Artistic principles in the 1960s and 1970s made it axiomatic that literariness and especially excessive literary narrative exerted a negative influence on good art. A piece of visual art did not need to be narrative. In other words, it did not have to be expressible in narrative form. Rather, it needed to make use of the visual means that were inherent to it. As a result, abstract art - art that consistently uses only the visual language unique to it - was seen as the purest form of art. It was already possible to perceive a certain parallel to this in the directions taken by art and theories of art and literature in the Post-World-War II period. In the 1950s and 1960s a less «literary» literature (poetry, laconic prose, literary theory) had emerged along with abstract art, minimalism etc. By the end of the 1970s, however, narrative had already begun to make a stealthy comeback in art. Its origins could already be seen in the period of Pop Art, and this process has continued and intensified to this day. In terms of theory itself, a concentration on poetics has been replaced by the expansion of narratology in many diverse disciplines. Consequently, the problems of the narrative in visual art need to be revisited; whether and to what extent the narrative is foreign to visual art or can be avoided, and to what extent it is unavoidable.

The basic assumptions behind this article can be summed up as follows:

1. Narrative has a greater role to play in visual art than is commonly believed. I will take it for granted that narrative is an integral part of both figurative as well as abstract art. In the scanty treatments of pictorial narrative that have been published to date (for example James Elkins (1991), Nelson Goodman (1980), and to a lesser extent Mieke Bal (1991), painting has been considered as narrative only when it portrays an actual event or truly «narrates a story» in the most direct sense. A narrative thread can, however, be found even in non-figurative art, for example in a repeatedly quoted black square (e.g. El Lissitzky Suprematicheskii Skaz About 2 Squares, Berlin 1922). Even a quote or an allusion to some existing well-known piece of art or literature could be seen as a retelling of that same work.
   
   
2. Current developments in art have made it necessary to consider the increasingly important role of narrative. Once again, art is being used for the purpose of saying something, of imparting some social significance. Understandably, the threshold established for it and the interpretation of its interdependence will limit its use. Here we can compare Roland Barthes' ideas:
   
   narrative can be supported by articulated speech, oral or written, by images, fixed or moving, by gesture, and by the organised combination of all these substances; it is present in myth, legend, fable, tale, tragedy, comedy, epic, history, pantomime, painting (think of Capraccio's Saint Ursula), stained-glass windows, the cinema, comic books, news items, conversation. Moreover, in these almost infinite forms, narrative occurs in all periods, places and societies; narrative begins with the very history of humanity; there is not, nor has there ever been, a people anywhere without a narrative ... narrative never prefers good literature to bad: international, transhistorical, transcultural, narrative is there, like life. (Barthes 1994: 95)
   
   
3. The success of visual narrative depends on one «reading» significantly more into it than in the case of written narrative. An analogy here might be a well-designed hypertext, in which the chronology of events or the order of their «reading» is not dictated by the author. In this respect the narrative will be the hypertext, as it has already been read or has taken place. The hypertext then itself becomes a potential narrative, a collection of motifs.
   
   
4. Nevertheless, one of the basic premises here would have to be the fact that narrative is by its very nature a form of verbal expression. Narrative can be considered to be the most extreme form of verbal expression, a way of creating a number of possible worlds, or, by using words, of creating something that does not exist. The other verbal extreme - poetry - is linked more to music in its sounds and rhythms. But narrative is not the essential manifestation of visual expression. Literariness is the quality a picture possesses that allows it to be expressed or «translated» through words. It is the approximation of a picture to words, that part of a picture that can be verbalised. This implies that it is possible for a pictorial narrative to exist. That is, a narrative need not exist only in verbal form. Only abstract art, abstract form, texture and architecture are visual to the extreme - and here too we see a connection to music. At the same time, visual narrative can also make reference to other senses such as sight or smell. As such, it is impossible to speak of purely visual or purely verbal forms of expression.

Visual art nevertheless often suffers under linguistic terrorism - language with its conceptions and categories, its established connections, forces itself on non-discrete visual experience. It hinders the instinctive experience gained from merely looking. This, the oppressive bed of language, is more or less what art has in this century repeatedly tried to liberate itself from (compare the Post-Stalinist identification of «bad art» with the above-mentioned notion of «literariness», the popularity of the title Untitled, abstract art as a whole, or the great love artists feel for Ludwig Wittgenstein's final proposition in his Tractatus «Whereof one cannot speak, thereof one must be silent.» At the same time we can find counter examples to these currents, such as conceptualism or the penetration of a picture by text.

Gotthold Ephraim Lessing (1729-1781) presented his doctrine in his most significant work on aesthetics, «Laokoon oder über die Grenzen der Malerei und Poesie» (Laocoön: An Essay on the Limits of Painting and Poetry) (1766), clearly distinguishing between classical poetry and painting. In essence this work targeted the so-called «shining antithesis» that 'poetry is painting that speaks', which was attributed to the ancient Greek lyricist Simonides. It became widespread through Horatius' maxim ut pictura poesis (poetry is like painting). Lessing's work, as its name implies, originates in the legend of Laocoön. The author bases his arguments on two earlier reworkings of the legend: the well-known group of statuary representing Laocoön's death as described by Winckelmann, and the description of the same in the second song of Virgil's Aeneid.

In this description by Winckelmann, Laocoön has already in a way become a rhetorical figure. Lessing continues in the same vein (though he might not entirely agree with all of Winckelmann's propositions).

Lessing's thoughts can be summarized as follows:

• in reproducing anything the material limitations of art confine it to the depiction of a single moment in time;
  
  
• in painting this one moment can only be used with reference to a single vantage point;
  
  
• a work of art is created to be contemplated repeatedly and at length, and therefore such a moment must be chosen with the greatest regard for its effect;
  
  
• but only that which gives free reign to the imagination is effective (Lessing 1984: 19).

