The Encyclopedia of the Novel - Peter Melville Logan 2014


John Frow

Fictional character takes many forms. We recognize Brer Rabbit or Reynard the Fox, the goddess Hera or Thomas the Tank Engine as quasi-persons, figures that move us in the way that only stories about a recognizably human destiny can move us. To “recognize” means both that we find a frame for understanding what kind of being this is, and that we see ourselves in these figures—“ourselves” only in the most general, anthropomorphic sense that we translate animals or gods or steam trains into human-like figures which will fit into our stories—and on that basis we project something more specific onto them, and draw something more specific from them. We like them or dislike them, identify with them or disapprove of them. We distinguish the white hats from the black hats, and we get emotionally involved with them.

The conditions for being a character are minimal because we have the capacity to turn just about anything into a quasi-person. Usually a character has a name; it speaks; and it performs an action or a series of actions, on the basis of which we impute intentionality to it. But even these minimal conditions need not all be met. The transmigratory soul in David Mitchell's Ghostwritten (1999) has no name; Bertha Mason, the madwoman in Charlotte Brontë's Jane Eyre (1847), never speaks; an anonymous, featureless Samuel Beckett character neck-deep in mud or garbage performs no action other than talking. What counts is less what they are than what we do with them: the historically, culturally, and generically various ways in which the reader or spectator or listener endows them with significance.

Humanist Theories of Character

Because the shape we give the raw materials of character is always a human shape, the easiest way of understanding character is as a displaced form of human being. This involves attributing unity, coherence, and psychological depth to the figures in a story and treating them as though they were separable from the texts which form them. Such a procedure has been heavily influenced by the literary techniques that work hardest at producing the illusion of setting in motion “real” human beings, and particularly by the techniques of the European novel from the eighteenth century onward.

A well-known image shows Charles Dickens daydreaming at his desk while his fictional creations float around him (see fig. 1). They have taken on a life of their own, becoming free to attract us or repel us regardless of their thematic function in the novels. The image exemplifies a widespread practice of reading; many of Dickens's readers wrote to implore him not to allow Little Nell to die, understanding her at once as a fictional construct and as a person to whom they were deeply attached. One of the great exponents of this way of reading is the Shakespearean critic A. C. Bradley, who reads the plays as psychological dramas and abstracts characters from texts to the extent that he can devote a long footnote to the question of whether Hamlet is too old to have been at University at Wittenberg at the time of his father's death: to examine this question as a matter of biographical fact, that is, rather than as matter of plot (403—9). Bradley's reading of Hamlet is concerned with what for him is the central question: that of explaining the motives for the hero's behavior and doing so in a way that makes it psychologically coherent and plausible.


Figure 1 “Dickens's Dream,” unfinished painting by R. W. Buss (1804—75) also known as “A Souvenir of Dickens,” 1875. Used with the permission of the Charles Dickens Museum

Character in this sense is a resource for moral analysis and is closely tied to literary pedagogies in which the analysis of ethical issues and dilemmas relating to literary characters—“what was the fatal flaw in Hamlet's character?”—forms the basis of an institutionalized practice for constructing “moral selves or good personal character” (Hunter, 233). The moral selves of fictional characters reflect and help shape our own. The theoretical challenge to which this humanist understanding of literary character responds is that of being able to explain with a single set of terms both the constitution of “real” moral subjects and the effects of unity which underpin literary character. The price paid for the continuity it posits between character and person is that both must be thought in terms of presence—of “real” personhood—rather than in terms of representation. Whatever the merits of this understanding and of the pedagogies that flow from it and support it (and it is arguable that uses of character are always bound up with practices of emulation), they work much better for the realist novel than for other kinds of text (see REALISM), and they offer little purchase for analyzing the textual and cultural conventions by which characters are constructed.

Although this ethical and humanist mode of criticism is still the dominant way of understanding character, both in literary criticism and in popular understandings, it has been extensively challenged. An influential essay by the British critic L. C. Knights, “How Many Children Had Lady Macbeth?” with its admonition that “in the mass of Shakespeare criticism there is not a hint that ’character’... is merely an abstraction from the total response in the mind of the reader or spectator, brought into being by written or spoken words; that the critic therefore—however far he may ultimately range—begins with the words of which a play is composed” (4), makes an important argument against the detextualization of character. This tradition in New Critical theory, which builds on an Aristotelian conception of character as a structural dimension of plot, works as a kind of bridge to structuralist accounts of character as a conventional construct, a textual effect rather than a quasi-real person.

