<<

. 23
( 82 .)



>>

of the events, even if initially it seems natural to suppose that they are occurring
sequentially.
“John jumps; Mary sings; Harold bleeds” is not self-evidently a narrative not
only because a unified subject eludes us, but also because it yields no reliable
insight into the order of these events. Likewise, “The Tartar hordes swept over
Russia; Socrates ate hemlock; No«l Carroll got his first computer; Jackie Chan
made his most successful movie; and dinosaurs became extinct” is not a narrative
not only because it lacks a central subject, but because it does not even conversa-
tionally implicate a perspicuous, reliable time ordering of the events it recounts.
The narrative connection requires both a unified subject and a perspicuous tem-
poral order where by a perspicuous temporal order I mean a retrievable one.This,
of course, does not mean that the temporal order must be stated outright in the
discourse string. Context and the knowledge that the intended, informed audi-
ence brings to the discourse may be enough for the intended reader, viewer, or lis-
tener to derive a perspicuous temporal ordering from an example.
ON THE NARRATIVE CONNECTION 121

So far I have alleged that a narrative connection is comprised of a series of
events that is both possessed of a unified subject and perspicuously time-ordered.
Moreover, it is important to realize that the requirements of a perspicuous time-
ordering and that of a unified subject do not amount to the same thing.This can
be seen by comparing the following two strings of discourse: “The Eastern
Roman Empire falls in 1453; the American Constitution is accepted in 1789;
Russia is defeated in 1905;” and “Hume publishes ˜Of the Standard of Taste;™
Hume fails to secure a professorship.” Neither is a narrative because the first lacks
a unified subject, though it is temporally ordered, whereas the second, though
apparently possessing a unified subject (viz., Hume), lacks a perspicuous time-
order “ which came first the failure to secure a professorship, or the publication?
For classificatory purposes, following Morton White, we might call our first
example “ “The Eastern Roman Empire falls ¦ etc. “ an annal.1 An annal is not
a full-blooded narrative. On the other hand, it is not altogether without temporal
order. So perhaps we should call annals members of the class of story forms, where
story forms are any sort of temporally ordered discourse. But with respect to the
annal, its principle of organization is simply temporal ordering.
However, the annal is not the only kind of story form.The chronicle is another.
Where temporal ordering is combined with a unified subject, we arrive at a
somewhat more complex structure than the annal, which has been called the
“chronicle.”2 For example:“The French Revolution occurred in 1789; Napoleon
became Emperor in 1805; Napoleon was defeated at Waterloo in 1815, after
which the Bourbons were restored.”This possesses a perspicuous temporal struc-
ture, as well as a unified subject, namely, French history.
In this example of a chronicle, each event is explicitly dated. However, in order
to qualify as a chronicle, it is not necessary that the temporal order be made
absolutely explicit in the text, so long as the reader, viewer, or listener is able to
derive a reliable temporal order from it. Had our earlier example stated “First the
President talked to his adviser; then he ate a piece of cheese; then he jogged; and,
finally, he waved to the reporters,” then this would be a chronicle, rather than an
annal, for even though the temporal structure is somewhat imprecise, it neverthe-
less affords a rough order of occurrence.
Or, in some cases, it may be appropriate to presume that the temporal order is
implicit for a target audience. Thus, “Kennedy is assasinated; Johnson becomes
President” qualifies as a chronicle for informed audiences who are able to fill in
the temporal order. A discursive representation that (temporally, but noncausally)
connects at least two events in the career of a unified subject such that a reliable
temporal ordering is retrievable from it (and/or from the context of enunciation)
is a chronicle. Chronicles are certainly more complex structurally than annals. But
are chronicles full-fledged narratives?
I conjecture that they are not, although I realize that this is open to termino-
logical dispute. The reason I think that a mere chronicle is not yet a narrative is
that it does not display any connection other than that of temporal succession
between the events it recounts. If I say “I woke up; later I dressed; still later I went
122 ART, HISTORY, NARRATIVE
AND


