Russian Formalism, Cognitive Poetics and Art as an Institution

Some 15 years ago I had the privilege of participating in a symposium entitled “fiction updated” held at the U of T on the occasion of Lubomir Dolezel’s retirement. At that time I took my clue from Lubomir’s work on fictional worlds’ semantics and spoke of characters and their versions across story worlds. Today, on another happy occasion, celebrating another milestone in Lubomir’s career, I thought that, following the theme of the conference, I should take my inspiration and point of departure alike from Dolezel’s other major line of research, that of meta theory and history of poetics. In Occidental Poetics then Lubomir sets out to “reconstruct the growth of theoretical knowledge in poetics” by tracing in it one or more research traditions. He goes on to say that logical and epistemological continuity in poetics involve not the re-assertion of the same claims but rather the constant reexamination of its major claims, and that in the course of a research tradition the issues of poetics appear as successive refinements, versions or variations on a limited number of fundamental themata which are being defined and redefined, enriched and reformulated.

My aim today is to define and relate to each other two major historical phases of one such tradition, namely, Russian Formalism and the Prague School on the one hand and Anglo-American poetics and analytic aesthetics from the 1970s onwards on the other. Three shared fundamental themata or notions and their transformations will form the focus of our discussion: the work of verbal art, literature or literariness, and the poetic cum aesthetic function. From the point of view of global disciplinary organization, the near identification of the three notions in early formalism and their resultant interchangeable discussion has evolved in the transition from one phase to the other into three distinct though interrelated sets of concepts and corresponding lines of enquiry.

As for each of them individually, the textual specificity of the work of verbal art, initially defined in terms of device and perceptibility and later as foregrounding, has become the object of a whole new area of enquiry, that of cognitive poetics or stylistics, systematically correlating textual structures with receptional effects. The assumption of early formalism that there is a constant, textually inherent quality called literariness and that it and it alone should be the object of literary theory was definitely rejected by 1924 by Tynjanov and Eikhenbaum, both of whom advocated instead a view of literature as a historically shifting cultural series whose boundaries at each point in time are determined by participants in the literary system, understood as a social action system. After many trials and errors, American aesthetics arrived at the same conclusion with respect to the notion of art in general, and then went on to develop a whole family of theories of art as an institution, all of them viewing art as essentially historical, situative, contextual and relational, rather than object based. A similar approach with respect to literature specifically has been developed by Peter Lamarque and Stein Haugen Olsen. Poetic function began with Jakobson’s 1921 notion of poetry as a message with orientation on the expression side (ustanovka na vyrazhenie), revised a generation later in 1958 as the poetic function, and defined as set on the message as such or for its own sake. The Prague school subsumed the poetic function under the broader category of the aesthetic one, with obvious Kantian foundation. But both poetic and aesthetic function remained abstractions in need of operationalisation and empirical content. This has been done since the early 1970s in the study of literature as cultural practice (Culler, Lamarque and Olsen), focusing on the codified, yet historically variable norms of literary processing handed down in the educational system and including rules of sense making, evaluation, and the appropriate cultural uses associated with the handling or treating of a text as literature.

Concerning disciplinary scope and depth, one notices that what started as a uniquely literary enquiry has by now acquired a broader and deeper foundation in terms of general textual, institutional and aesthetic theories respectively. What started as presumably inherent to the text in isolation (be it device, foregrounding, literariness or poetic function) is now considered instead as essentially related to individual participants in the literary system or to communities of such participants. It is the text-context crux which is by now the defining focus of all three lines of enquiry. In fact, the redefinition and transformation of this research tradition enables us to define its two poles as the formal – structural and the cultural – cognitive respectively.

