Porter Institute for Poetics and Semiotics

Towards a Theory of Space in Narrative

Author(s): Gabriel Zoran
Source: Poetics Today, Vol. 5, No. 2, The Construction of Reality in Fiction (1984), pp. 309-335
Published by: Duke University Press
Stable URL: http://www.jstor.org/stable/1771935
Accessed: 13-01-2016 08:16 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

Duke University Press and Porter Institute for Poetics and Semiotics are collaborating with JSTOR to digitize, preserve
and extend access to Poetics Today.

http://www.jstor.org

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY
OF SPACE IN NARRATIVE*

GABRIEL ZORAN
Hebrew and Comparative Literature, Haifa

The purpose of this study is to present a general model of the

structuring of space within the narrative text. The term space is used
here to mean specifically the spatial aspects of the reconstructed
world. This seems natural and rather obvious, but the term can be
applied to the literary text in various ways and is, itself, far from
unambiguous. It is.necessary, then, to examine the whole range of
problems arising from the use of the term with regard to the literary
text. If this task as a whole lies outside the scope of the present
study, we can at least clarify one essential aspect of it: the relation-
ship between space and time in the narrative text.

1. THE ASYMMETRY OF TIME AND SPACE IN THE NARRATIVE

In the extra-literary field, there are good reasons to combine space
and time: they are the two complementary aspects that together
cover all the dimensions of empirical existence. Accordingly, it
makes no difference whether one holds to a separative theory which
considers the aspects of space and time to be parallel yet indepen-
dent - e.g., the theories of Euclid, Newton, Kant, and Leibnitz -
or whether one supports a theory which regards these aspects as
interdependent, composing together a single, four-dimensional
complex (chronotopos, spacetime) - as does Einstein. Independent
or interdependent, space and time are perceived as complementary
aspects of equal status, belonging to a common field of debate.

* This is a somewhat revised version of a

paper presented at Synopsis II, "Narrative Theory
and Poetics of Fiction," an international symposium held at the Porter Institute for Poetics
and Semiotics, Tel Aviv University, and the Van Leer Jerusalem
Foundation, June 16-22,
1979.
It is based on parts of a thesis for the doctoral degree prepared under the
supervision of
Professor Benjamin Hrushovski.

Poetics Today, Vol. 5:2 (1984) 309-335

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
310 GABRIEL ZORAN

This conception of the relations between space and time is

occasionally applied to the narrative text, and is acceptable from
several viewpoints. From others, however, it is problematic. The
relationship between space and time in the narrative text lacks both
the clarity and the symmetry it possesses when applied to the field
of reality. Literature is basically an art of time. Although no one
today would state this as baldly as Lessing (1974:102-115) did, the
dominance of the time factor in the structuring of the narrative text
remains an indisputable fact.
The existence of space is pushed into a corner, so to speak. It is
not altogether discarded, but neither does it have a recognized and
clear-cut status within the text. It can be understood in various ways,
but none is as clear and unambiguous as the term time. This lack of
symmetry in the relationship between space and time is evident not
only in their status in the text, but also in the extent of the progress
of research on these concepts. Although the subject of space has
been dealt with more than once, research in general on the subject
is quite diffuse, and there are few assumptions that have become
generally accepted.
The difficulty apparently lies in one basic difference between
space and time in narrative. One may speak of time in terms of the
correlation between the structuring of the text and that of the world,
whereas it is impossible to speak about space in such terms. What-
ever specific terms are used in discussing time, they will always be
dominated by the basic opposition between the time of the text
and that of the world (narrated time and time of narration, erzdhl-
and to
zeit-erza'hlte Zeit, time of presentation and presented time,
a large extent also fabula and sujet [see especially Mtiller 1950;
The various
Liment 1955; Tomashevsky 1965; Even-Zohar 1968]).
possible relationships within these pairs of components can create
a wide range of categories, based on the modes of correlation
(whether it is a negative or a positive correlation [see Even-Zohar
19611), and on specific types of deviation from the "natural" struc-
turing of time (such as contraction, reversal of temporal order, etc.).
In principle, one may also distinguish between the application of
the term space to the reconstructed world and its application as a
dimension of the verbal text itself. The conception of the verbal
text as an exclusively temporal structure (as, for example, Lessing's
conception), is not acceptable today, and one may mention in this
connection Frank (1963) and Segre (1975). Nevertheless, despite
the possibility of distinguishing between the space of the text and
that of the world, one cannot point to any constant correlation
between them.

1.1. The spatial dimension of the text may be conceived of as its

graphic existence. There are texts, particularly those of concrete

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 311

poetry, in which graphic space is exploited and activated as an

inseparable part of the general structuring of the text. Still, this does
not alter the fact that language is a system of arbitrary signs; in other
words, the connection between the signifier and the signified is not
based on any real structural similarity between them, but simply on
convention. The fact that a certain poem may attempt to create
structural similarity between the described object and its graphic
organization (for example, Lewis Carroll's poem arranged in the
shape of a mouse's tail) does not signify any fundamental connection
or constant correlation between the verbal text and the world, but
rather the opposite - the correlation here is an unexpected thing,
a curiosity. It is completely different from the correlation which
exists between the two orders of time in the narrative. The latter is
a permanent phenomenon based on the temporal structure of
language, existing even when there is a contradiction between the
two orders of time (a negative correlation is not the same as a lack of
correlation altogether).

1.2. It is a more complicated matter when the dimension of space is

attributed to the structure of signifieds in a text. Here the reference
is to the structuring of meanings into a pattern not identical to the
temporal order in which they appear in the text. A spatial pattern is
any pattern perceived solely on the basis of the connection between
discontinuous units in a text, demanding therefore a perception of
the whole text or part of it as given simultaneously in space (which
is, for example, the case of analogies).1
Two problems arise when the term space is used to describe such
patterns: (a) the problem of whether the term has the same meaning
here as it has had thus far, and (b) the problem of whether we can
find a kind of correlation between spatial structure in this sense
and the space of the world.
As far as the first problem is concerned, I believe that we have
here a completely different usage of the concept of space. The
justification for the use of the word here is that we are speaking of
intemporality and coexistence.2 These may, of course, be regarded
as important aspects of space; however, if we consider the term as
it is generally used, and as we have used it thus far, it becomes clear

1. This common use of the adjective "spatial" is inspired mainly by the theory of Frank
(1963) about the spatial form in modern literature. Frank, however, used the concept
mainly to describe a property of an historical corpus of texts, while today it is used to
describe general textual properties. See, for example, Sternberg's (1973:228-230) dis-
cussion of analogy, as well as some later articles of Frank himself (e.g., 1978:275-290).
2. A definition based on such a conception is the classic Leibnitzian definition of space
as "the order of possible coexistences" ("ordre des coexistences possibles"), but it is not
mere chance that Leibnitz conceives of space as a subjective relative system, while Newton
conceives of it as something absolute and objective, a kind of "reservoir" of reality.

