To appear in special issue of The Journal of Narrative and Life History, 1997p>

Some Further Steps in Narrative Analysis

William Labov. University of Pennsylvania

The first steps in narrative analysis taken by Joshua Waletzky and myself were a by-product of the sociolinguistic field methods that had been developed in the survey of the Lower East Side (Labov 1966) and in the work that engaged us at the time -- the study of African American Vernacular English in South Harlem (Labov, Cohen, Robins and Lewis 1968). We defined the vernacular as the form of language first acquired, perfectly learned, and used only among speakers of the same vernacular. The effort to observe how speakers talked when they were not being observed created the Observer's Paradox. Among the partial solutions to that paradox within the face-to-face interview, the elicitation of narratives of personal experience proved to be the most effective. We were therefore driven to understand as much as we could about the structure of these narratives and how they were introduced into the every-day conversation that our interviews simulated. Labov and Waletzky 1967 laid out a framework which has proved useful for narrative in general, as this volume demonstrates.

Since that time I have published only a few studies of narrative (Labov 1972, Labov and Fanshel 1977, Labov 1981). This is not because I have lost interest in the subject, for I have written and delivered a great many unpublished papers in the area.[1] I may not have pursued these papers into the domain of publication because the analysis of narrative was competing with quantitative studies of language variation and change, where cumulative theories can be built upon decisive answers to successively more general questions. The discussion of narrative and other speech events at the discourse level rarely allows us to prove anything. It is essentially a hermeneutic study, where continual engagement with the discourse as it was delivered gains entrance to the perspective of the speaker and the audience, tracing the transfer of information and experience in a way that deepens our own understandings of what language and social life are all about. The most important data that I have gathered on narrative is not drawn from the observation of speech production or controlled experiments, but from the reactions of audiences to the narratives as I have retold them. In a regular and predictable fashion, certain narratives produce in the audience a profound concentration of attention that creates uninterrupted silence and immobility, an effect that continues long after the ending is reached. It is the effort to understand the compelling power of such narratives that brings me to the current essay, an abstract of a more extended treatment of narratives of personal experience to follow.[2]

Labov and Waletzky demonstrated that the effort to understand narrative is amenable to a formal framework, particularly in the basic definition of narrative as the choice of a specific linguistic technique to report past events. The L&W framework developed for oral narratives of personal experience proved to be useful in approaching a wide variety of narrative situations and types, including oral memoirs, traditional folk tales, avant garde novels, therapeutic interviews and most importantly, the banal narratives of every-day life. It allowed us to understand pseudo-narratives like recipes, apartment-house layouts, and other types of experience remodeled into narrative form. It gradually appeared that narratives are privileged forms of discourse which play a central role in almost every conversation.[3] Our efforts to define other speech events with comparable precision have shown us that narrative is the prototype, perhaps the only example of a well formed speech event with a beginning, a middle, and an end.

Narrative and the broader field of story telling has become a keen focus of attention in many academic and literary disciplines. Here the traditional trajectory of the story-teller runs athwart the main focus of this article. The classic image of the story-teller is someone who can make something out of nothing, who can engage our attention with a fascinating elaboration of detail that is entertaining, amusing and emotionally rewarding. From the first lines of such a narrative, we know that we are in the presence of a gifted user of the language. Credibility is rarely an issue here. Tall tales, myths and outright lies carry the day, and we normally do not know or care whether the events as told were the personal experience of the story-teller or anyone else. The narratives that are the central focus of my current work are altogether different. The tellers were not known as gifted story tellers; people did not gather to hear them speak. They were ordinary people in the deepest sense of the word. They did not manufacture events or elaborate the experience of others. Their narratives were an attempt to convey simply and seriously the most important experiences of their own lives. Sometimes the stories had been told many times, but very often they had not been, or were perhaps told for the very first time. They deal with the major events of life and death, including the sudden outbreak of violence; the near approach of death and the witness of it; premonitions of the future, often through communication with the dead; courage in the face of adversity and the struggle against overwhelming odds; cowardice and the betrayal of trust. I don't believe that this focus on serious and momentous issues will limit the scope of the analysis. Rather, the use of narrative to deal with issues of life and death will highlight the abilities displayed in more casual, humorous or even trivial accounts. In the less serious and more frequent deployment of narrative, techniques are practised to perfection; in the more serious domain, they are put to the test.

The narratives that form the focus of this work were normally told in the course of a sociolinguistic interview, where the interviewer formed an ideal audience: attentive, interested and responsive. Though they are fitted to some extent to the situation and often to a question posed by the interviewer, they are essentially monologues and show a degree of decontextualization. They exhibit a generality that is not to be expected from narratives that subserve an argumentative point in a highly interactive and competitive conversation. Such narratives are often highly fragmented and may require a different approach. Yet studies of spontaneous conversation also show a high frequency of monologic narratives that command the attention of the audience as fully as the narratives of the interview. The principles developed in this paper are exemplified most clearly in narratives of this type.