Turning our attention to the historical line of development, it is interesting to note that, with respect to all three themata, a significant portion of the Anglo-American work has been carried out independently of, or with at most nodding second hand acquaintance with the Slavic work. Nevertheless, its claims definitely constitute a further elaboration, transformation, and often improvement of the general assumptions first formulated by the formalist and structuralist schools. Such notional continuity and approfondissment amidst historical-factual discontinuity may suggest the existence of some inner logic in the elaboration of problem clusters. But what it reveals beyond any doubt in retrospect is the existence of a major research tradition unified in terms of its “set of general assumptions about the entities and processes in its domain of study and [even more so] the approaches and methods to be used for investigating the problems and constructing theories in this domain.” Now this definition of a research tradition stems from Larry Lauden, and is adopted by Dolezel in his book. The obvious theoretical progress made within this research tradition or paradigm over the last several generations suggests a few additional observations as regards the dynamics of literary theorizing;

• Literary theory is not, or at least not entirely, a chaos of ever changing incommensurable approaches or orientations.

• Theorising in this area need not and probably should not start ab ovo leda, clearing the slate from all previous work on a given topic on the grounds that it is supposedly antiquated, irrelevant, biassed etc.

• Literary theorizing, at least in part, can claim the status of rational enquiry, putting forth claims that can be learned and taught (Lernund lehrbar) and that are subject to standard rules of reasoning and argumentation. It can formulate sets of interrelated general claims which often possess significant empirical content, being open to historical or experimental examination/testing.

Having put all my general cards on the table, I should know heed the Quaker saying that God, or at least scholarly enquiry, is in the details, and provide some further specifications of each thema, all within the limited time frame available.

• Work of Verbal Art
Underlying Shklovski’s seminal essay on art as device are several textual and psychological assumptions. Thus, any message is supposed to have an expression and a content side, and the perception of either can be automatic or controlled, hence requiring mental effort, attention, conscious awareness, and the ability to modify habitual modes of perception. A message is perceptible only if it is received in a nonautomatic manner, and a non-automatic manner of reception, hence perceptibility (oshchutimost’) of a message is achieved on the expression side by foregrounding aspects of the verbal medium, thus making them prominent, and possibly central or dominant, in any act of text reception. In the formalist tradition, a whole array of such prominence enhancing textual patterns has been described and explored. In short verse texts this is accomplished by imposing on the text a whole grid of parallelisms and equivalences on all linguistic levels, a grid which is not required from any communicative-informational perspective. In longer prose texts an architectonic or compositional design of some kind is imposed on the incidents of the text: positive and negative parallelism, staircase, gradation, triple construction, embedding… Maximum perceptibility of the expression side is achieved in short verse in non-sense poetry, and in prose narrative if such patterns are devoid of thematic motivation and are expressly pointed out as formal game (bez’sjuzhetni roman, laying bare the device). The other major way of foregrounding textual elements is that of deviation from a previously established inner textual regularity or norm, be it by breaking a metrical regularity, shifting the level of style or introducing a radically different semantic element. This aspect was picked up in the early 1960s by Michael Riffaterre, who defines as stylistic device any inner textual deviation or break which makes the novel element highly prominent and hence perceptible.

On the content level, deautomatisation of content perception (cognitive de-automatisation) is achieved if an author refuses to call things by their habitual name, refuses to recognize them. In current terms, this is done by authors withholding the standard category name or label known to the reader for an object or action, thereby preventing the automatic activation of the appropriate schema or script stored in long term memory. Instead one is supplied with a correct detailed literal description without the overall semantic category, and comprehension is delayed or stymied. This kind of semantic riddling is what de-familiarisation is all about.

The foregrounded elements are considered to be inherently perceivable, and identifiable without reference to any external context. “Work of verbal art”, or better, artistry, as employed in this context, functions as a scholarly descriptive, quasi-theoretical, term used to designate texts in which the predominance of patterns of formal or semantic foregrounding can be identified supposedly independently and objectively. But both the identification of foregrounded elements and the description of their receptional effects as offered by the formalists lacked any empirical basis, and were ultimately claims made by the scholars on the basis of their own linguistic and literary intuitions. In addition, there were no theories of textual processing available in the 1920s, and Shklovski for example had to rely on vague speculations about reading effort and energy involved in text processing. And finally there was the implicit bi-conditional or equation of work of verbal art and work of literature. What was needed was hence (1) the development of experimental methods to test on groups of readers scholarly hypotheses concerning what is perceived as prominent or deviant in a given text (2) The formulation of cognitive-affective theories regarding the operations involved in text processing on various levels (sound to sense) and sizes (micro and macro patterns) (3) the dissociation of work of verbal art and work of literature, since texts with little linguistic foregrounding are sometimes considered literature, while others with a lot of it, such as advertising, are not.