Lessing feels that a moment of extreme passion cannot be conducive to such an effect. The apparent permanence of something transitory is not suitable for stimulating the imagination.

In Lessing's opinion, at the heart of a painting lies a story. The only matter open to debate is how the story is to be formulated. If we define storytelling and narrative as the setting forth of a story, then the existence of the story is the only condition, or rather the only possibility for the existence of storytelling and narrative. Looked at from the perspective established by Lessing, we must consider a painting as narrative. At the same time, a story exists at a point in time along with all of our prior knowledge of what has come before and what is to follow. And inevitably some sort of relationship develops between the MOMENT depicted and the moments preceding and following it. Through this, the temporal dimensions of the story emerge.

In Lessing's opinion the originality or novelty of the subject matter is not of prime importance to the artist. Subject matter that is already familiar will increase the impact of the painting and make it easier for it to have an effect. «Objects or parts of objects which follow one another are called action. Accordingly, actions are the true subjects of poetry.» (Lessing 1984: 78). In poetry bodies can be depicted only by suggestion and through their actions. In painting, everything happens at the same time, everything exists side by side and only one moment of the action can be depicted.

We can quite unambiguously state that in the case of a pictorial narrative, the only element that is missing is a definite, organised way of depicting the series of events sequenced according to the intentions of the author. All this is left open. Thus in the case of pictorial narrative the basic categories of classical narratology (compare Genette 1980) - TIME (sequence, tempo, frequency) and in part also EXPRESSION - are either vague or completely useless. A picture is the cross-section of a story; and the preceding and following events can be unravelled in the consciousness of the viewer only in the process of reading the story. The WAY something is retold - its temporal, emotional or physical distance assumes primary importance.

I suggest that a third fixed point in the case of pictorial narrative is the CHARACTER. I would maintain that while theoretically pictorial narrative depends on its reading, and as such the organization of events therein are weaker than in the case of verbal narrative, the system of the characters therein is nonetheless much more significant than in verbal narrative.

In the following I will discuss FISH as an example of the character, that is to say I will examine the portrayal of fish in visual art. Subsumed under this general heading or overall category are various species (flounders, ides, pikes) as mere concrete manifestations of fish. Why have I chosen to examine fish instead of people, which at first glance might seem to be the more logical choice?

Genre-painting as such has occupied a marginal position from the very origins of Estonian art. Prior to World War II, this would seem to have been the result of an intentional reticence. It was as if genre-painting had crawled into the shadows of other forms such as still-life or landscape painting. In the 1950s genre-painting was often perceived as representing a form of political oppression. As a result, humans enter into Estonian art as full-fledged elements only in the 1960s, and this coincides with the disfavour of literariness, as mentioned above. At the same time, it seems that the figure of the fish exhibits the characteristics and contains the essence of pictorial narrative much more clearly and surely than more obviously narrative art. As such it should then be possible to extend the relationships and types found for the fish character to other characters.

Fish have been portrayed in Estonian art with surprising frequency. There does not appear to be any such concentration of fish in the art of other nations or in world art as a whole. Why this should be remains unclear, but it is quite probable that there are other motivating factors for this than just being a coastal people and having a direct connection with fish. In examining how fish have been depicted over time, certain very clear periods and distinct iconographies emerge.

This type was most characteristic of the period of the end of the 1950s and the beginning of the 1960s. Perhaps the best-known and most often reproduced works are the following: Lydia Mei Still-life with Fish (1955, watercolour); Aleksander Vardi Still-life with Fish (1960, oil on canvas); Alo Hoidre Salmons (1962, autolito); Avo Keerend Fish on Platter (1964, plastic linoleum and intaglio); Herald Eelma Flounders (1964, coloured linoleum with cardboard cutouts).

The best known Estonian fish picture in this category and one that in its own way represents a transition is Peeter Ulas' The Great Turbot (1963, linoleum and veneer cut).

If fish were relatively rare at large retrospective exhibitions before the 1960s, in 1960 (December 1960 - January 1961) four still-life representations with fish were shown at an exhibition of Tallinn Artists. Records were achieved at the national exhibition of 1961 (September-October) and the spring exhibition of Tallinn artists in April of 1964, with five still-life representations with fish. After that fish began to recede quickly from exhibitions and by 1970 they had all but disappeared.

This type usually involves one big fish. In contrast to the previous period, this fish is alive and acts as an independent character. The time frame starts from the beginning of the 1960s. First and foremost in this group are the fish pictures by Ülo Sooster (see e.g. Kabakov 1996: 109-114). There were great similarities between Sooster's fish and those of a number of Tartu artists, in particular Elmar Kits and Valve Janov. A difference between the Tallinn and Tartu schools can readily be seen; the Tartu fish, done with oil in mixed media, were derived more from the spirit of abstract art, while the Tallinn fish were graphic works and were represented much more concretely. The best-known of these are: Silvi Liiva Fish (1973, etching), which portrays a large fish staring out from a pond while people surrounding the pond are trying to catch it; Marju Mutso December (1973, etching - a large fish depicted in a horizontal position). Elmar Kits was the first to depict fish in as early as the end of the 1940s and again in the 1960s, while Valve Janov's fish period falls in the 1960s.

Here we see an active fish in a specific world of its own. It is often depicted in an unusual vertical position. This period starts from the 1980s: Andres Tolts The Whale (1989, oil on canvas), Enno Hallek Fish (1959 -1985, oil on wood), Tiit Pääsuke Big White Bird and Kissing Fish (1989, charcoal, acrylic on canvas) and FishBird (1991, acrylic on canvas), Ando Keskküla Still-life with a Myth (1986, oil on canvas), Enn Põldroos Fish World (1989, oil on canvas), Jüri Arrak White Fish, Fish, Blue Fish, Fish Bearers and Fish in Figure (all 1992, oil on canvas). In the mid 1990s Raoul Kurvits, Kai Kaljo, Tiia Johannson and Jaan Toomik have used the motif of fish.