Structuralist Theories of Character

We could perhaps date the structuralist approach to character to the work of Vladimir Propp, who studied a corpus very different from that of the novel (see STRUCTURALISM). His Morphology of the Folktale, published in 1928, analyzed the basic elements of folkloric narrative into thirty-one functions (generalized forms of action such as interdiction, interrogation, or flight, considered in abstraction from the characters who perform them) and seven basic character types: the hero; the villain; the donor, who prepares the hero for his quest; the helper, who assists the hero; the princess and her father (the two are structurally merged) who set the hero on his quest and reward him with marriage when it is completed; the dispatcher, who sends the hero on his way; and the false hero or usurper. Propp's argument, influenced by linguistic theory, where grammatical categories work as empty slots that are filled with particular content, is that the multiplicity of characters appearing in folk tales can be reduced to this underlying typology (not all of the elements of which will necessarily be present, and some of which may be merged; see LINGUISTICS). As Mieke Bal puts it, “an actor is a structural position, while a character is a complex semantic unit. But as readers, we ’see’ characters, only reducible to actors in a process of abstraction” (115). Although Propp's typology is by definition reductive and thus does not attempt to do justice to the texture, tone, and particularity of the tales he analyzes, the power of this reading lies in its capacity to isolate general patterns in narrative: to move beyond the particularities of a text to the abstract formal structures composing it.

Following Propp, writers such as Algirdas Julien Greimas, Roland Barthes, and Tzvetan Todorov seek to construct a grammar of narrative which will specify a typology of roles from which characters are generated. Greimas influentially proposed the concept of the actant: the slot or character-class defined by a permanent group of functions and qualities and by their distribution through a narrative; the term (later taken up by social scientists like Bruno Latour) carefully does not distinguish between human and nonhuman actors. Greimas's schema is derived from Propp's and posits the following general narrative logic (180):


In its focus on a logic of actions, however, this unpromisingly general schema lacks the universality it claims, and like Propp's it neglects all of those thematic and structural functions performed by narrative agents that are not simply acts.

A fuller and more interestingly synthetic account of fictional character is to be found in Philippe Hamon. Hamon's starting point is an argument against the confusion of personne (person) with personage (character), and against the neglect of the verbal conditions of existence for character; this neglect is the reason why a banal psychologism is to be found even in otherwise sophisticated analyses. Insofar as character “is as much the reader's reconstruction as a textual construct” (119), Hamon proposes that the object of analysis should be the “textual character-effect” (120). The model he proposes is that of the relation of the phoneme to its distinctive features, in terms of which character is conceived as “a bundle of relations of similarity, opposition, hierarchy and disposition (its distribution) which it enters into, on the plane of the signifier and the signified, successively and/or simultaneously, with other characters and elements of the work” (125). The signified of character, its “value,” is constituted not only by repetition, accumulation, and transformation, but also by its oppositional relation to other characters (128).

This definition—purely formal as it is—sets up the possibility of establishing a calculus of the features, the “characteristics” that represent the basic components of character. To this end Hamon constructs a number of tables which yield a differential analysis of qualities, functions, and modes of determination of character. Let me paraphrase his summary. A character can be defined:

· by the way it relates to the functions it fulfills;

· by its simple or complex integration in classes of character-types, or actants;

· as an actant, by the way it relates to other actants within well-defined types of sequences and figures (e.g., “quest” or “contract”);

· by its relation to a series of modalities (“wanting,” “knowing,” “being able to”);

· by its distribution within the whole narrative; and

· by the bundle of qualities and thematic “roles” which it supports.

The theoretical consequence of this definition is that, insofar as character is “a recurrent element, a permanent support of distinctive features and narrative transformations, it combines both the factors which are indispensable to the coherence and readability of any text, and the factors which are indispensable to its stylistic interest” (141—42). This then leads to a final definition of character: as “a system of rule-governed equivalences intended to ensure the readability of the text” (144).

Yet in practice the theory does not account for the reader's interest and desire as they operate to establish characters as quasi-persons; it fails to explain the affective force of the imaginary unities of character. Jonathan Culler makes a similar criticism: that structuralist theorists view the usual primacy given to character as an ideological prejudice, rather than trying to account for it (230). I want therefore to turn to the question of how character works to construct the “interest” of a story, its affective hold; this is the question of the relation between the construction of character and the construction of the reader as a reading subject.