Contemporary cognitive stylistics or poetics (Stockwell, Gavin, Semino), has been developed precisely to address these desiderata. While many of the texts it analyses are taken from what is considered literature in our time and culture, the textual patterns and effects under study are not considered as either necessary or sufficient conditions for making the text in which they occur literary. Secondly, cognitive poetics places itself at the interface between linguistics, literary studies and cognitive science. According to Elena Semino, one of its leading practitioners, “it combines the kind of explicit, rigorous and detailed linguistic analysis of the stylistic tradition with a systematic and theoretically informed consideration of the cognitive structures and processes that underlie the production and reception of language”. What is new here is not so much the study of linguistic choices and patterns, but rather the way in which linguistic choices are related first to cognitive structures and processes and then to perceived effects. And, as Semino claims, this provides a more systematic and explicit account of the relations between texts on the one hand and responses and interpretations on the other. (Semino 2002, ix). Finally, informant testing is systematically employed to test the scholars’ specific hypotheses. The study of works of verbal art is thus absorbed into general cognitive text theory.

• “Literature” is understood already by 1924, at least by Tynjanov and Eikhenbaum, as designating a dynamic cultural and social series, system or institution. One cannot define in a non-historical, once and for all manner either “genre” or “literature” or what makes a work of verbal art a work of literature. Text elements, texts, and text models are not inherently literary or not. Rather, being literary is a status conferred upon a text or text model by the literary institution of the time (authors, critics, groups of readers belonging to the literary milieu [byt], publishers) for any of a number of heterogeneous reasons. It simply designates being accepted into the institution, forming part of it. A text does not change over time its nature as work of verbal art (istry), since this can be objectively defined, but its literary status can and does change on and off over time, depending on the criteria employed and the group in question. In logical terms, “work of verbal art” has a rather clear and permanent intension (even if no minimal conditions), and its extension depends on what texts there are out there which satisfy this definition. “Work of literature” on the other hand is in principle purely extensional: “whatever has been designated as such by any group within the institution at any time and for any reason,” and its intension would be differently defined for each group, as determined by the rules of use of the term adopted by it.

The literary historian does not consequently decide or define what is literature. Instead, he needs to describe what texts and kinds of texts have been considered literature, by whom in the institution, when and where and on what basis (ideal of literature). One focus of literary historiography is hence the study of the changing definitions, ideas and ideals of literature, as expressed in creative works, essays, manifestos and works of criticism of each group and their relations with actual literary production. As a discipline, literary history is also vitally concerned with the inner dynamics of the literary institution: changes in what is produced within the institution and their causes, the relation of centre and periphery, archaists and innovators, canonization of sub literary forms, the movement over time of the literature non-literature boundary as conceived of by participants, changes in the canon, etc. The struggle of innovative writers to be accepted could be understood as their claim to legitimacy because what they write is crucially pertinent to the literary system as it exists at that moment, even if they want to invert it completely.

“Literature” is thus a term designating a historically changing, socio- cultural institute defined by its practitioners, not by scholars, and whose practitioners as a group also decide what past and present works will be seen – as or regarded – as literature, depending on their criteria of membership. The rich line or lines of enquiry I have just outlined are a cumulative project starting with the formalists, continued and refined in Prague (Vodicka) and reiterated in the work of Juri Lotman. Cumulatively, they represent one of the major and enduring achievements of the formalist-structuralist tradition.