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
312 GABRIEL ZORAN

that the above usage is not the usual one. For one thing, space is not
necessarily an absence of time; the fact that things are not arranged
in chronological order, but rather in a simultaneous pattern, does
not make them necessarily spatial, except in a purely metaphorical
sense. Again, our perception of space has to do with concepts such as
volume, extension, and three-dimensionality, and all these are
unrelated to the concept of the spatial pattern. This pattern has no
particular location, no contours, no volume. It is a kind of abstract
organization, but it has nothing to do with the real existence of the
thing it organizes.
As to the second problem - whether one may speak of the
connection between the spatial pattern and the space of the world -
here the situation is more complex. It is impossible to reconstruct
the space of the world without structuring the information about it
into some kind of a "spatial" pattern, so there is a certain connection
between the two. But this characteristic is not necessarily limited to
reconstructed space alone, since the reconstruction of every aspect
of the world necessitates a "spatial" point of view - psychology,
characters, norms, and even, strange as it may seem, plot and time.
Still, there is no doubt that the reconstruction of space is especially
dependent on a "spatial" point of view.3 In any case, it is important
to distinguish between the "spatial" point of view and the spatial
object viewed. Moreover, although the connection between the
components is a permanent one, this connection can certainly not
be perceived as a correlation. The spatial pattern of the text does
not stand in any kind of correlation with the space of the world.
Finally, whatever the connection between the spatial pattern and
the world may be, it should be emphasized again that the spatial
dimension of the text has no autonomous existence. The text exists,
and is structured first and foremost in time. The so-called "spatial
pattern" is actually nothing other than a superstructure of a sub-
stance whose basic structure is in time. It is thus impossible to
"bypass" the time factor in the narrative. The narrative, with all
its components, is arranged in time, so that in a certain sense one
may speak of a temporal arrangement of space. We must therefore
identify the various principles of transformation from a world
existing in space to a medium structured in time.

2. THE TRANSFORMATION FROM SPATIAL OBJECTS TO

THE MEDIUM OF TIME
2.1. The possibility of describing the relation between the verbal
text and the world in terms of transformation is not a peculiarity

3. The importance of the "spatial" point of view for the reconstruction of fictional space
will be discussed in more detail in the section dealing with the field of vision. (See 4.1 and
especially 4.1.4.)

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 313

of the poetics of space. There are principles governing the trans-

formation of every component of the world into the structuring of
information about this component in the text. Characterization,
for example, can be conceived of as transformation from the physiol-
ogical-psychological existence of characters in the world to their
textual existence, i.e., dialogues, descriptions, actions, etc., all
arranged in a temporal verbal continuum. Needless to say, the
transformation is a two-way relationship from the text to the world
and vice versa, and does not necessarily reflect real processes of
reading or of creating.
The peculiarity of the transformation of the time factor in the
narrative is thus characterized by the fact of its being a transition
from one type of temporal structure to another. For this reason,
one can also speak of a correlative relationship between them. This
type of relationship is also possible for the dialogic component of
the text, but for most aspects of the reconstructed world it makes
no sense to speak of a correlation. The transformation takes place
between completely different levels of organization which have no
structural similarity between them. Space is unique in that here the
transformation from an object to a system of signs involves also a
transformation from a spatial arrangement to a temporal one. A
discussion of the problem of space in the narrative requires, first of
all, a confrontation with these principles of transformation.

2.2. Space as it appears in the narrative is a very complex pattern,

and only a small part of its existence in the text is based on direct
description. It is actually a combination of various kinds and levels
of reconstruction. Before dealing with such a complex problem, let
us examine how, in general, language is activated to describe an
object in space.
A spatial object is characterized by its being complete, full, and
existing simultaneously. In the attempt to give verbal expression to
the structure of such an object, the object must first lose some
of its "completeness," since it is impossible to give an identical
expression to all its parts and aspects: some of them may be
described explicitly, some of them implicitly, and some bypassed
altogether. Language cannot give full expression to the spatial
existence of any object. Second, when the simultaneous parts are
expressed as units of information, they must receive some kind of
temporal arrangement. One may begin from the overall pattern and
pass to the individual units or vice versa; the various units can be
arranged in different ways: from up to down, front to back, the
important to the subordinate, etc. In any case, the spatial aspects are
cut off, so to speak, from their spatial and simultaneous context,
and are arranged along a temporal line.
Thus far we have discussed descriptions of completely static

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
314 GABRIEL ZORAN

objects. However, space does not involve only static objects and
relationships - things may also move and change. Space is one aspect
of spacetime (chronotopos). Language can exploit this situation by
arranging its items of information by linkage to a movement. This
movement can be a real route of an object or a transfer of the look
or the thought from one object to another. In daily speech there is
even a marked preference for the spatial-temporal method of arrange-
ment (i.e., along a line of movement) over the static, map-like
structure, as was shown by Linde (1974).4

2.3. When we transfer these considerations from the discussion of

the possibilities of language in general to the discussion of the
structures of narrative texts, we must take into account two essential
differences. First, as far as the usual verbal usage is concerned, the
objects of space and of the world in general constitute an external
factor not dependent on language, whereas within the narrative
text neither space nor the world have an independent existence but
rather an existence derived from the language itself. In terms of
Benjamin Hrushovski's text theory (1974, 1975), this is an internal
field of reference. Thus, the important question to be posed at this
point is not only how is a given space expressed in language? but also
what is the nature of this space? The decisions of the text bear not
only on the verbal material, but also on the world.
The second difference which must be taken into account is that
the movement discussed above, which can serve in daily speech as
a vehicle for a certain arrangement of information, can have in the
narrative text a most central and dominant status. In fact, the entire
plot can be regarded as such a movement. A plot can be "space
oriented," serving as a motivation for the inclusion of space units; an
extreme example of this is "Nils Holgersson's Wonderful Journey,"
by Selma Lagerl6f, whose plot serves to present, in several ways,
a complete geography of Sweden. Obviously, plot is not usually
subordinate, especially in relation to space, but whatever the status
or functions of plot in the text, it must be seen as more than simply
a structure in time. It includes routes, movement, directions, volume,
simultaneity, etc., and thus is an active partner in the structuring of
space in the text.
The transformation of space into the temporal-verbal text may be
sketched as shown in Figure 1.
The different points of the verbal continuum may refer directly to
points in space or to points in the continuum of events, relating to
space through spacetime (the chronotopos). The text can refer to

4. In a study she carried out on the ways English speakers from a given group describe
apartments, Linde discovered that most people organize their description as a "walk"
through the apartment and only a few do it like an aerial "map."

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 315

space-tinme

verbal continuum of
continuum events in time
Figure 1

space and spacetime units which are large or small, complete or

partial, and can structure them in any order and manner it chooses.

3. THE THREE LEVELS OF STRUCTURING

From the above sketch, one may distinguish three different levels
of the structuring of space in the text:
The topographical level: space as a static entity (expressed in the
diagram by the bottom of the cube).
The chronotopic level: the structure imposed on space by events
and movements, i.e., by spacetime (in the diagram, the projections
of the inside of the cube on the lower square).
The textual level: the structure imposed on space by the fact that
it is signified within the verbal text.
These levels all belong to the reconstructed world, and can be
regarded as three levels of reconstruction. The most immediate level
of reconstruction is the textual one, in which the world still retains
several of the structuring patterns of the text. In the chronotopic
level, the reconstructed world is already independent of the verbal
arrangement of the text, but is still dependent on the plot. Finally,
on the highest level of reconstruction, the topographic one, the
world is perceived as existing for itself, with its own "natural"
structure, cut off entirely from any structure imposed by the verbal
text and the plot.
As mentioned in the discussion of transformation (section 2.1),
it must be remembered that no temporal order is implied here; the
three levels of reconstruction do not represent different stages in a
real act of reconstruction or of creation. The reader does not begin
at the textual level and then pass on to the others, or vice versa;
rather, he is continually moving back and forth among the three
levels and, moreover, he perceives them at once without being able
to separate them.