There is no distinct temporal limit for this type. It proliferated as subject matter in as early as the 1930s, and once again in the 1960s. It is difficult to distinguish this subject from the subject of fishermen in general: Arkadio Laigo The Fishermen (1935, woodcut),

Eerik Haamer The Eel Catchers (1942, oil on canvas), Herald Eelma How to catch Fish (1967, coloured linoleum) and Who Caught the Bigger One? (1967, coloured linoleum).

In the 1960s fish were generally depicted in close connection to seashore themes. This was particularly the case in graphic art (e.g. works by Ilmar Torn and Vive Tolli).

It is not yet clear whether the dominant role of fish in Estonian Art in the 1960s was to any extent caused by Ülo Sooster or whether Sooster was merely its precursor, one of the authors expressing the idea or one of the instigators of the whole avalanche. A search for parallels of any sort from world art reveals the following:

In twentieth-century art the fish motif appears primarily in works with a surrealistic bent. Fish might therefore be categorised as some sort of surrealistic figure. The fish has not been depicted in other art movements, perhaps with the exception of some cubist still-life representations of fish.¹ Of the better-known surrealist fish paintings one should mention: André Masson Battle of the Fishes; Paul Klee Goldfish Wife, The Golden Fish, Fish Magic, Around the Fish and Max Ernst Forest of Fish Bones. Above all one is reminded of fish in René Magritte - Collective Discovery, Exception, «Hommage à Alphonse Allais», The State of Grace, The Forbidden Universe etc.²

Kazimir Malevich has depicted fish in Aviator, in which a vertical white fish is the central figure. We encounter the same white fish in An Englishman in Moscow.³

The surrealist fish can be compared to the figure of the bird. Often fish and birds are intermixed in both form and theme. Constantin Brâncusi's birds have the shape of slender fish. Fish and birds are also connected by their «unanimality». In Genesis fish and birds were created on the same day and, from the human perspective, both move in an irreal space. Both are commonly used as generic nouns distinct from other living creatures and that cannot be called just ANIMALS. It is in terms of their general classes that BIRD and FISH have a deep symbolic meaning in art history. The bird as a parallel to fish is, however, very rarely encountered in Estonian art. The works of Tiit Pääsuke are exceptional in this respect.

In earlier periods of art (if we exclude the early Christian Fish as a symbol of Christ and fishermen as his disciples) fish have appeared prolifically in the art of the Low Countries. This is the only area where there is a situation comparable to the Estonian phenomenon. Even the same categories have been represented - the fish in still-life (particularly popular among painters from The Hague, such as Abraham van Beyeren), fishermen (which in the context of the time had a directly symbolic, Christian character) and mythological and eschatological fish in the process of doing something, above all in the paintings of Hieronymous Bosch and Pieter Brueghel the Elder, for example. We can see in the fish a connection between these two painters and surrealism. In the left panel of Bosch's Temptation of St. Anthony, depicting his battle with demons and subsequent rescue, we can see an air-battle with fish. In the right panel, where St. Anthony is meditating, we see a beautiful flying fish, a man and a woman on its back. We meet fish both in the air and in the foreground in the central panel where St. Anthony is surrounded by all manner of temptations. We also see many fish in the general confusion of Bosch's other paintings (for example, The Garden of Earthly Delights, The Haywain and The Last Judgement). Perhaps Brueghel's fish have largely been inspired by Bosch: e.g. his drawings for Hieronymus Cock's copperplate engravings, among them Big Fish Devouring Little Fish and the painting The Downfall of the Rebel Angels. We see fishermen in the works of Pieter Pourbus such as The Brugge Triptych of Fishermen. Here the everyday life of the fishermen with Jesus are depicted.⁴

In the better-known theories on the narrative, the character has always been placed in a rather secondary position. Beginning with Vladimir Propp's Morphology of the Folktale (Propp 1968), the character has always been subsumed under ACTION. The action itself and its effect on the progress of the story have been put in a primary position with respect to the doer:

What is significant is not the human side of the deed (love, hate) but rather its connection to the course of the action or its contrast with it. Function is seen as the basic unit of the narrative. It is defined through the action of an individual but nonetheless remains primary with respect to the character:

Propp reduced the characters of a folktale to a simple typology of seven archetypes; (VILLAIN, DONOR/PROVIDER, HELPER, PRINCESS (a person sought for) and her FATHER, DISPATCHER, HERO (seeker, victim), and FALSE HERO). The characters nonetheless retain their human dimensions and their allusions to concrete individuals. These characters, or more precisely spheres of activity, are connected to one of the 31 specific functions (outlined by Propp in Chapter III - Propp 1968).

The subsequent theories of narrative have, from the central period of structuralism on, simplified the scheme of characters and their roles to an even greater extent or made them more abstract. It seems that Roland Barthes is one of the very few who has been interested in the dismissal of the character, although he merely admits to the existence of the problem:

According to Barthes, in later periods the character, who up to now had only been a name, the agent of the activity, acquired some so-called psychological flesh on his bones and became a unique individual,

Thus the character was subordinated to the activity even before he had perpetrated any deed, before the action itself. At the same time, the perpetrator of the activity, independent of his category or name, has nonetheless always been an inevitable part of the story, at least as an agent or someone who does things.