Character and the Reader

Here the key concept is that of identification, the importance of which lies in its ability to mediate between character as a formal textual structure and the reader's structured investment in it. In Sigmund Freud's work, where the concept has most rigorously been analyzed, the concepts of identification and narcissism are closely linked, and I want briefly to investigate the relation between them (see PSYCHOANALYTIC). Narcissism is one of Freud's major explanatory categories, and he takes it to constitute the real basis of every object-choice: in every choice of a love-object “the libidinal energy is always borrowed from the ego, and always ready to return to it” (77). What is innovative in Freud's thought is not the postulation of a love of self, which in itself is a commonplace, but the fact that this is understood as occurring through the taking of the ego as a possible love-object, and the fact that the actual positions of subject and object may be less important than the fantasized positions (which may indeed both be internal to the ego).

Let me give an example taken from Freud's discussion of melancholia. In analyzing this neurosis Freud comes to the conclusion that the self-reproaches characteristic of melancholia are really “reproaches against a loved object which have been shifted away from it on to the patient's own ego” (1953, 14:248). This situation arises because of a withdrawal of libido from the lost object; however, the emotional energy is not then transferred to another object but rather is withdrawn into the ego, causing a splitting of the ego such that one part identifies itself with the abandoned object and is in turn judged and condemned by another part (249). It is to this dispersal of ego-identifications that I wish to liken the workings of fictional character. The “recognition” or “identification” of character would involve a mirroring of the semantic and libidinal processes of self-construction in an imaginary construction of “other,” quasi-unified selves.

This process of narcissistic dissemination of self-recognition, which I take to be the basis of all historically specific regimes of identification, is said by Freud to be characteristic of the language of dreams. The passage is worth quoting at length:

It is my experience, and one to which I have found no exception, that every dream deals with the dreamer himself. Dreams are completely egoistic. Whenever my ego does not appear in the content of the dream, but only some extraneous person, I may safely assume that my own ego lies concealed, by identification, behind this other person; I can insert my ego into the context. On other occasions, when my own ego does appear in the dream, the situation in which it occurs may teach me that some other person lies concealed, by identification, behind my ego. In that case the dream should warn me to transfer on to myself, when I am interpreting the dream, the concealed common element attached to this other person. There are also dreams in which my ego appears along with other people, who, when the identification is resolved, are revealed once again as my ego....Thus my ego may be represented in a dream several times over, now directly and now through identification with extraneous persons. (1953, 4:322—23)

Freud's work is rich in examples of such shape-shifting—think also of the analysis of the changing subject positions in “A Child is Being Beaten,” where the title sentence is transformed back through “my father is beating the child” and “my father is beating the child whom I hate” to “I am being beaten by my father” (17:179)—and I want to suggest that the play of dispersed identifications Freud uncovers can be mapped onto an account of the play of positions in discourse. What I mean by this is the way readers occupy (and thus “identify” with) distinct “voices” in the play of language, putting themselves in the place of the speaker of the text they are reading, and of the figured personages which both speak and are spoken about. Reading Joseph Conrad's Heart of Darkness (1902), I take on Marlow's perceptions and concerns (however far they might be from my own), even to the extent of thinking within the rhythms of his syntax; I place myself inside Kurtz's head (however concealed it is behind layers of rumor); I learn to loathe the administrators who thrive on the misery of the African laborers; and some part of me stands critically beyond Marlow's guiding of my judgment, making him in turn an object as well as a subject of understanding, assessing the limits of his sense of himself and the values he espouses, and placing him in the larger context of a shaped literary work.

Alex Woloch suggests that what he calls the “character system” of a novel, which regulates the relation between protagonists and “minor” characters, mirrors this process of scattering within the text as the minor characters work to develop different aspects of the protagonist by functioning as “foils, displacements, projections, and doubles” (127). Thus we could think of a text as involving multiple levels of projection and recognition through which the reader is bound into the text by working through his or her relation to characters that form part of an interwoven system of constantly shifting affective processes.

At the heart of the character system is the distinction between those who are subjects in their own right, with whom we identify, and those who are the objects of our perception. Heart of Darkness gives me no space to identify, except in the most general sense, with those nameless shapes of black men dying in a station in the Belgian Congo; whatever anger, disgust, or outrage we may feel is felt from Marlow's perspective, not theirs, since in the novel they have none. Although, as Woloch details, minor characters always have the potential to become full subjects—in Honoré de Balzac's Comédie humaine (1842—48, The Human Comedy) a minor character in one novel will become the center of another—and although we may react against authorial guidance to sympathize with an unsympathetic character (a Uriah Heep, for example, so clearly and essentially hateful that we may perversely take his side), it is only those characters whom we take to be fully subjects in whom we can “recognize” some dimension of ourselves.