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
316 GABRIEL ZORAN

The levels may be compared to three diapositive slides covering

each other. There is a difference in clarity among them: they are not
all receptive to the same extent, nor do they attract attention
equally. From the observer's point of view, however, they are always
perceived together, one through the other.5

3.1. The Level of Topographical Structure

As mentioned, this is space at its highest level of reconstruction,
perceived as self-existent and independent of the temporal structure
of the world and sequential arrangement of the text. The text can
express topographical structure by means of direct descriptions, e.g.,
as in Balzac's well-known openings, but in fact every unit of the text,
whether narrative, dialogic, or even essayistic, may contribute to
the reconstruction of the topographical structure.
This structure may be conceived as a kind of map based on
elements from the entire text, including all its components. True,
a map such as this cannot be entirely exhaustive. Some of its areas
are blank, and in the real world it may not be of much use in finding
one's way. For the purpose of reading, however, it provides a
sufficiently clear picture of the world.
The map is based on a series of oppositions, some of which are
general and typical, others of which are more specific. It encom-
passes the horizontal structure of the world, relationships such as
inside and outside, far and near, center and periphery, city and
village, etc. It may also include contours signifying the vertical
organization of the world and representing the opposition up-down.

5. The distinction between the three levels of structuring owes a lot to two existing
distinctions: that of Petsch (1942:162-189) and that of Kristeva (1970:191-197), but it
differs from these in some principal points.
First of all, these distinctions are binary, not ternary, and both seem to be a result of the
tendency to describe space in terms entirely symmetrical with those used to describe time,
and in describing time the binary distinctions are actually clearcut.
Petsch discriminates between Raum and Lokal. The Lokal is space given in itself, and is
thus more or less parallel to the level of topographical structure. Raum is space connected
with other levels of the text. Among others, it is connected with the time factor, being a
space revealed step by step. But this important aspect is pointed out only as an aside;
actually the Raum is connected with many aspects and properties of the text that have no
real connection with each other: it is also space experienced by the characters, it is
symbolic, it bears meanings, and so forth. Petsch's distinction is somewhat imbalanced
"in favor" of the Raum. The Raum is, in the final analysis, anything of "interest" one may
say about space, while Lokal is nothing but a neutral material, lacking any significance in
itself. In contrast, the distinction between the three levels of structure proposed here has
nothing to do with meanings: every level can be equally meaningful and function within the
text as a whole.
Kristeva's distinction is closer to my proposal. She discriminates between espace textuel
and espace geographique. The geographic space in her analysis may parallel both the topo-
graphic and chronotopic levels, and the espace textuel, of course, parallels the textual level.
But the development of the distinction and the description of the levels are quite different.

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 317

In addition, the map has patterns which refer not to the location of
things, but rather to their quality - patterns of colors, substances,
types of objects, etc.
Unlike topographical maps in reality, this map can structure space
on the basis of ontological principles as well; that is, space can be
divided up according to the modes of existence of its units. These
"modes of existence" sometimes overlap with the factor of topo-
graphical location: for example, the world of the gods - up; the
world of man - down. Yet they may relate to one another in rela-
tionships in themselves completely unspatial, such as the relationship
between the space of a dream and that of reality within the narrative.
Again, the ontological levels may be completely differentiated from
one another, or they may be mingled, appearing together in one
continuous space, such as in fantastic tales.
It is difficult to define beforehand all the different possibilities
of patterns in the topographical world, for these are not dependent
on the logic of the verbal text - on the contrary, as far as language
is concerned, every structure is possible. The possibilities open to the
writer are, instead, dependent on his personal outlook, tradition,
culture, individual qualities, etc.
Only one aspect of the structure of topographical space is
dependent on the logic of the narrative text: the special spatial
existence of the characters. The characters are generally perceived
as belonging to a separate narrative level with its own particular
problems. It should not be forgotten, however, that they also exist as
physical bodies in space, but the fact that they have many important
functions in other areas of the text makes them, spatially, a distinct
and exceptional entity. The formation of a character's external
appearance constitutes a special problem, different from the
formation of an inanimate object - although every text expresses
this difference in a different way. Imagine the grotesque effect that
would be created if a character was handled as a physical object.
In principle, therefore, one may state that the differentiation
between subject and object determines a basic differentiation within
space - between the external appearance of the characters and the
environmental objects.6

6. Some remarkablestudiesdealingwith spaceareactuallyconcernedwith questionswhich

may be regardedas belongingto the topographicallevel of organization.An outstanding
example is Bachelard's(1974) poetics of space, which is a discussionof the topographyof
space in the poetical imagination.However, Bachelard'sconception of the literarytext is
very problematicand quite remote from what is presupposedhere. Otherstudiesregardthe
topographicalstructureto a largeextent as a structureof signs,and connectit with the level
of meaning(e.g., Wilbur'sstudy of the house in Poe's tales [Wilbur1967]). This is typical
for studies based on a mythical conception of space, whether they deal explicitly with
mythic or folkloricmaterial,or whetherthey revealthe mythic level in written literature.
See, for example,Segal's(1974) discussionabout Soviet Structuralism.

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
318 GABRIEL ZORAN

3.2. The Level of Chronotopic Structure

As mentioned earlier, this plane has to do with the effect on the
structure and organization of space produced by the chronotopos,
the movement and the action of the narrative. Before continuing,
a clarification of the term chronotopos here is necessary. This
Einsteinian term was introduced into literary criticism by Bakhtin
(1978), who uses it to signify the entire complex of space and time
together, including physical objects, events, psychology, history,
etc. I, however, have not used the term to signify the totality of
space and time, but rather to describe a specific aspect; i.e., not to
signify all things that may be found in space or in time, but only
what may be defined by an integration of spatial and temporal
categories as movement and change. One may thus speak of the
effect of the chronotopos on the structure of space.
Within the overall chronotopos of the reconstructed world, one
should distinguish between synchronic and diachronic relationships,
which each have a different type of effect on the spatial structure.

3.2.1. Synchronic Relations: Motion and Rest. At every point of

the narrative, that is, at every synchronic situation, some objects
may be found at rest and others in motion. Naturally, the distribu-
tion can vary from point to point. One may generalize and state
that there are certain objects in space which are characterized by
their capacity for movement and others which remain at rest. This
is not the same as the relationship between subject and object, or
between characters and environmental objects; the differentiation
between the states of motion and rest may be determined among
"inanimate" objects and among characters alike. There are characters
which have a capacity for movement and there are those which are,
so to speak, tied to their places (see Uspensky 1973, Lotman 1973).
It is important to remember that movement and rest are relative
terms; rest is the state of being bound to a given spatial context,
while movement is the ability to cut oneself off from spatial context
and to switch over to different contexts. As to the nature of the
spatial context itself, this is determined by the narrative. For
example, the Cyclops in the Odyssey can move about freely on his
island, but the structure of the work - based on Odysseus's move-
ment from place to place - determines the Cyclops's island as a
single context, and the Cyclops as a character at rest.

3.2.2. Diachronic Relations: Directions, Axes, Powers. The dia-

chronic structure of the chronotopos also imposes its own structure
on space. Space, in its topographical structure, is all potential - it is
neutral, with regard to any specific movement or direction, and one
may seemingly move within it, from and to any point. In contrast,
the chronotopos determines defined directions in space: in the space

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 319

of a given narrative, one may move from point a to point b, but not
vice versa; in another narrative, the movement may be reversible. In
both cases, movement ceases to be potential; it is fully realized, and
is, so to speak, incarnated in space. Thus, for example, in the topo-
graphical space of the Odyssey, Troy and Ithaca are two places, and
there is a possibility of moving from one to the other. But the actual
direction of movement is determined by the chronotopic structure;
thus, one place is defined as the point of departure, another as the
target, and others as stations on the way, deviations, etc. Thus, axes
of movement in space are determined; one may state that space,
on the chronotopic level, is structured as a network of axes having
definite directions and a definite character.
Axes may or may not be determined by motions which actually
take place in the world of the text. An actual movement is a result of
several powers: will, obstructions, ideal, characters' intentions, and
so forth. These powers can also act in space when there is no real
movement. In Kafka's The Castle, for example, the line stretching
between the village and the castle is the central axis in the spatial
structure of the novel, focusing all the powers acting in the "world,"
despite the fact that it is never actively traversed by the main
character. Chronotopic structure of space does not mean an
occasional movement on a neutral scene, but rather a conception
of the entire space in terms of afield of powers.7

3.3. The Level of Textual Structure

To reiterate, this level encompasses the structure which is imposed
on space by the fact that it is formed within the verbal text. It
should be emphasized that the structure under discussion is not
that of the text itself as a verbal medium, nor that of its linguistic
materials, but rather an organization of the reconstructed world.
This structure, however, although applied to the "world," is not
derived from it. The objects structured belong to the reconstructed
world, but the structure itself is imposed on them by the linguistic
nature of the text.