Gérard Genette, who formulated one of the most well-known and systematic theories of the narrative, has all but left the character out of his orbit of investigation. More precisely, he has made him subordinate with respect to the perspective of the story and the narrator (Genette 1980, 1989). Tzvetan Todorov, Roland Barthes, Mieke Bal, and especially Algirdas Greimas have, to be sure, all attempted to resolve the question to a lesser or greater extent, but both Todorov's as well as Bal's character types derive directly from Greimas' model and Barthes, in the introduction to his analysis of the narrative mentioned above, also accepts this interpretation.

Algirdas Greimas' treatment of the character is connected to his concept of the actant, which has by now become well-established. Greimas developed his actant-theory in many of his works, starting with the most important of these, Sémantique Structurale (1966). In particular he treats it in his essay «Les Actants, les acteurs et les figures» (1973 - Greimas 1987).

At the heart of Greimas' theory of actants is a two-level organisation of narrative, connected to the general linguistic principles of the period of French structuralism at its height and the ideas deriving from these. There are three levels of characters. The archetypes are actants who participate in the syntactic level of the narrative. On the next level, the actors (acteurs), are recognised and presented on the discursive level of the narrative. Greimas finds it difficult to analyse this level owing to a lack of an appropriate theory. On the discursive level, one actant can be realised in terms of many different actors and the reverse is also possible - one actor embodies a number of actants. Finally, the characters (figures) of the actual story are realised from the actors. Greimas turns his attention primarily to the actants, 'the actant structure of the characters of the story', and this should create the appropriate foundation for the grammar of the narrative. Greimas is quite optimistic as far as the structural possibilities are concerned, being of the opinion that in this way it is possible to explain more fully the organisation of human imagination, which is the projection of individual as well as collective worlds (Greimas 1987: 106-107).

Actant is a class of characters in its widest definition, whose functions in narrative are one and the same despite their various manifestations in it. Greimas suggests that we describe and classify characters not according to who they are as individuals, but rather according to what they do (hence the term 'actant'). The actants appear in concrete texts as a force with a clearly defined aim. Actants are thus certain drives that can appear in the utterance and in no way correspond to actual characters of the story as they are mentioned in the utterance. This is for the following reasons:

1. an actant can be an abstraction (God, freedom), a collective character (such as the soldiers of an army), or characters with various interests who at one point react in tandem to achieve the same goal (such as all the «guilty» parties in Agatha Christie's Ten Little Indians);
   
   
2. one character can have different actant roles at various different times;
   
   
3. an actant might appear on the stage, but does not have to; it does not even need to be mentioned - an actant might be merely a general abstract notion or concept that only becomes apparent on the ideological level of the story.

Greimas offers three binary oppositions to unite six types of actant:

subject - object (sujet - objet)

sender - receiver (destinateur - destinataire)

helper - opponent (adjutant - opposant)

It is not necessary for all of these to appear in a story, although the presence of a subject is obligatory. Sender-receiver is the most ambivalent pair in Greimas' model. While it is only possible to categorise the actants a posteriori, after having become familiar with the utterance, it is nonetheless possible to perceive the sender, for example, as the source of the subject's knowledge, and the receiver as the group of individuals or humans in general who receive the message - the object sought by the subject.

These oppositions can be extrapolated from three basic patterns in the narrative: the desire or goal, the intercourse or communication, support or hindrance.

Syntagmatically it is possible to unravel the story, which Greimas considers the global utterance, as a chain of narrative utterances which in turn can be taken as the relationship between various actants. These correspond to Propp's functions. The notion of structure is implicit throughout the treatment, assuming the existence of this paradigmatic network (Greimas 1987: 108). A basic model that permeates Greimas' work - the semiotic square - establishes the positive and negative axes of the actants, who in turn double the number of actant archetypes:

positive subject - negative subject (antisubject)

positive object - negative object (antiobject)

positive sender - negative sender (antisender)

positive receiver - negative receiver (antireceiver)

Positive and negative are not qualitative assessments here and need not correspond to pairs such as hero - villain or good - bad, although they could do. An antisubject can react in tandem with the subject and at some point their actions might cross. On the level of the characters, the subject and antisubject might constitute a single whole (such as Faust's inner struggle, for example). According to Greimas it is possible to generalize the paradigmatic sequence of actants with this model and even apply it to stories with a minimal number (one) of actants. Even a hurdle or obstacle can be interpreted as the metonymic manifestation of an antiactant (Greimas 1987: 109).

The dimensions of the actant define the relationships between subject and object, subject and antisubject, subject and sender as well as receiver - but the subject is always the dominant actant. These relationships can be expressed through modal utterances. The first semantic constraint that defines the actant as a true operator is WANTING (vouloir). Other modalities or semantic constraints are KNOWING (savoir) and BEING ABLE (pouvoir) which define the subject's existence and operation.

Greimas replaced the fuzzy terms ATTEMPT or DIFFICULT TASK, in use at the time in narrative terminology, with the concept of PERFORMANCE (faire). Through performance it is possible to define a subject and an antisubject as an active subject. Performance requires reference to the competence that on the narrative level is expressed through modal utterances such as the desire and/or ability and/or knowledge to do something (Greimas 1987: 109). A competent and operating subject can be two examples of one actant. A subject actant can thus in the course of the narrative acquire a various number of actant roles.

The realization of actants into characters is a relatively complicated process and is only analyzable after the fact, after the realization of the story. Two extremes are possible [here] - the structure of the character is objectified, that is, as independent a character as possible corresponds to each actant, or the structure of characters is subjectified - as many actants and actant roles as possible are concentrated in only one character (Greimas 1987: 112).