The issues that are raised here go beyond the analysis of L&W, which dealt with temporal organization and evaluation. The framework I will present begins with these aspects of narrative, and then goes on to consider the further issues of reportability, credibility, objectivity, causality, and the assignment of praise and blame.

In this skeletal presentation, I will use one narrative to illustrate the principles involved:[4]

(1) Harold Shambaugh, Tape A-304, Columbus, Ohio, 7/28/70

(What happened in South America?)
a Oh I w's settin' at a table drinkin'
b And - this Norwegian sailor come over
c an' kep' givin' me a bunch o' junk
about I was sittin' with his woman.
d An' everybody sittin' at the table with me were my shipmates.
e So I jus' turn aroun'
f an' shoved `im,
g an' told `im, I said,
"Go away,
h I don't even wanna fool with ya."
i An' nex' thing I know I 'm layin' on the floor, blood all over me,
j An' a guy told me, says,
"Don't move your head.
k Your throat's cut."

This brief narrative has been proven to be paradigmatic in the ability to transfer experience from narrator to the audience. The reader is invited to commit these twelve lines to memory, and re-tell the story to an individual or group of others. Many listeners report the experience of viewing in the scene in a smoke-filled room in lines (a-h); that with lines (i-k) there is a sudden change of perspective, looking up from below; and after (k), about one third of the people in any audience make a sudden intake of breath, as if it were in fact their throat that was in question.

Over the course of some twenty years, I have dealt with the question of how this brief narrative commands attention and conveys experience as effectively as it does. The following pages are an outline of my attempt to provide an answer. The presentation is in the form of definitions; implications from those definitions; empirical findings from the study of a larger body of narrative; and theorems, which propose relations with emp;irical content that are more problematic. The reader is asked to accept the validity of those findings provisionally until a larger body of material can be presented.

0. Narratives of personal experience.

(0.1) Definition: A narrative of personal experience is a report of a sequence of events that have entered into the biography of the speaker by a sequence of clauses that correspond to the order of the original events.

This definition is based on the initial conception of L&W; a definition that separates narrative in this sense from other means of telling a story or recounting the past. It is an arbitrary segregation of one sense of narrative for technical purposes, but it proves to be a useful one. By here specifying that the experience must have entered into the biography of the speaker, I distinguish narrative from simple recounting of observations like the events of a parade by a witness leaning out a window. It will turn out that events that have entered into the speaker's biography are emotionally and socially evaluated, and so transformed from raw experience.

1. The temporal organization of narrative.

This discussion of temporal organization includes a new piece of terminology not present in L&W, a "sequential clause."

(1.1) Definition: Two clauses are separated by a temporal juncture if a reversal of their order results in a change in the listener's interpretation of the order of the events described.

Thus all the clauses in (1) are separated by temporal juncture with the following exceptions: (a) and (b) overlap so there is no juncture between them, and (i) overlaps (j) and (k), so there is no juncture between (i) and (j).[5]

(1.1.1) Implication [and definition of a minimal narrative]: A narrative must contain at least one temporal juncture.

As L&W point out, stories can be told without any temporal juncture by syntactic embedding, the use of the past perfect and other grammatical devices. Temporal juncture is the simplest, most favored or unmarked way of recounting the past.

(1.2) Definition: A sequential clause is a clause that can be an element of a temporal juncture.

Any temporal relation of a subordinate clause to its matrix clause will be indicated by its subordinate conjunction like before, after. Other subordinate conjunctions like about in (1c) can only indicate simultaneity. Subordinate (i.e., dependent) clauses cannot therefore enter into temporal juncture.

(1.2.1) Implication: All sequential clauses are independent clauses (but not all independent clauses are sequential clauses).

For an independent clause to be a sequential clause, its head must include a tense that is not only deictic, indicating a specific time domain, but identify sequential time relations. The English past progressive designates a time before the time of speaking but does not focus on the beginning or end points of that time. Can the progressive function as the head of a sequential clause? A number of cases like (1i) indicate that this is a possibility. The progressive in (1i) is simultaneous with (j,k) but appears to be sequenced after (h).[6]

(1.2.2) Implication: [In English], sequential clauses are headed by verbs in the preterit tense, past progressive, or the present tense with the semantic interpretation of a preterit (historical present).

Both the general definition of narrative and the definition of temporal juncture demand that the reports be reports of real events. It follows that modals, futures and negatives cannot serve as the heads of verb phrases which enter into temporal juncture. In English, this function is reserved for the indicative mood, which is our only realis mood.,

(1.2.3) Implication: All sequential clauses are in the realis mood.

(1.3) Definition: A narrative clause consists of a sequential clause [the head] with all subordinate clauses that are dependent upon it.

2. Temporal types of narrative clauses.

We can now employ these definitions to give a simpler and clearer picture of temporal ranges than L&W provided. With the narrative clause defined on the basis of sequential clauses -- clauses that can have temporal juncture -- it is possible to focus only upon temporal relations of the narrative clauses, and exclude others.[7]

(2.1) Definition: Therange of a narrative clause is the set of narrative clauses between the first preceding and next following temporal juncture.

