than fragmentation. This result was repeated at the sentence-level analysis: higher FIS led to higher co-presence. Worthy of further consideration and study, however, is the gradual unfolding of a reader's perception of this autonomous co-presence of subjectivities in FIS and how it establishes coherence rather than dispersion and fragmentation. What our segmentlevel findings begin to point toward is the importance of considering the temporal dimension of the reading experience. One related possibility in this regard is to consider the possibility that FIS introduces not two subjectivities, but three. Take, for instance, the following sentence from Middlemarch: "The really delightful marriage must be that where your husband was a sort of father, and could teach you even Hebrew, if you wished it" (Eliot 1997, 10). Arguably, the FIS features in this sentence, such as use of the inclusive second-person pronoun "you," invite the reader's own subjectivity into the mix during the unfolding reading experience. Such a proposal aligns with a phenomenological model of perception and reflection (e.g., Husserl 2004). Another relevant task for future empirical studies is the clarification of the role of irony in FIS and its effect on the kind of FIS-established subjectival copresence that builds coherence rather than dispersion, which we found in the studies presented here of readers' responses to Eliot's Middlemarch. Works Cited Bakhtin, Mikhail Mikhailovich. "Discourse in the Novel." The Dialogic Imagination. Four Essays. Ed. Michael Holquist. Transl. Caryl Emerson and Michael Holquist. Austin, TX: University of Texas Press, 1981. 259-422. Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) 26 PAUL SOPCAK, DON KUIKEN, AND DAVID S. MIALL Banfield, Ann. Unspeakable Sentences. New York: Routledge and Kegan Paul, 1982. Banfield, Ann. "L'écriture et le non-dit." Diacritics 21.4 (1991): 21-31. Bender, John. Imagining the Penitentiary: Fiction and the Architecture of Mind in Eighteenth-Century England. Chicago, IL: The University of Chicago Press, 1987. Bray, Joe. "The Effects of Free Indirect Discourse: Empathy Revisited." Contemporary Stylistics. Eds. Marina Lambrou and Peter Stockwell. London: Continuum, 2007. 56-68. Brinton, Laurel. "'Represented Perception:' A Study in Narrative Style." Poetics 9.4 (1980): 363-381. Ducrot, Oswald, Catherine Porter, Kara Rabbitt, and Linda Waugh. "Charles Bally and Pragmatics." Diacritics 21.4 (1991): 3-19. Eliot, George. Middlemarch. Oxford and New York: Oxford University Press, 1997. Fletcher, Angus, and John Monterosso. "The Science of Free-Indirect Discourse: An Alternate Cognitive Effect." Narrative 24.1 (2016): 82-103. DOI:10.1353/nar.2016.0004. Fludernik, Monika. The Fictions of Language and the Languages of Fiction: The Linguistic Representation of Speech and Consciousness. London and New York: Routledge, 1993. Fludernik, Monika. An Introduction to Narratology. New York: Routledge, 2009. Foucault, Michel. The Order of Things: An Archaeology of the Human Sciences. New York: Random House, 1970. Gunn, Daniel P. "Free Indirect Discourse and Narrative Authority in Emma." Narrative 12.1 (2004): 35-54. Hakemulder, Jèmeljan, and Emy Koopman. "Readers Closing in on Immoral Characters' Consciousness: Effects of Free Indirect Discourse on Response to Literary Narratives." Journal of Literary Theory 4.1 (2010): 41-62. DOI:10.1515/jlt.2010.004. Husserl, Edmund. Experience and Judgment: Investigations in a Genealogy of Logic. Evanston, IL: Northwestern University Press, 2004. Lodge, David. "'Middlemarch' and the Idea of the Classic Realist Text." Middlemarch: George Eliot. Ed. John Peck. Basingstoke: MacMillan, 1992. 45-62. MacCabe, Colin M.J. James Joyce and the Revolution of the Word. Basingstoke: Macmillan, 1979. Miall, David S., and Don Kuiken. "Aspects of Literary Response: A New Questionnaire." Research in the Teaching of English 29 (1995): 37-58. Miller, D.A. The Novel and the Police. Berkeley, CA: University of California Press, 1988. Miller, Hillis J. The Form of Victorian Fiction. Notre Dame, IN, and London: University of Notre Dame Press, 1968. Miller, Hillis J. The Ethics of Reading. New York: Columbia University Press, 1987. Palmer, Alan. "Intermental Thought in the Novel: The Middlemarch Mind." Style 39.4 (2005): 427-439. Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) THE EFFECTS OF FREE INDIRECT STYLE IN GEORGE ELIOT'S MIDDLEMARCH 27 Palmer, Alan. Social Minds in the Novel. Columbus, OH: The Ohio State University Press, 2010a. Palmer, Alan. "Large Intermental Units in Middlemarch." Postclassical Narratology: Approaches and Analyses. Eds. Jan Alber and Monika Fludernik. Columbus, OH: The Ohio State University Press, 2010b. 83-104. Pascal, Roy. The Dual Voice. Manchester: Manchester University Press, 1977. Roberts, Ruth. "Middlemarch and the New Humanity." George Eliot: A Centenary Tribute. Eds. Gordon S. Haight and Rosemary T. van Arsdel. London: Macmillan, 1982. 