Such an extreme structural approach was all the more understandable in its time, but today it is possible to interpret the system of actants formulated by Greimas more freely and more flexibly. The actants seem to offer possibilities for this particularly with the rediscovery of the character in narrative. The virtue of Greimas' model lies in its simplicity and systematicness as well as the seemingly endless possibilities for interpretation it seems to hold for many and very different types of narratives. Its primary weakness is that which always arises with generally accepted schemata - what can be gained by saying something more as regards the endless variability of the subject.

Roland Barthes has found that the common thread in all of the structural theories is the definition of the character in terms of a limited number of spheres of activity. At the same time Barthes acknowledges this approach himself by calling the second of his three narrative descriptions the level of actions (Barthes 1994: 107).

In Barthes' opinion linguistics provides the necessary basic theory best suited for describing a countless number of narratives. Linguistics, however, stops at the level of the sentence, since above the level of the sentence there are similarly more sentences (Barthes 1994: 82-3). Just as each sentence is a short narrative taken in the broadest sense of the word, it can also operate in the opposite direction - a narrative can be treated as a long sentence (Barthes 1994: 84). In this way we could say that sentence and narrative have a certain a fractal dimension: the smaller unit is reflected in the larger and the larger is similar to the smaller.

Barthes in his analysis of narrative structure prefers to distinguish three levels of description:

• the level of 'functions' (based on the traditions of Propp);
• the level of 'operation' (in Greimas' meaning, when he speaks of operators as actants; the roles are significant on this level);
• the level of 'the story' (which generally corresponds to the level of discourse).

These three levels are inevitably intertwined: function is meaningful only in the operation of an actant that derives its meaning from the telling of the story. The characters are, to be sure, the units of the operation level, but they are meaningful only on the third level, that of the story. Unfortunately, Barthes' analysis of this level, in common with other structuralists, remains superficial. Despite the fact that he noticed and introduced the problem of characters, Barthes nonetheless does not step outside the same structural circle, nor does he seems to want to.

Mieke Bal is the third to examine the problem of the character, the actor, and the actant in depth in her Narratology (1985). Already in her introduction Bal defines the text as a linguistic phenomenon: «... a text is a finite, structured whole composed of linguistic signs» (Bal 1985: 5) and «A narrative text is a story that is told in language; that is, it is converted into linguistic signs» (Bal 1985: 7-8). Despite the fact that our original hypothesis was different - that the text of a narrative need not necessarily be expressed in linguistic form - presumably many of Bal's conclusions can be used to expand this theme. At the same time, Bal finds that when events are transferred from one system of signs to another, such as film, their basic nature remains, to the extent that, broadly speaking, the same emotions are evoked in the viewer (Bal 1985: 5-6). This indicates that there is some sort of substantial linguistic structure between events and the understanding thereof (or perhaps they result from an inseparable penetration of linguistic structures into other systems). In Bal's theory of narrative, the story is transformed into linguistic signs only on the third level of the narrative text. Something else must then be significant in the inner structure.

Interestingly enough, Bal herself later backed away from such a strict linguistic approach and attempted to interpret works of visual art in a considerably more free manner, within the framework of narratology, reception theory and rhetoric (Bal 1991). Bal states:

Despite the fact that narratology is one of the sources of Bal's interdisciplinary treatment, she neglects the problem of the character in visual art.

In her Narratology, Bal finds that the tendency of structuralism to take events and their logic to be the only significant element while subordinating all other elements of the story to them is not correct. She considers events, actors, time and place to be equally important components in the fabula. Within this framework as it was elaborated right from the beginning, it is clear that in Bal's theory characters have a more significant part to play than in other theories of narrative. In spite of the fact she borrows terminology from Greimas (three-layered stratification of characters into actants, characters and actors, actant types), she uses these in a significantly broader way. Nevertheless, Bal's character is firmly connected to the action and is defined by it. Actors are agents that perform actions, while acts and actions are defined by cause or experience (1985: 5). In this way Bal has interpreted the event by means of the actors («events are the transition from one state to another state caused or experienced by the actor»; Bal 1985: 13).

Getting back to visual art and our fish character, we need to examine the problem of the character in art more broadly. From the preceding it was clear that it is possible to approach the character from opposite directions - to treat him with respect to his relationship to the narrator (the individual who is telling the story, how, from what distance and point of view) or to take the nature of his action as the point of departure. The first of these possibilities, i.e. the narrator's position and the clarification of the relationship between it as well as that of the author, is very fuzzy and does not seem very promising in art. Even during those periods when it was necessary to represent the sentiments of the narrator (for example socialist realism, naturalism), clarifying these gave us very little information about the story. This aspect falls into that layer of visual narrative which, in comparison with a verbal representation, becomes all the more inferior. The second possibility - treating the character with respect to his actions - seems much more promising. In any case, it must be emphasised that in a visual narrative the characters do not need to be directly recognizable nor even have human proportions.

The primary reason for this is that in visual art, usually some moment in the action is represented. Even a work of abstract art can be treated as the fixing of a state or process. On the basis of this fixed moment, it is possible to restore the preceding and following moments, particularly if the story is familiar. In art, the number of characters is always more limited than in a verbal text. All real, literally depicted characters are seen at the same time (naturally not the roles, given that these can be abstract forces). Moreover, depicted action is relatively easily recognised and defined, and the number of characters is limited. This a good reason to take the system of action and actor as a starting point.

Let us try to characterise the various periods of fish depiction that have already been discussed above within the framework of Greimas' actant model.