In the transcription conventions followed here, the narrative range is indicated by a left subscript indicating the number of preceding narrative clauses the particular clause is simultaneous with, and the right subscript the number of following clauses. The range is then the sum of the two.

(2.2) Definition: A free clause is a clause which refers to a condition that holds true during the entire narrative.

A free clause is not then defined syntactically, but semantically. A past progressive which serves as a restricted clause in one narrative may be a free clause in another.

(2.2.1) Implication: A free clause cannot serve as a sequential clause in that narrative in which it is free.

(2.3) Definition: A [temporally] bound clause is an independent clause with a range of zero.

(2.3.1) Implication: All bound clauses are sequential clauses.

(2.4) Definition: A narrative clause with a range greater than 0 is a restricted clause.

(2.4.1) Implication: Narratives are sets of bound, restricted and free clauses.

We can then rewrite (1) with temporal ranges and classes of narrative clauses indicated. Quotations with multiple clauses are resolved into individual sequential actions. In narrative, an important distinction between actions and quotations is that the actions frequently overlap, while quotations rarely do so. The rule that one person talks at a time is never flouted in personal narrative.

0a2 restrictedOh I w's settin' at a table drinkin'
1b0restrictedAnd - this Norwegian sailor come over
0c0boundan' kep' givin' me a bunch o' junk
about I was sittin' with his woman.
d freeAn' everybody sittin' at the table with me were my shipmates.
0e0boundSo I jus' turn aroun'
0f0boundan' shoved `im,
0g0boundan' told `im, I said, "Go away."
0h0 bound[and I said] "I don't even wanna fool with ya."
0i2An' nex' thing I know
restrictedI 'm layin' on the floor, blood all over me,
1j0restrictedAn' a guy told me, says, "Don't move your head."
0k0 bound[And he said,] Your throat's cut."

Here the subscripts for (a) indicate that (a) is not simultaneous with any preceding events, but does overlap with the two following (and with the free clause (d), which is not counted). But it is not simultaneous with (e), since at that point Shambaugh is no longer simply sitting at the table drinking.

3. Structural types of narrative clauses.

We now consider the structural types of narrative clauses introduced by L&W. The chief addition to this part of the framework is that complicating action clauses are necessarily sequential clauses, that is, they can participate in temporal junctures; this is not true of abstracts, orientations and codas.

(3.1) Definition: An abstract is an initial clause in a narrative that reports the entire sequence of events of the narrative.

(3,2) Definition: An orientation clause gives information on the time, place of the events of a narrative, the identities of the participants and their initial behavior.

(3,3) Definition: A clause of complicating action is a sequential clause that reports a next event in response to a potential question, "And what happened [then]?"

(3.3.1) Implication: All sequential clauses are clauses of complicating action and all clauses of complicating action are sequential clauses.

(3.4) Definition: A coda is a final clause which returns the narrative to the time of speaking, precluding a potential question, "And what happened then?"

These are all qute self-explanatory, but also qite incomplete. Missing as yet is the notion of a conclusion or resolution, which cannot be defined until the concept of 'most reportable event' is introduced. More important is the fact that many clauses in narratives do none of these things. The major thrust of the L&W analysis is to raise the question, "If a narrative is a report of events that occurred, why do we find sentences headed by negatives, futures and modals in narratives?" Or to put it another way, under what conditions is it relevant to talk about what did not, but might have occurred? The primary contribution of that paper was to link this structural question with the socio-emotional concept of 'evaluation.'

4. Evaluation.

I begin here with a non-linguistic definition of this basic concept.

(4.1) Definition: Evaluation of a narrative event is information on the consequences of the event for human needs and desires.

(4.2) Definition: An evaluative clause provides evaluation of a narrative event.

Though this is quite straightforward, it does not by itself relate to the structural features of narrative in section 2. L&W discussed many types of linguistic structures that served the function of evaluating narrative events, including emphasis, parallel structures and comparatives. By far the most important of these were the modals, negatives and futures that were questioned above. Our proposal that these references to events that did not occur, might have occurred, or would occur served an evaluative purpose was the main theoretical thrust of L&W, and might be described as the:

(4.3) L&W Hypothesis 1: A narrative clause in an irrealis mood is an evaluative clause..

More generally, looking at comparatives as well, we can advance this proposal to the level of a theorem:

(4.4) L&W Theorem 1: A narrator evaluates events by comparing them with events in an alternative reality that was not in fact realized.

From work that followed immediately after L&W, we know that the frequency of irrealis clauses in narrative increases rapidly with age, as speakers gain the ability to evaluate their experience (Labov 1972).

Finally, we have to distinguish between an evaluative clause and an evaluation section. Evaluative material is frequently spread throughout a narrative, but more frequently it is concentrated in a way that suspends the forward movement of the action. More generally, we define such sections:

(4.5) Definition: A section of a narrative is a group of clauses of a common functional type.

and by this means re-state a second L&W Theorem:

(4.6) L&W Theorem 2: Evaluation is characteristically concentrated in an evaluation section, placed just before the most highly evaluated action, or 'point' of the narrative.

