Focalization, a term coined by Genette (1972), may be defined as a selection or restriction of narrative information in relation to the experience and knowledge of the narrator, the characters or other, more hypothetical entities in the storyworld.
Genette introduced the term "focalization" as a replacement for "perspective" and "point of view" (Niederhoff → Perspective – Point of View). He considers it to be more or less synonymous with these terms, describing it as a mere "reformulation" ( 1988: 65) and "general presentation of the standard idea of 'point of view'" (84). This, however, is an underestimation of the conceptual differences between focalization and the traditional terms.
Genette distinguishes three types or degrees of focalization—zero, internal and external—and explains his typology by relating it to previous theories:
"The first term [zero focalization] corresponds to what English-language criticism calls narrative with omniscient narrator and Pouillon 'vision from behind,' and which Todorov symbolizes by the formula Narrator > Character (where the narrator knows more than the character, or more exactly, says more than any of the characters knows). In the second term [internal focalization], Narrator = Character (the narrator says only what a given character knows); this is narrative with 'point of view' after Lubbock, or with 'restricted field' after Blin; Pouillon calls it 'vision with.' In the third term [external focalization], Narrator < Character (the narrator says less than the character knows); this is the 'objective' or 'behaviorist' narrative, what Pouillon calls 'vision from without'" ( 1980: 188–89).
The passage synthesizes two models: a quasi-mathematical one in which the amount of narrative information is indicated by the formulas derived from Todorov; and a more traditional one based on the metaphors of vision and point of view, which is derived from Pouillon and Lubbock. That these two models are not equivalent has been shown by Kablitz (1988). If a novel begins by telling us who a character is, to whom she is married, and for how long she has been living in a certain town, it will reveal no more than the character knows herself, but no one would describe such a beginning as an example of "vision with" or character point of view. To tell a story from a character's point of view means to present the events as they are perceived, felt, interpreted and evaluated by her at a particular moment.
Genette himself leans in the direction of the Todorovian, information-based model. On occasion, he talks about focalization in terms of the point-of-view paradigm, e.g. when he describes it as placing narrative focus at a particular "point" ( 1988: 73); but in general, he thinks of focalization in terms of knowledge and information. He thus defines it as "a restriction of 'field' [...], a selection of narrative information with respect to what was traditionally called omniscience" ( 1988: 74). This emphasis is also implied by the very term itself and the preposition that goes along with it. Genette consistently writes "focalisation sur" in French: while a story is told from a particular point of view, a narrative focuses on something. This preposition indicates the selection of, or restriction to, amounts or kinds of information that are accessible under the norms of a particular focalization. If focalization is to be more than a mere "reformulation" of point of view, it is this aspect of the term, the information-based model, which should be emphasized.
Genette's emphasis on knowledge and information is also revealed by his extensive treatment of alterations ( 1980: 194–98), defined as a transgression of the informational norm established by the focalization of a text. Alterations take two forms: paralepsis, the inclusion of an event against the norm of a particular focalization; and paralipsis, a similarly transgressive omission of such an event. According to Genette, the norms that are violated by these transgressions cannot be defined in advance (e.g. by commonsensical inferences as to what a particular narrator may have learnt about the story he or she tells). Instead, the norms are established by each particular text: "The decisive criterion is not so much material possibility or even psychological plausibility as it is textual coherence and narrative tonality" (208). Shen disagrees with this view, arguing that it boils down to a merely quantitative approach, a measurement of the relative length of the normative and the transgressive portions of the text; she suggests that there is a more general "legitimacy" that is violated by alterations (2001: 168–69). However, her examples and her analyses show that "legitimacy" in matters of focalization is far from self-evident. In her case, it rests on rather arbitrary assumptions about the limited knowledge of first-person narrators and the unlimited knowledge of third-person narrators.
A major point in Genette's theory is his rigorous separation between focalization and the narrator (referred to with the grammatical metaphor of "voice"). Most previous theories analyze such categories as first-person narrator, omniscience, and camera perspective under one umbrella term, usually point of view. Genette believes that such cavalier treatments of the subject "suffer from a regrettable confusion [...] between the question who is the character whose point of view orients the narrative perspective? and the very different question who is the narrator—or, more simply, the question who sees? and the question who speaks?" ( 1980: 186). What follows from the separation of the two questions is a plea for a relatively free combination of narrator types and focalization types, a position that has ignited a considerable amount of controversy.