The first fish type, the fish in still-life, appears at first glance to be an object. Something has been done to the fish, the fish itself does not do anything, it does not operate with any concrete aim. The objectivity of the fish in still life instead emphasizes activity on the part of someone else. The fish in still life is the object of fishing, the object of eating, the object of depicting. If we consider one of the basic conditions of a narrative to be the existence of an active subject, then still life seems to be subjectless and must therefore be non-narrative. If, however, we extend our action and events outside the framework of the picture itself and include the artist (for example the author as narrator), we could then treat the author as a subject and the creation of the picture as the action. The creation of the picture is actually the author's story and the picture as a narrative tells this story (if it has nothing else to say and superficially the still-life pretends to do at least this). In this way the fish serves as a means, i.e. it functions in the role of the helper; the object is the picture as a finished work; the desire to depict something, to create a picture, is the sender; and the public, critic, art council, buyer is the receiver. The opponent in this model could be the bounds of human ability, human creativity, the subordination of matter to the imagination. It is clear that every still-life «tells some story» or reflects a given epoch as well as the author's standpoint at that time. In Soviet art of the 1950s the still-fish reflected sobriety and realism, whereas those of 17th-century Holland reflected an abundant catch, if nothing else.

In the case of the still fish there is one more character - the fisherman. His participation in this conditional story is considerably more problematic. The fisherman can enter the list of characters only in exceptional instances - for example should he wish to perpetuate an extraordinarily big fish that has been caught. In such instances, however, the fisherman is evidently only a second parallel subject. The still-fish was used in Soviet art to refer back to fishermen as a collective subject and related themes in general. In referring to fish as food, the fisherman is placed in the same position as a character that appears only occasionally - the eater, humanity as someone who enjoys eating fish. His role in the system of characters is, however, secondary. With respect to the fish, eating is undoubtedly adversarial in nature, but not with respect to the artist, since in the case of the still-fish he is the only possible subject. Clearly the eater can be a manifestation of a parallel subject (with the depiction of fish as a favourite food) or as is more probable, the eater as an onlooker might be equated with the object.

Since in one of its aspects the still fish participates as the object of being caught, we should therefore next examine the fourth type that we have discussed - that is, the fisherman's fish. In the case of the fisherman's fish, we can speak of the fish as an object with complete definitiveness. This fish is no longer just an excuse to create a picture, instead the desires and expectations of the fisherman to catch fish are reflected therein. The picture itself already directly conceals the story, and it is not necessary to move the actants outside the picture, although even here this possibility exists. The object status of the fish is primary in the case of the caught fish - someone wants to catch the fish, the fish is the object of capture. Nevertheless, it is possible to go further. The fish as an object of capture can be depicted more or less conditionally, with a greater or lesser degree of empathy. One can relate to the fish as an equal adversary of the fisherman or even with a greater degree of sympathy than with the fisherman. One can recognize the fish as an independent creature. Perhaps this is found more in the sphere of literature than in art, but it certainly exists as one possibility here as well. The fish as such can become even more significant than the fisherman, thereby becoming an antisubject.

But the existence of the still-fish and the fisherman's fish are not yet convincing reasons to speak of the fish as a character. The proliferation of the still-fish could still be explained away by attributing it to the inexplicable waves that are always hidden away in the bottom layers of fashion, or it could merely be motivated by the texture of the fish's surface and its exterior form. The depiction of fish made it possible little by little to depart from the framework of realism and covertly to deal with that which is characteristic of abstract art. The laconic form of the fish's surface is most appropriate for graphic art and the majority of Estonian fish pictures are in fact graphic representations. One could say that the texture of the fish has some substantial link with the whole of Estonian art in the 1960s.

The form of the fish is relatively universal. It is reminiscent of birds, lips, eyes, a slender tree; the skeleton of a fish resembles a bare tree. Ülo Sooster's flat fish were related to his egg and bird as well as his roundish juniper trees. The phallic shape of the slender fish standing on end is connected to other analogous forms (the more distant psychoanalytic background, however, lies outside the scope of the present treatment). Something mysterious for the human senses is hidden in the fish; it is silent, lives in water, and breathes by means of gills. In the art of the 1960s, however, the surface of the fish assumed importance. The treatment of the fish remained superficial. In this guise, techniques for expression were developed. At the same time it was possible to maintain a connection to the obligatory themes that were acceptable at that time - work, shore life etc. Although one might not want to believe that the framework of the still-fish or fisherman themes had any direct symbolic significance, it was not possible to avoid oblique subconscious symbols completely and these might in fact be the reason for the proliferation of the second fish type. In this way the still-fish was partly just an excuse or a smoke screen for something else.

At the end of the 1960s, and in the case of Tartu artists even earlier, the fish became independent, separate from still life and became the independent fish. This differentiation is smooth and almost imperceptible. One transition period work was The Great Turbot by Ulas already mentioned above. Actually nothing more than a change in backdrop took place. Anything that referred to a concrete foundation (a platter or dish) disappeared and the fish began to be surrounded by some sort of environment, presumably water. Ülo Sooster's and Elmar Kits' fish could be considered the most typical fish of the period. The fish is generally alone in the centre of the picture and in a horizontal position. It is most probably acting in its natural environment. In the case of this fish we are presumably dealing with a being, a character, a subject. Because of its very independence, a second source for this type might be the fish as the object of capture, hence the fish that we recognize and whose experience we share. Often, and in particular in the fish pictures of Sooster, the solitary fish can fill the whole space and itself become an environment.

Ilja Kabakov describes just this as the inspiration behind Sooster's fish pictures:

The fish motif retreats in the 1980s, only to reappear again at the beginning of the 1990s. Now, however, it is in a completely different form; as the mythological fish either static, solitary, horizontal or altogether active in a standing position. Consequently the basic characteristic of this new fish is a certain un-fish-like quality - the vertical posture of the fish is, after all, just that. The surrounding environment can also be understood in many ways. A fish on its end is anthropomorphic and humanoid. Tolts' Whale is mythical. Its name already makes reference to its connection with the biblical whale. So too is the case with the fish in Keskküla's Still-life with a myth. Within the framework of the structure of the character, mythical fish belong to the realm of subject, and more rarely to the realm of object and to helper/opponent categories (such as in the Keskküla example). These sorts of pictures contain a clear narrative which can be very static.