The application of this conception of evaluation to (1) is straightforward except for the problem of quotations. On the one hand, (1g) can be viewed as a simple action which is a bound event: Shambaugh said something to the Norwegian sailor. On the other hand, what was said represents two distinct speech actions: a bare imperative which represents an unmitigated command; and a negative scalar (='if I don't want to do a minimal action like fooling with you, it follows that I don't want to do any more important action.') From everything we know about the connectivity of speech acts, the analysis must ultimately rise to this more abstract level of action. Yet it is at the level of sentence grammar that we find our most direct clues to evaluation. When an actor in the narrative is animated to speak directly, no matter what the topic or the addressee, the current situation is open to evaluation. The use of negatives, comparatives, modals or futures is therefore to be read as a form of evaluation. In this sense, (1h) evaluates the narrative situation by comparing it with one in which Shambaugh would want to fool with the other, and in (1j) the other evaluates the situation in comparison with one where it would be safe for Shambaugh to move his head. The assignment of structural categories to the 12 clauses is accordingly that of (1").

OR0a2 restrictedOh I w's settin' at a table drinkin'
CA1b0 restrictedAnd - this Norwegian sailor come over
CA0c0 boundan' kep' givin' me a bunch o' junk
about I was sittin' with his woman.
ORd freeAn' everybody sittin' at the table with me were my shipmates
CA 0e0 boundSo I jus' turn aroun'
CA 0f 0boundan' shoved `im,
CA0g0boundan' told `im, I said, "Go away,"
EV0h0 bound [and I said,] " I don't even wanna fool with ya."
An' nex' thing I know
CA 0i2restricted I 'm layin' on the floor, blood all over me,
EV 1j0restrictedAn' a guy told me, says, "Don't move your head."
CA 0k0 bound[and he said], "Your throat's cut."

5. Reportability.

One of the most difficult yet essential concepts in narrative analysis is reportability. The original concept is that the telling a narrative requires a person to occupy more social space than in other conversational exchanges -- to hold the floor longer, and the narrative must carry enough interest for the audience to justify this action. Otherwise, an implicit or explicit "So what?" is in order, with the implication that the speaker has violated social norms by making this unjustified claim. The difficulty is that there is no absolute standard of inherent interest, and it has been proposed that in some relaxed circumstances with no competing topics, a narrative can be told that is thoroughly banal and ordinary. Given the difficulty of measuring the interest of the narrative or the competing claims, this approach to reportability is itself of limited interest. Yet the concept of "the most reportable event" is central to the organizational structure of the narrative, as we will see below.

One approach to this problem is to turn to a more objective aspect of the narrator's social situation, as developed in Sacks' approach to the insertion of narrative into conversation (1992, Vol. 2, p. 3-5). In Sacks' approach, the problem is not seen as one of "holding the floor," but rather of controlling the assignment of speaker. For Sacks, a narrative is rarely told as a single turn of talk, since the frequent back channel signals of the addressee are themselves taken as turns of talk. I summarize his discussion as:

(5.1) Sacks Assignment Theorem: In free conversation, speakers have no control over the assignment of speaker in the second or third turn following their turns, but the performance of the narrative is effectively a claim to return the assignment of speakership to the narrator until the narrative is completed.

This Sacks principle has four implications that lead to a new definition of reportability.

(5.1.1) Implication 1: Since a narrative requires a series of narrative units longer than the normal turn allows, the successful completion of the narrative requires automatic re-assignment of speaker role to the narrator after the following turn of talk if the narrative is not completed in that turn.

(5.1.2) Implication 2: A narrative must be introduced by a speech act which informs listener than automatic reassignment to the narrator will be required if the narrative is not completed within that stream of speech.

(5.1.3) Implication 3. Listeners have a reliable means of recognizing the ends of narratives.

(5.1.4) Implication 4: To be an acceptable social act, a narrative must be accepted as justifying the automatic re-assignment of turns to the narrator.

We can now re-introduce a definition of reportability in terms not of the general concept, but of a reportable event in the narrative.

(5.2) Definition: A reportable event is one which justifies the automatic reassignment of speaker role to the narrator.

(5.2.1) Implication: To be an acceptable social act, a narrative of personal experience must contain at least one reportable event.

It is clear that the reportability of the same event will vary widely depending on the age, experience, cultural patterns of the speakers, and even more importantly, the immediate social context with its competing claims for re-assignment of speakership. The universal principles of interest which underlie this approach to narrative dictate that certain events will almost always carry a high degree of reportability: those dealing with death, sex and moral indignation. Yet one step outside of these parameters leads us to a such a high degree of contextualization of reportability that only a person intimately acquainted with the audience and the recent history of the social situation can be sure of not making a misstep in introducing a narrative. This relativization of reportability does not however prevent us from recognizing within a narrative degrees of reportability with some confidence. In fact, the creation of a narrative and the ensuing narrative structures are dependent upon the recognition of a unique event which is the 'most reportable.'