Anglo American poetics and aesthetics spent a lot of time and energy looking for intrinsic textual features which could serve as a-historical necessary and sufficient conditions for a text being considered literature. The failure of all such attempts led initially to Wittgensteinian skepticism about the possibility of any definition of art. This was superseded by a radical reorientation, where one believes once again in this possibility, but where the relevant features are now essentially historical, contextual, situational and relational. Art and literature are now regarded as socio-cultural practices that must be understood within the contexts of their time and place (Davies 1991, 111). Art exists only in a community, which determines what kinds of features are criterial or decisive for something being a work of literature. So far USA 1967 sounds very much like Prague 1937 (Mukarovsky: Aesthetic Function, Norm and Value as Social Fact). But now comes the innovative part: in the absence of any intrinsic defining properties, the crucial question becomes who, at any given historical situation, decides that certain texts should be certified as works of literature, on what grounds, and according to what rules and procedures. Even more crucial is the question who decides, and on what grounds, to admit any new work into the literary series, or accept it into the literary domain. The question what is literature is replaced by when does a text acquire the status of a work of literature, and what are the grounds, practices, mechanisms and institutions involved in such a decision. The anecdotal, case oriented study of such issues undertaken by the formalists, especially with regard to 19th century Russian literature, is now complemented and given a foundation in terms of general socio-historical theories.

Arthur Danto has pointed out the major role often played by dominant theories, such as normative poetics, in deciding upon what new works the literary status will be conferred, and George Dickie has put forth the institutional theory of art. Put in a nutshell, his argument is that there is in our culture a loosely defined but widely recognized socio cultural institution, the Artworld, with various defined roles, such as writer, critic, editor etc. Some agents within this system are authorized, because of their socially recongised role in it, to confer the status of work of literature upon a new text. These agents have the authority to confer such status on behalf of the artworld because of their acknowledged knowledge and understanding of literature, its history, dominant theories, the current state of practice, and how literature works. Furthermore: such acts of conferral are not arbitrary, and the agents’ decisions are taken on the basis of shared conventions or criteria which are time and place bound and need to be reconstructed by literary historians. While G. Dickie’s institutional approach focuses on power relations (role, authority), Lamarque and Olsen focus more on the constitutive rules which establish and maintain literature as a cultural practice. To understand the institution of literature, says Olsen, is to master the constitutive rules that create the possibility of literature and regulate the expectations and performances of artist and audience roles (Olsen 2005, 3).

The irreducibly historical nature of any account of art or literature is further stressed by Jerrold Levinson in his intentional-historical theory of art. To him, arthood is projected upon a work when it is regarded or treated in the same way as some earlier objects that are already taken at that point to be works of art. Something thus becomes an artwork by being related to a given tradition or practice, and art making involves an agent’s intentionally relating his newly created object to the body of already existing art. The theory thus posits a backward looking act of reference, which, by linking present objects to past artworks makes the former artwork as well. (Levinson 25). “The concrete history of artmaking up to a given time is ineliminably implicated in any artmaking undertaking at that time” (28). This backward looking aspect has, curiously enough, been anticipated by Eikhenbaum who, in “Literatura i pisatel” (Moi Vremennik 1929) says: “Every work of literature has not only its fate, but also its past.” Stephen Davies, in his discussion of definitions of art (Davies 2003) argues in the same vein that “What makes something an artwork is that a historically appropriate reflexive relation, such as reference, continuation, amplification or rejection, holds between it and prior works create in the same artworld. In a phrase, art (now) is determined through its relation to art (past).” (175).

The historical theory has been given its most explicit and convincing formulation by Noel Carroll (Carroll 1999), who says “what is crucial is the descent or genetic link of a new work”. No matter how different, disruptive etc a new work is, it can legitimately raise the claim to be considered a work of literature if it can be shown to be pertinent or significantly related to the existing (current) artworld context or situation, to constitute an intelligible outcome of or response to it. Says Carroll: “to justify a new work’s claim to the status of work of art its proponents need to provide a historical narrative of how it came to be produced as an intelligible response to an antecedently acknowledged art-historical situation” (253). This situation may include current artistic practices but also issues, problems, aims, theories and ideals of art, and the relevant response may consist of continuation and variation, but also of disagreement, repudiation and radical break. Carroll is right in remarking that historical narrative as procedure for identifying works of art can well incorporate the mutations of the avant-garde into the continuous evolution of art, since the avantgarde is of course an extreme reaction to and rejection of the current situation. The view that avant-garde works, no matter how disruptive and different from their immediate predecessors, have a strong claim to belong to the same literary or artistic system since they are created deliberately as a reaction to its existing situation was of course central to the Russian formalists’ understanding and defence of the radical innovators of their own time, such as futurism and zaumni jazyk.