The actant model for the four fish can be represented schematically:

In the case of two categories (the mythological wanderer and the fish as the object of capture), humans must inevitably step into the role of characters. The human being has always been a metaphor for any character. Even animal characters are usually anthropomorphic or more or less brought closer to human notions of what constitutes action. Mythical heroes resemble humans in their acts, they differ only with respect to their powers. Plant or rock characters, in contrast, are relatively rare phenomena.

Human beings are distinguished from lifeless nature by their ability to act, think and feel independently, with volition or intentionality. The fish as a character is similarly capable of independent action or at least of directing the action. Thus we cannot avoid the relationships between the character and his actions even in the following.

The unresolved problem in the case of the character/actor/actant was their relationship to intentional action. Which is primary, and can the character be separated from the action at all? Structural narratology is convinced that they are not, and prefers action over the enormity of characters. The reason for this is clear - there is an endless number of characters, while the aims of their action provides a more certain basis for their categorization.

It is possible that the character in art does not perform a real act in the actual picture itself. Is there in such instances then no story or narrative? In figurative art it is certainly much more complicated to analyse intentional activity based on the depiction of the event at one moment in time. That moment often represents a static state. If we do not know the story that belongs to it and the title does not give any hints as to what it might be, the system of actants becomes obscured. Among the types we have examined, intentional activity is relatively easy to determine in the case of the fisherman's fish that are caught and the mythological fish. More complex instances are the cases of the still-fish and independent fish. Can we see action in just being, in so-called independent existence? The broader treatment of action in current narratology will help in this.

Donald E. Polkinghorne believes that the current narrative approach sees action as an expression of existence. Acting is like writing a story and understanding action is like arriving at an interpretation of a story. The cause of an action has changed from physical cause to expressive cause. Expressive cause is, for example, when alphabetical letters on a page «cause» the reader to understand the meaning embodied in them. Thus there is no need for the movement of physical bodies, there is no need to account for the fact that something either pushes or pulls them. Human action is actually competence; the understanding of physical movement and the ability to do it. In the same way the acts of agents can be symbolically represented in either spoken or written form (Polkinghorne 1988: 142, 143).

Paul Ricoeur believes that people possess the ability to comprehend the world of action just as they understand which phrases can form meaningful sentences. The recognition and composition of a meaningful plot requires an understanding of the kinds of human activity, how these kinds of activity can be gathered together into a plot and their temporal ordering. Ricoeur identifies six conditions for recognizing action:

1. Actions imply goals, they are carried out to achieve results or accomplish an end.
2. Actions refer to motives which explain why someone did or does something. This explanation is clearly different from the causes of physical movement.
3. Actions are carried out by agents. A person performs or is the author of an action and thus the actions are taken to be the work or deed of someone who can be held responsible.
4. Actions take place in situations that consist of closed physical systems and the agent recognizes that the circumstances set favourable and unfavourable conditions for them.
5. Actions take place in interaction with other persons whether as cooperation, competition or struggle. Cohorts can either help or hinder the accomplishment of an act.
6. The outcome of an action may bring about a change in one's fortunes or feelings.

In short, unlike in the case of a simple physical event, we are competent to recognize with human actions such questions as «what?», «who?», «how?», «with whom?» and «against whom?» (Ricoeur 1984: 55).

Among other things, it is worth recalling Yuri Lotman's definition of an event: «An event in a text is the movement of an actor over the boundary of the semantic field» (Lotman 1990: 109). In every instance this crossing of the boundary is a volitional act. Lotman further categorises characters as being «motionless or moving», that is, those who do not perform actions and those who do or in other words those that cross the boundaries. If no action is performed, the text is athematic. At the same time, Lotman emphasises the relativity of events, and consequently actions as well. At some level, depending on some normative situation, an act can be an event, and in other instances is not (Lotman 1990: 109).

We can successfully use the propositions outlined above to gain a broader understanding of narrative actions as the basis for the fish character and the third of the types we have delimited. Does someone act with respect to the fish, does the fish itself act or cause something, or is the fish a helper/adversary. First of all, the still fish and the independent fish acquire the potential to become active characters by their very being. At the same time, we can recall Lessing's ideas: the most important aspect of portraying activity in art is finding the most effective moment to do so. This moment need not be the climax of the activity. Lessing finds the opposite to be true - the best opportunity need not be fixed in the event, activity, or emotion itself so much as portray the tension and expectation involved. It must stimulate the viewer with respect to the preceding and following moments. Even stativity can contain these possibilities. In conclusion, the still fish can be an object or helper, the independent fish a subject, the mythological fish can presumably fulfil all of these functions and the fisherman's fish can be an object or adversary.

If, to follow the example set by Kabakov, we ignore the symbolic meaning of fish found in Christianity or other cultures (with the assumption that it would be difficult to find such a uniform symbolic background in Estonian art in the period under scrutiny, as mentioned earlier), then the next stage is to compare the fish types we have found in art with other images of fish found in Estonian language and folklore. These areas are naturally distinct and are not directly connected to art. The goal of such a comparison would be to find out whether the fish types found in art are characteristic to this one area alone or do they represent human attitudes toward fish more broadly. As concerns the choice of materials used for such a comparison, the possibilities are endless. Here, however, we will make only a very tentative and cursory comparison. Since the abundance of fish was a unique feature of Estonian art, we will limit our comparison solely to Estonian material. The first source material used is the Standard Dictionary of Estonian Literary Language (Eesti kirjakeele seletussõnaraamat), which should reflect quite well the language usage of the past few decades. Under the entry for fish, it is defined as: «a cold-blooded vertebrate living in water, which breathes by means of gills, moves with fins and a tail and is usually covered with scales.» The following examples illustrating the entry reflect the following activities connected with fish (Eesti kirjakeele seletussõnaraamat II 1992: 63):