(5.3) Definition: A most reportable event is the event that is less common than any other in the narrative and has the greatest effect upon the needs and desires of the participants in the narrative [is evaluated most strongly].

A narrative of personal experience is essentially a narrative of the most reportable event in it. This is normally reflected in the abstract, if there is one. As we will see, the construction of narrative must logically and existentially begin with the decision to report the most reportable event. Narrative (1) is introduced as a narrative about a situation where Shambaugh was close to dying. The most reportable event in it is that the Norwegian sailor cut Shambaugh's throat. The problem of narrative construction is how to construct a series of events which include in a logical and meaningful way this most reportable event. But before considering how this is done, we must recognize another dimension, orthogonal to that of reportability. Given the constraints of social situations, and the pressure to assert claims to speakership, it is normal for speakers to put forward narratives of the most reportable event in their immediately relevant biography. It follows that:

(5.3.1) Implication. The more reportable the most reportable event of a narrative, the greater justification for the automatic reassignment of speaker role to the narrator.

This creates the paradox of the next section.

6. Credibility.

At the outset, it was pointed out that this approach to narrative is based upon serious and straightforward accounts of events that are asserted to have actually taken place, rather than jokes, tall tales, dreams or other genres of a less serious nature.[8] The narrative is then heard as an assertion that the events narrated did take place, in roughly a forming corresponding to the verbal account. This immediately involves the concept of the credibility of the narrative.

(6.1) Definition: The credibility of a narrative is the extent to which listeners believe that the events described actually occurred in the form described by the narrator.

Remembering that the reportability of an event is related to its frequency, as well as its effects upon the needs and desires of the actors, it follows almost automatically that as reportability increases, credibility decreases. This in fact may be termed the Reportability Paradox, which may be stated as a theorem.

(6.2) Theorem: Reportability is inversely correlated with credibility.

The further understanding of how narrators create narratives, and what structures they erect as they produce them, depends upon an understanding of this paradoxical relationship. The next proposition is not an obvious implication, but one that proceeds from the observation of social life. It is limited to 'serious' narratives in the sense stated at the beginning of this section, and may be relative to various social contexts.

(6.3) Theorem: A serious narrative which fails to achieve credibility is considered to have failed, and the narrators claim to re-assignment of speakership will then be seen as invalid.

An "invalid claim to re-assignment" is a technical way of stating that the narrator has suffered a loss of status which will affect future claims of this sort as well as other social prerogatives. It is an outcome normally to be avoided.

(6.3.1) Implication: The more reportable the events of a narrative, the more effort the narrator must devote to establishing credibility.

The nature of that effort must now command our attention.

7. Causality.

Given the fact that the narrator has decided to produce a narrative about the most reportable event, considerations of credibility lead logically and inevitably to the following mechanism of narrative construction:

(7.1) Theorem: Narrative construction requires a personal theory of causality.

1. The narrator first selects a most reportable event e0, which the narrative is going to be about.

2. The narrator then selects a prior event e-1 which is the efficient cause of e0, that is answers the question about e0, "How did that happen?"

3. The narrator continues the process of step 2, recursively, until an event e-n is reached for which the question of step 2 is not appropriate.

The question "How did that happen?" is not appropriate when the answer is, "Because that's the kind of thing we always (usually) do." Event e-n is the Orientation or the narrative, more specifically, the behavioral context of the Orientation. In (1), such orientation is provided by (a). Shambaugh need not explain why it came about that he and his shipmates were sitting around a bar drinking: this is the kind of thing that they always do in port, and the listener is understood to know this.

We have no direct evidence of the actual sequence of steps 1...3; the view of narrative as a folk theory of causality does not rest upon observation. It is a necessary implication of all the definitions and implications of sections 5 and 6. There are many intricacies and complications in the full description of the options open to the narrator in constructing this causal theory. In the case of narrative (1), the causal sequence of events reconstructable from the form of Shambaugh's account may be given as follows:

e4Orientation: Shambaugh and his shipmates were sitting at a table drinking.
e3[For no known reason,] a Norwegian sailor came to complain to Shambaugh about a non-existent condition.
e2[Because there was no basis for the complaint,] Shambaugh rejected the complaint.
e1[Because there was nothing further to be said,] Shambaugh turned his back on the sailor.
e0[Because Shambaugh had turned his back to the sailor,] the sailor was able to cut Shambaugh's throat.

The causal relationships are not given as explicitly in the narrative as in (7.2). The causal basis of e2 is given in the word junk of (1c) and (1d). e1 is implied but not stated. The causal link between e1 and e0 is actually given by Shambaugh in a discussion that followed the narrative. The moral he drew from the story is that the next time he shoves someone he will stand up and hit him. Shambaugh's theory of the events is therefore that he got his throat cut because he turned his back on someone who was behaving in an incomprehensible way.

There are many intricate and difficult issues in the reduction of a narrative statement to a causal one, and undoubtedly there will be wide variations in such acts of interpretation. The essential construction is that there is a proposed chain of events linking the orientation to the most reportable event. It will turn out eventually that the selection of the Orientation is a crucial act of interpretation of the stream of events, and a necessary step in the next aspect of narrative, the assignment of praise and blame.