to class,” I suspect that most people would agree that this falls short of a full-
fledged narrative, although the events cited might be turned into ingredients of a
narrative. But why isn™t it a narrative properly so called? To put it vaguely “
because the connection among the events alluded to by it is not tight enough.
Of course, as noted, this might be a controversial conclusion. Some theorists, like
Gerald Prince, might be willing to call this a narrative.3 However, I think that it is
more profitable theoretically to discriminate between different sorts of temporal dis-
courses or, as I call them, story forms “ between, for example, annals, chronicles, and
narratives proper. Having more distinctions at least has the advantage of affording
more fine-grained structural insights. And what we have been talking about are
important structural differences no matter what labels we finally settle on.
On my view, narratives proper require more than simply temporal ordering as a
principle for connecting events and/or states of affairs. But in order to motivate my
taxonomy, it is necessary for me to specify the nature of the connection I have in
mind rather than simply saying it is more than merely a matter of temporal ordering.
What is the connection that obtains between events and/or states of affairs in
a narrative proper? One popular candidate is causation. This seems eminently
plausible, since narratives typically represent changes in states of affairs, and change
implies some subtending causal process.
“Creon had Antigone executed; consequently, his son committed suicide,
which caused his wife to commit suicide, and, as a result, Creon felt anguish.”This
is a narrative. It has a unified subject, Creon, and it is perspicuously time-ordered.
But in addition to being time-ordered, the events are also causally linked. Earlier
events in the discursive string are the causes of the later events in the string in the
sense that the earlier events supply sufficient grounds, all things being equal, for
the occurrence of the later event. This is certainly a “tighter” connection than
mere temporal ordering. Is it the secret of the narrative connection?
It is not, if what we are after is the generic or comprehensive connection
between earlier events and/or states of affairs and what follows. Admittedly, in
some narratives, causation of this sort is what supplies the connection between
earlier and later events. But causation in this sense “ the sufficency sense “ is too
strong a relation to hypothesize as the relevant connection operative in all narra-
tives linkages. Were the relation causal on this understanding, that would suggest
that earlier events in narratives causally entail later events. And although this may
obtain in some cases, it does not obtain in all cases, nor does it seem to me to
obtain in even most of the typical cases.
Most narratives are not strings of causal entailments. In most narratives, the
earlier events in a sequence of events underdetermine later events.We read that a
thief enters a bank and robs it; in the next scene, as he exits the bank, he is appre-
hended by the police whom we subsequently learn have been watching him all
along.The first event did not entail the second event. Indeed, we may even be sur-
prised by the appearance of the police when the thief leaves the bank. Neverthe-
less, “the thief enters the bank to rob it, but subsequently, as he exits, he is
apprehended by the police” is a narrative, though, strictly speaking, the first event
ON THE NARRATIVE CONNECTION 123

is not the fully determining cause of the second event.That is, although robbing
the bank is causally relevant to the arrest, it does not causally entail the second
event. That is why we cannot think of all narratives on the model of a series of
causal chains in which earlier events and/or states of affairs function as the effi-
cient causes of succeeding events and/or states of affairs.
Looking at the preceding counterexample concerning the thief and the bank
may suggest a way of modifying the causal account.The first event by itself does
not causally entail the second event, but the second event brings with it the causal
information “ the presence of the police observers “ that explains the second
event “ the arrest “ when it is combined with the first event “ the robbery.We can
call this addition, following William Dray, a causal input.4 On this model, the nar-
rative connection is a matter of the first event and/or state of affairs + a causal
input + the second event. Does the causal input model yield a general picture of
the narrative connection?
I think not, and again the problem is that it is too strong. Consider this narrative:
Because the town refused to surrender, the invading army laid siege to it;
because a message reached them that their base of operations was under
attack, the besiegers soon withdrew, but, in withdrawing, their campaign
fell into confusion.5
This seems to be an acceptable narrative. However, not every successive event in it
can be subsumed under either the causal model or the causal input model. The
first connection in the story “ the refusal to surrender/the siege “ is arguably
causal; the second connection “ the message/the withdrawal “ possibly involves a
causal input; but the last connection “ the withdrawal/the confusion “ is causally
underdetermined.Withdrawing from the siege might have integrally contributed
to the campaign falling into confusion. But it does not necessitate it.Withdrawing
troops does not causally entail falling into confusion. Nevertheless, this last con-
nection seems like a legitimate narrative connection. Indeed, this sort of connec-
tion appears quite often in narratives. But it does not fit either the deterministic
causal model or the causal input model. Thus, we need a characterization of the
narrative connection that is not as strong as those provided by the causal model or
the causal input model.
Since narratives represent change, it is natural to think that the narrative con-
nection has something to do with causation. But it is too demanding to expect
that, in all cases, the narrative connection involves an earlier event that causally
necessitates the succeeding state. But though the earlier event need not be the
cause, in this sense, of succeeding states, it is not causally irrelevant either. What
might the narrative connection be if it need not be the sufficient cause of succes-
sive states, but is also somehow causally relevant to successive states? One obvious
relation fitting these desiderata is that the earlier event and/or state of affairs in a
narrative connection is a causally necessary condition for successive states (or a
contribution to such a causally necessary condition).6 That is, more precisely, the
earlier event in a narrative connection is at least a necessary or indispensable con-
124 ART, HISTORY, NARRATIVE
AND