• Aesthetic function
Implicit in Jakobson’s work is the assumption that the structure and make up of a text dictate what communicative function will be dominant in its reception, and that certain text patterns will dictate that the aesthetic or poetic function be dominant in the reception of the corresponding text. This view, implying textual isolationism and determinism, has by now been replaced by the notion of the work of literature encompassing text, reader and the historical cultural situation. In each generation there are accordingly literary institutional conventions stipulating that, ceteris paribus, if a text possesses certain structures it should be received adopting the aesthetic attitude. But the specific content of this form-function correlation is historically variable (Lotman). Moreover, co-textual, para-textual and contextual factors, in this order, may on particular occasions invalidate a current convention or conversely make the reader adopt the aesthetic attitude towards a text devoid of the stipulated structures (Stanley Fish’s famous example of random lines of text on the blackboard in the context of a literature class).

But what does it mean specifically to have the aesthetic function or attitude as dominant in the reception of a text, beyond the general philosophical distinction between intrinsic value and purposefulness vs. instrumental ones? Here too much important expansion and improvement has been achieved in the last 30–40 years. One aspect suggested would be treating the text as display text, framed and decontextualised in terms of person, tense, and deicitics. One could further claim that once a text is treated as display text it can model or put on display the working of all other communicative functions or of the communicative process itself, a line of thought begun by Mukarovsky in the 30s and followed independently by Mary-Louise Pratt in the 70s. Or it may mean focusing one’s attention during textual reception on the text’s formal properties, what Noel Carroll has termed “design appreciation.” It may also mean special rules of use, such as setting aside questions of factual truth and practical usefulness, and allowing or even encouraging a multiplicity of different interpretations (SJ Schmidt’s rules of literariness). But the major significance of having the aesthetic function dominate one’s reception of a text would be to adopt and apply to it a whole series of special rules of interpretation and evaluation developed in the literary institution and acquired in the process of literary education. Jonathan Culler has suggested several such rules in Sturcturalist Poetics. For example, make semantic capital out of formal features and patterns, treat individual statements as having universal or symbolic significance, etc. To have internalized these rules and to be able to use them correctly means to possess literary competence and be a full fledged participant in the literary action system. But we must remember that these rules themselves are time and place bound, and that they are seldom formulated in an explicit and systematic manner. Rather, in the educational process they are acquired through example and repeated practice, and in critical practice they are presupposed. It is hence the task of the literary historian to extract them from the practice of leading critics and other authority figures in the literary institution, and make them explicit. Much more could be said on this subject, especially with regard to the particualrhierarchy of criteria of value employed in acts of literary evaluation, but ars longa vita brevis, and it is time to stop.


Carroll, Noël: Philosophy of Art: A Contemporary Intruduction. Routledge, London 1999.

Carroll, Noël (ed.): Theories of Art Today. The University of Wisconsin Press, Madison–Wisconcin 2000.

Davies, Stephen: Definitions of Art. Cornell University Press, Ithaca 1991.

Davies, Stephen: Definitions of Art. In: Routledge Companion to Aesthetics. 2003.

Lamarque, Peter: Literature. In: Routledge Companion to Aesthetics. 2003.

Levinson Jerrold: Contemplating Art. Oxford University Press, USA 2006.

Livingston, Paisely: Literature. In: Oxford Handbook of Aesthetics. Oxford University Press, USA 2003.

Olsen, Stein Haugom–Petterssson, Andars (eds.): From Text to Literature. Palgrave Macmillan, USA 2005.

Semino, Elena–Culpeper, Jonathan (eds.): Cognitive Stylistics. Benjamins, Amsterdam 2002.

Van Peer, Willie: Foregrounding. In: Encyclopedia of Languages and Linguistics. Ed. Keith Brown. 2nd edition. Elsevier Science, 2005.