8. The assignment of praise and blame.

In section 7, narrative construction is equivalent to assigning a theory of causality. In accounts of conflict between human actors, or the struggle of human actors against natural forces, the narrator and the audience inevitably assign praise and blame to the actors for the actions involved. The ways in which this is done are too varied to be reduced to a simple set of propositions. They include the use of linguistic devices of mood, factivity and causativity, evaluative lexicon, the insertion of "pseudo-events,"[9] and the wholesale omission of events. Narratives may be polarizing, where the antagonist is viewed as maximally violating social norms, and the protagonist maximally conforming to them; or they may be integrating, where blame is set aside or passed over by a variety of devices. The study of how narrators assign praise and blame is a major aspect of narrative analysis, which lies beyond the scope of the present article.

It may enough at the moment to see that in (1) the antagonist is assigned to a social type conventionally associated with Scandinavian sailors in port: a large, violent, drunken and irrational person (see O'Neill 1956). Shambaugh views himself as a rational being who made a mistake in underestimating the extent of the irrationality of the other.

An understanding of how the underlying events are presented can be obtained from a broader view of the most likely cause of the events involved. The scene in the bar in Buenos Aires reflects a common source of violence in working class society, reflected in many narratives gathered over the years. The situation is most clearly analyzed in the following extract from an interview with Joe Dignall, 20, of Liverpool:

A lot of fellas, if they're with a gang, they let their birds sit with their mates, while he stands at the bar with his mates, talkin' about things. And you could go up, start chattin' this bird up, an' next thing--y'know, you're none the wiser. An' she's edgin' yer on, on, you're a nice fella, you've got a few bob. Great! And--you're chattin' it up there, you're buyin' her a few shorts. . . Nex' thing, eh, a fella comin' there over there, "Eh ay lads. . . what are ya doin'?" Well you don't know he's goin' with her, so you tell him to push off. Nex' thing he's got his friends--his mates on to you, an' uh. . . you're in lumber! You've either got to run, or fight!

Thus one can see behind these events the possibility that there was in fact a woman sitting at the table who had originally come with or been with the Norwegian sailor; that she had joined Shambaugh's group, or that his group had joined her, seeing no connection between her and the Norwegian sailor; that she had stayed to pick up a few drinks or customers; and that the Norwegian sailor had in his own eyes a legitimate complaint. The sequence of causal factors from the point of view of the Norwegian sailor would be quite different, and the motivation for his behavior might be less incomprehensible. This is not our problem here, however. It is sufficient to see that Shambaugh has presented a causal sequence beginning with the Orientation (1a) which implements his own causal theory. It is equally clear that if the story had been presented from the viewpoint of the Norwegian sailor, a different orientation would have been selected, one considerably earlier in the sequence of events.

The assignment of praise or blame certainly reflects the point of view of the narrator. But it is not usually a conscious part of the information conveyed by the narrator to the audience; it is rather the ideological framework within which events are viewed. By making that ideology overt, we are departing from the dramatic mechanism that is the essence of the narrative speech event: the transfer of experience from the narrator to the audience. That experience is certainly colored by the moral stance taken by the narrator. One might think that those who take the same moral stance will be more impressed by the narrative than those who do not, and therefore find it more credible as well as more interesting and engaging. So far, it appears that this is true only to a limited extent. Shambaugh's narrative appears to have an equal impact on those who accept his point of view and those who find him insensitive to the social reality around him. The effect of the narrative in transferring experience is relatively independent of the narrator's assignment of praise and blame. Through one means or another, the narrator induces the audience to see the world through the narrator's eyes.

9. Viewpoint.

In trying to understand how experience is transferred from the narrator to the audience, we encounter the most characteristic feature of narratives of personal experience, the particular point of view or viewpoint from which the action is seen.

(9.1) Definition: The viewpoint of a narrative clause is the spatio-temporal domain from which the information conveyed by the clause could be obtained by an observer.

One feature of oral narratives of personal experience that distinguishes them most sharply from literary narrative is that in literature, one can switch viewpoints, take an impersonal viewpoint, and enter into the consciousness of any or all of the actors. In oral narratives of personal experience, there is only one option. The events are seen through the eyes of the narrator.

(9.2) Finding: The viewpoint in oral narratives of personal experience is that of the narrator at the time of the events referred to.

This finding applies consistently to all the narrative and free clauses in the narrative. It applies to all of the clauses of the narrative (1), which are consistently an account of events as they become known to Shambaugh.

(9.2.1) Implication: The temporal sequence of events in oral narratives of personal experience follows the order in which the events became known to the narrator.

In literary narrative, it is not uncommon for the viewpoint to shift to give information about events that occurred at an earlier point in time. Classically, this is expressed as a flash-back: "Meanwhile, back at the farm..." It is an empirical findings of some weight that flashbacks are not used in the type of narrative that we are dealing with here.

(9.2.2) Finding: There are no flashbacks in oral narratives of personal experience.

