Propp’s functions are certainly fundamental, but they may not be incidents. Fractals, being mathematical forms, can replicate the same patterns from the infinitely large down to the level of the infinitesimal, but incidents, according to many thinkers in Athens, Leningrad, Paris, and elsewhere, are more like water. Unless we know the scale of a map, we cannot always know whether it represents a gulf, a bay, or a tiny inlet, but in all of these—and even when we descend to the level of buckets, drops, and molecules—we are still talking about water. But when we divide a molecule of water, we are no longer dealing with water because we have lost the basic internal structure of H2O. In the same way, an incident may be so small that it cannot be reduced to a combination or an expansion of other incidents, but even this smallest incident has an internal structure of elements that, taken separately, are not incidents any more than hydrogen and oxygen separately are water. The critical community today would usually say that “Night fell” lacks the internal structure that would warrant the name “incident” and that Propp’s functions may need to be taken not as incidents but as parts of a larger unit that follows a fixed formula. Propp himself expressed this very idea:
Morphologically, a tale (skazka) may be termed any development proceeding from villainy[,]…a lack…through intermediary functions to marriage[,]…or to other functions employed as a denouement….
This type of development is termed by us a move (khod). Each new act of villainy, each new lack creates a new move. One tale may have several moves, and when analyzing a text, one must first of all determine the number of moves of which it consists. One move may directly follow another, but they may also interweave; a development which has begun pauses, and a new move is inserted. Singling out a move is not always an easy matter, but it is always possible with complete exactitude.
(Morphology of the Folktale, 92)
Bremond argues that if this basic pattern for a “move” is indeed fixed, if Propp’s labors have proved that functions can only take fixed positions in a fixed pattern (which Bremond doubts), then the real base unit, the narrative atom, is not the function but the series. Like Propp, Tomashevsky distinguishes three kinds of very basic units, but for him the most important seems to be what Propp passes over as intermediary in the passage just quoted. For Tomashevsky, “motifs are easily distinguished according to their importance. Dynamic motifs take the first position, then preparatory motifs, and then motifs defining the situation, etc.” These three kinds of motif correspond—in reverse order—with Algirdas Julien Greimas’s three types of narration: “les énoncés descriptifs, les énoncés modaux, et les énoncés translatifs.” The French Structuralists took many of their ideas from the Russian Formalists, but laid them out with greater elegance than any of the Russians except perhaps Shklovsky. Greimas, an Eastern European operating in France, defines the first kind of motif as something that can be expressed in a sentence with the verb “to be” and an adjectival word or phrase; the second can be expressed with a modal verb such as “wants to,” “ought to,” “needs to,” etc.; and the third with an active verb. The hero is in a special state; he needs something; he achieves it. This sequence seems to define linguistically what the Russian Formalists were struggling to convey but never found the terminology to explain succinctly.
Shklovsky expresses this internal structure for a plot more clearly than anyone before the French Structuralist clarification of Formalist concepts:
I have already said that if we take some typical story-anecdote, we will see that it represents something complete [zakonchennoe]. To take for example as one’s material: a shrewd answer which gets the person out of a tight spot, we have the explanation of the tight spot into which the character has fallen, his answer, and a particular resolution; such is the structure of “trickster novels” in general. For example, the man who is marked after a crime by having a lock of hair cut off cuts it off from all his friends and thus saves himself; similarly in a story of the same type about a house marked with chalk (1001 Nights, Andersen). Here we see a definite…cycle which sometimes unfolds with descriptions, characterizations, but which in itself represents something resolved. Several such stories can form a more complicated structure, being set into a single frame, so to speak, united into a single plot cluster.
(Theory of Prose, 68)
To sum up this traditional definition of an incident: it may be as large as a whole novel or as small as three sentences, it may be treated from the creator’s point of view as one of the building blocks for a text or as the generating entity whose transformation forms the text, it may be treated from the reader’s point of view as one of many components discovered by analysis or as the simple outcome of a summary, but the tripartite internal structure of an incident emerges near the center of the best investigations of this question, as it did when Aristotle characterized the plot of a good drama as something with a beginning, a middle, and an end. A beginning is something which demands no prior material, like the exposition of a situation using the verb “to be.” An end needs no subsequent material, like the action that resolves the problem. A middle needs both prior and subsequent material, as does the entire interplay of needs and desires which can be presented in modal verbs.
The most sophisticated modern theorists seem to converge with Aristotle upon the observation that the plot of a successful literary work, and each of its component incidents, has this tripartite internal structure whose components may themselves be incidents or may be smaller units which lack this internal structure, just as oxygen or hydrogen lack the internal structure to be water.