Character and Type

The question of the affective binding-in of readers to texts is, however, inseparable from that of the historically shifting regimes that govern our identification with or against fictional characters: learning how to read character is directly bound up with the practice of the self, of recognition of other selves, and of forming an emotional bond with fictional “selves,” and these practices work in distinctively different ways in different GENREs and in different historical and cultural formations.

A description of character as an effect of historically specific operations of reading would analyze how forms of literary character have drawn upon and fed back into folk psychologies and typologies such as the doctrine of humors, of the ruling passion, or of the racial or psychological or historical “type,” and the stock characters of the Greek New Comedy or European commedia dell'arte or Javanese Wayang theater. One form of this theory is supplied by a historicist aesthetics. In the work of Georg lukÁcs, a neo-Hegelian conception of the literary type is related to the development of the commodity form and of reification, such that characters can be said to “correspond” or not to the particular state of development of the historical process. But this kind of generalization from particular to type has become more generally embedded in our reading of the “realist” novel, where both moral and social characteristics tend to be raised to a higher power: Lord Dedlock “stands for” the values of a near-obsolete aristocracy; Jay Gatsby is the type of the nouveau-riche.

In a very different sense of the “type,” Erich Auerbach discusses the concept of figura, a patristic and medieval rhetorical device whereby an event or character is said to have been prefigured by an earlier event or character, and in turn to be their fulfillment, and both ultimately foreshadow “something in the future, something still to come, which will be the actual, real, and definitive event” (58). Thus Dante's figural reading of Cato in the Purgatorio assimilates him to the Christian tradition on the basis of “a predetermined concordance between the Christian story of salvation and the Roman secular monarchy” (66), because “for Dante the meaning of every life has its place in the providential history of the world” (70).

Northrop Frye's fivefold classification of fiction in terms of the hero's power of action (33) seeks to give a comprehensive account of such historical typologies (see MYTHOLOGY). Frye's descending scale is, roughly, a classification by rank and power, and he argues that in European literature the center of gravity has shifted progressively down the scale from a mythic mode to ROMANCE, the high mimetic, the low mimetic, and the ironic; but this is true only in the most sweeping perspective, and it is unclear how far this corresponds to real social history, or indeed whether it is intended as a historical or as a structural description.

This model has been reworked by Hans Robert Jauss in terms of a scale of norms of reception which reflect historically distinct (but overlapping) regimes of identification. The classification of “interactional patterns of aesthetic identification with the hero” (298) is again fivefold, and it ranges between the extremes of cultic participation and aesthetic reflection. The first level, that of “associative identification,” is structured upon the interactions of the game or ceremony. It is realized in religious cults and in various forms of literary game-playing. The second level, that of “admiring identification,” puts into play the category of the exemplary and various techniques of emulation, such as those by which the collective memory of the Christian Middle Ages or of the Communist state is constituted. The third level, “sympathetic identification,” corresponds to Frye's low-mimetic mode. It relies on the category of pity, and it works through either moral interest or sentimentality. It is realized in such bourgeois genres as the eighteenth-century domestic novel and domestic drama. The fourth level is that of “cathartic identification,” and it is said to “place the spectator in the position of the suffering or hard-pressed hero in order, by means of tragic emotional upheaval or comic release, to bring about for him an inner liberation which is supposed to facilitate the free use of his judgement rather than the adoption of specific patterns of activity” (297). Cathartic identification corresponds to the patterns of classical French tragedy and COMEDY. The fifth level, that of “ironic identification,” is roughly equivalent to Bertolt Brecht's conception of the alienation-effect, but it is here polemically subsumed within the category of “identification” itself, and its privileged examples are Miguel de Cervantes Saavedra's Don Quixote(1605, 1615), Charles Baudelaire's Les Fleurs du Mal (1857, The Flowers of Evil), and in general the poetics of MODERNISM. Jauss's account is flawed, however, by his privileging of one mode of identification, catharsis, in a move that displaces historical categories into a quasi-universal norm of reception.