7. The concept of the axis appears in Meyer's (1957) analysis of space in Goethe's Novelle,
where looking through a telescope creates an axis connecting between the two main
localities. However, Meyer's use of the word is rather local and metaphoric, and I believe
it can be made systematic and consequent. Brown (1967) tackles, though not in detail,
questions which seem to me to belong to the chronotopic level, when he tries to classify
spaces according to the directions of movements taking place within them. The conception
of space as a field of power is inspired mainly by the theory of Kurt Lewin, who tried to
apply the physical concept of field and the principles of topology to psychology and social
sciences (Lewin 1936, 1938, 1957). This method seems to me applicable for textual
analysis
as well, but for the time being its poetic potentialities have not been developed to a
large
extent. (For a few exceptions, see Lotman 1973, O'Toole 1980).

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
320 GABRIEL ZORAN

One may compare this to the relationship between the fabula and
the sujet. Both belong to the reconstructed world, both may be
regarded as levels of organization of elements of reality (events),
but the fabula retain their natural arrangement whereas the sujet
forces on the motives the verbal order of the text. Still, the sujet in
itself is not identical to the verbal level. Here, too, at the level of
textual structure, there are patterns of organization imposed on the
reconstructed world which are not natural to it, neither as space
nor as spacetime, but are rather forced onto it because of its being
signified in a verbal text.
These patterns of organization have to do mainly with three
aspects of the verbal text: (1) the essential selectivity, or the in-
capacity of language to exhaust all the aspects of given objects;
(2) the temporal continuum, or the fact that language transmits
information only along a temporal line; (3) the point of view, and
the perspective structure of the reconstructed world due to it.

3.3.1. The Selectivity of Language, and its Effects. The fact that
language cannot express all aspects of space results in a certain
measure of selectivity. It may express some things in a concrete way,
others in a vague or general way, and may ignore still others al-
together. What selection actually takes place is of course up to
the specific text, but in any event there must be some selection.
Language is not able to give a complete and continuous report on
space and, moreover, the reader does not always demand such a
report. The reader is much more demanding about the filling in of
gaps on the narrative plane than he is about the filling in of gaps in
space.8 There are many gaps in the information about space, and it is
not essential to fill them all. They do not always attract attention
during the process of reading. Yet their very existence causes a
permanent distinction in space between absolute, clear, specific
elements and unclear, unspecific elements. This distinction has
nothing to do with the "real" existence of space in the reconstructed
world, but rather with its verbal existence in the text. Thus, entire
areas in space may be differentiated from one another by the type
of verbal selection carried out in them. In the Odyssey, for example,
the events taking place on earth are sketched in great detail, while
the scenes on Olympus do not materialize and lack detail. This
distinction is congruent with that between two ontological areas,
but the congruence is not an automatic one. In the Iliad, the nature

8. The concept of informational gap connected with the structure of plot has been
developed to a large extent by Perry and Sternberg (1968) and Sternberg (1973). But, as
will be shown in the following discussion, I tend to focus rather on what they call "gaps
the filling of which is automatic and unnecessary," i.e., my discussion makes more use of
Ingarden's (1965) broader and more neutral concept of spots of indeterminacy.

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 321

of the verbal selection is identical in both areas: this, of course,

creates a completely different image of the relationships between
the world of gods and that of man (see also Hellwig 1964).

3.3.2. The Linearity of the Text. Recognition of the principle of

the temporal continuum of language, and the necessity of structuring
information about space in a temporal continuum, leads us to ask:
(1) what is the principle of segmentation of spatial information;
that is, how does the text pass from one unit of space to another;
and (2) what effects does the order of transmission of information
have on the image of space and the way it is reconstructed.

3.3.2.1. The principle of sequential structure, or the ordering of

spatial units in the text, may be borrowed from the spatial structure
discussed above. It may be based on the chronotopic level - on the
tracing of movement in space: the movement of a character, of an
object, of the eye, etc. It may be based on the topographical level,
proceeding from one object to the object nearby, from the sur-
rounding to the surrounded, from the upper to the lower, etc. But
it may also follow orders which are not spatial in themselves:
catalogues of items belonging to a similar category, functional
relationships, or scales of various kinds, etc. Naturally, all the above
principles may intermingle, overlap, dominate in various degrees,
and connect units of different scopes.

3.3.2.2. The Effects of the Order Chosen. In the same way that
different orders existing in space can motivate the arrangement of
the text continuum along a certain line, the same orders can be
especially stressed by means of the continuum structure. When, for
instance, the text passes from high objects to low ones, the vertical
dimension of space is stressed more than its other dimensions. The
text continuum can also impose kinds of direction upon space. This
process is similar to that of the axes previously discussed, but here
the directions are not determined by powers or motions in space,
but only by means of the verbal arrangement. One should also take
note of the different effects of the spatial image if the text chooses
to move from the internal to the external or vice versa, from the high
to the low or vice versa, and so forth.
Another effect worth noting concerns the order in which informa-
tion of various scopes is rendered. The text supplies information
about concrete local items which compose space as well as about the
wider global contexts within which these items are structured. When
the global information appears at an early stage in the description,
the concrete items join in later on, and the picture takes on a unified
character. On the other hand, it is possible to delay the appearance
of this global information, in which case the individual items appear

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
322 GABRIEL ZORAN

- at least for a while - without a clear-cut context, and one receives

the impression of a non-unified, disconnected space.
3.3.3. The Perspectival Structure. The point of view in the text
forces upon the reconstructed space a perspective structure. This
structure differs from the perspective organization of space in a
drawing, although it may sometimes aspire to similar effects. In
principle, the spatial perspective of a drawing (or a photograph) is
based on a continuous line stretching from the onlooker out to the
horizon along which the sizes of objects get continually smaller.
In contrast, the spatial perspective of language is not based on a
continuous line but on a binary opposition of here and there.
Here-there relationships occur in two ways: between the spatial
location of the act of narration and the "world" as a whole; and,
within the "world," between things perceived at a certain instant
as in the foreground and those perceived as in the background. These
two here-there relationships are parallel to two types of coordination
systems in language (Miller, Johnson-Laird 1976): the deictic system,
whose center ("hero") is the spatio-temporal location of the speech
act; and the intrinsic system, whose center is any point in the world
chosen for that purpose. These two centers exist side by side
throughout the text as a matter of principle. But the relationships
between them can change at any point: they may come closer
together, one may become prominent at the expense of the other,
the objects which are here and there may change, and the relation-
ships between here and there may be reversed.
4. THE "HORIZONTAL" STRUCTURE OF SPACE
4.0. So far we have discussed the three levels of space in the text.
The differentiation between these levels is to a large degree "ver-
tical." In order to proceed further with this analysis, it is necessary
to take account of a "horizontal" viewpoint as well; that is, to
examine the parts of space, its boundaries, its scope. Up to now we
have not discriminated between units of different scopes, and have
considered the space of the narrative as a complete whole. We must
now examine the nature of this whole. It may be seen as a complex
entity comprised of parts or, alternatively, it may be regarded as
a single unit forming part of some spatial totality extending beyond
it. Thus, one may speak of three possible scopes of spatial units:
the total space which encompasses the world of the text; the spatial
complex which the text actually presents; and the spatial units which
compose this complex.
It should be evident, in the light of the distinction between the
three levels of structure, that the different scopes of spatial units
are not expressed identically at each level, nor do they necessarily
correspond from level to level. The discrimination between the
scopes is a logical one rather than a clear-cut boundary in space.