The no flash-back condition holds for a very large number of narratives of personal experience that have been collected and studied over the years. In narrative (1), Shambaugh's throat is cut without him realizing it. We do not learn that this has happened until Shambaugh learns of it. One can easily construct a plausible narrative in which the opposite technique is used. For example, "The next thing I know I'm lying on the floor, blood all over me. He had pulled a knife and cut me before I knew what happened. A guy told me, says...." Yet no such examples have been found. The past perfect is used, but only to report events that the narrator knew at the time they are reported. Perhaps examples will be found if the search continues long enough, but at this time it seems to be an empirical fact that the no flashback condition is binding on oral narratives of personal experience.

10. Objectivity.

Among the thousands of personal narratives that have been recorded and studied over the past few decades, we find a great deal of variability in the degree of objectivity. L&W presented a scale of the objectivity of evaluative statements which range from reports of internal emotions to observations of material objects and events.

In general, we find that narratives of upper middle class, university-educated speakers tend to report on the narrators' emotions. In contrast, many working class narrators are sparing in their reporting of subjective feelings. It was somewhat surprising to me to find that the "subjectivity"characteristic of middle class speakers is considered a positive quality by therapeutically oriented writers, and people who do not report their emotions in narratives of bereavement, for example, are considered to be suffering from an impairment of normal and desirable abilities. My own experience in retelling narratives to various experiences indicates a somewhat different scale of values. Those narratives that have the greatest impact upon audiences in the sense outlined above -- which seize the attention of listeners and allow them to share the experience of the narrator -- are those that the use the most objective means of expression. To follow this argument more exactly, some definitions are required.

(10.1) Definition: An objective event is one that became known to the narrator through sense experience. A subjective event is one that the narrator became aware of through memory, emotional reaction or internal sensation.

The general observation that narratives that report experience objectively are more effective than those that report subjectively is not a strongly evidentiary statement; no hard data supports it. Nevertheless, some experimental evidence supports the belief that objectivity increases credibility.

(10.2) Theorem: Since it is generally agreed that the narrators' observations can be affected by their internal states, reports of objective events are more credible than reports of subjective events.

To exlore the many ways in which objective and subjective approaches to narrative differ would go beyond the scope of this brief summary. The clauses of narrative (1) are entirely objective. At no point are there any statements that describe how the actors feel. We are told what they said and what they did. The concluding statements of (1j,k) are those of a third person witness to the events, more objective than any statements of the main actors. It is proposed here that this objectivity is a necessary condition for the capacity of narratives like (1) to transmit experience to the audience. To develop this point clearly, a further implication may be drawn from finding (9.2):

(10.3) Theorem: The transfer of experience of an event to listeners occurs to the extent that they become aware of it as if it were their own experience.

The as if condition may be expanded to mean that listeners achieve awareness of the event in the same way that the narrator became aware of it. The essential No Flash-Back condition follows immediately from this condition. If theorem (10.3) holds, two consequences immediately follow to mark the limits of the transfer of experience in personal narrative.

(10.3.1) Implication: The transfer of experience from narrator to audience is limited, since the verbal account gives only a small fraction of the information that that the narrator received through sight, sound and other senses.[10]

(10.3.2) Implication: To the extent that narrators add subjective reports of their emotions to the description of an objective event, listeners become aware of that event as if it were the narrator's experience.

These two implications lead to a further proposition that is more than an implication. It combines the experience derived from re-telling stories and observations of social class differences in narrative structure with the logic of 10.3.1-2 to derive the following theorem:

(10.3.3). Theorem: The objectivity of the description of an event is a necessary condition for the transfer of experience in personal narrative.

There is more than a hint of a paradox here. The transfer of experience is a subjective phenomenon, which is not easy to observe or measure. The theorem argues that this subjective experience is obtainable only through the objective presentation of events.

11. Resolution.

In the L&W treatment, the resolution of a narrative was simply the ending or outcome; there was no very precise way of distinguishing it from the last complicating action. But this situation is sharply altered with the introduction of the most reportable action as a structural unit in section 5. The resolution can be seen to be logically the series of complicating actions that follow rather than precede e0.

(11.1) Definition: The resolution of a personal narrative is the set of complicating actions that follow the most reportable event.

In narrative (1), there appears to be no distinct resolution section. At first glance, the narrative ends with the most reportable event, so that the resolution coincides with the e0. Yet closer examination shows that (1j,k) is not the most reportable event, but a report of the most reportable event, which itself is implied rather than reported. Is it then a resolution?

L&W defined the coda as a the clause or clauses that bring the narrative back to the time of telling, so that the question "What happened then?" is no longer appropriate. This does not mean that the listener is automatically satisfied with all of the information given on the outcome of the most reportable event. If a resolution is not satisfactory in this respect, the listener will have the impression that the narrative is incomplete. I have probed a number of audiences for their reaction to narrative (1) on this point, and the consensus seems to be that the narrative does come to an end. Shambaugh paused at this point long enough for me to ask a question, and in a series of exchanges I learned that the Norwegian sailor's knife had cut his throat but missed the jugular vein; that Shambaugh in fact had the knife upstairs in his room; that one of his friends had hit the Norwegian sailor with a chair; and that the blow had killed him. These facts are interesting, and their absence from the narrative throws light on Shambaugh's approach to the matter. But they do not form part of the narrative as it is now constituted, and we must infer that (1j,k) is indeed its resolution.