tribution to a sufficient, though nonnecessary, condition for the occurrence of the
relevant later event in the narrative complex (or, in other words, it is, at least, what
J. L. Mackie calls an INUS condition).7
Referring to our robbery example, had not the earlier events cited
obtained, the later events would not have obtained, though the earlier events in
question do not guarantee the occurrence of the later events. In other words,
had the robber not been pilfering the bank, ceteris paribus, the police would not
have arrested him, but his robbing the bank, at that point in the story, did not
guarantee his being arrested. (Here we are obviously talking about token cau-
sation, not type causation.)
Likewise, in the preceding story about the withdrawal of the army, the retreat
from the town can be understood as a causally necessary condition for the army in
question falling into confusion when it did. Other factors may also have conspired
to cause the disarray of the army. But the preceding story tells us that a causally
necessary condition for this particular instance of disarray was that the army was
in the process of withdrawing.That is, if that army at that time had not been with-
drawing, it would not have fallen into disarray.
Similarly, that the thief robbed the bank is a causally necessary condition for
his arrest. Had he not robbed the bank, all things being equal, he would not have
been arrested by the police observers when he was.The robbery was a necessary
condition for his arrest.
On the strict, deterministic causal model of the narrative connection, the ear-
lier events in the narrative necessitate the succeeding events and/or states of
affairs. On the causal input model, the earlier event plus some causal input neces-
sitate the succeeding event. But, as I have noted, the narrative relation is often
weaker than that of necessitation or causal entailment. Rather, in a great many
cases, earlier events merely function to make later events causally possible. Insofar
as the earlier event is generally of the nature of a causally necessary condition (or
a contribution to such a condition), the later event is often not predictable.
Many events share the same causally necessary conditions.The hero is rushing to
save a child from an oncoming truck.This opens up the possibilities that either the
child will be saved from the truck or not.That the oncoming truck is rushing at the
child is a necessary condition for both the event of the child™s destruction or its being
saved by the hero. The earlier event does not necessitate the outcome, but it is, all
things being equal, a necessary condition for either of the alternative outcomes.
The reason that we hesitated to identify strict deterministic causation as the
relevant connective for all cases of narrative connection was that we noted that
very often the earlier events and/or states of affairs in stories underdetermine the
events that follow them. Hypothesizing that the earlier events and/or states of
affairs are causally necessary conditions, however, cannot be defeated by this
objection, since x™s being a causally necessary condition for what follows is consis-
tent with the appearance of a wide range of events and/or states of affairs. More-
over, hypothesizing that the earlier event is at least a causally necessary condition
(in the sense of an INUS condition)8 coincides with the intuition that many have
ON THE NARRATIVE CONNECTION 125