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 323

the total space

the complex of space

the unit of space

--- 1 levelof textual

---t structure

-- ---- level of chronotopic

structure

---- I - - level of topographic

t structure
-

4.1. Units of Space: The Field of Vision

4.1.1. The Basic Units Composing Space. The largest units that can
be conceived as parts in relation to the entire spatial complex would
be something in the scope of a scene, presuming that space is a series
of such scenes, and that each such unit consists of many smaller
units. Within the three levels of spatial structure discussed above,
a scene on the topographical level is a place, on the chronotopical
level a zone of action, and on the textual level a field of vision.

4.1.1.1. Places may be houses, cities, streets, fields, mountains,

forests, etc. A place is a certain point, plane, or volume, spatially
continuous and with fairly distinct boundaries, or else surrounded
by a spatial partition separating it from other spatial units.

4.1.1.2. A zone of action is not defined by spatial continuity or a

clear topographical border, but rather by the proportions of the
event taking place within it.
The event itself has nothing to do with given spatial borders nor
does it necessarily take place in a defined topographical unit; it is
defined, rather, by its relationship to other events which occurred
before or after it. Several simultaneous events may take place within
one place, for example, in a room; or, a single event may take place
in a discontinuous space. A telephone conversation, for example,
is one event which takes place in two nonadjacent, disconnected
places (ignoring for the time being the assertion that telephone
wires or radio waves are a part of space; they are a part of the
physical space, but not of the human event defined as the telephone
conversation).9

9. This claim is based on the well-known distinction between mathematical-physical space

and experienced space, in which human life actually takes place (erlebter Raum, espace
vicu). See Minkovski (1933, 1936); Binswanger (1955); Bollnow (1963); Bachelard (1974).

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
324 GABRIEL ZORAN

4.1.1.3. The concept field of vision may be understood if we

examine for a moment the example of the telephone conversation,
considering the different ways in which such an event can take shape
in a text. The text may attach itself to one of the conversants in his
room, leaving the reader to reconstruct the words of the other one:
in this case, we would say that the field of vision of the text, or of
this section, is attached to a place (room) but not to an event. Or the
opposite: The text may choose to describe the conversation from the
point of view of both conversers, seeing each one as he acts in his
own place: in this case, the field of vision attempts to attach itself
to the disconnected zone of action. But a field of vision is not
necessarily limited to places or zones of action: it may actually
encompass any spatial unit.
In our discussion on the perspective structure of space, we stated
that every point in the text has certain elements perceived as "here"
and other elements perceived as "there." One may define the field
of vision as that part of the world perceived as being "here." Other
fields of vision which preceded it in the continuum, or which will
follow it, and spatial units indirectly formed or unrealized as fields
of vision - all these are perceived as "there."
Although we perceive the field of vision as being here, it should
not be understood as a focalized, spatial unit given to clear localiza-
tion. We have already seen that from the topographical point of view,
it does not necessarily have to encompass a solid, continuous, or
closed unit. The text refers to this spatial unit, however, as if it
were something continuous and defined, and it surveys it in a single
view, independent of real conditions of vision or of perception
existing in the world.
The field of vision of the text is thus different from the ordinary
optical field of vision. The text may refer to an entire city as a field
of vision, to a split event (such as a telephone conversation), to a
complete battlefield, to a complete house (disregarding the walls
which divide its rooms), etc. Naturally, there is also the possibility
of following optical rules of perception in a field of vision, but this
is only one convention among others and it is no more "natural"
for the field of vision than the convention of linear perspective is
for a drawing.10

4.1.2. Two Examples of Field of Vision. To illustrate the Field of

Vision, let us pick two concrete examples - the opening sentences

The importance of this distinction for the history of the concept of space in literary
criticism is highly important.
10. Examples of analyses of visual and sensual structures of fields of vision may be found
in Alewin (1957) and Iskra (1967), although their discussions are in terms of description
and of course not of field of vision.

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 325

of two short stories: Wolfgang Borchert's "Die Drei Dunklen

K6nige" ("The Three Dark Kings") and Heinrich von Kleist's "Das
Erdbeben in Chili" ("The Earthquake in Chile").
Er tappte durch die dunkle Vorstadt. Die Hiuser standen abgebrochen gegen
den Himmel. Der Mond fehlte und das Pflaster war erschrocken fiber den
spaten Schritt. Dann fand er eine alte Planke. Da trat er mit dem Fuss gegen,
bis eine Latte morsch aufseufzte und losbrach. Das Holz roch miirbe und
siiss. Durch die dunkle Vorstadt tappte er zuriick. Sterne waren nicht da.
(Borchert 1949)

(He groped his way through the dark suburb. The houses stood in a broken
line against the sky. The moon was absent and the pavement was frightened
by the late step. Then he found an old plank. He kicked against it with his
foot until a lath gave a rotten sigh and broke loose. The wood smelled rotten
and sweet. Through the dark suburb he groped his way back. There were no
stars.)

In St. Jago, der Haupstadt des K6nigreichs Chili, stand gerade in dem Augen-
blicke der grossen Erdeschiitterung vom Jahre 1647, bei welcher viele tausend
Menschen ihren Untergang fanden, ein junger auf ein Verbrechen angeklagter
Spanier, Namens Jeronimo Rugera, an einem Pfeiler des Gefangnisses, in
welches man ihn eingespert hatte, und wolte sich erhenken. (Kleist 1923)

(In Santiago, the capital of the kingdom of Chile, at the very moment of the
great earthquake of 1647 in which many thousands of lives were lost, a young
Spaniard by the name of Jeronimo Rugera, who had been locked up on a
criminal charge, was standing against a prison pillar, about to hang himself.)

In the first example, the reader "sees" in one glance an area which
is about the size of a suburb, and which could be similarly surveyed
in reality. The scene includes a perceptible background (the houses)
and a foreground (the man kicking at the plank). The topographical
place (the suburb) is entirely overlapped by the zone of action
(defined by his walking) and by the field of vision. The second
excerpt presents a different kind of field of vision. Here, too, there is
a perceptible and rather concrete place (the cell), but its background,
in contrast to that of the Borchert excerpt, is not a series of objects
seen together with the character but an immense space which cannot
be shown as vividly as the prison. This is due not only to the interior-
exterior relation, but also to the fact that there are two different
kinds of perception here: a concrete and visual perception, and a
conceptual "sight" from a historical-geographic point of view. But
the condensed structure of the sentence forces these two domains -
although perceived in entirely different ways - to be surveyed in
one field of vision.
These two examples cannot, of course, exhaust the large range of
possible structures of fields of vision, nor do they represent all the
aspects of such structures. An extensive analysis is beyond the scope

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
326 GABRIEL ZORAN

of this paper, which proposes to analyze the concept "field of

vision" rather than its concrete phenomena.

4.1.3. Field of Vision and Description. One should bear in mind the
difference between a field of vision and a scenic description. Both are
verbal units defined by their reference to the fictional world. But a
scenic description is a particular instance of a field of vision, and
only one of its possible components. A field of vision may consist
of a scenic description, an action, a dialogue, a summary, an essay,
etc. The concept of field of vision solves, in my opinion, the
ambiguity caused by the classical dichotomy between description
and narration, and its automatic parallelism with the pair space and
action. This set of concepts is to a great extent responsible for the
false identification of space in the narrative with the descriptive
sections, and for excluding action as well as most of the other
components of the text from the phenomena relevant to space. A
field of vision is not a phenomenon confined to those specific
sections in the text that contain direct information about space; each
section in the text constitutes a field of vision from the point of view
of its spatial reference, although this spatial reference can be of
several kinds and degrees. Thus, fields of vision may differ in the
amount of information about space they contain and in the im-
portance of this information, but they do not differ in their basic
relevance to space.