38-46. Rundquist, Eric. Free Indirect Style in Modernism: Representations of Consciousness. Amsterdam: John Benjamins, 2017. Schabert, Ina. "The Authorial Mind and the Question of Gender." Telling Stories: Studies in Honour of Ulrich Broich on the Occasion of His Sixtieth Birthday. Eds. Elmar Lehmann and Bernd Lenz. Amsterdam: Gruner, 1992. 312-328. Seltzer, Mark. Henry James and the Art of Power. Ithaca, NY, and London: Cornell University Press, 1984. Sotirova, Violeta. "Reader Responses to Narrative Point of View." Poetics 34.2 (2006): 108-133. Sotirova, Violeta. D.H. Lawrence and Narrative Viewpoint. London: Continuum, 2011. Sotirova, Violeta. Consciousness in Modernist Fiction: A Stylistic Study. New York: Palgrave Macmillan, 2013. Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) 28 PAUL SOPCAK, DON KUIKEN, AND DAVID S. MIALL Appendix A Brinton's (1980) Formal Criteria for Identifying Free Indirect Style, as Presented in Sotirova (2006, 111-112) (1) shifted pronouns: the third person she or he is the subject of consciousness; it is the referent of the expressive content of the sentence; use of pronoun rather than proper name; third person reflexive pronoun may occur even when there is no third-person sentence subject; (2) the narrative past tense and present and future time deictics are cotemporal to suggest simultaneity of the moment of consciousness with an event in the narrative past; special verbal past tense – imparfait, past progressive; shifted tenses of modals with past meaning, otherwise only found in Indirect Discourse where sequence of tenses is observed; (3) pronouns, demonstratives, definite articles, and definite noun phrases which have no antecedent in the previous discourse may occur; (4) contains the non-embeddable, independent clauses of direct quotation, such as direct questions and imperatives; contains rhetorical questions and clauses with preposed adverbs (never) or initial conjunctions (and); interjections, exclamations, lexical fillers, repetitions, hesitations, optative or incomplete sentences; (5) lexical items which express the character's emotions, attitudes, judgements, evaluations and beliefs; qualifying adjectives, generally prenominal (dear, good, damned), epithets, qualifying adverbs (probably, miserably), nicknames or petnames and attitudinal nouns (fool); (6) verbs of consciousness or of communication occur in parentheticals. Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) THE EFFECTS OF FREE INDIRECT STYLE IN GEORGE ELIOT'S MIDDLEMARCH 29 Appendix B Middlemarch Passage 5 Segmented (Eliot 1997, 9-10) 1. Yet those who approached Dorothea, though prejudiced against her by this alarming hearsay, found that she had a charm unaccountably reconcilable with it. Most men thought her bewitching when she was on horseback. 2. She loved the fresh air and the various aspects of the country, and when her eyes and cheeks glowed with mingled pleasure she looked very little like a devotee. 3. Riding was an indulgence which she allowed herself in spite of conscientious qualms; she felt that she enjoyed it in a pagan sensuous way, and always looked forward to renouncing it. 4. She was open, ardent, and not in the least self-admiring; indeed, it was pretty to see how her imagination adorned her sister Celia with attractions altogether superior to her own, 5. and if any gentleman appeared to come to the Grange from some other motive than that of seeing Mr. Brooke, she concluded that he must be in love with Celia: 6. Sir James Chettam, for example, whom she constantly considered from Celia's point of view, inwardly debating whether it would be good for Celia to accept him. 7. That he should be regarded as a suitor to herself would have seemed to her a ridiculous irrelevance. 8. Dorothea, with all her eagerness to know the truths of life, retained very childlike ideas about marriage. 9. She felt sure that she would have accepted the judicious Hooker, if she had been born in time to save him from that wretched mistake he made in matrimony; 10. or John Milton when his blindness had come on; or any of the other great men whose odd habits it would have been glorious piety to endure; 11. but an amiable handsome baronet, who said "Exactly" to her remarks even when she expressed uncertainty, – how could he affect her as a lover? 12. The really delightful marriage must be that where your husband was a sort of father, and could teach you even Hebrew, if you wished it. Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org) Anglistik, Volume 31 (2020), Issue 1 © 2020 Universitätsverlag WINTER GmbH Heidelberg Powered by TCPDF (www.tcpdf.org)