1. fish as food: to smoke, salt, dry fish. Marinated, boiled, fried fish. There is fresh fish available. Garnished fish. Lots of fish was eaten;
2. fish as an independent creature: the fish made a splash. A fish looks for where it is deeper, a human where it is better;.
3. fish as a symbol (referring to something else): Swims like a fish. Dumb as a fish. Cold as a fish. Experienced as a fish. That man is a slippery fish. That sort of fish he is does get caught on a hook; an old fish, like a fish in water, like a fish on land, to go fishing in muddy water.
4. fish as an object of capture: to catch, angle for, clean, scrape, gut fish. The boat was full of fish. Fish are raised in ponds. The fishermen got lots of fish. Much fish was exported. To go fishing, to be fishing, to come from fishing.

including fish as the adversary of fishermen: Fish were biting. A fish got caught on the hook. The fish is not biting.

The examples cited reveal four attitudes toward fish; fish as food, fish as the object of capture, fish as a symbol, other activities under the guise of fish or the symbolic fish. These types are nevertheless perhaps not as clear-cut here as they are in art. In places, the boundaries between the various relationships seem fuzzy. For example, a clear boundary is lacking between the mixed types, fish as the adversary of the fisherman and fish as an independent creature as it is lacking between the latter and the symbolic fish. Differences can be seen in the scope of the examples. A large proportion of sayings are concerned with the symbolic fish. Fish as food is very significant and also very well represented. On the basis of these examples, all of the existing types that have been observed are represented, and no new types have been added to the scheme. The same actant relationships are also present. The type «fish as food and the object of capture» correspond for the most part to the object (also to the adversary if the fish being caught is considered an equal) the fish as a creature and a symbol for the most part correspond to the subject.

The four classes we have found are also very well represented in folklore. Below we will examine only proverbs and folksongs. There is a surprising lack of fish in riddles, while in folk religion they are mainly connected with fishing itself and to a lesser extent with supernatural denizens of the water. Certainly one of the reasons for this is simply a matter of pragmatics - sources of livelihood are most frequently represented in folk religion. Humans are interested in fish above all as a source of sustenance and only then as creatures. There might also have been some Christian background as well - the fisherman of the New Testament held a revered and positive vocation. Ivar Paulson has emphasised the respect and esteem displayed in attitudes to the whole marine world (Paulson 1997: 70). Of course fish examples can also be found in other types of folklore (fairy tales, legends), but this would take us too far from the framework of the immediate comparison. Longer tales are often devoted to stories of the origins of different fish, the flounder being one of the most popular of them. A great amount of quantitative emphasis is placed on the fisherman and fish as an object of capture as compared to other genres of folklore. We can find the same types of fish in proverbs with fish as we found in the dictionary. There are altogether 113 types of proverbs with fish. Some of the more characteristic of these are listed below (Krikmann & Sarv 1980-1988, Electronic archive of proverbs⁵):

Fish are most often mentioned as objects of capture in folksongs, just as they are in proverbs and folk religion. At the same time the attitude to fish in folksongs is as to equal creatures who act independently according to their own wishes. Fish have to be cajoled and coaxed. Fish are treated as individual creatures whose appearance, species and activities are significant. Of course, the element of magic incantations here is quite apparent. The following examples are taken from Ülo Tedre's (ed.) Estonian Folksongs (Tedre 1969-1974).

When we compare the folkloric examples with the fish in Estonian art, the similarity of approach is apparent and even surprising. The still fish corresponds to fish as food. The attitude toward it is relatively pragmatic, almost as if it were a lifeless object. The fisherman's fish is the parallel of the fish as an object of capture - the attitude is one of esteem. This is the usual attitude toward an equal adversary. The most respected of all is the fish as an independent creature. It was referred to as the independent fish in the art examples. And the last one - the symbolic fish, the upright humanoid fish, corresponds to the fish as a metaphor for other things. A common feature of both Estonian art as well as linguistic and folkloric material is the lack of an evil fish, a fish as an enemy or destroyer, something that should be feared (although such a creature exists in Estonian Animated films, for example in Avo Paistik's Klaabu, Nipi and the Angry Fish (1979). Such fish have been encountered in the art of the Low Countries in the pictures of Bosch and in the legend of Jonah in the Old Testament (if we take the biblical whale to be a fish). Perhaps the reason for this is the selfsame Estonian spirit that has renounced the figure of Satan and replaced it with Vanapagan, a bumbling, rather foolish giant in Estonian folklore. The obvious reason, that Estonians are a coastal people who have engaged in fishing, cannot be appropriate here, given that the inhabitants of the Netherlands are also a coastal people.

The system of four characters that we have found in post-World War II Estonian art is on the one hand a strong interface between human primeval imaginings and language and folklore and on the other hand seems to contain the developmental potential for further examination of the problem of the characters in visual art and other areas of art. We have set the fish in literature aside for the time being, but a superficial glance is enough to convince us of the existence of similar types there too. In terms of visual art, the next step is to examine the character of the human. The representational possibilities for humans - the nude (i.e. simply the representation of a physical object), the portrait, humans in activity, the human figure as an icon or a symbol - all hint at a similar system.

The four types we have found for fish do not correspond completely to the six structural aspects of the actant scheme, the sender/receiver pair does not seem to be found in the works that have been examined. At the same time all narratologists have admitted the possible abstract nature of this pair, and that they can embody some general scheme of background ideas. Evidently the same abstract ideas may be transferred to the fish.