Genette's theory was welcomed as a considerable advance on the previous paradigm of perspective or point of view, and the neologism of focalization has been widely adopted, at least by narratologists. Genette himself claims that his term is preferable because it is less visual and metaphorical than the traditional ones ( 1980: 189). Other critics prefer it because it is not part of everyday speech and thus more suitable as a technical term with a specialized meaning (Bal  1997: 144; Nünning 1990: 253; Füger 1993: 44). However, the main argument is that the term dispels the confusion of the questions who sees? and who speaks? This argument has become a veritable commonplace (e.g. Bal  1997: 143; Edmiston 1991: x; O'Neill 1992: 331; Rimmon-Kenan  2002: 71; Nelles 1990: 366; Nünning 1990: 255–56). Finney states it as follows: "'Focalization' is a term coined by Gérard Genette to distinguish between narrative agency and visual mediation, i.e. focalization. 'Point of View' confuses speaking and seeing, narrative voice and focalization. Hence the need for Genette's term" (1990: 144). It is true that Genette introduces the term focalization immediately after his polemics against the typological conflation of who sees? and who speaks?, but he does not establish a connection between these polemics and his neologism—nor is there such a connection. As a term, focalization dispels the confusion of seeing and speaking no more than the traditional terms do. On the contrary, the connection between the question who sees? and point of view should be a little more evident than between who sees? and focalization. It is perfectly possible to embrace Genette's scheme, including the separation and free combination of narrator and focalization types, while referring to his three focalizations as points of view.
The case that the advocates of focalization have made for its superiority to point of view is by no means beyond dispute. Nor is it improved by the fact that some of them use the new term while still thinking along the lines of the old, overlooking the semantic differences between them and neglecting the new conceptual emphasis of the neologism. Füger, for example, explains that internal and external focalization can be distinguished by the "situation of the agent of the process of perception" (1993: 47), which is nothing but a roundabout paraphrase of point of view. A characteristic instance of the reinterpretation of focalization in terms of point of view is a change of preposition in the English translation of Genette's study: "[L]e mode narratif de la Recherche est bien souvent la focalisation interne sur le héros" (1972: 214). "[T]he narrative mood of the Recherche is very often internal focalization through the hero" ( 1980: 199). The rendering of sur as through speaks volumes. It seems that the translator is under the spell of the point-of-view paradigm. Instead of thinking about focalization as a selection of or a focusing on a particular region of the storyworld—in this case the mind of the protagonist—the translator regards this mind as a kind of window through or from which the world is perceived.
Bal's influential revision of Genette's theory is another example of the reinterpretation of focalization in terms of point of view, although she is more aware of this than others. Thus she admits that perspective "reflects precisely" what she means by focalization ( 1997: 143), and she points out that Genette ought to have written "focalisation par" instead of "focalisation sur" (1977: 29). The continuing influence of the point-of-view paradigm also seems to underlie Bal's reconceptualization of Genette's typology in terms of focalizing subjects and focalized objects. According to her, the distinction between Genette's zero focalization and his internal focalization lies in the agent or subject that "sees" the story (the narrator in the first case, a character in the second); the difference between Genette's internal and external focalization, however, has nothing to do with the subject that "sees" but with the object that is "seen" (thoughts and feelings in the first case, actions and appearances in the second). Thus she ends up with a system of two binary distinctions that replace Genette's triple typology. There are two types of focalization: character-bound or internal (Genette's internal focalization) and external (Genette's zero and external focalization combined into one). Furthermore, there are two types of focalized objects: imperceptible (thoughts, feelings, etc.) and perceptible (actions, appearances, etc.).
At least some of the elements in this reconceptualization result from Bal’s adherence to the point-of-view paradigm, notably the elimination of the distinction between Genette’s zero and external types (merged by Bal into external focalization). Within the point-of-view model, this change makes some sense. If one thinks about Genette’s zero and external focalization in terms of a point from which the characters are viewed, this point would appear to lie outside the characters in both cases. However, if one thinks in terms of knowledge and information, zero and external focalization are worlds apart. The first provides us with complete access to all the regions of the storyworld, including the characters’ minds, whereas in the second the access is extremely limited and no inside views are possible.
While it is possible to explain the motivation of Bal's modifications of Genette's theory by pointing out her adherence to point of view, it must be said that, in themselves, these modifications are hardly compelling. It is simply erroneous to claim that Genette's zero and internal types are distinguished by the focalizing subjects, whereas his internal and external types differ in the focalized objects. All of Genette's focalizations vary, among other things, in the range of objects that can be represented; his zero focalization and his internal focalization (distinguished in terms of the focalizing subjects by Bal) are also dissimilar in this respect. Furthermore, the "focalized object" is a misleading concept: the crucial distinction concerning such objects is between "perceptible" and "imperceptible" ones, which means that the subjective element of perception that Bal has previously eliminated is reintroduced by way of the adjective. As Edmiston writes: "[T]he focalizer can be characterized by his objects of focalization, despite Bal's efforts to separate them [...]. Subject and object [of focalization] may be analyzed separately, but they cannot be dissociated totally, as though there were no correlation between them" (1991: 153).
Another feature of Bal's theory, pointed out and criticized by Jahn, is "that [...] any act of perception (brief or extended; real, hypothetical or fantasized) presented in whatever form (narrated, reported, quoted, or scenically represented) counts as a case of focalization" (Jahn 1996: 260). This is a problematic premise, which perhaps stems from taking Genette's question who sees? rather too literally. It ultimately reduces the analysis of focalization to a paraphrase of narrative content, to identifying acts of perception. However, if a narrative tells us that Mary sees John, we cannot be certain that the narrative is also focalized "by" (to use Bal's preferred preposition) Mary. Whether this is the case depends on how Mary's act of perception is narrated and on the context in which it occurs. Admittedly, Bal is not the only one to equate focalization with perception. This premise is also shared by Herman & Vervaeck (2004), Margolin (2009) and Prince, who explicitly states that his "discussion links focalization only to the perception of the narrated by (or through, or 'with') an entity in that narrated" (2001: 47).