Frye and Jauss are concerned less with objective typologies than with historically different ways of dealing with characters—although of course these shifting forms of aesthetic identification will in turn tend to favor certain forms of character construction. A more detailed and historically specific investigation of changing forms of involvement with characters is to be found in Deirdre Lynch's account of the transition from a neoclassical to a romantic regime of characterization. Pitching her argument against histories of the novel in which the genre moves from the “flat” and formulaic characters of Daniel Defoe or Henry Fielding to achieve its full realization in the “round,” psychologically complex characters of Jane Austen, Lynch posits instead that what is at stake is the transition from one set of material, rhetorical, and affective practices to another (see HISTORY). Charting a set of changing practices of self-cultivation, of shopping, of fashion, of character “appreciation” and many others, she posits that “with the beginnings of the late eighteenth century's ’affective revolution’ and the advent of new linkages between novel reading, moral training, and self-culture, character reading was reinvented as an occasion when readers found themselves and plumbed their own interior resources of sensibility by plumbing characters' hidden depths” (10). In Tobias Smollett, for example, we can see a gradual shift from viewing the protagonist as an empty position to seeing him as an object of identification, so that we come to be involved with “a being that, through its capacity to prepossess, can train the reader in sympathizing and so in participating in a social world that was being reconceived as a transactional space, as a space that held together through the circulation of fellow feeling” (89).

Character and Person

This shift in part reflects (and in part helps to form) that larger movement in the late eighteenth century in which the grounding of personal identity in “an essential core of selfhood characterized by psychological depth, or interiority” (Wahrman, xi) becomes dominant, in which childhood takes on a new status as the foundation of “the unique, ingrained, enduring inner self” (282), and in which an organic model of the realization of an essential selfhood displaces an older model in which the self is less an essence than a set of publicly appropriate roles. This new understanding breaks radically with older presuppositions. As Charles Taylor puts it:

We have come to think that we “have” selves as we have heads. But the very idea that we have or are “a self,” that human agency is essentially defined as “the self,” is a linguistic reflection of our modern understanding and the radical reflexivity it involves. Being deeply embedded in this understanding, we cannot but reach for this language; but it was not always so. (177)

While character in the novel is not reducible to forms of social selfhood, since it is produced by means of specifically literary conventions of representation, it is nevertheless closely bound up with the transformations of selfhood since the 1700s. This is less a matter of reflection of a preexisting reality than of the way the reading of character actively helps shape readers' sense of what it might mean to be a person. In this sense reading character is what Michel Foucault calls a technology of the self: a machine for modeling behavior, an exercise in self-cultivation through a recognition of and identification with other (represented) selves. Reading novels has for much of the time since the 1700s been a matter of learning to become a self within the regime of expressive interiority of which the novel has been so crucial a support.


1. Auerbach, E. (1998), “Figura,” in Scenes from the Drama of European Literature.

2. Bal, M. (1997), Narratology.

3. Bradley, A.C. (1905), Shakespearean Tragedy, 2nd ed.

4. Culler, J. (1975), Structuralist Poetics.

5. Foucault, M. (1988), “Technologies of the Self,” in Technologies of the Self, ed. L.H. Martin, H. Gutman, and P.H. Hutton.

6. Freud, S. (1953), Standard Edition of the Complete Psychological Works, ed. J. Strachey, et al.

7. Frye, N. (1957), Anatomy of Criticism.

8. Greimas, A.J. (1966), “Réflexions sur les modèles actantiels,” in Sémantique structurale.

9. Hamon, P. (1977), “Pour un statut sémiologique du personnage,” in R. Barthes, W. Kayser, W. Booth, et al., Poétique du récit.

10. Hunter, I. (1983), “ Reading Character,” Southern Review 16: 226—43.

11. Jauss, H.R. (1974), “ Levels of Identification of Hero and Audience,” New Literary History 5: 283—317.

12. Knights, L.C. (1951), “How Many Children Had Lady Macbeth?” in Explorations.

13. Laplanche, J. (1976), Life and Death in Psychoanalysis, trans. J. Mehlman.

14. Lukács, G. (1962), Historical Novel, trans. H. and S. Mitchell.

15. Lynch, D. (1998), Economy of Character.

16. Propp, V. (1968), Morphology of the Folktale, trans. L. Scott.

17. Taylor, C. (1989), Sources of the Self.

18. Wahrman, D. (2004), Making of the Modern Self.

19. Woloch, A. (2003), The One vs. the Many.