12. Conclusion.

The L&W analysis introduced the definition of narrative as a technique of reporting past events through temporal juncture, and established a basis for the understanding of the temporal organization and evaluation of narrative. This contribution has explored further the concept of reportability, arguing that the most reportable event is the semantic and structural pivot on which the narrative is organized. Given an initial inverse relation between credibility and reportability, it follows that narrators who command the attention and interest of their audience will normally maximize credibility by the objective reporting of events.

The second half of this contribution focuses upon the capacity of a narrative to transfer the experience of the narrator to the audience. This capacity is seen to depend upon the unique and defining property of personal narrative that events are experienced as they first became known to the narrator. It is proposed that the transfer of experience of an event to listeners occurs to the extent that they become aware of it as if it were their own experience. It follows that this is only possible if the narrator reports events as objective experience without reference to the narrator's emotional reactions.

It is also proposed that a narrative can be viewed as a theory of the causes of the most reportable event, so that the crucial interpretive act is the location of the orientation as the situation that does not require an explicit cause. The chain of causal events selected in the narrative is intimately linked with the assignment of praise and blame for the actions reported. This view of narrative as a theory of moral behavior and the narrator as an exponent of cultural norms will be pursued in later publications.


Labov, William. 1966. The Social Stratification of English in New York City. Washington D.C.: Center for Applied Linguistics.

Labov, William. 1972. Language in the Inner City. Philadelphia: Univ. of Pennsylvania Press.

Labov, William. 1981. Speech actions and reactions in personal narrative. In D. Tannen (ed.), Analyzing Discourse: Text and Talk. Georgetown University Round Table. Washington, DC: Georgetown University Press. Pp. 217-247.

Labov, William and Joshua Waletzky. 1967. Narrative analysis. In J. Helm (ed.), Essays on the Verbal and Visual Arts. Seattle: U. of Washington Press. Pp. 12-44

Labov, William and David Fanshel. 1977. Therapeutic Discourse: Psychotherapy as Conversation. New York: Academic Press.

Labov, William, P. Cohen, C. Robins and J. Lewis.. 1968. A study of the non-standard English of Negro and Puerto Rican Speakers in New York City. Cooperative Research Report 3288. Vols. I and II. Philadelphia: U.S. Regional Survey (Linguistics Laboratory, U. of Pa.)

O'Neill, Eugene. 1956. Long Day's Journey into Night. New Haven: Yale University Press.

Sacks, Harvey 1992. Lectures on Conversation. Volume I and II. Edited by Gail Jefferson. Cambridge: Blackwell.


[1]Particularly, The re-organization of reality, the lectures given at the University of Rochester, October 1977; The vernacular origins of epic style, the W.P. Ker Lecture at the U. of Glasgow, May 4, 1983; On not putting two and two together: the shallow interpretation of narrative, Pitzer College, March 10, 1986; and a number of presentations of narrative under the title of "Great Speakers of the Western World."

[2] The most recent form of the appreoach to narrative presented here is the producvt of a class on Narratigve Analysis that I taught at Penn in the fall of 1976, and is much indebted to the contributions of members of the class. I would like to thank in particular Trevor Stack, Matt Rissanen, Kirstin Smith, Pierette Thibault and Dr. Herbert Adler.

[3] In fact, one approach to the definition of conversation is to see it as a way of instancing general principles by means of narrative.

[4] This narrative is one of those discussed in detail in the exploration of the relation of speech to violence in Labov 1981.

[5] The question of ordering events within quotations is a difficult one; my best judgment is that the utterances of (g) and (h) could have been uttered in either order without change in their logical or interactive force, and so for (j) and (k). But for the moment, I take the narrative literally saying that the two sets of utterances were so ordered. The question as to whether (e,f,g) are well ordered within themselves is also open.

[6] If we interpret the grammatical construction as derived from "The next thing I know is..." then the verb to be would be a sequential clause, which is not likely. The next thing I know must then be interpreted as equivalent to Next. . ..

[7] Thus it is no longer necessary to engage in a lengthy tabulation of the range of free clauses, including all preceding and all following clauses, numbers which may change each time we revise the transcription.

[8] This does not preclude the presence of humorous elements in a narrative that is fundamentally serious: all that counts here is that the narrative is understood as asserting that the events actually took place in roughly the form reported.

[9] The use of turn around in (1e) is one of the many verbs of motion that are used to amplify the degree of activity of a narrator, which do not necessarily involve an observable action.

[10] In Sacks' view, listeners are not in fact entitled to the experience of the narrator; it cannot become their experience (1992 II:242-248). While no one can doubt Sacks fundamental insight that experience cannot be transferred as easily as information, the article take here is based on the empirical finding that some degree of experience is in fact transferred.