that narrative has something to do with causation (if only because it is concerned
with processes of change).
Of course, in saying that the earlier event in a narrative connection is at least a
causally necessary condition in the sense stipulated, one leaves open the possibility
that sometimes the relation may be stronger causally. Sometimes the earlier event
may be the cause of a subsequent event and/or state of affairs in the sense of
necessitation, or sometimes a series of earlier events and/or states of affairs may be
jointly sufficient for the production of a later event. I do not preclude these possi-
bilities in some cases. Rather, I deny that they supply an account across the board
of the narrative connection. A better candidate for that account is that the earlier
event and/or state of affairs in a narrative connection is at least a causally necessary
condition or ingredient for bringing about later events (or a contribution to such
a condition). This way of approaching the matter holds only that earlier events
must figure in the causal network that gives rise to later events, but allows that it
can figure in that network not only as a sufficient condition but as weakly as sim-
ply a causally necessary condition (or as a contribution thereto) that is indispens-
able to the sufficient but nonnecessary cause of the relevant later events.
Once it is proposed that the earlier events in a narrative are at least causally
necessary conditions (in the sense of INUS conditions) of later events (or contri-
butions thereto), it is natural to wonder whether this hypothesis isn™t itself too
strong. Perhaps it is too stringent to require that the earlier events be causally nec-
essary conditions. Might they not be just necessary conditions?
Consider this putative narrative: “Aristarchus hypothesized the heliocentric
theory thereby anticipating Copernicus™ discovery by many centuries.” This
appears to have a unified subject “ the heliocentric theory “ and it is time-ordered
“ first Aristarchus hypothesized the heliocentric theory and then Copernicus had
the same insight centuries later. Moreover, if this is what we might call a narrative
of anticipation (or prefiguration), then the first event, Aristarchus™ discovery, is a
necessary condition for the anticipation (or prefiguration) of the second event,
Copernicus™ discovery. Thus, if this is a narrative, then it appears that a necessary
condition, rather than a causally necessary condition is all that we need to require
of a narrative connection.
Yet if there is no line of influence stretching from Aristarchus™ discovery to
Copernicus™, I, at least, find it strained to think that this is narrative. It is an interesting
series of events. Indeed, mention of the second event in this series retrospectively
reveals something of the significance of the earlier event, and, as we shall see, retro-
spective significance is a frequently recurring feature of narrative. However, where
the events bear no sort of causal relation to each other, they seem more of the order
of coincidence than of narrative, at least if you agree that narrative involves changes in
the career of a unified subject, where change is a function of causal processes.9
Assuming that the narrative connection involves the earlier event in the sequence
playing the role of at least a causally necessary condition, the narrative connection is
structurally more complex than the chronicle. However, the narrative connection
bears an obvious relation to the chronicle, since a given narrative connection will
126 ART, HISTORY, NARRATIVE
AND


imply a chronicle of events, that is, a time-ordered sequence of events. From any nar-
rative connection proper, one can retrieve a time-ordering of the relevant events. A
narrative connection, in other words, entails some chronicle, although a mere chron-
icle does not entail a narrative. Similarly, both a narrative proper and a related chron-
icle will imply some annal, though no annal strictly entails a chronicle or a narrative
connection. In this way, our story forms are ordered hierarchically.
Of course, the order of presentation, as it is locally articulated, of a narrative con-
nection “ that is, the story as told (or what the Russian Formalists called the syuzhet)
“ may deviate from the chronological-causal structure of the narrative connection
itself (from what the Russian Formalists called the fabula), since the exposition may
involve all sorts of devices such as the flashback, the flashfoward, the hysteron pro-
teron, and so on.10 Nevertheless, in the narrative proper, the narrative connection (or
fabula) should be retrievable from the exposition. If it is not, there is no narrative.
In the preceding paragraph, I introduced the idea of a local articulation, like a
flashback. However, narratives as a matter of discursive practice are globally artic-
ulated in what might be called a forward-looking manner. Once we are in a flash-
back, for example, the events are told in a temporally progressive way.They are not
told backwards, so to speak.This differentiates narratives from mere explanations

<<

. 23
( 82 .)



>>