4.1.4. The Problem of Identification and Delimitation of the Field

of Vision. If every unit of the text constitutes a field of vision, the
question may then be posed: What is it that causes the reader to
identify a particular spatial unit as a field of vision, to differentiate it
from another unit and at the same time demarcate a specific unit
within the reconstructed world? Here we must deal with the main
problem of the concept field of vision - the fact that it is a unit of
the reconstructed world, determined not by properties of the
"world" as such but rather by the perception of the world through
language.
At this point it is necessary to examine the function of the reader
in the course of his reading. It was stated earlier that a field of vision
is what the reader can perceive as being "here." Strictly speaking,
this means that at any one moment of reading, there is only one tiny
unit, or aspect, "in front of" the reader, and that during the reading
process he passes from one tiny unit to the next, so that the overall
image of space must be like a chain of tiny objects. This is what
Lessing (1974, Laokoon, Chap. 16) implied when he negated the
possibility of reconstructing the appearance of an object in one's
mind from a detailed description of its parts. He argues that by the
time the list of parts is finished, the first items will already have been

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 327

forgotten. This, to a certain extent, is an atomistic conception of the

process of reading, reducing it to a row of points, each of which is
connected only to the point just before or after it. The function of
the memory is reduced to that of merely connecting adjacent units.
In a conception such as Lessing's, there is no place for a field of
vision: when he speaks of space, he is referring only to solitary
objects - Agammemnon's sceptre, Hera's chariot - never to an
entire system of spatial relationships. When the act of reconstruction
is identified with the verbal decoding, space cannot be perceptible
other than as a series of small particles.
Space can be truly perceptible only in the framework of a con-
ception which assumes that the reconstruction of the world is not
parallel to the verbal interpretation alone, but also has to do with
accumulation in the memory and with various acts of linking: A
conception of this type was proposed by Segre (1975), for example.
He describes the point at which the reader finds himself at a certain
moment of reading as being continually in a system of relationships
with synthesis in memory, which preserves all which has already been
read, including the possibilities that have been eliminated, while at
the same time aspiring towards the open possibilities in the con-
tinuation of the text. Only in the framework of such a model can the
concept field of vision be explained. The field of vision is the com-
bination of the present moment of reading with the synthesis of the
memory. Here, however, backward synthesis refers not to the whole
complex of passages read, but only to those passages which relate
in some way to the spatial framework of the item about which one is
presently reading. This is a combination of the element perceived
at this moment of reading, together with other items and informa-
tion, in such a way that they may be perceived as spatially con-
tinuous and as forming one spatial wholeness, given to a single
survey. This synthesis of memory may encompass a scene extending
over many pages, or may be limited to a short description. In any
case, the backward synthesis needed for the field of vision acts on
a consecutive textual unit: the overall synthesis can of course
encompass information transmitted in earlier textual units, relating
to other parts of space, or even to the same part within a different
context, but this information will not belong to the same field of
vision. Its function may be that of an invisible background which is
also an important element in the general complex of space, but no
longer belongs to what is perceived as "here." The field of vision is
thus to a certain extent the point of intersection between the "here"
of space and the "now" of the text. It is a unit of reconstructed
space which has a correlative in the verbal text: it may be located
and identified both within the text and within the world.

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
328 GABRIELZORAN

4.2. The Complex of Space

We have observed in the foregoing that the textual existence of space
is like a series of fields of vision. We have defined and demarcated
a single field of vision, but it remains for us to understand how
different fields of vision combine to create the complex of space
as a whole. This process takes place in two dimensions: the dimen-
sion of the text continuum - how fields of vision change as the
reader progresses through the text; and the "world" dimension -
the arrangement of fields of vision within the reconstructed world
itself.
Fields of vision may shift from one to another in various ways.
Most obviously, there may be a break, such as a chapter or section
ending. However, this is not necessarily the most characteristic
method. Unlike scenes in a naturalistic theater, textual fields of
vision do not always occur in complete, closed units. They may be
much more fluid, they may widen or narrow in scope, as with a
movie camera, or move gradually from one place to another, making
their demarcation less clear-cut than our previous discussion might
indicate (see section 4.1.4). They may also shift by way of projection:
one field of vision may be constructed from pieces of information
supplied by another. A character in field A may relate something
that occurred in another place, thereby creating field B.
There are also various ways for fields of vision to combine within
the reconstructed world. The first possibility that comes to mind is
the perspectival one, in which one field is perceived as being "in
front" and another serves as background. This is the situation of
almost every field of vision in relation to the other fields in the text:
what is perceived at any given moment is foreground; the other
fields of vision form the "unseen" background. This foreground-
background relation is especially apparent in cases of projection.
Although projection usually results in a complete substitution of
one field of vision for another, the perspectival aspect can be
preserved, creating the effect of two simultaneous fields of vision.
This can happen when the projected area is not structured as an
autonomous field of vision - for example, in the case of discon-
tinuous mentioning of certain places by characters in a given field
of vision (i.e., not a continuous narration by one of the characters,
which would mean a complete displacement to the new field of
vision). Here the field of vision, rather than being autonomous,
becomes a kind of background existing outside the area of the
primary field.11 But two concrete and fully autonomous fields of

11. A similar effect, although much more rare, is sometimes gained when the projected
field of vision is dominant and autonomous but the text mentions insistently the primary
field of vision. For instance, take the description of Achilles's shield in the Iliad: there are,
of course, autonomous fields of vision encompassing the reality depicted on the shield,

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 329

vision cannot occur simultaneously because by definition a field

of vision encompasses all the items of space surveyed at a given
moment.
So we return to the question, what does join the fields of vision
to form a map of reality. I propose that it is the materials of reality
themselves, in a further degree of reconstruction; that is, at the
chronotopic and topographic levels. At these levels of reconstruction,
the materials are abstracted from their fields of vision and re-
organized in the topographic and chronotopic patterns of the text
(the horizontal and vertical structures, the system of axes, etc.;
see sections 3.1, 3.2).

4.3. Total Space

Having constructed a scheme of the complex of space, based on a
series of fields of vision, we still may discover that some spatial
information exists which is not structured within any field of vision;
that is, spatial elements that the text presupposes, or provides
indirectly, but does not "show." This information belongs to the
total space - that spatial information which exists beyond the
boundaries of the actually presented space.
The concept of total space in a text is necessary because of the
way we generally think about space. It is impossible to imagine space
as anything other than total. Of course, we do consider limited
sections of space, but at the same time we regard them as parts of
a larger space encompassing them.
This tendency is not necessarily connected with the Euclidean
conception of empty space;12 it may also be connected with the
semantic properties of words. Thus, for example, the meaning of
the word room includes the possibilities for additional rooms, a
house, a place of settlement, and so forth. Any spatial object may
also be perceived as a synecdoche for a more comprehensive space.
Total space, however, is not merely a vague duplication of space
actually shaped in a literary text: it is an essential component with
its own functions and modes of existence, as can be shown with
regard to the three levels of structuring.

4.3.1. Total Space from the Topographic Point of View. From the
point of view of the topographic level, the concept "total space"
is needed because it enables us to locate the events, to answer the

but the text continues insistently - by means of naming materials and verbs of producing -
to refer back to the primary field of vision, Hephaestus' workshop, which is in the fore-
ground.
12. The irrelevance of the Euclidean conception of space to the literary work of art was
claimed by Ingarden (1965). Ingarden, however, concluded from it the finiteness of the
literary space, which seems to me a wrong conclusion.