The equation of focalization with perception is also made by David Herman in "Hypothetical Focalization" (1994), an article that I will use here to point out the problems inherent in this equation. Drawing on possible-worlds semantics, Herman examines passages that explicitly describe what might have been seen at a particular point in the story if anyone had been there to see it. Thus, in Poe's "The Fall of the House of Usher," the narrator invokes an imaginary onlooker of this kind when he describes the house: "Perhaps the eye of a scrutinizing observer might have discovered a barely perceptible fissure, which, extending from the roof of the building in front, made its way down the wall [...]" ( 1956: 97–8). The problem with Herman's article is that it analyzes hypothetical perception rather than hypothetical focalization. The discovery of the fissure by Poe's imaginary observer is hypothetical only in comparison with the case of a character actually seeing this fissure. In terms of the focalization of Poe's story, the discovery is not hypothetical at all for the simple reason that the narrator mentions it. It has an effect on the focalization in that it contributes to the distancing of the narrating I from the experiencing I: the narrating I knows there was a fissure because he saw it very clearly at the end of the story, whereas the experiencing I seems to be unaware of it when he approaches the house for the first time. Generally speaking, instances of hypothetical perception would appear to point in the direction of zero focalization (or narratorial point of view in the traditional paradigm), just like the "report [of] what a character did not in fact think or say" discussed by Chatman ( 1980: 225). Hypothetical focalization in the strict sense is a focalization option that is conceivable but not realized in a text, such as an internally focalized version of Fielding's Tom Jones. Whether a text itself can achieve or suggest such hypothetical focalization is an interesting question awaiting an answer.
While Bal's revision of Genette's theory involves deletions such as "external focalization," it also contains additions, notably the "focalizer," i.e. the "agent that sees" in a given focalization (Bal 1997: 146). This concept has spawned a considerable amount of controversy, including a more specific debate about the question of whether narrators can be focalizers. Bal, Phelan (2001) and many others assume that both characters and narrators can be focalizers; Chatman (1990) and Prince (2001) argue that characters can focalize while narrators cannot. Genette, on the other hand, rejects character focalizers but concedes, with some reluctance, the possibility of regarding the narrator as a focalizer ( 1988: 72–3). However, he does not see any great need for the term, an attitude shared by Nelles, who considers it redundant (1990: 374). The skepticism of the latter two critics seems to be justified. To talk about characters as focalizers is to confuse focalization and perception. Characters can see and hear, but they can hardly focalize a narrative of whose existence they are not aware. This leaves us with the narrator (or the author?) as the only focalizer, an inference whose interest is primarily scholastic. If all types of focalization can be attributed to one agent, this attribution does not provide us with any conceptual tools that we can use in distinguishing and analyzing texts.
Furthermore, the concept of focalizer is misleading because it suggests that a given text or segment of text is always focalized by one person, either the narrator or a character. But this is a simplification. Consider the famous beginning of Dickens's Great Expectations, in which Pip, the first-person narrator, tells us how, as a little orphan, he visited the graves of his family and drew some highly imaginative conclusions about his relatives from the shape of their tombstones. This passage focuses on the thoughts and perceptions of the boy, but it also communicates the knowledge and the attitude of the adult narrator, primarily through style (elaborate language, ironically inflated lexis, etc.). It makes little sense here to ask whether or not the boy is the focalizer in this passage. It is more appropriate to analyze focalization as a more abstract and variable feature of the text, wavering between the knowledge and the attitudes of the adult narrator and the experience of the child character.
To sum up, the various theoretical innovations introduced by the advocates of focalization are fraught with considerable problems; focalization is hardly so much superior to point of view that the old term can be discarded. Niederhoff (2001) compares the meanings and merits of the terms, making a case for peaceful coexistence of and complementarity between the two. There is room for both because each highlights different aspects of a complex and elusive phenomenon. Point of view seems to be the more powerful metaphor when it comes to narratives that attempt to render the subjective experience of a character; stating that a story is told from the point of view of the character makes more sense than to claim that there is an internal focalization on the character. Focalization is a more fitting term when one analyses selections of narrative information that are not designed to render the subjective experience of a character but to create other effects such as suspense, mystery, puzzlement, etc. If focalization theory is to make any progress, an awareness of the differences between the two terms and of their respective strengths and weaknesses is indispensable.
(a) The most pressing need is for an analysis of the specific conceptual features of the focalization metaphor in comparison with related metaphors such as perspective, point of view, filter, etc. This needs to be complemented by a thorough, non-dogmatic analysis of texts that shows which of these terms is more appropriate to which kind of text. (b) The question raised by Herman's (1994) article remains to be investigated: Is there such a thing as hypothetical focalization? In other words, can a text suggest or imply a focalization that is not present in this text?