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
330 GABRIEL ZORAN

question, "where does it happen?" All the internal relationships

taking place on the topographical level answer only the question,
"what does it look like?" or "how is it made?" - but in order to
reconstruct space, the answer to this question alone is not sufficient.
It is essential to locate the entire system within some larger space.
Texts may be differentiated from one another by the degree of
importance which they attach to the question of localization, as well
as by the degree of exactness with which they answer this question.
Place of events may be located precisely by street and house number,
or in a very general way - by the name of the city or country, or
even less than this. Each of these cases implies a different way of
coordination within the total space.
Apart from the question of specific location, total space also has
to do with the assumptions of the text about the nature of the world
in general, and thus it is strongly connected with the external field
of reference (see Hrushovski 1976). The text refers the reader to a
certain model of external reality by means of which he must re-
construct the world. As regards the spatial complex of the text,
there is constant play between this model and the internal field of
reference, whereas in total space the external field of reference
becomes the prominent factor. The external field of reference may
be of several types: historical, geographical, mythical, science-
fictional, fantastic, etc. Naturally, the clarity of the localization and
reconstruction is dependent on the type of field of reference referred
to in the text. In any event, it should be emphasized that the con-
nection between total space and the external field of reference in
no way signifies that we are dealing here with something which
depends for its validity on something outside the text. The text itself
determines the nature of its total world, and the model of external
reality, although not necessarily created by the text, is chosen,
modified, and fully controlled by it.

4.3.2. Total Space from the Chronotopic Point of View. Here one
must examine the possible connections in the domain of plot
between total space and the complex of space. One can imagine an
absolute separation between them in the case where total space has
no significant relation to the plot, serving only as a static background
against which the plot unfolds. In this case, the narrative constructs
an entirely autonomous chronotopos. Examples of this possibility
can be found in the stories of Edgar Allan Poe. Even if one could
locate his stories in some geographical-historical space, it would be
insignificant; one can conceive of his total space as empty space. At
the other extreme, there is the possibility of active connection with
total space, such as when general historical events determine the
course of action. In this case, total space may be schematized as a
field of powers, the center of which is outside the actual complex

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 331

of space, but does, of course, affect it. The connection is not neces-
sarily made by means of abstract power lines: one can also imagine
real axes of movement in the text that continue into total space
or emerge from it. The best example of chronotopic connection
between total space and the complex of space occurs in the epics; in
fact, the connection is so close that it almost annuls the differentia-
tion between them. The voyages of Odysseus encompass all the areas
of the world fixed in the conscience of its contemporary reader,
the early Greek, just as Paradise Lost covers the entire cosmological
system of the Renaissance. In these cases, the text totally exhausts
the world. These are spaces which can no longer be imagined as parts
of something larger. From the chronotopic aspect, the complex of
space here is almost identical to total space.

4.3.3. Total Space from the Textual Point of View. From the point
of view of the textual level, the nature of total space and its relation
to the space complex should be described by means of two distinc-
tions: one between presentation and representation, and the other
between determinacy and indeterminacy.'3 These two distinctions,
of course, are applicable not only to the relation between the
complex of space and total space, but also to other areas of the text
and the world. Within the complex of space and even within a single
field of vision, there are also spots of indeterminacy and represented
elements, but total space is the most complete and comprehensive
area which can be described by these two attributes.
It should be stressed, however, that these two distinctions are not
overlapping and that these attributes are not applicable to the same
aspects or parts of total space.
The quality of representation belongs to all that can be positively
said about total space. All these things differ from the information
about the complex of space because they are rendered indirectly:
they are not presented but represented. Actually, this is what we said
at the beginning of our discussion (4.3) when we claimed that total
space is not structured in the fields of vision. Fields of vision are
the modes of presenting space; the materials of total space are
not structured within them but scattered in various indirect ways:
mentioning of places by characters and even by the narrator, materials
of metaphors and similes, synedochic items which enable to rebuild
the world, and so forth.
But this information, apart from being indirect, is sparse in
relation to the information about the complex of space. One of the
prominent qualities of total space is the immense domain of missing

13. The distinction between presentation and representation was suggested by B.

Hrushovski. The distinction between determinacy and indeterminacy is, of course,
Ingarden's (1965).

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
332 GABRIEL ZORAN

information. All that can be said about it, both on the basis of
textual hints and on the basis of a general knowledge of the external
field of reference - all that does not abolish its indeterminacy. Total
space is an immense area of indeterminacy, and relative darkness,
in which the complex of space appears as an island of determinacy
and clarity.
Total space is also an essential assumption for determining the
perspectival structure of the world. To a certain extent it constitutes
the absolute there, because it is always conceived as being beyond
the horizon of the field of vision. But similarly, it is necessary when
locating the primary here - the act of narration. If the act of
narration is not actually concretized as a part of the narrated world,
its spatial existence and the coordinates connecting it with the world
are also one of the aspects of total space.

4.3.4. The Ontological Opacity of Total Space. Finally, an important

characteristic of total space, which concerns all the levels of struc-
turing, to a certain extent, is lack of ontological clarity. This is of
course related to the general property of indeterminacy which exists
in total space, but here it is not only the result of an absence of
information, but rather of certain contradictions or fundamental
lacks of clarity which are covered, so to speak, by the general in-
determinacy of this area: total space is a kind of no man's land
bridging different ontological areas. It is perceived not only as the
direct continuation of the reconstructed world in the text, but also
as a continuation of the real space of the reader, of the external
field of reference, the act of narration, and possibly more. All these
areas, which could be completely incompatible from the ontological
point of view, are somehow swallowed up in the total space, and are
perceived as existing on one continuous plane of being. This situation
is best typified by the frequent location of fairy tales in distant
lands: The ontological passage from the reader's world which is
dominated by realistic probability to a world of imagination is thus
expressed by a physical remoteness within an indeterminate space.14

14. One should point out three conceptions relevant in some way to the concept of total
space proposed here. Timpe (1971) distinguishes between actual space and the potential
one, but the description of potential space as the dimensions the actual space can achieve
is rather vague. Hrushovski, in his study on Warand Peace (1976), raises the problem of
total space while discussing the possibility of locating the fictional salon of Anna within
the historical Petersburg. But, in this discussion, the concept of external field of reference
is used for what I call total space. I believe that these are different concepts, although they
have a close connection (see section 4.3.1). And, finally, Rokem's discussion about the
off-stage world in the theater (1979) may throw some light on the problem of total space
from a very interesting angle.

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 333

5. CONCLUSION
In this paper I attempted to set forth a model which would indicate
the central aspects of space in the narrative and determine their
mutual relationships. The aspects discussed have to do mainly with
the inherent structure of space and its mode of existing as a part of
the reconstructed world. This may clarify as well the issues I did not
deal with in the framework of this study - the functionality of space
within the overall structure of the text. The discussion was limited
to the mode of existence of space and did not deal with its functions.
It goes without saying that space is not a neutral material just
existing in the world; it has various functions relating to other planes
of the text. Every element in space - actually every element in the
text - has to be regarded, to use the term of Hrushovski (1976), as
a juncture, in which patterns from all the textual planes may inter-
sect: patterns of space together with patterns of characterization,
ideas, mythology, and so forth.
The functions of space may appear clearly when dealing with
single texts and pointing out the system of relations within the
complex of its components. One could, perhaps, even describe and
analyze its possible functions within the framework of a theoretical
discussion; this, however, is a domain of questions entirely different
from those dealt with here. The fact that space here is neutralized -
for methodical purposes - from its specific functions does not mean
that space in itself is conceived here as a neutral factor. On the
contrary, the assumption that all the textual components have
functional relations with each other is rigorously maintained; what is
lacking here is an assumption about a hierarchy: I did not stipulate
in advance which element is the means and which is the end. Despite
any functional approach, one tends to regard space as subordinate
to characters rather than characters to space, and the same about
the relation of space to other aspects of the text: it is always
regarded as a means to certain ends. It was exactly this situation that
prompted many discussions devoted to space to deal with characters,
ideas, or general interpretation, neglecting their specific issue.
I, however, tried to begin with space and end with it, without
turning aside even at junctions with roads that seem, from an
hierarchical point of view, more like highways. No discussion about
the functions of space could be worthwhile without analyzing first
its mode of existence and its several aspects. Furthermore, apart
from the way it may be exploited in specific texts and the meanings
which may be conveyed through it, it should not be forgotten that
space is first and foremost a central aspect of the world, whether real
or fictional and in whatever medium it may be transmitted. The
purpose of the model suggested here was to throw light on some
problems raised by the status of space within a fictional world
conveyed through a verbal medium.

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
334 GABRIEL ZORAN

RE FERENCES
Alewyn, Richard, 1957. "Eine Landschaft Eichendorffs," Euphorion 51, 42-60.
Bachelard, Gaston, 1974. La poetique de l'espace (Paris: Presses universitaires de France).
Baxtin, M. M., 1978. "The forms of Time and the Chronotopos in the Novel: From the
Greek Novel to Modern Fiction," PTL 3:3, 493-528.
Binswanger, Ludwig, 1955. "Das Raumproblem in der Psychopathologie," in: Ausgewdhlte
Vortrdgeund Aufsdtze von Ludwig Binswanger, Vol. II (Bern: Franke Verlag), 174-
225.
Bollnow, Otto Friedrich, 1963. Mensch und Raum (Stuttgart: Kohlhammer Verlag).
Borchert, Wolfgang, 1949. Das Gesamtwerk (Hamburg: Rowohlt).
Brown, Russell E., 1967. "On Classifying the Setting of the Novel: Hans Henny Jahnn's
Fluss ohone Ufer," Neophilologus LI, 395-401.
Even-Zohar, I., 1968. "Correlative Positive and Correlative Negative Time in Strindberg's
'The Father' and 'A Dream Play,' " Ha-Sifrut 1:3-4, 538-568. (In Hebrew; English
summary pp. 770-771.)
Frank, Joseph, 1963. "Spatial Form in Modern Literature," in: The Widening Gyre (New
Jersey: Rutgers UP), 1-63.
1978 "Spatial Form: Some Further Reflections," Critical Inquiry 5:231-252.
Hellwig, Brigitte, 1964. Raum une Zeit in homerischen Epos (Spudasmata: Studien zur
Klassischen Philologie und ihre Grenzgebieten, Vol. II. Georg Olms Verlagsbuch-
handlung, Hildsheim).
Hrushovski, Benjamin, 1974. "A Unified Theory of the Literary Text," in: Z. Ben-Porat
and B. Hrushovski, Structuralist Poetics in Israel (Tel Aviv Univ.: Institute for Poetics
and Semiotics), 13-23.
1976 Segmentation and Motivation in the Text Continuum of Literary Prose: The First
Episode of "Warand Peace" (Tel Aviv Univ.: Papers on Poetics and Semiotics).
Ingarden, Roman, 1965. Das Literarische Kunstwerk (Tiibingen: Max Niemeyer Verlag).
Iskra, Wolfgang, 1967. Die Darstellung des Sichtbaren in der dichterischen Prosa um 1900
(Miinster: Aschendorff).
Kleist, Heinrich von, 1923. Gesammelte Werke, Vol. III (Berlin: Tillgner Verlag).
1962 (1923) "The Earthquake in Chile," in: The Marquise of 0 and Other Stories,"
trans. Martin Greenberg (New York: Signet).
Kristeva, Julia, 1970. Le texte du roman: approche semiologique d'une structure discursive
transformationnelle (The Hague/Paris: Mouton).
Liamert, Eberhard, 1955. Bauformen des Erzdhlens (Stuttgart: J. B. Metzlersche Verlags-
buchhandlung).
Lessing, Gotthold Ephraim, 1974. Werke. Sechster Band. Kunsttheoretische und Kunst-
historische Schriften (Miinchen: Carl Hanser Verlag).
Lewin, Kurt, 1936. Principles of Topological Psychology (New York/London: McGraw-
Hill).
1938 The Conceptual Representation and the Measurement of Psychological Forces
(Durham, N.C.: Duke UP).
1957 Field Theory in Social Science (New York: Harper and Brothers).
Linde, Charlotte, 1974. The Linguistic Encoding of Spatial Information (unpublished
dissertation, Columbia Univ.).
Lotman, Jurij M., 1973. Die Struktur des Kiinstlerischen textes (Frankfurt am Main: Suhr-
kamp Verlag).
Meyer, Herman, 1957. "Raumgestaltung und Raumsymbolik in der Studium
Erzahlkunst,"
Generale, Vol. 10, 620-630.
Miller, George A. and Philip N. Johnson-Laird, 1974. Language and Perception (Cambridge/
London/Melbourne: Cambridge UP).
Minkowski, E., 1933. Le temps vecu: Jtudes phenomenologiques et psychopathologiques
(Paris: Collection de l'evolution psychiatrique).
1936 Vers une cosmologie: fragments philosophiques (Paris: Fernard Aubier).

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions
 TOWARDS A THEORY OF SPACE IN NARRATIVE 335

Muller, Giinther, 1950. "Uber das Zeitgeriist des Erzghlens: am Beispiel des Jiirg Jenatsch,"
in: Deutsche Vierteljahrschriftfiir Literaturwissenschaft und Geistesgeschichte 24:1,
1-31.
O'Toole, Lawrence M., 1980. "Dimensions of Semiotic Space in Narrative," Poetics Today
1:4, 135-149.
Perry, Menakhem and Meir Sternberg, 1968. "The King Through Ironic Eyes: The Nar-
rator's Devices in the Biblical Story of David and Batsheba and Two Excursuses on
the Theory of Narrative Text," Ha-Sifrut 1:2, 263-292. (In Hebrew; English sum-
mary pp. 449-452.)
Petsch, Robert, 19422. Wesen und Formen der Erzdhlkunst (Halle: Max Niemeyer Verlag).
Rokem, Freddie, 1979. "The Off-Stage World" (manuscript).
Segal, Dmitry, 1974. Aspects of Structuralism in Soviet Philology (Tel Aviv Univ.: Papers
on Poetics and Semiotics 2).
Segre, Cesare, 1975. "Space and Time of the Text," 20th Century Studies, 37-41.
Sternberg, Meir, 1973. "Delicate Balance in the Story of the Rape of Dinah: Biblical
Narrative and the Rhetoric of the Narrative Text," Ha-Sifrut 4:2, 193-231. (In
Hebrew; English summary p. XIII.)
Timpe, Eugene F., 1971. "The Spatial Dimension: A Stylistic Typology," Yearbook of
Comparative Criticism III: Patterns of Literary Style, ed. Joseph Strelka. (University
Park, Penn./London: The Pennsylvania State UP), 179-197.
Tomashevsky, Boris, 1965. "Thematics," in: Russian Formalist Criticism, Lee T. Lemon and
Marion J. Reis, eds. (Lincoln, Nebraska: Univ of Nebraska Press).
Uspensky, Boris, 1973. A Poetics of Composition: The Structure of the Artistic Text and
Typology of a Compositional Form (Berkeley/Los Angeles/London: Univ. of
California Press).
Wilbur, Richard, 1967. "The House of Poe," in: Poe: A Collection of Critical Essays, R.
Regan, ed. (Englewood Cliffs, N.J.: Prentice-Hall).

This content downloaded from 208.95.48.254 on Wed, 13 Jan 2016 08:16:02 UTC
All use subject to JSTOR Terms and Conditions